SUU, Alm.del - 2021-22 - Endeligt svar på spørgsmål 1026: MFU spm. om, hvilke karakteristika der er ved modellerne for tilladt brug af cannabis i blandt andet Tyskland, Canada, Uruguay, Portugal, USA og Holland, til sundhedsministeren

Sundhedsudvalget 2021-22
SUU Alm.del
Offentligt

TECHNICAL REPORT

Monitoring and evaluating changes

in cannabis policies: insights from

the Americas

January 2020

SUU, Alm.del - 2021-22 - Endeligt svar på spørgsmål 1026: MFU spm. om, hvilke karakteristika der er ved modellerne for tilladt brug af cannabis i blandt andet Tyskland, Canada, Uruguay, Portugal, USA og Holland, til sundhedsministeren

Legal notice

This publication of the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) is

protected by copyright. The EMCDDA accepts no responsibility or liability for any consequences

arising from the use of the data contained in this document. The contents of this publication do

not necessarily reflect the official opinions of the EMCDDA’s partners, any EU Member State or any

agency orinstitution of the European Union.

PDF

ISBN 978-92-9497-461-7

doi:10.2810/151487

TD-02-20-009-EN-N

Luxembourg: Publications Office of the European Union, 2020

Reproduction is authorised provided the source is acknowledged.

Recommended citation: European Monitoring Centre for Drugs and Drug Addiction (2020),

Monitoring

and evaluating changes in cannabis policies: insights from the Americas,

Technical report, Publications

Office of the European Union, Luxembourg.

About this report

This report provides an overview of the changes in cannabis policies in the Americas

and the evidence emerging from evaluations of their impact. Highlighting the

challenges in monitoring and evaluating regulatory changes in the drugs field, it will

be of particular interest to those involved in planning or evaluating any changes to

cannabis regulation.

About the EMCDDA

The European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) is the

central source and confirmed authority on drug-related issues in Europe. For over

20 years, it has been collecting, analysing and disseminating scientifically sound

information on drugs and drug addiction and their consequences, providing its

audiences with an evidence-based picture of the drug phenomenon at

European level.

The EMCDDA’s publications are a prime source of information for a wide range of

audiences including: policymakers and their advisors; professionals and researchers

working in the drugs field; and, more broadly, the media and general public. Based in

Lisbon, the EMCDDA is one of the decentralised agencies of the European Union.

Praça Europa 1, Cais do Sodré, 1249-289 Lisbon, Portugal

Tel. +351 211210200

[email protected] www.emcdda.europa.eu

twitter.com/emcdda facebook.com/emcdda

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Contents

Acknowledgements ............................................................................................................................. 3

Executive summary ............................................................................................................................. 4

Foreword.............................................................................................................................................. 6

1. Introduction ...................................................................................................................................... 7

Background to cannabis policy developments in the Americas ...................................................... 9

A description and typology of different policy approaches ............................................................ 13

3.1

US approaches...................................................................................................................... 14

3.2

Uruguay ................................................................................................................................. 15

3.3

Canada .................................................................................................................................. 15

4. An assessment of early evidence of the impact of policy changes in the United States ............... 16

4.1

Literature review selection criteria......................................................................................... 17

4.2

Prevalence of use ................................................................................................................. 17

4.3

Consumption patterns ........................................................................................................... 23

4.4

Product differentiation and price ........................................................................................... 24

4.5

Treatment admissions ........................................................................................................... 26

4.6

Adverse medical events ........................................................................................................ 26

4.7

Impaired driving ..................................................................................................................... 28

4.8

Consumption of other substances......................................................................................... 30

4.9

Criminal justice and public nuisance outcomes .................................................................... 30

4.10

Tax revenues......................................................................................................................... 32

4.11

Public opinion ........................................................................................................................ 33

5. Issues to consider when establishing a comprehensive monitoring and evaluation framework for

changes to cannabis regulations .......................................................................................................... 34

5.1

Potential objectives of cannabis policy change and metrics of evaluation ........................... 34

5.2

Establishing the data infrastructure ....................................................................................... 40

5.3

Thinking seriously about outcome evaluations ..................................................................... 42

6. Concluding thoughts ...................................................................................................................... 45

References ............................................................................................................................................ 46

Appendix A: Identification and commentary on ongoing/planned studies of legalisation ..................... 52

US studies ......................................................................................................................................... 52

Uruguay ............................................................................................................................................. 53

Canada .............................................................................................................................................. 53

Appendix B: Research papers on the impacts of recreational cannabis laws ...................................... 54

Appendix C: Additional trend data ........................................................................................................ 69

Prevalence ......................................................................................................................................... 69

Hospital and poison control centres .................................................................................................. 73

Traffic fatalities .................................................................................................................................. 75

Arrests for cannabis possession and supply offences ...................................................................... 76

Acknowledgements

The EMCDDA would like to thank Bryce Pardo, Beau Kilmer and Rosalie Liccardo Pacula, RAND

Europe/RAND Drug Policy Research Center for authoring this report. The EMCDDA is also grateful

to Rebecca Jesseman, Director of Policy, CCSA and members of the EMCDDA Scientific Committee,

who reviewed and provided comments on the manuscript.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Executive summary

Over the past 50 years, several jurisdictions in Europe, Australia and the Americas have reduced the

penalties associated with using or possessing small amounts of cannabis. As of December 2019,

Canada, Uruguay and 10 US states have gone further and passed laws that license the production

and retail sale of cannabis, mostly by private companies, to adults for non-medical — sometimes

referred to as recreational — purposes (

). With discussions about alternatives to cannabis

prohibition becoming more

common in some parts of the world,

there is a growing interest in learning

from the cannabis policy changes in the Americas.

To learn more about these new cannabis regimes and their consequences, the European Monitoring

Centre for Drugs and Drug Addiction (EMCDDA) commissioned a review of the changes governing

recreational cannabis policies in the Americas and an overview of preliminary evaluations. Findings

from this research are intended to inform discussions about the development of a framework for

monitoring and evaluating policy developments related to cannabis regulatory reform. Key insights

include the following.



In addition to the populations of Canada and Uruguay, more than 25 % of the US population

lives in states that have passed laws to legalise and regulate cannabis production, sales and

possession/use for recreational purposes. In the US, allowing licensed production and sales

is often at the discretion of sub-state jurisdictions, which may impose further zoning

restrictions on cannabis-related activities. This variation can complicate analyses that attempt

to compare legalisation and non-legalisation states, especially when the outcome data are not

representative at state level.

The peer-reviewed literature on cannabis legalisation is nascent, and we observe conflicting

results depending on which data and methods are used, as well as which implementation

dates and policies are considered. It is important to remain sceptical of early studies,

especially those that use a simple binary variable to classify legalisation and non-legalisation

states. This scepticism should extend to the many studies that fail to account for the existence

of robust commercial medical cannabis markets that predate non-medical recreational

cannabis laws. Even if a consensus develops on certain outcomes, it does not mean that a

relationship will hold over time. Changes in the norms about cannabis use and potentially

other substances, the maturation of markets and the power of private businesses (if allowed)

could lead to very different outcomes 15 or 25 years after recreational cannabis laws have

passed. Evaluations of these changes must be considered an ongoing exercise, not

something that should happen in the short term.

One area in which there seems to be a consensus — so far — is with respect to cannabis-

involved hospitalisations and emergency department (ED) visits in Colorado. While studies

have observed increases in the number of adverse events after changes in medical or

recreational supply, they tend to utilise a simple pre-post design without a control group.

Thus, they do not produce causal estimates, and it is possible that some increases could be

due to changes in reporting or measurement. For example, it is unclear if increases in the

number of ED visits and hospitalisations are due to a greater willingness on the part of

individuals to report the use of cannabis and/or if doctors are more aware of acute cannabis

intoxication after the policy change and are now more likely to screen or confirm with

urinalysis.

One insight arising from the evaluations of the regulatory changes in the Americas to date is

the importance of the amount and range of data collected before the change; simply



( ) Two other US jurisdictions, the state of Vermont and the District of Columbia, also allow the cultivation and possession of

cannabis for adults but do not permit its commercialization.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

comparing past-month prevalence rates will not tell us much about the effect of the change on

health. While US jurisdictions have been moving quickly to legalise the use of cannabis, the

data infrastructure for evaluating these changes is limited. In contrast, Canada has made

important efforts to field new surveys and create new data collection programmes in

anticipation of legal changes. This highlights the importance of any jurisdictions that are

considering changes to the regulatory framework for cannabis starting to think about

improving data collection and analysis systems in advance.



While there is much to learn from what is happening in the Americas, policy discussions

should not be limited to approaches that have been implemented there. There are several

regulatory tools (e.g. minimum pricing, potency-based taxes) that receive very little attention

— if any — that could have important consequences for health, public safety and/or social

equity. It needs to be recognised that all decisions of this nature involve trade-offs and

acknowledging that individuals (and governments) have different values and preferences for

risk when it comes to cannabis policy is important for productive debates on this controversial

topic.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Foreword

The EMCDDA exists to facilitate an evidence-informed understanding of issues that are important for

developing better drug-related policies and actions across Europe. In our series of reports,

Cannabis:

controversies and challenges,

we focus attention on an illicit drug with a long history of use that has

recently been an area of rapid policy development and intense European and wider international

drug policy debate.

some countries, questions on what constitutes an appropriate policy response

to cannabis have become both

topical and important.

The aim of this set of publications is to

explore, in an objective and neutral manner, some of the complex issues that exist in this fast-

changing area.

The paper

Medical use of cannabis and cannabinoids: questions and answers for policymaking,

published in December 2018, provided an overview of developments in the way in which countries

and jurisdictions are regulating the use of cannabis and cannabinoids for medical purposes. In this

report, the latest in the series, the focus is shifted to recent changes to regulatory systems in the

Americas that permit the consumption of cannabis by adults for non-medical, recreational purposes.

In order to improve our understanding of these new cannabis regimes and their consequences, the

EMCDDA commissioned RAND Europe and the RAND Drug Policy Research Center to undertake a

review of the changes to recreational cannabis policies in the Americas and produce an overview of

the evidence emerging from preliminary evaluations.

At this stage the evidence base is still insufficient to comment with any certainty on the impact of the

changes that are occurring in the Americas. However, this is a rapidly developing area and this

review provides a detailed summary of the current evidence base. The EMCDDA will continue to

follow up on developments to help inform the European debate in this area.

Importantly, this review, in addition to providing EU audiences with a clearer picture of the

developments occurring in the Americas, also highlights some of the challenges associated with

monitoring and evaluating regulatory changes in the drugs field. We trust that this report will be of

particular use to those involved in planning or evaluating any changes to cannabis regulation.

Alexis Goosdeel

Director, EMCDDA

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Introduction

The cannabis policy landscape in the Americas is dramatically changing. Individual states in the US

began liberalising their cannabis laws as far back as the 1970s, when several of them reduced or

removed criminal penalties for possession of small amounts of cannabis. Starting in the 1990s, US

states began to allow individuals to possess cannabis for qualified medical conditions. In contrast to

subsequent European systems, medical cannabis in the US was primarily voter initiated and allowed

people to grow and smoke non-standardised cannabis herb for a wide range of medical indications.

These early medical cannabis reforms maintained state-level prohibition, except for a class of patients

who obtained recommendations from authorised medical practitioners.

Distinct from, and in marked contrast to, these earlier medical cannabis reforms, voters in the US

states of Colorado and Washington passed ballot initiatives in 2012 to repeal the prohibition on adult,

non-medical cannabis and to license for-profit firms to produce and sell the drug to adults aged 21

years and over. Since the 2012 election, 10 more US jurisdictions have approved commercial models

for non-medical cannabis, and in two others (Vermont and the District of Columbia — DC) non-

commercial approaches have been adopted.

However, reforms to cannabis supply laws in the Americas are not limited to the US. In late 2013,

Uruguay became the first country in the world to repeal prohibition on cannabis supply for non-

medical markets; however, its approach is more restrictive than the regulated commercial regimes

adopted in the US. Registered adults in that country could begin growing a small number of plants at

home or join a social club in late 2014; by mid-2017, individuals could buy rationed amounts of

cannabis, grown by state-authorised producers, in licensed pharmacies that chose to sell it. In June

2018, Canada became the second country to legalise cannabis production and supply for non-

medical purposes; retail stores opened in some provinces, starting in October 2018. In addition, other

countries in the Americas have passed laws to permit access to cannabis or cannabis-derived

products for medical purposes.

Apart from the regulatory changes in the Americas, which are the focus of this report, changes are

occurring globally. For example, the Dutch government is implementing its 2017 commitment to

experiment with a closed supply chain to coffee shops, while Malta’s government has taken steps to

launch a national debate on whether or not there could be recreational cannabis use, and how this

should be implemented. In 2018,

the parties forming the government of

Luxembourg

reached an

agreement that may allow

for

the future sale of

cannabis to residents, while the highest courts in

South Africa and Georgia have initiated reforms based on human rights that permit the consumption

of cannabis in private settings, but not its sale. Other countries have not formally legalised cannabis

but may have reduced the emphasis that they place on control or introduced some formal or informal

tolerance for personal use of the drug.

The focus of this report is on changes that are relevant to recreational cannabis, sometimes referred

to as adult or non-medical cannabis, with an emphasis on the implementation of commercial models

in US states. The report does not separately evaluate medical cannabis reforms in the US or

elsewhere. These are discussed only insofar as they help to explain the context in which subsequent

recreational cannabis reforms took place. Cannabis and cannabinoids are made available for medical

use in a wide variety of ways; hence the term ‘medical cannabis’ can be used to describe many

different products and forms of supply. For reviews of the evidence on the various medical regimes,

see Pacula and Smart (2017), the EMCDDA (2018) report

Medical use of cannabis and cannabinoids

— questions and answers for policymaking

and the accompanying summary of reviews (Hall, 2018).

Similarly, this report does not examine reforms aimed at reducing criminal penalties for possession or

use but focuses on the broader set of policy changes aimed at regulating the recreational cannabis

market.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

To learn more about these new cannabis regimes and their consequences, the EMCDDA

commissioned a review of changes governing recreational cannabis policies in the Americas and an

overview of preliminary evaluations. Findings from this research are intended to facilitate a better

understanding among the EU audience of developments occurring in the Americas and to inform

discussions about the framework necessary for monitoring and evaluating policy developments

related to cannabis policy changes that might increase the availability of this substance within Europe.

This report is structured as follows: following this introduction, Section 2 provides the background to

and context of recent developments in cannabis policies in the Americas, with a particular focus on

the more mature state commercial markets in the US; Section 3 offers a description and typology of

different policy approaches with supporting case summaries; Section 4 presents an assessment of

early evidence of the impact of policy changes in this area; and Section 5 provides guidance for

European policymakers considering regulatory changes in the cannabis area on the establishment of

a comprehensive, monitoring and evaluation framework. Appendix A highlights some of the ongoing

and planned studies of legalisation implementation and its consequences in the Americas.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Background to cannabis policy developments in the Americas

Changes in cannabis policy that introduce greater tolerance can take various forms. Countries in

Europe and elsewhere have reduced penalties for low-level offences, have removed criminal

sanctions for possession or use or have introduced formal or informal procedures that reduce the

likelihood of sanctions being applied for some forms of personal possession or use. However, up to

now these policy changes have taken place within the overall context of maintaining the prohibition of

cannabis supply and not establishing a regulated market for the drug for recreational users. This is an

important distinction, because efforts to repeal prohibition and replace it with rules governing

production, distribution and possession/use of cannabis are expected to have a greater impact on

public health and safety than amending laws related to possession or use. However, many

jurisdictions in the Americas have begun to repeal the prohibition of cannabis, allowing licensed

businesses to produce and supply it for a narrow set of qualified medical patients (i.e.

medical

cannabis)

or any adult over a minimum age (i.e.

recreational cannabis).

The main focus of this report

is on recent efforts aimed at replacing cannabis prohibition with regulated regimes that permit

recreational cannabis. However, some information relating to medical cannabis regimes is also

provided, as they have been a very common step in the evolution of cannabis regulatory

developments in the Americas (Kilmer and MacCoun, 2017) and are therefore relevant to

understanding the context of subsequent reforms in relation to recreational cannabis use.

The move towards changing cannabis laws to allow medical access is driven by multiple factors.

Three common arguments are (1) the desire to improve the lives of patients who could immediately

benefit from these products, particularly when they are produced in a manner that can ensure

consistency and quality; (2) the desire to learn more about the medical benefits and risks of these

products; and (3) the desire to take action to soften prohibition and improve civil liberties. Another

driver of change that is becoming increasingly important in some jurisdictions is the desire to generate

legal economic activity and tax revenues from cannabis businesses (Subritzky et al., 2016; Hall and

Kozlowski, 2018).

It is also important to consider these medical laws when thinking about the move towards legalising

recreational cannabis. As Table 2.1 makes clear, only a small minority of countries that allow medical

cannabis have made the move towards permitting adults to access cannabis for recreational

purposes. In contrast, Uruguay did not start developing a medical programme until after legislation for

recreational cannabis was passed. This legislation was pushed by the government largely to reduce

drug-related crime and violence associated with the cannabis trade (Aguiar and Musto, 2015).

Despite its initial unpopularity with the citizens of Uruguay (Garat, 2015), it was signed into law in

December 2013.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

TABLE 2.1

Changes to cannabis supply laws in the Americas

Country

Implemented

nationally

Year

passed

Comments

Medical

United States

1996

onwards*

In total, 33 states and DC have passed laws to allow cannabis flowers and other

products to be produced and distributed for qualifying conditions. Thirteen other states

allow physicians to recommend or individuals to obtain CBD oils, but do not necessarily

permit their production. This all remains illegal under federal law.

Currently, all medical cannabis is supposed to be produced by federally licensed private

companies and delivered by post. Efforts are being made to eliminate the illegal

dispensaries that still operate in some jurisdictions. Registered patients may also

cultivate a limited amount at home for their own personal needs.

Production is licensed by state, and extracts are available in pharmacies for those with a

physician’s prescription. Currently products containing CBD and a low THC

concentration are available for a set of conditions.

There is limited access to cannabis-derived products approved by health authorities,

which currently include only imported high-CBD/low-THC oils available in pharmacies.

There is an ongoing pilot project with one domestic cultivator.

Licences have been granted since 2017. As at mid-2018, there are a handful of

cultivators and processors as well as two medical dispensaries. Regulations allow small-

and large-scale production of medical cannabis for residents as well as tourist and export

markets.

The decree signed in December 2015 permits medical cannabis under the national drug

law. The decree was superseded by law. The Ministry of Health has licensed cultivators,

although it is unclear how much cannabis production will be permitted. The law and

accompanying regulations focus on export markets.

The health agency resolution allows the importation of cannabis oils for epilepsy and

multiple sclerosis.

A law and a ministerial decree have directed the Ministry of Health to draft regulatory

guidelines to allow the state to produce and patients to inscribe into a programme to

obtain or import cannabis oils and other derivatives for treatment of qualifying conditions.

A legislative act and regulatory decree were passed in 2017 to permit the importation of

medicinal products to treat qualifying conditions. Regulations have not yet been finalised.

A law was passed to guide the Ministry of Health with regard to drafting appropriate

regulations to allow doctors to prescribe and patients to obtain cannabis and cannabis

derivatives.

Recreational**

United States

2012*

Ten states have passed laws to allow for-profit companies to produce and sell cannabis

(Alaska, California, Colorado, Illinois, Massachusetts, Maine, Michigan, Nevada, Oregon

and Washington) to anyone aged 21 years and over; DC and Vermont allow only home

production and sharing of cannabis. Production and possession remain illegal under

federal law.

Residents aged 18 years and over must register with the government to either grow

cannabis at home, join a collective or purchase cannabis from pharmacies. Home

growing of cannabis and social clubs started operating in the latter half of 2014. State-

licensed companies produce retail products sold in pharmacies. Sales began in mid-

2017. The government determines the price and potency of cannabis sold in

pharmacies. Pharmacy sales have slowly been introduced, partly because of issues

pertaining to the access of financial institutions with branch offices in the US.

In mid-2018, the Canadian parliament passed the Cannabis Act to permit adults to

cultivate, process and use cannabis for recreational purposes. The law came into effect

in October 2018. Under the law, the provinces are given regulatory power with regard to

distribution, but federal law provides general rules on production, promotion, packaging

and licensing.

Canada

Yes

2001

Uruguay

Yes

2013

Chile

Yes

2015

Jamaica

Yes

2015

Colombia

Yes

2015

Brazil

Argentina

Yes

2016

2017

Mexico

Peru

Yes

2017

Uruguay

Yes

2013

Canada

Yes

2018

* Federal law still prohibits cannabis. The date reflects the date when the first state passed a law.

** In October 2018, the Mexican Supreme Court ruled that prohibiting cannabis for personal use is unconstitutional; however, this does

not technically legalise use and supply. The legislature is working to amend the law, to codify the ruling. For more information, see Lopez

(2018).

THC, tetrahydrocannabinol — the principal intoxicative agent in cannabis; CBD, cannabidiol — a cannabinoid that has the potential to

reduce certain maladies and attenuate some of the effects of THC.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Most of the changes have been limited to supplying cannabis for medicinal purposes, and there is an

important variation in what these legal frameworks allow. These regulatory systems and the products

that are available to patients differ from traditional medicinal formulations that contain synthetic

cannabinoids (e.g. dronabinol or nabilone) or plant-derived preparations approved by the government

(e.g. Sativex or Epidiolex). Some jurisdictions make it easy for patients to access a myriad of

cannabis preparations, including raw or herbal products that can be smoked or vaporised (e.g. many

US states, Canada), as well as various cannabis-derived preparations, such as vaporisable

concentrates, tinctures, edibles and topicals. More restrictive markets only allow cannabis extracts

with high levels of cannabidiol (CBD) and a trace or modest amounts of tetrahydrocannabinol (THC)

(e.g. about a dozen US states, Chile, Brazil, Mexico) or limit the types or forms of cannabis available

to patients (e.g. prohibiting smoking of herbal or raw cannabis).

However, in US states, recreational legalisation has followed the provision of medical cannabis, and

multiple hypotheses have been offered about how the latter may have smoothed the transition to the

former. Kilmer and MacCoun (2017) offer five potential explanations for the transition:

(a) it demonstrated the efficacy of using voter initiatives to change cannabis supply laws;

(b) it enabled the psychological changes needed to destabilise the ‘war on drugs’ policy stasis;

legalisation;

(d) it created a visible and active cannabis industry;

(e) it revealed that the federal government would allow state and local jurisdictions to generate

tax revenue from cannabis.

Of course, some of these reasons are specific to the US (e.g. state ballot initiatives and federalism).

With respect to the visibility of the cannabis industry, Kilmer and MacCoun (2017) note the following:

The proliferation of medical dispensaries in the 2000s introduced the public to the idea of

stand-alone stores selling [cannabis] products (and not just to those residing in states that

allow medicinal [cannabis], as dispensaries are regularly featured in media stories across the

country). Subsequent advertisements in alternative weekly newspapers and occasionally on

billboards also exposed voters to this new quasi-legal industry.

Of course, not everyone is thrilled with the establishment of medical dispensaries and the

related advertising, and there is tremendous variation in what states allow and how they are

regulated (Pacula et al. 2015). Further, even in states that allow dispensaries, some local

jurisdictions have decided to prohibit them. But for those who see these dispensaries on a

regular basis, they not only have a sense of what legalisation may look like but also may be

desensitised to the idea of retail [cannabis]. In fact, some may conclude that allowing any

adult to enter those stores instead of only those who have recommendations — which in

some states are very easy to obtain (e.g., High Times 2016) — may not be a big change.

Medical [cannabis] laws also created an industry looking to expand its market beyond medical

patients. Indeed, it is hard to imagine that some of those developing, e.g., brands, new

methods of ingestion, and improved production methods did not have their sights on a larger

market. Although those in the advocacy community criticised the industry for not donating

more to the campaigns (e.g., see quotes from Ethan Nadelmann in Freedlander 2016), some

with a financial interest did make contributions.

In Canada, a series of provincial and Supreme Court cases during the early 2000s gradually legalised

the medical use of cannabis. The initial medical system developed at the national level slowly

unfolded. In 2001, Canada enacted enabling legislation via the Controlled Drug and Substances Act

to pass the Marihuana Medical Access Regulations (MMAR). This new set of regulations permitted

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

individuals with qualifying illnesses (which were listed by the regulation) to obtain medical cannabis.

The initial programme was overly restrictive, so that many patients wishing to take cannabis for their

condition had to source it from the illicit market (Fischer et al., 2015). Individuals were allowed to

obtain cannabis by three different means: (1) by applying to access Health Canada’s supply of dried

cannabis via the postal system; (2) by applying for a personal-use production licence (i.e. to grow

cannabis at home); or (3) by designating someone to cultivate on their behalf with a ‘designated-

person production licence’. This last provision was criticised for permitting unregulated and outright

illicit sales (Fischer et al., 2015).

In response to concerns from law enforcement and other policymakers that the system governed by

the MMAR was open to abuse, and after extensive consultations, the Harper government introduced

the Marihuana for Medical Purposes Regulations (MMPR) in 2013 to further restrict and regulate the

market by eliminating home-growing of cannabis and limiting designee production. Under the MMPR,

only licensed producers could cultivate medical cannabis. Justin Trudeau became prime minister in

2015 and prioritised cannabis policy reform, including the legalisation of recreational cannabis. In

2016, the MMPR was struck down on constitutional grounds and was later superseded by the Access

to Cannabis for Medical Purposes Regulations (ACMPR), which built on the existing framework for

licensed commercial producers established under prior regulations while permitting patients to register

to cultivate a small amount for personal medical use. Existing producers licensed under the ACMPR

would later be deemed as such under the Cannabis Act.

While bricks and mortar medical dispensaries (

) were technically illegal under Canadian law, some

jurisdictions allowed them to operate; other jurisdictions, such as Vancouver, eventually established

some licensing and regulatory frameworks to control the growing retail industry (Johnson, 2015). As in

multiple US states, this probably introduced some Canadians to the idea of stand-alone stores selling

cannabis products as well as advertising such products. Many Canadian medical cannabis producers

anticipated supplying the much larger recreational market. As cannabis is federally legal in Canada,

some of these Canadian firms are listed on major US stock exchanges, in contrast to firms from the

US, and some investors believe that they are ‘poised to surge’ (Chang, 2018). Yet projections of

earnings and market valuation have been revised downwards, as sales have not met analyst

expectations (Subramaniam, 2019).

Another important difference between Canada and most of the US states that have legalised

recreational cannabis is the amount of public discussion and deliberation that took place before the

policy change (

). After Prime Minister Trudeau announced his party’s intentions to legalise cannabis,

the Liberal government convened a task force that would spend nearly a year collecting information

about regulating cannabis from several sources, ranging from international experts to domestic

cannabis users. The task force’s resulting report was released in December 2016. The report and

subsequent recommendations helped shape the legislation introduced in April 2017. Soon after, there

were several hearings in parliament about the bill. In the Senate’s Standing Committee on Social

Affairs, Science and Technology alone, there were 18 hearings. After more than 1 year of

deliberation, the final bill received Royal Assent in June 2018, and legal sales began on 17 October

2018.

( ) Shops with a physical location, as opposed to online or virtual businesses.

( ) The most obvious exception is Vermont. Since 2014, the state legislature has seriously deliberated various legalisation

proposals, from commercial models to the grow-at-home model, the latter being implemented in 2018. Other state authorities

have commissioned expert reports assessing the impacts of legalisation (e.g. California) (Caulkins et al., 2015; Newsom et al.,

2015).

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

A description and typology of different policy approaches

Jurisdictions considering alternatives to prohibition on cannabis supply are faced with a wide range of

options from which to choose. Figure 3.1 presents some alternatives to the status quo prohibition of

cannabis supply, including eight ‘middle-ground’ options between traditional prohibition and the

standard for-profit commercial model, which is being adopted in an increasing number of US states.

However, as will be made clear in this section, the for-profit option is not the only option that is

pursued in the US, and both Uruguay and Canada have created regulatory regimes intended to limit

the power of businesses.

It is also the case that these options are not mutually exclusive. Jurisdictions can allow both home

production and commercial sales, a scenario that has occurred in most places that have legalised

cannabis. Furthermore, the choices need not be the same at the various levels of cannabis markets

(Wilkins, 2016). For example, it is possible to allow commercial businesses to produce cannabis but

require that its distribution occurs via state-run retail stores, as is the case with alcohol in some

jurisdictions, or via postal delivery, as is done in some parts of Canada.

FIGURE 3.1

Some alternatives to status quo cannabis supply prohibition

Source: Caulkins et al., 2015

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

3.1

US approaches

More than 25 % of the US population lives in 10 states where voters have passed laws permitting the

commercialisation of cannabis production and distribution. Table 3.1 reports dates of passage for both

medical and recreational legalisation as well as when retail dispensaries began selling recreational

cannabis to adults.

While the commercial models in Colorado, Washington and California receive a great deal of

attention, not all US states that have legalised cannabis have allowed commercial sales. Both DC and

Vermont have opted for non-commercial access, allowing adults to privately grow and give away up to

an ounce (28 g) of cannabis. In 2014, DC voters approved Initiative 71, which permits adults to grow

up to six plants and to give away cannabis without remuneration (

). The law maintains a prohibition

on selling cannabis and made public consumption a minor offence. The initiative has not been

formally evaluated, but early media reports suggest that the new law has spurred unregulated

transactions under the guise of giving cannabis to someone as a gift (Garcia, 2017).

In addition to DC, the state of Vermont legalised recreational cannabis in January 2018, which came

into effect on 1 July of the same year. After more than two years of legislative deliberation and a

government-sponsored report on the options and issues regarding legalisation, Vermont became the

first state to repeal cannabis prohibition through legislation, rather than an initiative approved by

voters. Similar to DC, Vermont’s law permits adults aged over 21 years to grow six plants (up to two in

flower) and to possess or give away up to an ounce of herbal cannabis. It does not permit sales. As in

the case of DC, ‘grey market’ sales of cannabis, though technically illegal, may be occurring in

Vermont’s unregulated non-commercial system (Goldstein, 2018).

TABLE 3.1

Passage of recreational and medical cannabis laws in US states that permit these

State

Alaska

California

Colorado

Illinois

Maine

Massachusetts

Michigan

Nevada

Oregon

Vermont

Washington

Medical cannabis law

passed*

1999

1996

2000

2010

2013

2002

2013

2008

2000

1998

2004

1998

Recreational cannabis

law passed

4 November 2014

8 November 2016

6 November 2012

4 November 2014

25 June 25 2019

8 November 2016

6 November 2018

8 November, 2016

4 November 2014

22 January 2018

6 November 2012

Start of recreational

sales

29 October 2016

1 January 2018

1 January 2014

Not applicable

Not yet open

20 November 2018

2 December, 2019

1 July 2017

1 October 2015

Not applicable

8 July 2014

*Medical cannabis laws have varied over the years. Here, we report the first year in which the state adopted a

medical cannabis law, not necessarily when voters approved the initiative. Initially, these laws permitted

qualifying individuals to possess cannabis. Over time, dispensaries and collective cultivation sites sprung up to

supply growing patient demand. Since their inception, states have adopted various regulatory approaches,

formalising commercial medical markets.

Source: National Conference of State Legislatures (2019)

( ) Many but not all jurisdictions in the US allow citizens to propose initiatives for consideration by voters provided they obtain

enough signatures.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

3.2

Uruguay

In 2013, Uruguay became the first country in the modern era to legalise recreational cannabis use.

Under the law, resident adults aged over 18 years can register with the state regulatory authority to

access cannabis through one of three available supply channels. Registrants can switch from one

supply channel to another, after changing their status with the regulatory authorities. Regardless of

which supply stream users opt into, they are technically limited to 480 g per year. The three types of

access are home cultivation, membership at a cannabis social club or retail sale via licensed

pharmacies.

Individuals who decide to grow cannabis at home cannot cultivate more than six plants. As of May

2019, there were some 7 000 registered home growers. Clubs are limited to 45 members and can

cultivate only up to 99 plants. Currently, there are 119 clubs in operation across the country, servicing

about 3 600 members. Registered members are prohibited from obtaining more than their allowance

of 480 g. Those opting to purchase at pharmacies must also register with authorities and are limited to

purchasing no more than 10 g per week, with an annual cap of 480 g. There are just a few varieties of

the plant on offer at pharmacies (i.e. regulators currently limit retail pharmacy sales to only a few

strains), and authorities have aimed to keep potency below 10 % THC. According to regulators, there

are currently 17 licensed pharmacies, mostly in the capital region, that sell cannabis to about 36 000

registered adults (

). There are no taxes on retail sales, but the government has set a price of UYU 40

(about EUR 1) per gram.

3.3

Canada

As a federal system, the new recreational cannabis law removes most prohibitions on cannabis from

national drug control statutes (criminal prohibition on supply to or possession by minors or

unregulated production remains). In most provinces, adults aged over 18 years can possess up to

30 g of herbal cannabis (or its equivalent in weight in other forms, such as concentrates (

)). The law

allows the cultivation, distribution and possession of cannabis by government-authorised entities.

Federal law provides standards and a general framework for regulating cannabis, including

requirements for an inventory tracking system, fees, quality and testing rules, restrictions on labelling

and packaging, prohibitions on promotions aimed at enticing young people and bans on the types of

products for sale. Federally licensed producers can be publicly traded, for-profit businesses, and

there are no restrictions on alcohol and tobacco companies receiving a production licence.

Though the federal government has lifted prohibition, provinces have considerable power in designing

rules governing distribution, taxation, and establishing age limits for purchase (some provinces opted

to increase the minimum age to 19 years). Provinces are also tasked with enforcing their own

regulations. Though federal law permits adults to cultivate up to four plants at home, some provinces

have banned home cultivation. Provincial retail systems are still evolving, with some opting for public

distribution channels similar to government-run alcohol stores.

What distinguishes Canada from the US states that have legalised cannabis is that the federal

government led the effort to repeal prohibition, with the aim of reducing youth access, protecting

public health and improving public safety by shrinking the illicit market. Unlike states in the US,

Canadian provinces do not maintain their own competing criminal drug laws. Another important

variation found in Canada’s law is that it allows for mail order supply of cannabis, minimising the need

for bricks and mortar stores. The federal government is also committed to public health and education

campaigns aimed at reducing problematic use, including underage use.

( ) See

https://www.ircca.gub.uy/farmaciasadheridas/

( ) Under Canadian law, 0.25 g of concentrates is equivalent to 1 g of dried herbal cannabis, and 15 g of solids containing

cannabis is equivalent to 1 g of dried herbal cannabis.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

An assessment of early evidence of the impact of policy

changes in the United States

This section summarises the early evidence of the consequences of policy changes concerning

recreational cannabis laws in the US. Here, we focus on the commercial cases that are specific to the

US, given that these policy changes have unfolded over longer time periods, allowing for more

evaluations of their impacts. They are also likely to have larger social impacts than non-commercial

reforms such as home cultivation. Recreational cannabis laws are frequently referred to as

‘recreational marijuana laws’ and ‘medical marijuana laws’, as in US statutes and regulations

cannabis is more commonly referred to as ‘marijuana’. We refer to these as REC (recreational) or

MED (medical) cannabis. Colorado and Washington were the first states in the US to pass REC laws,

followed by Oregon and Alaska. Though other states, such as Maine and Illinois, have passed similar

laws, they have yet to fully implement market regulations, or too little time has passed since they have

been in place to allow the evaluation of the impacts. This analysis largely focuses on the pioneering

states of Colorado and Washington, where more time has passed and more data have been

collected.

Though REC laws have existed only in the last seven years, their adoption follows the passage of

MED laws and, in some cases, robust commercial markets (see discussion in Section 2). The

variation in MED laws and associated outcomes has been documented in scientific literature (Cerdá

et al., 2012; Hasin et al., 2015; Pacula et al., 2015; Hall, 2018); as a result, we do not assess MED

laws or their policy impacts here.

The consequences of the REC legal changes analysed here are likely to vary depending on the

design and implementation of pre-existing medical cannabis laws and the size and scope of state

commercial markets. For example, jurisdictions with commercial medical markets, lax restrictions on

patient access and minimal regulatory burdens may not observe as many post-REC changes as

states that restricted medical access to a narrow set of qualifying conditions or limited the number of

dispensaries (

). In some cases, the adoption of REC laws is a marginal change (Colorado and

California had large, loosely regulated medical markets for more than a decade before the adoption of

REC laws, and the regulation of medical markets in Washington state did not occur until after the

passage of REC laws); for other states, recreational laws are more novel (Massachusetts and Maine

have had smaller or limited commercial medical markets for shorter time periods).

An assessment of the adoption of REC laws (and MED laws before them) must consider the

underlying economic and social trends. Legalisation — but more importantly commercialisation — is

likely to affect social norms by promoting use, normalising behaviours, increasing availability,

introducing new products and lowering price (Caulkins et al., 2015; Hall and Lynskey, 2016; Subritzky

et al., 2016; D’Amico et al., 2018). The extent to which any or all of these consequences resulted in a

net benefit or net cost to society is unclear. Nevertheless, many consequences may depend on the

timing of REC legalisation and its implementation and the shape in which both unfolded. Therefore,

examining changes in commercial access (e.g. when stores opened) is perhaps a more accurate

exposure variable than the date when a law was adopted. In addition to timing, an ideal measure for

REC exposure should account for retail store density, something most of the literature has omitted in

its analysis.

( ) As the US federal government maintains the scheduling of cannabis as a Schedule I drug and because pharmacies require

a federal licence to distribute controlled substances for medical purposes, medical cannabis products cannot be distributed

through traditional pharmacies, even in states that permit cannabis for medicinal use. This is, in part, how and why the

cannabis dispensary system started in the US.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

4.1

Literature review selection criteria

We reiterate that the contemporary literature base on REC law impacts is thin, given that policies are

still developing, markets are still maturing, and there are significant data gaps. Furthermore, states

are adopting these legal changes against a backdrop of federal prohibition. Therefore, the full extent

of these effects may change considerably in an environment in which the substance is legal at the

national level (e.g. in Canada or Uruguay). Interstate trade and federal tax policy could have broad

implications on the outcomes assessed here. Looking to the future, the immediate outcomes

evaluated in this review may not reflect long-term outcomes, which may take decades or longer to

develop. We therefore urge caution when interpreting the short-term measures and findings reported

below.

That said, we assessed the emerging evidence on the outcomes of REC law changes. We excluded

studies specific to MED, though the context of REC adoption and design is relative to a state’s

medical framework. These underlying legal and social factors related to MED may extend to

differences in the outcomes reported post REC adoption. We also note that many studies specify

various policy interventions, sometimes unclearly. Several peer-reviewed articles focus on REC

enactment (after a law was passed to effect a change; in most cases, the prohibition on possession or

home cultivation was lifted before implementation) or REC implementation (once the law is in effect

and retail stores are open). Some ignore the distinction entirely, blending the two, rather than

evaluating impacts in policy change after enactment or implementation.

We prioritise peer-reviewed articles over the grey literature or reports published by advocacy groups,

but we have included a few well-designed working papers. In terms of the research design of articles,

we highlight findings from quasi-experimental studies that include a control group or employ an

interrupted time series method to assess policy impact (

). Studies that use state representative

samples are prioritised over studies that have a non-representative or convenience sample. We also

avoided including studies that evaluate measures at only one point in time or fail to include measures

from pre- and post-implementation periods. Study design and methodology were an important

component of this analysis. Our initial scan of the literature yielded many methodologically weaker

studies that assessed outcomes of interest but were excluded for further review because they did not

meet our methodological selection criteria. We have included all studies captured in our search in

Appendix B, but we only assess outcomes of more rigorous studies in the body of this review.

Our search strategy focused on databases of peer-reviewed articles and working papers. We

searched Google Scholar and PubMed for peer-reviewed articles in English and the National Bureau

of Economic Research for working papers. We identified papers written and/or published up until July

2018, focusing on REC legalisation in the Americas; in addition six more recent peer-reviewed

publications with credible control groups were added to the review during the publication process to

ensure the findings were as up-to-date as possible (Kerr, D. et al., 2018; Anderson et al., 2019;

Aydelotte et al., 2019; Cerdá et al., 2019; Everson et al., 2019; Lane and Hall, 2019) and insights from

a new review (Smart and Pacula, 2019). Given that REC is a recent phenomenon, we concentrated

our screening on articles published after 2012. The following search terms were used:

recreational;

marijuana

cannabis;

and

legalisation.

We screened articles, focusing on evaluations relevant to

REC and those that utilised methods that allowed for making inferences (e.g. pre-/post-analyses,

interrupted time series, differences-in-differences).

4.2

Prevalence of use

Some analysts hypothesise that cannabis prevalence rates will change after recreational

commercialisation (e.g. Caulkins et al., 2015; Hall and Lynskey, 2016). There are many reasons for

( ) The one exception was a series of studies looking at adverse events pre/post REC legalisation in Colorado. We opted to

include these studies, as they provide unique assessments of health outcomes relative to MED and REC legislation adoption in

that state.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

this, such as changing norms regarding cannabis use behaviours, product innovation, reductions in

price and advertising (if allowed). There is greater concern regarding potential increases in youth

prevalence, given the impacts of cannabis on the developing brain and the finding that early initiation

is associated with poorer outcomes later in life (Fergusson and Boden, 2008). However, the

consequences of REC legalisation will probably depend on how policy is implemented and the degree

to which underlying attitudes and trends may be associated with recent changes in a state’s medical

cannabis policy. In turn, both medical and recreational policy changes may be a latent factor of

underlying social attitudes regarding cannabis. Nevertheless, changes in supply, especially as it

relates to

access

to cannabis through retail establishments and

price,

may have a greater impact on

prevalence than a mere change in the law, which is why several have cautioned that the longer term

effects of liberalisation policies may not be consistent with short-term effects (Caulkins et al., 2015;

Cerdá et al., 2017).

Prevalence estimates are some of the most referenced statistics about the consequences of REC

legalisation. The prevalence of cannabis consumption measures the proportion of a particular

population that self-report using cannabis within a certain period, typically in the last month, in the last

year or during their lifetime. The federal and state governments regularly estimate prevalence rates

for different age cohorts using household or school surveys. The National Survey on Drug Use and

Health (NSDUH) and the Monitoring the Future (MTF) surveys are two well-known, nationally

representative surveys that regularly report cannabis prevalence, frequency of use and perception of

risk estimates by age groups. The NSDUH also reports pooled 2-year, state-level prevalence

estimates by age group. In addition to these surveys, Colorado and Washington (

) have conducted

their own state-wide prevalence surveys of high school students (Healthy Kids Colorado Survey —

HKCS; and Washington state’s Healthy Youth Survey — HYS). Our scan of the literature also

identified studies that evaluated changes in other repeated cross-sectional surveys, such as the

National College Health Assessment survey or the National Alcohol Survey.

Figure 4.1 plots the past-month cannabis prevalence rates for the five states that passed REC laws

before 2016 and for the entire country (

). Table 4.1 is included as a reference, indicating when

MED/REC laws were adopted and when stores opened (except DC, which remained ‘grow and give’

only). Almost all states report increases in prevalence rates among the general population post REC

legalisation, though prior trends suggest an increase in states’ predated REC laws. Of course, simply

examining trend data does not allow for strong causal inferences, and there is an emerging quasi-

experimental literature attempting to control contextual factors and isolate these policy effects.

TABLE 4.1

Dates of MED/REC adoption and implementation for selected states

Medical cannabis

State

Alaska

Colorado

Oregon

Washington

Adoption

November 1998

November 2000

November 1998

Year stores legally

opened

Not applicable

2005

2010

2009

Not applicable

Recreational cannabis

Adoption

November 2014

November 2012

November 2014

November 2012

Implementation

October 2016

January 2014

February 2015

October 2015

July 2014

( ) We focus on these two states, given that there are more available post-treatment data, because recreational markets have

been operating for over four years.

( ) Additional trend data, including plots for individual states, are included in Appendix C.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

FIGURE 4.1

Prevalence of cannabis use in the past month (%) among people aged 12 years and older

2009

2010

Alaska

2011

Colorado

2012

2013

2014

Oregon

2015

Washington

2016

2017

Source: NSDUH

So far, the peer-reviewed literature reports mixed findings regarding changes in self-reported

prevalence after the adoption of REC laws. Studies in this nascent literature focused on adults are

more likely to find an increase than those examining younger populations; indeed, there are some

studies suggesting that youth prevalence may have declined. Table 4.2 compiles and summarises

these findings described below.

Youth

In terms of changes among students, Cerdá et al. (2017) reported significant declines in the

perception of harm and significant increases in past-month use of cannabis among 8th and 10th

graders in Washington compared with states that did not adopt REC laws. The study employed a

difference-in-difference approach to compare past-month prevalence rates between 2010-12 and

2013-15 in Colorado and Washington against rates in the rest of the US. The authors found that, in

Washington, perception of harm reduced by 9 % for both grades (p < 0.02) and that past-month use

increased by 5.0 (p = 0.03) and 3.2 (p = 0.007) percentage points for 8th and 10th graders,

respectively (Cerdá et al., 2017). However, no significant difference in perception or use was reported

for 12th graders in Washington or for any of the grades in Colorado.

The study treated REC enactment and implementation as one policy change. Though the law

changed after voters approved the initiatives, stores were not authorised to sell cannabis products to

customers until after the implementation. The different trends in prevalence may also be an artefact of

the different nature of the medical cannabis policy environments in Washington and Colorado, the

latter having a robust commercial market and allowing un-registered caregivers to cultivate a large

number of plants for registered patients. This finding may also be an artefact of the data used. Using

survey data from a much larger sample of students in Washington (Washington’s Healthy Youth

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Survey; HYS), Dilley et al. (2019) found that, if anything, there was a decrease in the past-month

prevalence among 10th-grade students in Washington after REC enactment, according to state-wide

surveys. Whether these results hold up when specific subgroups of students are examined remains to

be seen.

Using the individual-level NSDUH data with state identifiers from 2008-2016, Cerdá et al. (2019)

found no evidence that REC enactment influenced the past-month cannabis prevalence for those

aged 12-17. Although it should be noted that only three states (Colorado, Washington and Oregon)

had implemented REC markets with open stores during the period of evaluation, and those had only

been open for a relatively short period of time during their window of evaluation. Anderson et al.

(2019) examined a longer time-frame (1993-2017) using the national and state Youth Risk Behavior

Survey (YRBS) and reported there was no association between MED laws and past-month

prevalence among students; however, they did find evidence suggesting prevalence may have

declined for students after REC laws were enacted (

). We expect additional analyses will be

conducted with these datasets that exploit additional changes in implementation of the state laws,

information about per capita outlets and retail sales, and other aspects of the regulatory

environments. The results thus far suggest that the mere passage of the laws does not affect youth.

More research is needed to understand whether youth consumption is influenced by features of more

mature cannabis markets, which are characterised by lower prices, product proliferation, and higher

outlet density.

College students

In two studies of college students, D. Kerr et al. (2017, 2018) investigated changes in cannabis

prevalence rates after REC enactment in Oregon. In their first study, authors collected responses from

online surveys of students (aged 18 to 26) across seven universities, only one of which was in

Oregon, between 2012-2016. Using these repeated cross-sectional surveys, authors found that

cannabis use rates had increased across six of the seven universities over the time period, but the

increase in use rates was significantly greater for students in Oregon that also reported recent heavy

alcohol use (Kerr, D. et al., 2017).

In another study using a different repeated cross-sectional survey of university students (ages 18-26)

from 2008 to 2016, authors compared changes in self-reported cannabis use rates from two

universities in Oregon and 123 universities and colleges from non-REC states. Authors found that,

compared to the students in the other states, Oregon students showed a relative increase in rates of

past-month cannabis use after REC law passage (odds ratio = 1.29 [1.13, 1.48]) (Kerr, D. et al.,

2018).

Adults

W. C. Kerr et al. (2018) analysed self-reported past-year cannabis use in Washington with a repeated

cross-sectional random sample of adults aged 18 years and over from before the state implemented

REC legislation but after it enacted it (wave 1: January to April 2014) and post implementation

(wave 2: March to May 2015; wave 3: August to October 2015). Participants were also asked to recall

their past-year cannabis use for the 12 months prior to REC enactment, which started in December

2012. The authors found that respondents reported a small and non-significant increase in past-year

use (from 24.3 % to 25.6 %) from the responses relating to the pre-enactment period to the combined

average of all three post-enactment/implementation waves (Kerr, W.C. et al., 2018). They go on to

state that there was no statistically significant change in the prevalence of the simultaneous use of

alcohol. However, the study relies on participants’ recall of past-year use, prior to REC law enactment,

which may not be accurate and is likely to be influenced by the state’s medical cannabis market.

( ) However, critics have raised concerns about merging national and state-representative samples from the YRBS for

conducting analyses of the effects of any state policy, including medical and recreational cannabis laws. Doing so can lead to

distortions of the sample in some states, which may affect the results obtained using these data (Jones et al., 2019; Rapoport

et al., 2019).

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Furthermore, the study does not clearly distinguish between REC enactment and implementation. The

null effect reported may be due to the study’s design, such as relying on respondents’ memory.

Without a suitable control group for comparison, it is hard to draw strong inferences.

One cohort study analysed self-reported past-year use data from the repeated cross-sectional

National Alcohol Survey of adults across the country between 1984 and 2015. The authors used an

age-period-cohort design to assess changes in respondents’ past-year use across states that passed

REC or MED laws or allowed home cultivation or operational dispensaries. The authors found that the

underlying period effects were the main factor contributing to increases in self-reported cannabis use;

the passage of REC laws, or laws allowing home cultivation of cannabis and the opening of

dispensaries were not significantly associated with increases in prevalence (Kerr, W. C. et al., 2017).

They noted that, though REC/MED policies were not associated with changes in prevalence, survey

data show a decline in use during the 1980s and 1990s but a sharp increase from 2005 to 2015,

varying by demographic and age cohort. The authors attributed recent increases in self-reported

prevalence to general period effects that influence the greater population, rather than specific REC

policies. Nevertheless, the authors noted that their design does not allow for an analysis of state-

specific trends. Furthermore, as their analysis covers a period of more than 30 years, it may be

confounded by changes in social attitudes towards cannabis use. They noted that self-reported use

may be affected strongly by legal and social norms, which might bias responses.

The previously mentioned NSDUH analysis by Cerdá et al. (2019) looking at cannabis use among

youth also examined whether enactment of REC laws influenced the prevalence rates for adults,

broken down into two age groups: 18-25 and 26+. Similar to the findings for those under 18, they

found no effect on the prevalence rates for cannabis use among those 18-25 after REC enactment;

however, the results were different for the oldest group. The authors found that REC enactment was

associated with an increase in the prevalence of past-month cannabis use among those aged 26 and

over (odds ratio [OR], 1.28; 95 % CI, 1.16-1.40).

While the significant within-state variation in how legalisation is being implemented can complicate

state-year panel analyses, it does provide opportunities for learning more about how store density and

other factors may influence cannabis prevalence and other outcomes. Merging information about

recreational store addresses with location information for adults participating in Washington’s

Behavioral Risk Factor Surveillance System from 2009-2016, Everson et al. (2019) found that past-

month cannabis use ‘increased among adults living in areas within 18 miles of a retailer and,

especially, within 0.8 miles (odds ratio [OR] = 1.45; 95 % confidence interval [CI] = 1.24, 1.69).’ Of

course, it would make sense that some stores would try to locate in places with more cannabis users,

and the authors attempted to account for this by estimating generalised mixed models with a random

intercept by community effect and a random time by community effect as well as controlling for a host

of community-level (and individual-level) variables. The authors acknowledge that an important

limitation is that the analysis does not account for the number of unlicensed medical cannabis outlets

or ‘community gardens’ operating in the state before and after the REC law was passed; however,

they argue this may have biased their results toward zero (

Finally, one study examined self-reported past-month rates of use in two cohorts of pregnant women

and young mothers in an alcohol and drug case management intervention programme before and

after REC law enactment and implementation in Washington (comparing wave 1, February 2001 to

November 2012, with wave 2, December 2012 to July 2015). The authors compared self-reported

past-month cannabis use in samples of participants from nine populous counties in the state and

found that women who completed the programme post REC law enactment (in or after December

( ) Everson et al. note: ‘Because we did not account for alternate sources, we overestimated the net proportional increase in

cannabis access attributable to the retail market (although not for the law-abiding general public). Our findings may thus

underestimate the potential effect of introducing retail access as compared with locations with smaller or less easily accessible

medical and illicit markets.’

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

2012) were significantly more likely to report cannabis use on exit (OR = 2.1,

< 0.0001) (Grant et al.,

2017). The authors went on to note that women who used cannabis during pregnancy in the post-

REC group were significantly less likely to quit than their counterparts in the pre-REC group (33.1 % v

45.4 %,

< 0.01). However, the sample of this cohort study limits interpretability to at-risk young

mothers and pregnant women attending a drug and alcohol outreach programme and does not

include the larger population of young mothers. Furthermore, researchers ignored any different policy

impacts between REC enactment and implementation. The reliance on self-reported data may

introduce bias in a population of individuals referred to a drug and alcohol abuse intervention

programme. The pre-REC population may have under-reported their cannabis use to avoid stigma

and legal sanctions.

The mixed results reported in the literature may be, in part, due to the variation in study design

methods, sub-population selection and which cannabis policy change researchers are measuring over

which time period. As noted, several studies seem to conflate REC law adoption, enactment and

implementation. Even when voters approve a REC initiative, it can be one or two years before any

cannabis is sold in stores. Along with lags in survey data, researchers, eager to evaluate the impacts

of REC, may be incorrectly specifying when a change in policy occurred, particularly given the existing

medical cannabis policies that are in place in all of these states that had legalised REC cannabis.

Furthermore, the time periods analysed may yield different results. The inclusion of long pre-REC

legalisation periods, which pick up the effects of MED cannabis provision, may return different results

from studies over shorter pre-REC periods. Finally, studies that combine young people and adults, or

at-risk populations and general household populations, may fail to pick up important heterogeneous

effects that exist across these populations, as shown in studies examining the effects of medical

cannabis policies (Chu, 2014; Pacula et al., 2015; Wen et al., 2015; Smart, 2016).

In Table 4.2 we summarise the main findings of the four articles examining changes in prevalence

rates post REC introduction that were discussed above. The remainder of the prevalence studies did

not meet our inclusion criteria because of their research design, which lacked a suitable control, or

because they were taken from convenience samples, they were cross-sectional or point-in-time

measures or they failed to compare pre-REC outcomes with post-REC outcomes.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

TABLE 4.2

Summary table of peer-reviewed studies examining the effect of REC laws on cannabis

prevalence

Study

Anderson et

al. (2019)

Cerdá et al.

(2017)

Outcome evaluated

Prevalence rates in high school

students

Perception of harm and youth

prevalence

Policy change

REC/MED enactment

across states

REC enactment and

implementation in

Colorado and Washington

Changes associated with policy

No change in reported use after MED,

possible decline after REC enactment

Reductions in perception of harm and

increases in reported use for 8th and 10th

graders in Washington

No changes in perception or prevalence

among 12th graders in Washington or

young people in Colorado

Cerdá et al.

(2019)

Dilley et al.

(2019)

Everson et

al. (2019)

Grant et al.

(2017)

Prevalence rates and self-reported

cannabis use disorder rates for youth,

young adults, and adults

Prevalence rates for secondary school

students

Prevalence rates for adults

REC enactment in

Colorado, Washington,

Alaska and Oregon

REC enactment in

Washington

REC enactment and

implementation in

Washington

REC enactment and

implementation in

Washington

REC enactment in

Oregon

No changes in prevalence rates for those

ages 18-25. Increase in respondents 26

years or older

Decrease in prevalence rates for 8th and

10th grades

Increase in current use of cannabis for

those that reported proximity to a retail

outlet

Significant increases in self-reported use

Changes in past-month rates of use in

pregnant women and young mothers

in an alcohol and drug case

management programme

Prevalence rates in college students

Kerr, D. et

al. (2017)

No changes in prevalence rates for

students in Oregon. An increase in

cannabis prevalence rates for those also

reporting recent alcohol binge drinking

Increase in cannabis prevalence rates for

students in Oregon

Increases in prevalence are attributed to

underlying age-period-cohort factors and

not a change in policy

No significant change

Kerr, D. et

al. (2018)

Kerr, W. C.

et al. (2017)

Kerr, W. C.

et al. (2018)

Prevalence rates in college students

Prevalence of adults

REC enactment in

Oregon

REC and MED enactment

and implementation

REC implementation in

Washington

Prevalence of adults aged 18 years

and over

4.3

Consumption patterns

The implementation of REC laws may affect cannabis consumption patterns in two important ways:

mode of administration and intensity of use. Commercial REC provision in the US allows promotion,

product innovation and price competition. Each of these factors is expected to shape the way

individuals consume the drug — how, where, when and how much. Product innovation allows the

substance to be consumed in edibles or vaporised through the use of an electronic cigarette.

Improved cultivation methods are shaping the potency of cannabis sold in REC markets, resulting in

consumption of higher THC levels and with variations in cannabinoid profiles. The rate and amount of

THC that regular users consume is likely to be shaped by changes in price and availability made

possible by commercial REC markets.

Nationally, the average number of days of use in the past month reported by past-month users

increased from 12.4 to 14.5 between 2003 and 2017 (SAMHSA, 2018). Figure 4.2 displays the

percentage of past-month users who reported daily or near-daily use (i.e. > 20 use days in the past

month) by state and for the entire country. For the entire country, the rate increased from 33.4 % to

41.7 % between 2003 and 2017. State estimates of daily or near-daily use have not been tabulated

beyond 2014, but a similar increasing trend is reported for most states. In 2014, approximately half of

past-month users in Colorado and Oregon reported using cannabis on 20 or more days, up from 38 %

and 37 %, respectively, in 2009.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

FIGURE 4.2

Past-month cannabis users aged 12 years and over using on a daily or near-daily basis (%)

2003

2004

2005

Alaska

2006

2007

Colorado

2008

2009

2010

2011

2012

2013

2014

2015

2016

Oregon

Washington

Source: NSDUH

The previously mentioned study by Everson et al. (2019) which exploited variation in recreational

store density in Washington state not only found that those who lived within 0.8 miles (1.3 km) of a

store were more likely to report use in the past month, they also found that these individuals were

much more likely to report daily or near-daily cannabis use.

In addition to using individual-level NSDUH data to examine the association of REC enactment with

prevalence of cannabis use, Cerdá et al. (2019) also examined the association with frequency of past-

month use and past-year cannabis-use disorder (CUD). With respect to frequency of use, they only

found a statistically significant increase associated with those aged 26 and older. As for CUD, there

appears to be a small and statistically significant increase for youth, but the authors cautioned that

they could not rule out that unmeasured time-varying confounders may explain this finding (

). But for

those aged 26 and over, they report a statistically significant increase in past-year CUD and that there

was less concern that unobserved confounding may explain the association.

4.4

Product differentiation and price

As mentioned in the previous section, commercial REC has brought with it product innovation. Such

product differentiation and promotion, in turn, shape user behaviour and consumption. Cannabis

flower is still the dominant form of cannabis consumed in REC markets, but trends in data collected as

part of the seed-to-sale traceability systems in REC states suggest that new products that allow for

various modes of administration (vaporising, edibles and topicals) account for an increasing

proportion of sales. In one analysis of market trends in Washington, researchers analysed all

recorded retail sales (36 million observations) for a little more than two years (July 2014 to September

2016) (Smart et al., 2017). Researchers found that there were important changes in product type,

potency and price. Figure 4.3 is an updated reproduction from Smart et al. (2017), showing the

( ) Specifically, Cerdá et al. (2019) note: ‘E-value analyses suggested that unmeasured time-varying confounders

hypothetically more prevalent in [REC]states that increase the risk of cannabis use slightly (1.08-1.11 times) may explain this

finding. The extent to which such confounders exist is unclear because our difference-in-difference design accounted for

unmeasured time-invariant sources of confounding and also adjusted for measured time-varying individual- and state-level

demographic characteristics. However, the small E-values warrant a conservative interpretation of the increase in CUD among

participants aged 12 to 17 years.’

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

FIGURE 4.3

Washington’s cannabis sales by product type after 37 % excise tax (million USD)

Source: RAND analysis updating Smart et al. (2017)

increasing trend in sales of edibles and vaporisable products, which grew to roughly one third of total

sales expenditures by mid-2017.

In terms of potency, an analysis of transaction data (included in the article but not shown here) shows

that the reported potency of cannabis flower, as measured by state-accredited laboratories, has

increased. The proportion of strains with reported THC concentration greater than 15 % grew to 93 %

by September 2016; cannabis flower of less than 10 % THC potency accounted for less than 2 % of

expenditures. The authors noted that cannabis flower with more than 20 % reported THC has

increased by almost half since October 2014, accounting for 57 % of the retail expenditures (Smart et

al., 2017) (

). As for price, it is hard to know what consumers would have faced in the absence of

REC legalisation but, as predicted, prices are falling. Updated analyses of Smart et al. (2017) and by

Davenport (unpublished) show that the average post-tax price for 1 g of cannabis flower fell from

roughly USD 20 in October 2014 to USD 10 in October 2015 and USD 7 in October 2017, and it

continues to decline (

A smaller study of MED and REC cannabis users from Washington and Colorado (n = 317) and a

web-scraping analysis of prices listed on Weedmaps (n = 3 802) compared user-reported and

dispensary-advertised prices of cannabis in both states during REC enactment periods with prices

after implementation (wave 1, October 2013; wave 2, May 2014; wave 3, October 2014). The authors

found that there was little price variation in the initial months post REC implementation but that

individuals reported paying higher prices for in-store purchases than for social purchases (Hunt and

Pacula, 2017). However, the authors noted that a limitation of the study is the very short time period

post REC implementation (within five months of stores opening). As noted above, other researchers

(Hansen et al., 2017a; Smart et al., 2017; Orens et al., 2018) have shown that prices do decline over

( ) We note that the potency reported by retailers may not be accurate. For example, initial studies of results from laboratories

in Washington show that there are clear systematic differences in potency results across laboratories but that general potency

results are inflated (Jikomes and Zoorob, 2018). State regulators have made efforts to address reporting biases.

( ) Price drops have also been documented in Colorado. A new study by Orens et al. (2018) reports that, from 2014 to 2017,

‘the price of one gram of adult use flower exhibited a steady downward trend, decreasing 62.0 percent, from $14.05 to $5.34

per gram’ (pre-tax).

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

long periods of time, providing a strong indication that the immediate effects of these policies may be

different from outcomes in the longer term.

Our literature review did find two studies examining the promotion of REC cannabis products, but

these were excluded from further analysis, as they did not evaluate trends before and after REC

implementation.

4.5

Treatment admissions

It is hypothesised that commercial REC legalisation will change the intensity and frequency with which

individuals consume cannabis. This includes consuming cannabis more frequently as well as

consuming products with a higher potency. These factors may lead to cannabis use disorder and thus

increase the number of treatment admissions. However, it may take years for a cannabis use disorder

to develop, suggesting that any immediate changes in admissions may reflect the impacts not of REC

legalisation, but perhaps of MED use. At the same time, cannabis possession and use are no longer

criminal offences once REC laws have been adopted. Therefore, we would expect reductions in

criminal justice referrals to treatment for cannabis after the adoption of REC. A more general point is

that treatment admissions data are also influenced by the availability and perceived attractiveness of

services and reporting and diagnostic practices, all of which can vary over time. However, no studies

investigating changes in treatment admissions were identified during our review.

4.6

Adverse medical events

Under the assumption that REC legalisation increases consumption, a proportion of this type of

cannabis use may result in an adverse event (e.g. an overdose that results in an emergency

department (ED) visit or hospitalisation) (

). Accidental ingestions or overdoses, especially by young

children, may result from the supply of certain products, such as edibles that look like traditional

sweets. Regulators in Colorado and Washington promulgated emergency regulations to improve

labelling and warnings as well as educate the public about the potential effects of new products, some

of which are meant to look like benign consumables.

We identified eight articles that looked at changes in various adverse events (such as hospitalisations

or ED episodes) before and after (pre and post) MED/REC and their relationship with MED/REC

legislation. All articles examine outcomes in Colorado over various periods, some including MED

treatment periods. All were pre/post studies without any comparison groups, and may be subject to

self-reporting biases (e.g. individuals may be more likely to report cannabis use or seek medical

attention post REC). Therefore, caution should be exercised when interpreting their findings.

Significant differences were assessed using standard non-parametric hypothesis testing for all

studies. Two of the eight studies examined paediatric or child exposures to cannabis. All of them

examined hospital data assessing the impact of changes to Colorado’s MED and REC laws. These

studies report an increased frequency of adverse events post REC enactment or implementation, and

some have observed an increase prior to the adoption of REC, attributing this to lenient MED policies.

However, several studies compare different pre and post periods, which in some cases are not

immediately before or after the adoption of REC (comparing, for example, per capita rates of events in

2009 and in 2015).

Two studies compared paediatric cannabis exposures necessitating medical attention in Colorado

before and after MED or REC implementation. Both studies found statistically significant increases in

the frequency of such events following changes in legislation. Wang et al. (2016) report that the rate

of cannabis-related visits to the children’s hospital increased from 1.2 per 100 000 residents 2 years

prior to REC implementation (2012/2013) to 2.3 per 100 000 residents 2 years after (2014/2015)

( ) An ED visit is recorded when an individual presents to a hospital for emergency care but is not admitted for observation or

treatment. A hospitalisation is recorded when a patient is admitted for observation and treatment.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

(p = 0.02). The same study found a fivefold increase in annual poison control (

) paediatric cases for

cannabis exposures (from nine in 2009 to 47 in 2015) post REC implementation (

). This amounted

to an average annual increase in calls to Colorado poison control centres for cannabis of 34 %,

significantly more than the national average of 19 % (p < 0.001). In another retrospective study of

cannabis-related ED visits, Wang et al. (2018) found that cannabis-related ED visits increased from

1.8 per 1 000 in 2009 to 4.9 per 1 000 in 2015 (p

< 0.0001). Though rates increased, it is hard to

say what policy mechanism led to their increase. Cannabis policy in Colorado went through several

transformations between 2009 and 2015, including the adoption of several laws and voter initiatives

that formalised commercial markets.

Kim and Monte (2016) examined hospitalisation incidents coded for cannabis exposure between 2001

and 2014, collected by the Colorado Hospital Association, to compare before and after MED and REC

implementation. The authors found that cannabis-related hospitalisations doubled from 15 per

100 000 between 2001 and 2009 (the pre-MED commercialisation period, when dispensaries were

operating) to 28 per 100 000 between 2010 and 2013 (the post-MED commercialisation period)

(p < 0.001) (Kim and Monte, 2016). The authors noted that this statistically significant doubling

occurred after MED implementation. They then examined rates of cannabis-related ED visits between

pre- and post-REC implementation, reporting that the figure nearly doubled from 22 per 100 000

(2010-13) to 38 per 100 000 (January-June 2014) (p < 0.001).

Using a shorter time period, Kim et al. (2016) examined ED visits for cannabis from 2011 to 2014. The

authors reported no significant change from 2011 to 2012 in the rate of ED visits with ICD-9 codes of

cannabis use among out-of-state residents, which is possible given that MED is restricted to in-state

residents. However, from 2012 to 2014, the state-wide rate among out-of-state residents rose from 78

per 10 000 visits in 2012 to 112 per 10 000 visits in 2013 and 163 per 10 000 visits in 2014 (p < 0.001

for all comparisons). Among Colorado residents, the rate of ED visits possibly related to cannabis use

per 10 000 visits increased from 61 in 2011 to 101 in 2014 (p < 0.001) (Kim et al., 2016). By

comparing year-on-year changes, the authors sought to determine the association with changes in

REC enactment and implementation. ED data suggest that the rate of adverse events continued to

rise over the period, starting in 2011, perhaps because of the expansion of commercial medical

markets.

One study looked at annual rates of hospitalisations, ED events and calls to Colorado’s poison control

centres between 2000 and January to September 2015. Wang et al. (2017) reported that cannabis-

related hospitalisations increased from 274 to 593 per 100 000. Examining trends over time, Wang et

al. (2017) found a year-on-year increase in cannabis-related hospitalisations over a period starting in

2009 and ending in September 2015. The rates of ED visits also increased year on year, starting in

2011 (the first year for which data were available). The rates of ED visits significantly increased from

2012 to 2013 (358 v 443 per 100 000;

= 0.003) and from 2013 to 2014 (443 v 554 per 100 000;

= 0.0005). The prevalence of mental illness diagnoses in cannabis-related ED visits was five times

higher than the prevalence among those whose ED visits were not cannabis-related, and the

equivalent figures for hospitalisations showed a ninefold higher prevalence of mental illness

diagnoses among hospitalisations that were cannabis-related. Calls to poison control centres,

although flat between 2000 and 2009, increased significantly, from 44 to 93 in 2010 (p < 0.0001), and

again jumped from 123 to 221 (p < 0.0001) between 2014 and 2015 (Wang et al., 2017). The sharp

and sustained increase in adverse events began right around the time when MED commercial

dispensaries were formalised (around 2010) and continued throughout the adoption and

implementation of REC in 2014.

( ) Calls to poison control centres can be made by patients, the general public or medical professionals. In the US, there are

about 2 million exposure calls per year, approximately a quarter of which come from a healthcare facility.

( ) As a point of reference, Colorado Poison Control reported over 25 000 exposures for those under the age of 19 years

between July 2015 and June 2016.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

ED visits for cannabis-related mental health diagnoses also increased in Colorado post REC. Similar

to Wang et al. (2017), K. E. Hall et al. (2018) reported that, post REC implementation, the prevalence

of mental health diagnoses in cannabis-associated ED visits was five times greater than the

prevalence in ED visits not related to cannabis (prevalence ratio 5.35, 95 % confidence interval (CI)

5.27 to 5.43). The rate of ED visits in Colorado associated with both cannabis and mental health

significantly increased from 224.5 per 100 000 in 2012 to 268.4 per 100 000 in 2014 (p < 0.0001)

(Hall, K. E. et al., 2018).

In an analysis of ED events in Denver between 2012 and 2015, Sokoya et al. (2018) reported that the

number of maxillary and skull base fractures increased post REC implementation (2012-13 v 2014-15)

(p < 0.001 for both outcomes); however, counts of other facial/cranial fractures were unchanged. Bell

et al. (2015) examined cases of burns related to hydrocarbon exposure likely to be related to the

home production of butane hash oil extractions from January 2008 to August 2014, before and after

Colorado implemented its MED and REClaws. The authors reported no cases prior to MED

implementation, 19 cases between October 2009 and December 2013 and 12 cases in 2014, after

REC implementation (Bell et al., 2015). Although not reported, the average monthly rate increases

from zero in the pre-MED period to 0.37 during MED implementation and to 1.75 per month during

REC implementation. Unlicensed extractions using an ‘inherently hazardous substance’ were

subsequently criminalised in Colorado.

Though studies report increases in adverse events, it is possible that some increases could be due to

changes in reporting or measurement. For example, it is unclear if increases in ED visits and

hospitalisations are due to a greater willingness on the part of individuals to report use of cannabis

and/or if doctors are more aware of acute cannabis intoxication post REC and are now more likely to

screen for or confirm cannabis use using urinalysis.

4.7

Impaired driving

Changes in cannabis policy may affect impaired driving outcomes (such as accidents or citations for

driving under the influence) in multiple ways. While the bulk of the research suggests that driving

while under the influence of alcohol is more dangerous than driving under the influence of cannabis, it

also suggests that driving under the influence of cannabis is more dangerous than driving sober

(Caulkins et al., 2016). It is possible that impaired driving outcomes may change after the

implementation of REC, depending on whether cannabis becomes a substitute for or complements

alcohol consumption; however, there is mixed evidence as to the relationship. In our review, we

identified two studies that evaluated motor vehicle fatalities and one that examined reported insurance

claims.

One peer-reviewed study examined the impact of REC on motor vehicle crash fatalities in Washington

and Colorado and neighbouring states that did not pass REC but were substantially similar in terms of

traffic and roadway characteristics from 2009 to 2015. Researchers analysed changes in the annual

number of vehicle fatalities reported in the Fatality Analysis Reporting System (FARS), using a

standard difference-in-difference approach with random effects. They then compared four years of

pre-REC adoption (2009-12) outcomes with three years of post-REC enactment and implementation

outcomes (2013-15). The analysis found no statistically significant difference in fatal crashes between

REC and non-REC states (+0.2 fatalities/billion vehicle miles travelled; 95 % CI

−0.4

to +0.9)

(Aydelotte et al., 2017). However, the authors noted that their policy specification conflates REC

enactment and REC cannabis commercialisation, as they used the date the REC law was passed, not

the date when stores opened.

A National Bureau of Economic Research working paper (Hansen et al., 2018) employing synthetic

controls also assessed the impact of REC on motor vehicle crash fatalities in Washington and

Colorado and all states that did not adopt a REC law between 2000 and 2016. The authors found

that, between 2013 (the last year before REC implementation) and 2016, drivers testing positive for

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

THC (

) increased in Colorado and Washington by 92 % and 28 %, respectively. However, the

authors concluded the following: ‘We find the synthetic control groups saw similar changes in

marijuana-related, alcohol-related and overall traffic fatality rates despite not legalising recreational

marijuana’.

Two more recent studies find evidence of an increase in traffic fatalities after legalisation. In a follow-

up study, Aydelotte et al. (2019) added additional years from the FARS database, comparing monthly

crash data from 2007 to 2017 in Colorado and Washington, which passed REC laws in 2012, against

five other states with MED laws, and four that did not change any cannabis laws. The authors also

examined changes in rates after opening of stores in Washington and Colorado in 2014. Aydelotte

and colleagues reported that the fatal crash rate in Washington and Colorado increased by 1.2

crashes (p

0.087) per billion vehicle miles travelled after REC enactment, but that there was a

statistically significant and larger increase after opening of stores (an increase of 1.8 crashes per

billion vehicle miles;

= 0.02). The results are sensitive to stores opening, supporting evidence that

REC implementation, rather than mere passage, may correlate with outcomes.

In another quasi-experimental analysis, Lane and Hall (2019) examined monthly traffic fatalities

between 2009 and 2016 in Colorado, Washington and Oregon compared to nine neighbouring

jurisdictions that did not change cannabis laws. Authors specified the opening of dispensaries as the

treatment condition instead of REC passage or enactment. They found that store openings were

associated with an immediate increase in 1.08 traffic fatalities per million residents followed by a trend

reduction of 0.06 fatalities per month (both

< 0.001). There was a similar step-up increase and

declining trend in both treatment states (step: 0.90,

< 0.001; trend:

−0.05,

= 0.007) and

neighbouring jurisdictions (step: 1.15,

= 0.005; trend:

−0.06,

= 0.001). They concluded that:

The results suggest that legalizing the sale of cannabis for recreational use can lead to a

temporary increase in traffic fatalities in legalizing states that can spill over into neighbouring

jurisdictions.

In another widely cited report, published by the Highway Loss Data Institute, authors compared auto

insurance collision claim rates (not necessarily fatalities) in Colorado, Washington and Oregon with

those in neighbouring states (Nebraska, Utah, Wyoming, Montana, Idaho and Nevada) that did not

adopt REC, from January 2012 to October 2016. The analysis found that collision claim frequencies

increased significantly by a combined rate of 2.7 % after REC implementation (Highway Loss Data

Institute, 2017).

The mixed results reported in the earlier studies identified are likely to be due to the difference in

outcomes evaluated, the specification of the REC policy change or the analytic methods employed.

Aydelotte et al. (2017) and Hansen et al. (2018) examined motor vehicle crash fatalities but with

different exposure variables (the latter was REC implementation, the former REC enactment); the

report by the Highway Loss Data Institute (2017) examined insurance claim data and the relationship

with REC implementation. Neither of the motor vehicle fatality studies found a relationship between

the fatality rates and REC. However, the report examining insurance claims did find a significant and

positive relationship, but one that does not negate findings from other papers. It is possible that non-

fatal accidents increased over this period, even if fatal accidents did not.

In the more recent analyses reported here, Aydelotte et al. (2019) and Lane and Hall (2019) examine

the relationship with motor vehicle fatalities and the opening of REC dispensaries. Both studies found

a positive relationship between opening of stores and motor vehicle deaths. These findings suggest

that the appropriate policy change to consider in evaluations is opening of stores, when product

becomes commercially available, instead of mere changes in the law.

( ) However, testing positive for THC is not an appropriate measure for impairment, as THC metabolites may remain in the

system long after the subjective and objective effects of intoxication have abated (EMCDDA and CCSA, 2018).

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Another possible reason for the inconsistency is the difficulty in identifying impairment at the time of

an accident. Current methods reported in traffic fatality and insurance claims data are not sufficient to

identify, among drivers who may have consumed cannabis within a specified time period before the

accident, those who were actually impaired. Until such technology is developed and widely adopted in

current reporting transport databases, we may not be able to sufficiently identify effects from

observational data.

4.8

Consumption of other substances

The use of cannabis may substitute for or complement other psychoactive substances. There have

been studies examining the relationship between the use of cannabis and alcohol, tobacco or other

drugs; however, most of these studies have evaluated the relationship in users of illicit cannabis. REC

and MED legalisation may substantially change the relationship or cannabis users’ access to other

drugs. There are relatively few studies examining the effect of licit cannabis on the use of other

substances or associated behaviours. Several studies have examined the population-level outcomes

associated with access to medical cannabis. It is important to consider the changes that REC may

have on the use of other substances. Social norms and commercial promotion of cannabis may make

the use of this drug more appealing to individuals who would have otherwise consumed alcohol in a

social setting. Likewise, there is an ongoing discussion in the US on using cannabis as a substitute for

riskier opioid analgesics to treat chronic pain (Hall, W. et al., 2018).

Several studies that examined the relationship with cannabis and other drugs are discussed in the

table in the Appendix. Yet an extensive and recently published review of this literature by Smart and

Pacula (2019) found complex and inconclusive results depending on the substance evaluated and

measure used. They concluded:

Evidence of the impact of cannabis liberalization on the use of other substances is

inconclusive. We have limited evidence of how alcohol or tobacco use has been impacted,

and despite a broader literature evaluating the impact of cannabis laws on opioid-related

outcomes, the findings from this literature are puzzling. Studies assessing impacts on self-

reported misuse and distribution of opioids show no impact of [MED laws], yet studies

evaluating opioid-related adverse events and opioid prescribing show reductions. Opioid-

related mortality, which early studies suggested was reduced by [MED laws], now appears to

be positively correlated with these policies and the adoption of [REC laws]. The significant

policy action being taken to combat the opioid crisis as well as the evolution of the types of

opioids driving opioid-related harm likely contributes to the lack of robust findings for this

outcome.

4.9

Criminal justice and public nuisance outcomes

The implementation of REC is likely to affect criminal justice and other public nuisance outcomes in

multiple and divergent ways. There are concerns about drug-induced crime and disorder, such as

acquisitive crime, vagrancy and public consumption. Likewise, REC may alter the broader systemic

elements associated with crime, including illicit trade and violence. For example, REC legalisation

could potentially lead some illicit distributors to other forms of crime, and unregulated processing that

causes explosions pose a public safety hazard; however, we are not aware of any studies examining

this. For some of these concerns, it is more than likely that the effects are mixed. For example, police

are no longer tasked with enforcing laws prohibiting the possession of cannabis but are now required

to stop and detain cannabis-impaired drivers. Some of these criminal justice outcomes are specific to

the policy choices found in the US. For example, under federal prohibition, cannabis businesses

cannot access the banking system and must deal almost exclusively in cash, making them a target for

robberies.

However, there are limits to interpreting law enforcement data. Reports may not be complete or

reflective of the change in outcomes. They also may be confounded by changes in policy directives or

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

attitudes of individual officers. Post REC, public consumption may become more prevalent but under-

enforced as the public and law enforcement feel less compelled to intervene in minor infractions. As a

result, evaluating counts of public consumption citations may not reflect the true magnitude of

violations of the law with regard to using cannabis in public. Many other common ‘quality of life’

policing tasks are under-enforced. Cities have laws against littering, but these are often arbitrarily

enforced by police officers, as this is sometimes seen as a poor use of police time. Likewise, changes

in the estimates of intoxicated motorists may be a result not of REC per se but of changes in law

enforcement directives and training aimed at detecting impaired driving.

The studies evaluated here report mixed findings. Some suggest that REC is associated with

increases in some crimes but unrelated to changes regarding other crimes. This largely depends on

the stage at which the policy is evaluated (enactment v implementation) and where. Several studies

found that REC is associated with increases in cannabis-specific crimes (e.g. possession, sales) in

neighbouring non-REC states, suggesting that individuals may be moving cannabis over state lines.

Using Denver police data on crime reported by census tract, one study compared rates of violent,

property and cannabis-related crimes (defined as public consumption, robbery of a cannabis facility,

unauthorised distribution, etc.) during MED/REC in Colorado (from January 2013 to October 2015)

and examined whether or not they were related to the density of cannabis outlets. Freisthler et al.

(2017) reported that the density of cannabis outlets was not associated with property or violent crimes

in local areas, but the density of outlets was positively correlated with property crime in spatially

adjacent areas over time. Likewise, the density of outlets was associated with higher rates of reported

cannabis-specific crime in the same and neighbouring census tract (Freisthler et al., 2017).

Nevertheless, this analysis was not able to assess the relationship between unregistered caregiver

growers or home growers and incidents of crime. Furthermore, without an interrupted time series and

in the absence of controls, authors can only report a correlation with crime and dispensary density. It

is also possible that changes in law enforcement capacity and training may have confounded the

analysis.

At a more aggregated level, an interrupted time series study examined violent and property crime

clearance rates (i.e. rates in which a charge was made for a reported crime) each month from the

Federal Bureau of Investigation’s (FBI’s) uniform crime report data for Colorado and Washington,

compared with the US as a whole, from 2010 to 2015. Makin et al. (2018) estimate that REC

enactment in Colorado and Washington is associated with increases in police clearance rates for both

property and violent crime. This is the appropriate policy indicator, as criminal penalties for

possession and use were removed from the penal code on enactment. The authors employ an

autoregressive integrated moving average (ARIMA) model to control for seasonality effects in

robustness checks, reporting that their findings hold. Without measuring the mechanism of action

precisely, they attribute REC enactment to potentially increasing clearance rates by shifting

enforcement priorities in the field.

One working paper compared FBI arrest data from the uniform crime report in counties in Washington

and Colorado with those in neighbouring states from 2009 to 2014. The authors looked at arrests for

driving under the influence (DUIs) and various drug-related offences in relation to REC enactment and

implementation (2009-12 v 2013-14). They found that counties bordering Colorado saw an increase of

about eight cannabis possession arrests per 100 000 compared with non-bordering counties post

REC enactment and implementation; this was greater for counties bordering Washington, which saw

an increase of 22.9 arrests per 100 000 compared with non-bordering counties (Hao and Cowan,

2017). Reported arrest rates diminish the further the county is from the state borders of Washington or

Colorado, suggesting that cross-state diversion is occurring. However, law enforcement agencies

near the border may have shifted enforcement priorities (e.g. targeting out-of-state motorists),

potentially confounding such an analysis.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

When examining arrests for other drug-related offences (possession of other narcotics, distribution of

cannabis or DUIs), the authors reported no association between counties bordering a REC jurisdiction

and non-bordering counties post REC enactment and implementation. Nevertheless, the study

conflates the REC periods of enactment and implementation. It is more likely that cross-state

diversion would have occurred during the implementation period, after stores were opened than

immediately following enactment. However, their findings show mixed results. Rates of cannabis

possession arrests in counties bordering Colorado increased just prior to enactment in 2012, declined

in 2013 and increased again during implementation in 2014. In Washington, arrest rates in

neighbouring counties declined prior to enactment in 2012, increased in 2013 and then declined

during the first year of implementation in 2014. Such divergence could be explained by the fact that

jurisdictions neighbouring Washington have greater access to cannabis, with Oregon adopting a REC

law in 2014.

Another working paper examined Washington state recreational cannabis transaction retail data to

assess potential interstate trafficking between Washington and Oregon (Hansen et al., 2017b).

Washington implemented its REC in July 2014, whereas in neighbouring Oregon stores did not offer

recreational cannabis to adults until October 2015. Hansen et al. assessed the impact that REC

implementation in Oregon had on sales of cannabis on the Washington side of the border. They also

reported that retail sales near the border in Washington declined by 41 % immediately after REC

implementation in Oregon (Hansen et al., 2017b). The authors suggest that a substantial amount of

demand for cannabis in Washington may have originated in Oregon, resulting in considerable

interstate trafficking prior to REC implementation in that state.

4.10

Tax revenues

No studies have evaluated changes in cannabis policy and its impact on

overall

tax revenues (not just

those that are associated with cannabis excise taxes and fees); however, state regulators regularly

report sales and revenue data post REC. In Table 4.3, we report annual sales revenue and tax

receipts. Colorado’s figures are for the calendar year (CY) and Washington’s are for the fiscal year

(FY), which runs from July to June (2016 = July 2015 to June 2016). Retail cannabis sales surpassed

a billion dollars in Colorado in 2016 and in Washington in 2017. Sales and tax receipts have

increased year on year (with the exception of 2018, for which the information is partial).

TABLE 4.3

Cannabis sales and tax revenue

Colorado

Year

2014

2015

2016

2017

2018

2019*

Sales (CY)

(million USD)

683.5

995.6

1,307.2

1,507.7

1,545.7

386.7

Tax receipts (CY)

(million USD)

67.6

130.4

193.6

247.4

266.5

63.5

Sales (FY)

(million USD)

259.5

786.4

1,371.9

Not reported

Washington**

Tax receipts (FY)

(million USD)

64.9

185.7

314.8

362.0

Not reported

*January to March for Colorado.

**Recreational only.

Sources: Colorado Department of Revenue, ‘Marijuana tax data’; Washington State Liquor and Cannabis

Board, Marijuana Dashboard reports sales and tax receipts throughout 2017; Washington State Liquor

and Cannabis Board annual report FY 2018 reports total tax receipts for that year.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

4.11

Public opinion

Our search found several studies that evaluated public opinion of REC. However, none of them met

our inclusion criteria, because they were point-in-time estimates, they failed to include pre-REC

measures or they were non-representative. That said, the longest running nationally representative

survey asking Americans if they think the ‘use of marijuana should be made legal’ has shown a

steady rise in positive responses since the mid-1990s. In 1996, one out of four Americans supported

making the use of cannabis legal. By 2012, support had hit 50 %, and this rose further, to 66 %, in

October 2018 (McCarthy, 2018). However, the poll asks about ‘use’, giving less indication as to

respondents’ opinions on commercial legalisation. Nonetheless, most commercial legal REC models

have been a result of popular voter initiatives.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Issues to consider when establishing a comprehensive

monitoring and evaluation framework for changes to cannabis

regulations

As discussions about cannabis policy intensify (Hughes et al., 2017), policymakers are in a unique

position to learn from the experiences of cannabis reform in the Americas. Doing so, of course,

requires an understanding of the motives or goals for the policy changes that occurred in that region

and the extent to which the policy changes have in fact achieved their stated goals. However, such an

understanding can only be achieved through actual measurement of the outcomes associated with

these stated goals of the policy reform, both before and after implementation. States in the US that

are adopting recreational cannabis laws often argue that such reforms will improve civil liberties and

generate tax revenues (Hall and Kozlowski, 2018). There is less emphasis placed on public health

and safety goals in the US than in the national discussions in Uruguay and Canada. Researchers

note that many jurisdictions have not paid adequate attention to the measurement and monitoring of

critical outcomes that are relevant to understanding the impact that REC may have on public health

and safety. Without establishing a robust baseline measurement prior to the policy changes, it is

difficult to accurately determine the causal impact of such reforms (including unintended

consequences) or if they met their intended objectives.

This is perhaps the greatest lesson that policymakers can learn from the early assessments

conducted thus far — the need to consider, measure and monitor public health and safety objectives

associated with cannabis policy reform before policy changes are implemented. Doing so requires

considering all of the potential objectives of cannabis reform that a country may wish to achieve as

well as the possible unintended consequences of such changes. Only then is it possible to identify

metrics that can be used to assess changes and begin collecting baseline data. Moreover, as

discussed in Section 3, many models of cannabis reform are available, and the regulatory model

adopted will depend on the specific objectives of the policy change and will also influence the

baseline and follow-up data that will be appropriate to assess if the objectives have been achieved.

This section offers insights into creating such a comprehensive framework. We begin with a taxonomy

of potential objectives that jurisdictions might seek to achieve (or harms to avoid) with cannabis

reform, mapping them to specific outcomes that can be used as metrics for monitoring and evaluation.

Since many of these measures are not currently collected, Section 5.2 offers some ideas for creating

a data infrastructure to collect them. Section 5.3 offers insights into evaluating changes.

5.1

Potential objectives of cannabis policy change and metrics of evaluation

Discussions of cannabis policy change should begin with a clear understanding of objectives.

Table 5.1 provides a taxonomy of potential policy objectives mentioned in jurisdictions that have

recently reformed cannabis laws. These are based on communications with policymakers as well as

statements made by proponents of ballot initiatives in US states that have recently voted on such

initiatives. The stated objectives are to reduce crime and improve civil liberties, to promote public

health and to generate economic activity and tax revenues. While not exhaustive, the examples

provided in Table 5.1 provide a sense of what might motivate a jurisdiction to reform its cannabis laws

and give us the opportunity to consider metrics for measuring the outcome of such a policy change.

However, it should be noted that, as illustrated above, the evidence of whether or not these regulatory

changes can achieve these outcomes is not yet available.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

TABLE 5.1

Statements commonly made by proponents of cannabis law reform

Crime/public safety and civil

liberties

Reduce/eliminate the illicit market and related crime

Re-prioritise law enforcement resources

Reduce burden on criminal justice system

Reduce criminalisation of non-violent drug offenders

Reduce racial or ethnic disparities; correct for injustices of drug prohibition

Health

Make cannabis-based products available for medicinal purposes

Allow for more research on medicinal benefits

Reduce youth access and consumption

Minimise contaminants and ensure product quality

Offer more information to users on potency and harms

Improve precision in dosing and potency

Make it easier to talk about cannabis-related problems

Reduce use of other potentially more dangerous products (alcohol, opioids)

Prevention

Economic/budget

Limit availability by regulating licensees and store operations

Increase government revenue through taxation and licensing fees

Reduce cost of prohibition on criminal justice system

Create new jobs in the legal economy (eliminate illicit market jobs)

Normative

Government should not control what someone puts in their own body

Prohibition lacks legitimacy

Similarly, a comprehensive evaluation of any policy change must consider the possibility of undesired

outcomes. In Table 5.2, we offer a list of statements commonly made by opponents of relaxing

cannabis laws.

TABLE 5.2

Statements commonly made by opponents of cannabis law reform

Crime/public safety

Will not decrease, and may possibly increase, illicit market activity

Burden on law enforcement and public safety

Increase drug-induced crime, accidents, and negligence

Increase in drugged driving

Health

Increase prevalence

Increase potency, reduce price, and increase access to new products

Entices youth and new consumers

Increases addiction and substance use disorder

Encourages experimentation and use of other drugs

Normalises drug use, potentially re-normalisation of tobacco use

Prevention

Economic/budget

Makes prevention campaigns ineffective

Increases costs to state

Reduces productivity

Normative

Sends the wrong message

Incompatible with international treaties

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Table 5.3 lists some potential metrics that could be used for monitoring and evaluating the impacts of

cannabis legalisation on particular policy objectives. This list is based on the studies listed in

Section 4 as well as our own experiences evaluating other cannabis policy changes. It also builds on

some of the suggestions offered in Kilmer and Pacula (2017). In the table, we have placed these

items into four general categories that are not mutually exclusive (

): health, crime and criminal

justice, economics and other. We are not aware of any country that collects all the variables listed in

the table, and it is likely that additional variables of importance are missing.

TABLE 5.3

Variables for monitoring implementation and evaluating changes in cannabis policy

Variables

HEALTH

Prevalence

General population surveys (GPS),

school surveys, medical surveys

Adult annual and 30-day prevalence rates are expected to

increase with cannabis liberalisation; however, the health

implications of this are unclear, as simple increases in

prevalence rates indicate nothing about the potential health

risks associated with consumption. More information about

frequency of use, mode of administration and product

consumed is needed to fully understand the potential health

impacts of the changes in prevalence rates.

The effect of legalisation on youth prevalence rates is

ambiguous. The rates could increase, as we expect they will

with adults, but they could decrease (1) if the regulated

market makes it harder for young people to access cannabis

and/or (2) if there is a forbidden fruit effect (MacCoun,

1993). These effects may also offset one another.

If cannabis liberalisation policies increase frequency of use

— especially daily/near-daily use — then there are greater

potential health effects. Specific health risks depend on the

demographics of the user group (young people v adults,

pregnant women or other at-risk users), the amount

consumed per dose, the potency and the cannabinoid

profile of the products consumed, and the method of

ingestion — all of which will influence the actual health risk.

The rates of CUD might rise or fall with cannabis

liberalisation, depending on what happens to heavy use and

treatment admissions.

On the one hand, a decrease in criminal justice referrals

could reduce the probability of someone with CUD receiving

treatment. On the other hand, legalisation may make it

easier for those with CUD to talk openly about their

problems and obtain help.

Sources for obtaining these data

Implications for evaluating policies

Frequency

GPS, school surveys, medical surveys

Cannabis use disorder

(CUD)

GPS or school surveys; voluntary

treatment, and ED and hospital

admissions; qualitative interviews with

selective populations — all of these

sources capture a subset (sometimes

overlapping) of people with CUD

For duration, longitudinal studies

would be strongly preferred, but it may

be possible to get some of this

information from retrospective

surveys.

GPS, school surveys, medical

surveys, traceability data and

qualitative surveys of key user groups

Products used

There are different risks associated with different types of

cannabis products. Some adverse health events are tied

directly to acute overconsumption of THC, while other

longer term health effects are related to the method of

administration (i.e. the long-term effects of smoke on the

respiratory system). As a result, information on the products

used will be useful for thinking about the potential health

consequences observable among the population.

( ) For example, traffic accidents could be included in either health or crime.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Variables

Amount consumed

Sources for obtaining these data

Questions about the amount of

cannabis consumed may be included

in GPS or school surveys as well as

medical surveys. Some web-based

surveys ask about the amount

consumed and use picture prompts to

help guide respondents. In such

cases, respondents are randomly

shown a photo of different amounts of

cannabis and asked to quantify how

much they think each amount weighs.

This improves researchers’

understanding of the amount of

cannabis consumed. Capturing the

potency of the product (actual or

perceived) is also useful but may best

be obtained through qualitative

interviews with selective populations.

GPS, school surveys, medical

surveys; treatment admissions;

hospital and ED admissions; and

qualitative studies of key user groups

Implications for evaluating policies

The amount consumed per day or per episode of use is

important for understanding (1) the risk of an acute harm

and (2) the risk of developing a CUD. Ideally, it would be

preferable to move beyond grams of herbal cannabis or

puffs from a vape pen towards collecting information about

the consumption of THC and other chemicals in the

cannabis plant.

Prevalence and

consumption of other

substances

The net effects of cannabis liberalisation on public health

depend on whether cannabis becomes a substitute for or

complements other potentially harmful substances. For

example, if cannabis is used most often with tobacco (i.e.

they are ‘complements’), then cannabis legalisation might

hamper efforts to reduce tobacco use. If cannabis and

alcohol are substitutes, then having users switch from

alcohol to cannabis might generate net health gains (e.g.

reduced accidents, reduced domestic violence, reduced

liver cirrhosis). If cannabis and alcohol are complements,

then some health risks may be made worse with cannabis

liberalisation policies (e.g. accident risk, heart risk). Thus, it

is important to understand if liberalisation policies lead to the

use of cannabis in lieu of other substances (tobacco,

alcohol, opioids, new synthetic cannabinoid agonists) or to

joint consumption. This might differ among different

segments of the population.

Even if cannabis use does not directly cause mental health

disorders but simply increases the likelihood of onset at an

earlier age, cannabis liberalisation policies may still increase

the prevalence of some mental health disorders (psychoses,

schizophrenia, etc.). The correlation may depend on the

typical amount consumed, frequency of use and THC

exposure.

If cannabis liberalisation policies lead to an increase in

consumption such that they increase the prevalence of

CUD, it is possible that a jurisdiction will see a rise in

cannabis-related treatment admissions. However, if

treatment referrals for cannabis come largely through the

criminal justice system, then legalisation may lead to an

overall decline in treatment admissions.

ED visits and hospitalisations may emerge from one of three

potential mechanisms: (1) immediate unexpected reactions

to cannabis products consumed by naive users (e.g.

accidental poisoning from edibles); (2) risky modes of

consumption (e.g. dabbing) or behaviours while intoxicated

(e.g. driving); and (3) longer term health problems

aggravated by prolonged cannabis use, particularly in

certain forms of products (e.g. respiratory issues from

smoking).

Calls to poison control centres typically result from

overexposure to THC, which most frequently occurs through

the overconsumption of edibles (multiple standardised

doses) or exposure to high-potency products.

Other mental health

disorders

GPS, school surveys, medical

surveys; longitudinal surveys;

treatment admissions; hospital and ED

admissions; qualitative studies

involving key populations

Treatment admissions

Treatment facilities; healthcare

systems; GPS surveys; qualitative

studies on selective populations

Possibly from government healthcare

records

Hospitalisations and ED

visits

Hospital databases; insurance claims

Calls to poison control

centres

Poison control centres; emergency

service calls

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Variables

Education and

employment outcomes

Sources for obtaining these data

GPS or school surveys; longitudinal

surveys; education and employment

agencies; standardised testing results

in secondary schools or for

graduation; graduation rates;

employee drug testing data

Implications for evaluating policies

Absenteeism and presenteeism are both potentially affected

by increases in cannabis use, particularly if cannabis is

consumed prior to or during work or school. Such behaviour

could ultimately influence longer term productivity and tax

revenues of the government. Tracking changes in education

and employment outcomes before and after the adoption of

a change in cannabis policy can help identify the need for

additional workplace/school policies or targeted messaging

regarding appropriate use.

CRIME AND CRIMINAL JUSTICE

Cannabis-related arrests

and seizures

Police statistics

The legalisation of cannabis should generate a net decrease

in the number of arrests and seizures, but these will not fall

to zero for at least two reasons: (1) most jurisdictions retain

the prohibition of use/purchases by young people and (2)

illicit markets may persist for some time after a policy

change. Information on arrests can provide insights into

criminal justice savings associated with legalisation, while

seizure data can provide some insight into profit motives for

the illicit market to remain (e.g. specific products not

available in the legal market, or specific populations unable

to participate). Nevertheless, law enforcement data may be

biased because of internal policy directives and changes in

resources, agency capacity and citizen reports. Caution

should be used when evaluating these data.

Enforcing new regulations adopted by a jurisdiction around

legal cannabis will require resources, so tracking information

on cannabis-related police contacts (including those that

simply lead to confiscation) is important for accurately

understanding the net budgetary impacts on crime and

public safety resources. This information can also be useful

for understanding how changes in cannabis laws affect

racial/ethnic disparities in criminal justice outcomes.

Though some individuals may already be processing illicit

cannabis flower and turning it into concentrates, the

legalisation of cannabis might encourage more people to

extract cannabinoids from dried plant matter using volatile

chemicals and solvents. Unsafe processing practices may

pose a public safety hazard should the volatile chemicals

combust.

In the US, arguments have been made that cannabis

legalisation will improve civil liberties, reduce racial

disparities in the criminal justice system and improve police-

community relations. Monitoring how the criminal justice

system responds to changes in the law will allow us to

evaluate these arguments.

As previously stated, the enforcement of new regulations

adopted within a jurisdiction will require some criminal

justice resources, particularly if some violations of the new

regime still involve criminal penalties. Tracking this

information is necessary for understanding the cost of

enforcing the new policy and net criminal justice savings.

This is one of the most likely infractions to arise in response

to the legalisation of cannabis within a jurisdiction, although

the rate of increase depends on a variety of factors,

including a jurisdiction’s reliance on cars, the deterrent

effect from law enforcement and the effectiveness of

prevention campaigns to promote sober driving/responsible

use. However, not all impaired driving will be detected, so,

given limited policing resources, these data will be useful for

understanding which groups are likely to drive under the

influence of drugs and where and at what times such

behaviour is likely to occur.

Cannabis-related police-

citizen contacts

Perhaps police databases — new

record keeping may be required

After legalisation, civilians could also

track consumption in public spaces

Unregulated cannabis

processing involved in

explosions

Police or public safety statistics

Penalties for cannabis

offences

Court records; corrections agencies,

police databases

Criminal justice spending

on enforcing cannabis

laws

Police databases; court records;

corrections agencies; regulatory

agency data (compliance checks,

citations, fines)

Self-reported impaired

driving (or travelling with

someone who is

impaired)

GPS, school surveys, medical

surveys; could be added to treatment

intake surveys

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Variables

DUI prosecutions

Sources for obtaining these data

Police statistics; insurance claim data;

qualitative data collection in particular

areas

Implications for evaluating policies

DUI arrests and prosecutions are a function of (1) the

amount of impaired driving and (2) the level of enforcement.

Higher DUI arrests without additional enforcement suggest a

rise in impaired driving and the need for additional

messages to be targeted at at-risk users. It also raises the

criminal justice costs associated with legalisation. However,

total DUI arrests might fall with a change in policy if

cannabis becomes a substitute for alcohol. Thus,

information on DUI arrests (for alcohol, cannabis and other

substances if available) can help inform the jurisdiction

about how cannabis is being used by specific segments of

the population. Nevertheless, changes in enforcement

priorities and training may confound analyses that use law

enforcement data.

Changes in traffic crashes and fatalities may indicate a

change in impaired driving that is not reflected in arrest

statistics. Objective analyses will focus on the total number

of crashes/fatalities, not just those in which the driver tested

positive for THC. This can offer a better understanding

regarding the potential substitution of cannabis for alcohol or

their concurrent use.

Tracking changes in arrests/convictions for other crimes

helps to identify the public safety benefits/costs associated

with legalisation — in so far as arrests reflect changes in

crime patterns (that may be associated with use of cannabis

and/or its substitutes or complements). Nevertheless,

changes in enforcement priorities and training may

confound analyses that use law enforcement data.

Since not all crimes result in an arrest, tracking information

from calls to emergency services may identify changes in

the need for law enforcement intervention because of a

change in the consumption of cannabis and/or its

substitutes and complements.

Information on the changes in victimisation with the

legalisation of cannabis is relevant for considering public

safety impacts. Again, it is unclear if victimisation would rise

or decline with legalisation; much will depend on how the

policy change affects the use of other substances.

Traffic crashes or

fatalities

Transport and insurance claims

databases

Arrests/convictions for

other crimes

Police databases; court records

Calls to emergency

services

Police databases, possibly other

government agencies

Victimisation surveys

Victimisation surveys; medical reports

ECONOMICS

Total expenditures on

purchasing cannabis

products

GPS; retail sales (medical or other

outlets); combining consumption

estimates from wastewater testing with

THC price estimates

Information on cannabis expenditure serves at least three

purposes: (1) it helps identify tax evasion, particularly if

taxes are tied to the amount sold; (2) it is helpful for

understanding which products are sold (by form and

potency if information is collected by product); and (3) it is

useful for constructing an estimate of the total size of the

cannabis market (demand based), which can then be used

as a way of measuring the impact of further policy

refinements on the market.

Information on the source of cannabis is very helpful for

understanding the extent to which the legal market is able to

replace the illicit market.

Traceability systems are valuable for a range of regulatory

purposes in a legal cannabis market, including (1)

measuring the total amount supplied to a market (which,

when combined with a demand estimate, can identify the

presence and size of an illicit market); (2) projecting

estimates of tax revenue from sales (which can then be

compared with actual tax revenue to help identify tax

evasion); and (3) identifying and tracking products sold,

which is useful for understanding the potential health risks

that consumers might face (this is relevant to prevention

messaging as well as the issuing of product recalls).

Source of cannabis

supply

Amount of cannabis

produced in the legal

market

GPS; web surveys of heavy users

Traceability systems; visual

inspections

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Variables

Prices

Sources for obtaining these data

GPS, medical surveys, arrestee

surveys; web surveys of heavy users;

menu analyses; traceability systems;

law enforcement sources;

crowdsourcing websites such as ‘Price

of Weed’

Regulatory agency overseeing

cannabis legalisation; local

government zoning boards

Implications for evaluating policies

Prices are important, because they can influence many of

the outcomes that are discussed in cannabis policy debates:

size of the illicit market, tax receipts, business revenues and

consumption.

Presence of retail

establishments

The illicit market for cannabis can be reduced if there is a

licensing system identifying legitimate sellers from illegal

sellers. Enforcement against illegal sellers is a necessary

component of a licensed system. However, the density of

retail establishments can also influence the perceived

availability (changing norms about the substance) and can

reduce the effectiveness of prevention programmes aimed

at reducing use among particular groups. The regulation of

retail establishments will be a cost of legalisation, affecting

net revenues from a change in policy. More retail

establishments increase the number of entities that need to

be policed (and hence the cost of doing so).

Tracking sales and tax revenue is important for measuring

the size of the market and whether or not tax revenue is

growing. These taxes are often earmarked for prevention

efforts, regulatory oversight and other costs imposed by

users on society.

Information on employment in the cannabis industry can

indicate the economic value of the industry to a community.

It also provides insights into the relative importance of this

industry vis-à-vis other industries in an area.

Of course, legalisation can affect other segments of the

economy, besides those directly involved in the cannabis

trade.

While there are likely to be criminal justice savings

associated with cannabis legalisation, there are also agency

costs associated with regulating a new business. The

economic impact of legalisation is a function of the jobs and

tax revenue that comes with the new market minus the cost

of regulation and law enforcement monitoring and regulating

the new market. This information is vital for a proper

estimation of the net economic impact of legalisation.

Licit sales and tax

revenues

Regulatory agency overseeing

cannabis legalisation; other

government agencies

Employment in the

cannabis industry and

other related industries

Could try to deduce this from total

production estimates and qualitative

methods pre legalisation; after

legalisation, it should be part of regular

labour statistics

Cost of regulation and

law enforcement

Regulatory agencies overseeing

cannabis legalisation; police

databases; court records; corrections

OTHERS

Consumer characteristics

Administrative or regulatory data;

qualitative studies of key target

populations

Jurisdictions that collect information on patients or

registered adults (e.g. Uruguay) may allow for a further

understanding of the user base. This could include

descriptive information on patient ailments or demographic

information regarding a typical registrant and how much

they purchase over the course of a year.

These polls provide a source for evaluating the normative

justification for a change in cannabis policy.

The degree to which cannabis products are advertised and

promoted might influence the type and amount of prevention

material needed by a community to reduce the potential

harms associated with legalisation.

Public support

Advertising/promotion

Public opinion polls

Could require all cannabis

advertisements be reported to

regulatory agencies; could also

conduct studies by the European

Medicines Agency

5.2

Establishing the data infrastructure

Once a jurisdiction has identified the objectives for considering a policy change as well as specific

metrics that would be useful for monitoring impacts, it can then consider appropriate data collection.

The second column of Table 5.3 offers ideas for collecting or generating relevant data. Keep in mind

that there are strengths and weaknesses of different data sources and various approaches. In this

section, we discuss in greater detail some of the issues to consider when relying on some of these

sources.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

General population surveys (GPS) and student surveys

GPS and student surveys will be useful for some of these items, but care must be taken when using

these data to draw inferences about cannabis policy changes, for several reasons. Most importantly,

there is strong evidence suggesting that survey respondents under-report the prevalence of drug use

in GPS (e.g. Harrison et al., 2007). While many studies have relied on assumptions that under-

reporting — although unknown — is fairly stable year on year, these assumptions will not necessarily

hold if a policy changes. For example, if a policy change reduces the stigma around admitting

cannabis use, then more people will be willing to honestly report their prior behaviour, making it hard

to identify the extent to which any change detected is attributable to a true increase in cannabis use

as opposed to more honest reporting, or both.

One way to assess the degree to which under-reporting might exist and change with policy would be

to validate the GPS with a biological test (e.g. saliva tests or urinalysis), to determine the proportion of

respondents who are misreporting at a given point in time. This has been done in the US, but we are

not aware of any attempts to do this in Europe (Kilmer et al., 2015). Another alternative that might be

considered is to incorporate the use of the randomised response technique for key estimates into the

questionnaires (Blair et al., 2015). Such validation does not have to be done every year, but doing it

now and possibly a few years after a significant policy change would provide extremely useful

information about baseline rates of under-reporting and how these change over time. The use of

wastewater epidemiology may potentially generate useful confirmatory measures for cannabis

prevalence rates (Zuccato et al., 2008). Countries in Europe and elsewhere may not want to rely on

US under-reporting estimates from nearly 20 years ago to inform these adjustments.

There are also some questions that could be added to GPS regarding consumption and expenditure

that would improve our understanding of these markets and how they could change (e.g. asking about

the types of cannabis products used and the amount consumed). Adding questions about market

transactions is critical for understanding expenditures and revenues generated by illicit suppliers,

which is important if the goal is to understand the impact of policy changes on the size of the illicit

market. Since it may not be possible to add an entire cannabis market module, as was done in the US

from 2002 to 2014 (and reinstated in 2018), adding two questions concerning (1) the amount spent

during the last purchase and (2) the number of purchases in the previous month would be enough to

provide the foundation for market estimates.

It would also be helpful to collect information about the types of products purchased. Information on

the types of products is especially valuable if a jurisdiction decides to limit the legal sale of cannabis

to particular forms (e.g. herbal cannabis), which Uruguay does. Responses can indicate if individuals

are sourcing other products (e.g. edibles) from illicit markets.

Web surveys

While it is possible to use internet-based surveys with strong sampling frames to generate population-

representative inferences about cannabis use and purchasing patterns (e.g. Pacula et al., 2016),

many of the web surveys about cannabis fielded in Europe and the US are convenience samples (e.g.

Kilmer et al., 2013; van Laar et al., 2013; Matias et al., 2019). While these relatively inexpensive

surveys do not allow for precise estimates about a representative sample, they can provide plausible

ranges for key variables if the samples are large enough. If targeted correctly, they can help provide

information about heavy users, who may be less likely to be included in the sampling frame for GPS

or student surveys but account for the largest part of consumption and expenditures (Kilmer et al.,

2014).

Indeed, combining estimates from GPS and web surveys is becoming an increasingly popular

approach to generating market estimates for consumption and expenditures. Typically, the GPS

estimates of the number of use days are combined with the web survey estimates of the amount

consumed per use day for various types of users (e.g. weekly, daily/near daily). In some cases, these

estimates are generated only for an entire country (e.g. Caulkins and Kilmer, 2013), while other

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

studies generate these estimates for specific age or gender groups and then add them together to

generate an aggregate estimate (e.g. Caulkins et al., under review; van Laar et al., 2013). Another

approach to using unrepresentative web surveys to generate population-wide estimates is to use a

raking algorithm to reweight web survey respondents to match GPS data (see Caulkins et al., 2019).

Cannabis medical surveys

In the light of the changes happening within US jurisdictions, several health insurers are beginning to

elicit information about cannabis use (both medicinal and recreational) in their regular patient surveys

so as to monitor use of the substance in a fashion similar to alcohol and tobacco. While this

information is not publicly available from private insurers, it could easily be collected by public insurers

in various jurisdictions. Distinguishing medicinal consumption from recreational consumption would be

key, as the particular forms of cannabis, the amounts consumed and frequency of use are likely to

differ, depending on the subpopulation considered (Pacula et al., 2016; Lankenau et al., 2017).

Cannabis tracking systems

In the shadow of federal prohibition, states that passed legalisation have implemented ‘seed-to-sale’

traceability systems to signal that they are taking the issue of diversion seriously. These systems track

every plant throughout the supply chain, ending at retail sale. They can include information about the

prices paid by processors and retailers as well as information about cannabinoid content and the

types of products sold on the licit market. While these systems were not necessarily designed for

research purposes, they are being used by an increasing number of researchers to help understand

the markets and other cannabis-derived products (Smart et al., 2017), to estimate the price elasticity

of demand (Hansen et al., 2017) and the level of competition in the market at the wholesale and retail

levels (Caulkins et al., 2018).

In the US, there are no systems to track sales made to specific individuals (i.e. the identities of

purchasers is not tracked), but there are in Uruguay. In addition to a seed-to-sale system, Uruguay

tracks and limits purchases by individuals. Retail purchases can be made only at participating

pharmacies and are limited at 40 g/person/month. To make sure individuals do not exceed these

limits, a biometric system requires registered buyers to submit a thumbprint before making a purchase

(Miroff, 2017). The pharmacy receives immediate feedback from the government database about

whether or not the person has exceeded the allowable retail transaction limit; if not, the sale is made.

The total weights and numbers of sales per location are then reported to the national monitoring

agency.

Other countries considering changes in cannabis policy are not constrained in the same way as the

states in the US, and may want to limit the amount of cannabis that can be purchased in a given time

period, or consider implementing seed-to-sale systems to obtain data that are relevant to other policy

objectives, including identifying leakages to the illicit market, contaminated products for recalls and

valuable information regarding the types (and amount) of various products sold. This information can

be important for monitoring compliance and providing information for economic and other analyses.

Such information will allow regulators and researchers to triangulate with other indicators, such as

self-reported user behaviours from population surveys.

5.3

Thinking seriously about outcome evaluations

The third column of Table 5.3 provides an explanation for why particular types of data are useful for

generating inputs that can be used to measure many outcomes of interest, including (1) the actual

demand for cannabis in a market; (2) the size of the legal cannabis supply and industry; (3) the size of

the illicit market supply of cannabis; (4) the net impact of cannabis legalisation on the criminal justice

system (arrests), the public safety system (calls to emergency services, victimisation) and the health

system (hospitals, poison control centre calls, accidents); and (5) the net economic impact of the

market on the government in terms of jobs created, tax revenue, regulatory costs and health/law

enforcement net costs (or savings). It is worth mentioning that this report and many evaluations of

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

REC to date exclude any evaluation of the benefits — including mere pleasure — of legal cannabis

reforms. It is difficult to assess these consequences, as they are often subjective and difficult to

quantify, but any serious evaluation of these legal changes ought to consider these outcomes

(Caulkins et al., 2016).

Collecting the appropriate outcome data is necessary but insufficient for evaluating these policy

changes. Serious thought must be given to defining precisely what is evaluated and what is the

counterfactual, to allow the determination of causal inference about policy change. This section

addresses some of the issues for those considering an evaluation.

Understanding the cannabis market and expected sanction for a cannabis violation

before

the

policy change

There are a number of mechanisms that may influence cannabis use after a policy change, such as

stigma, price, promotion, availability, expected sanction and the ‘forbidden fruit’ effect (MacCoun,

1993). Collecting the baseline data for price and expected sanction is important for understanding not

only the casual mechanisms at play but also how different forms of legalisation may have different

effects on use and other outcomes. For example, if a jurisdiction that does not criminally sanction

cannabis possession chooses to legalise the supply of cannabis, the criminal justice cost savings of

doing so will not be as great as a jurisdiction that criminally sanctions cannabis users. Alternatively, if

a government implements a regulatory regime that keeps the price per unit of THC similar to its pre-

legalisation price, it may not detect as much of an effect on use compared with other regimes that

have experienced price declines as a result of market competition (Kilmer et al., 2010).

Collecting market information (e.g. information on prices, products, promotion, the number of outlets)

is important, especially for jurisdictions that had vibrant commercial medical markets before

recreational legalisation. It could also be useful in countries such as Spain, where, despite national

prohibition, significant numbers of cannabis social clubs are tolerated in some regions, some of which

may become increasingly commercialised (Pardal, 2018).

Canada has made an effort to obtain baseline data for future evaluations. For example, building on

the Canadian Student Tobacco, Alcohol and Drugs Survey, Health Canada designed the Canadian

Cannabis Survey, aimed at understanding the pre-legalisation cannabis market. In 2017, the national

government fielded the phone-initiated survey designed to obtain a sufficient representation among

key population groups across all provinces and territories. The survey not only included questions that

are relevant to cannabis consumption but asked respondents about their purchasing habits (how

much they bought, how often, from what sources and at what prices). The findings informed

discussions on the cannabis law and will also be used to evaluate its implementation in the future

(examining changes in prevalence, price, frequency of use, etc.). In addition, Public Safety Canada

commissioned work to estimate the price of cannabis throughout the country (Ouellet et al., 2017),

and Statistics Canada created a ‘Cannabis Stats Hub’, which has many similarities to the EMCDDA’s

Statistical Bulletin but includes more detailed information about the cannabis economy.

Considering the counterfactual and underlying context

Section 4 highlights the growing number of studies recreational legalisation in the US that include a

control group, but much attention is still paid to simple pre-post analyses of trend data (for more on

trends, see Appendix C). While cross-jurisdictional comparisons seem like an obvious approach to

applying rigorous research methods to learn about these policy changes, caution must be exercised,

as important differences in definitions, survey methodologies, the frequency of data collection and

other factors can limit inferences from these exercises (MacCoun, 1993; Kilmer et al., 2015;

Giommoni et al., 2017). More importantly, the assumption that any single jurisdiction adopts a uniform

policy, which is required and embedded in many cross-national studies, is clearly problematic (e.g.

states in the US vary considerably in terms of their cannabis regulations and enforcement). Similarly,

context matters. Different jurisdictions may have different starting points for the adoption of REC. This

may be less of a transition for a jurisdiction with a robust medical cannabis market or some social

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

access (e.g. the Dutch coffee shops or cannabis social clubs in Spain) than for a jurisdiction that

aggressively enforces cannabis prohibition.

Another approach is to take advantage of within-state variation in how policy is being implemented.

For example, although the states of Colorado and Washington passed legalisation in 2012, there are

still a number of localities in these states that do not allow retail stores. This has allowed researchers

to exploit the variations in retail cannabis outlet density to learn more about the effect of

commercialisation (e.g. Dilley et al., 2017). Similarly, studies exploiting variations in medical cannabis

access within California have been carried out to examine the impacts on use, hospitalisation and

crime (e.g. Mair et al., 2015; Freisthler et al., 2016; Hunt et al., 2018). An advantage of using intra-

state variation is that there is less concern about data comparability and policy comparability, and

there may be shorter time lags. Of course, there may be important issues of policy endogeneity that

will need to be considered in the statistical models (Pacula and Smart, 2017).

Paying close attention to dates and isolating policy exposure

The date used to denote when the policy change occurred obviously has important implications for

the results of outcome analyses. There can be not only lags between the date of passage and the

date the policy goes into effect but also important lags between passage and the ability to purchase

cannabis from a store, as it can take time to build a regulatory system. For example, the voters in

Colorado and Washington passed legalisation in November 2012, but the recreational stores did not

open until January 2014 and July 2014, respectively (

). Thus, outcome analyses of legalisation that

use dates when laws were passed, rather than when stores opened (or when a majority of the adult

population had access to cannabis), are at serious risk of drawing the wrong conclusions if changes

are driven principally by market activity and not shifts in norms (Pacula and Smart, 2017). The dates

can also be useful for learning more about the causal mechanisms underlying possible changes in

use (e.g. the period when it is legal to use or possess but before the stores have opened) and

understanding the short- and long-term implications of these policies.

It is imperative that when regulatory changes are introduced, governments, or possibly agencies such

as the EMCDDA, document what the legal changes allow/require and whether or not local variation is

permitted (e.g. if shops allowed). Although this seems obvious, it can be difficult to obtain this

information, as enacted laws may not include regulatory details that can influence outcomes of

interest (e.g. testing requirements). Better-prepared jurisdictions will require regulatory agencies to

publicly document this information and note when changes are made to regulations. This type of

documentation is especially important if the new policy allows local variation.

Acknowledging that the short- and long-term effects of legalisation could be different

It is entirely possible that the short-term effects of a policy change differ from the long-term effects.

This is likely to be true, as knowledge about products and how they might be used (or misused) takes

time to develop, as do changes in social norms concerning the appropriate (or inappropriate) use of

the products. Even existing markets can see the introduction of new products (e.g. hash oils and

concentrates) that capture different properties of the original cannabis plant (e.g. higher THC content)

and which can also change the known health impacts of the substance. It is certainly the case that we

have very little knowledge of the health implications of the products currently sold in US legal markets

(e.g. Kilmer, 2018). Most of the research on the health benefits and consequences of cannabis are

based on cannabis plant material with far lower THC and higher CBD concentration levels than those

observed on the legal market in the US today. Thus, it is difficult to infer the long-term effects of these

newer, less researched, products.

( ) Even though some stores opened in Washington in July 2014, it was some time before most of the licensed stores opened

and had enough inventory to push prices below the pre-legalisation levels.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Concluding thoughts

This document is intended to inform discussions about the value of having tools in place to evaluate

the impact of any major changes to the regulatory approach to cannabis used. By highlighting the

various approaches unfolding in the Americas and some of the emerging evidence from states in the

US, namely Colorado and Washington, we construct an analytic framework and empirical foundation

that we hope will contribute to productive and dispassionate policy discussions.

We cannot stress enough that the peer-reviewed literature review on legalisation of adult cannabis

use is nascent, and that results are likely to be conflicting, depending on which data and methods are

used as well as which policies are evaluated. It is important to remain sceptical of early studies,

especially those that use a simple binary variable to classify legalisation and non-legalisation states.

This scepticism should extend to the many studies that fail to account for the existence of robust

commercial medical cannabis markets that predate recreational cannabis laws. Even if a consensus

develops on certain outcomes, this does not mean that that relationship will hold over time. Changes

in norms concerning cannabis use and potentially other substances, the maturation of markets and

the power of private businesses (if allowed) could lead to very different outcomes 15 or 25 years after

recreational cannabis laws have been passed. Evaluations of these changes must be considered an

ongoing exercise, not something that should happen in the short term.

Insights from evaluations of possible supply changes in Europe will depend on the number of data

collected before the change; simply comparing past-month prevalence rates will not tell us much

about the effect of the change on health. While US jurisdictions have been moving quickly to legalise

cannabis, the data infrastructure for evaluating these changes is limited. In contrast, Canada has

made important efforts to field new surveys and create new data collection programmes in

anticipation of legal changes. If

some

jurisdictions are considering a major adjustment to cannabis

supply policies in the near future, then it is prudent to start thinking as soon as possible about

improving data collection and analysis systems to support subsequent evaluation exercises. Given

the open borders that exist within the EU, the wider impact of policy changes made within any one

country will also deserve consideration. Section 5 of this report should serve as a useful resource for

some of these discussions.

While there is much to learn from what is happening in the Americas, policy discussions should not be

limited to approaches implemented there. There are several regulatory tools (e.g. minimum pricing,

potency-based taxes) that receive very little attention — if any — that could have important

consequences for health, public safety and/or social equity (

). Acknowledging that individuals (and

governments) have different values and preferences for risk when it comes to cannabis policy can

make for more productive discussions on this controversial topic.

( ) Uruguay is one exception here. The government does not set a minimum price for cannabis sold in the pharmacies — it

sets the price. Some provinces in Canada are contemplating THC potency taxes, while others have the authority to set a price

floor.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

References

Aguiar, S. and Musto, C. (2015), ‘El rayo verde: regulación de la marihuana en Uruguay’, in Labate B.

C. and Rodrigues T. (eds),

Drogas, política y sociedad en América Latina y el Caribe,

CIDE,

México, pp. 297-316.

Anderson, D. M., Hansen, B., Rees, D. I. and Sabia, J. J. (2019), ‘Association of marijuana laws with

teen marijuana use: new estimates from the youth risk behavior surveys’,

JAMA Pediatrics

173(9), pp. 879-881.

Aydelotte, J. D., Brown, L. H., Luftman, K. M., Mardock, A. L., Teixeira, P. G. R., Coopwood, B. and

Brown, C. V. R. (2017), ‘Crash fatality rates after recreational marijuana legalization in

Washington and Colorado’,

American Journal of Public Health

107, pp. 1329-1331.

https://doi.org/10.2105/AJPH.2017.303848.

Aydelotte, J. D., Mardock, A. L., Mancheski, C. A., Quamar, S. M., Teixeira, P. G., Brown, C. V. and

Brown, L. H. (2019), ‘Fatal crashes in the 5 years after recreational marijuana legalization in

Colorado and Washington’,

Accident Analysis & Prevention

132, p.105284.

Bell, C., Slim, J., Flaten, H. K., Lindberg, G., Arek, W. and Monte, A. A. (2015), ‘Butane hash oil burns

associated with marijuana liberalization in Colorado’,

Journal of Medical Toxicology

11,

pp. 422-425. https://doi.org/10.1007/s13181-015-0501-0.

Bierut, T., Krauss, M. J., Sowles, S. J. and Cavazos-Rehg, P. A. (2017), ‘Exploring marijuana

advertising on weedmaps, a popular online directory’,

Prevention Science

18, pp. 183-192.

https://doi.org/10.1007/s11121-016-0702-z.

Blair, G., Imai, K. and Zhou, Y. (2015), ‘Design and analysis of the randomized response technique’,

Journal of the American Statistical Association,

pp. 1304-1319.

Borodovsky, J. T., Lee, D. C., Crosier, B. S., Gabrielli, J. L., Sargent, J. D. and Budney, A. J. (2017),

‘U.S. cannabis legalization and use of vaping and edible products among youth’,

Drug and

Alcohol Dependence

177, pp. 299-306. https://doi.org/10.1016/j.drugalcdep.2017.02.017.

Caulkins, J. P. and Kilmer, B. (2013), ‘Estimating the size of the EU cannabis market’, in Trautmann,

F., Kilmer, B. and Turnbull, P. (eds),

Further insights into aspects of the EU illicit drugs

market,

European Commission, Brussels.

Caulkins, J., Kilmer, B., Kleiman, M., MacCoun, R., Midgette, G., Oglesby, P., Pacula, R. and Reuter,

P. (2015),

Considering marijuana legalization,

RAND Corporation, Santa Monica, CA.

Caulkins, J. P., Kilmer, B. and Kleiman, M. A. R. (2016),

Marijuana legalization: what everyone needs

to know,

2nd edition, Oxford University Press, New York.

Caulkins, J. P., Bao, Y., Davenport, S., Fahli, I., Guo, Y., Kinnard, K., Najewicz, M., Renaud, L. and

Kilmer, B. (2018), ‘Big data on a big new market: insights from Washington State’s legal

cannabis market’,

International Journal of Drug Policy

57, pp. 86-94.

Caulkins, J., Davenport, S., Doanvo, A., Furlong, K., Siddique, A., Turner, M. and Kilmer, B. (2019),

‘Triangulating web and general population surveys: how well do results match legal cannabis

market sales?’,

International Journal of Drug Policy

73, pp. 293-300.

https://doi.org/10.1016/j.drugpo.2019.06.010.

Cerdá, M., Mauro, C., Hamilton, A., Levy, N. S., Santaella-Tenorio, J., Hasin, D., Wall, M. M., Keyes,

K. M. and Martins, S. S., (2019), ‘Association between recreational marijuana legalization in

the United States and changes in marijuana use and cannabis use disorder from 2008 to

2016’,

JAMA Psychiatry.

doi:10.1001/jamapsychiatry.2019.3254.

Cerdá, M., Wall, M., Feng, T., Keyes, K. M., Sarvet, A., Schulenberg, J., O’Malley, P. M. et al. (2017),

‘Association of state recreational marijuana laws with adolescent marijuana use’,

JAMA

Pediatrics

171, pp. 142-149. https://doi.org/10.1001/jamapediatrics.2016.3624.

Cerdá, M., Wall, M., Keyes, K. M., Galea, S. and Hasin, D. (2012), ‘Medical marijuana laws in 50

states: investigating the relationship between state legalization of medical marijuana and

marijuana use, abuse and dependence’,

Drug and Alcohol Dependence

120, pp. 22-27.

https://doi.org/10.1016/j.drugalcdep.2011.06.011.

Chang, E. (2018), ‘Canadian cannabis stocks set to surge in fall’,

US News and World Report,

https://money.usnews.com/investing/stock-market-news/articles/2018-07-09/canadian-

cannabis-stocks-set-to-surge-in-fall, accessed on 29 May 2019.

Cheng, C., Mayer, W. J. and Mayer, Y. (2018), ‘The effect of legalizing retail marijuana on housing

values: evidence from Colorado’,

Economic Inquiry

56, pp. 1585-1601.

https://doi.org/10.1111/ecin.12556.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Chu, Y.-W. L. (2014), ‘The effects of medical marijuana laws on illegal marijuana use’,

Journal of

Health Economics

38, pp. 43-61.

Cohn, A. M., Johnson, A. L., Rose, S. W., Rath, J. M. and Villanti, A. C. (2017), ‘Support for marijuana

legalization and predictors of intentions to use marijuana more often in response to

legalization among U.S. young adults’,

Substance Use and Misuse

52, pp. 203-213.

https://doi.org/10.1080/10826084.2016.1223688.

D’Amico, E. J., Rodriguez, A., Tucker, J. S., Pedersen, E. R. and Shih, R. A. (2018), ‘Planting the

seed for marijuana use: changes in exposure to medical marijuana advertising and

subsequent adolescent marijuana use, cognitions, and consequences over seven years’,

Drug and Alcohol Dependence

188, pp. 385-391.

https://doi.org/10.1016/j.drugalcdep.2018.03.031.

Davis, K. C., Allen, J., Duke, J., Nonnemaker, J., Bradfield, B., Farrelly, M. C., Shafer, P. and Novak,

S. (2016), ‘Correlates of marijuana drugged driving and openness to driving while high:

evidence from Colorado and Washington’,

PLoS One

11.

https://doi.org/10.1371/journal.pone.0146853.

Dilley, J. A., Hitchcock, L., McGroder, N., Greto, L. A. and Richardson, S. M. (2017), ‘Community-level

policy responses to state marijuana legalization in Washington state’,

International Journal of

Drug Policy

42, pp. 102-108. https://doi.org/10.1016/j.drugpo.2017.02.010.

Dilley, J. A., Richardson, S. M., Kilmer, B., Pacula, R. L., Segawa, M. B. and Cerdá, M. (2019),

‘Prevalence of cannabis use in youths after legalization in Washington State’,

JAMA

Pediatrics

173, pp. 192-193.

EMCDDA (2018),

Medical use of cannabis and cannabinoids — questions and answers for

policymaking,

Rapid Communication, Publications Office of the European Union,

Luxembourg.

EMCDDA and Canadian Centre on Substance Use and Addiction (CCSA) (2018),

Cannabis and

driving — questions and answers for policymaking,

Publications Office of the European

Union, Luxembourg.

Estoup, A. C., Moise-Campbell, C., Varma, M. and Stewart, D. G. (2016), ‘The impact of marijuana

legalization on adolescent use, consequences, and perceived risk’,

Substance Use and

Misuse

51, pp. 1881-1887. https://doi.org/10.1080/10826084.2016.1200623.

Everson, E. M., Dilley, J. A., Maher, J. E. and Mack, C. E. (2019), ‘Post-legalization opening of retail

cannabis stores and adult cannabis use in Washington State, 2009–2016’,

American Journal

of Public Health

109(9), pp.1294-1301.

Fergusson, D. M. and Boden, J. M. (2008), ‘Cannabis use and later life outcomes’,

Addiction

103,

pp. 969-976. https://doi.org/10.1111/j.1360-0443.2008.02221.x.

Fiala, S. C., Dilley, J. A., Firth, C. L. and Maher, J. E. (2017), ‘Exposure to marijuana marketing after

legalization of retail sales: Oregonians’ experiences, 2015-2016’,

American Journal of Public

Health

108, pp. 120-127. https://doi.org/10.2105/AJPH.2017.304136.

Fischer, B., Kuganesan, S. and Room, R. (2015), ‘Medical marijuana programs: implications for

cannabis control policy — observations from Canada’,

International Journal of Drug Policy

26,

pp. 15-19.

Freisthler, B., Ponicki, W. R., Gaidus, A. and Gruenewald, P. J. (2016), ‘A micro-temporal geospatial

analysis of medical marijuana dispensaries and crime in Long Beach, California’,

Addiction

111, pp. 1027-1035.

Freisthler, B., Gaidus, A., Tam, C., Ponicki, W. R. and Gruenewald, P. J. (2017), ‘From medical to

recreational marijuana sales: marijuana outlets and crime in an era of changing marijuana

legislation’,

Journal of Primary Prevention

38, pp. 249-263. https://doi.org/10.1007/s10935-

017-0472-9

Garat, G. (2015),

El camino: cómo se reguló el cannabis en Uruguay según sus actores políticos y

sociales 2014,

Junta Nacional de Drogas, Montevideo.

Garcia, M. (2017),

D.C. marijuana market: stuck in a gray zone,

NPR,

https://www.npr.org/2017/07/30/537324044/d-c-marijuana-market-stuck-in-a-gray-zone,

accessed on 29 May 2019.

Giommoni, L., Reuter, P. and Kilmer, B. (2017), ‘Exploring the perils of cross-national comparisons of

drug prevalence: the effect of survey modality’,

Drug and Alcohol Dependence

181, pp. 194-

199. https://doi.org/10.1016/j.drugalcdep.2017.09.027.

Goldstein, S. (2018),

Vermont AG: it’s illegal for businesses to ‘gift’ marijuana,

Seven Days,

https://www.sevendaysvt.com/OffMessage/archives/2018/07/23/vermont-ag-its-illegal-for-

businesses-to-gift-marijuana, accessed on 29 May 2019.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Grant, T. M., Graham, J. C., Carlini, B. H., Ernst, C. C. and Brown, N. N. (2017), ‘Use of marijuana

and other substances among pregnant and parenting women with substance use disorders:

changes in Washington state after marijuana legalization’,

Journal of Studies on Alcohol and

Drugs

79, pp. 88-95. https://doi.org/10.15288/jsad.2018.79.88.

Hall, K. E., Monte, A. A., Chang, T., Fox, J., Brevik, C., Vigil, D. I., Van Dyke, M. and James, K. A.

(2018), ‘Mental health-related emergency department visits associated with cannabis in

Colorado’,

Academic Emergency Medicine

25, pp. 526-537.

https://doi.org/10.1111/acem.13393.

Hall, W. (2018), A summary of reviews of evidence on the efficacy and safety of medical use of

cannabis and cannabinoids, EMCDDA, Lisbon.

Hall, W. and Lynskey, M. (2016), ‘Evaluating the public health impacts of legalizing recreational

cannabis use in the United States’,

Addiction

111, pp. 1764-1773.

https://doi.org/10.1111/add.13428.

Hall, W. and Kozlowski, L. T. (2018), ‘The diverging trajectories of cannabis and tobacco policies in

the United States: reasons and possible implications’,

Addiction

133, pp. 595-601.

https://doi.org/10.1111/add.13845.

Hall, W., West, R., Marsden, J., Humphreys, K., Neale, J. and Petry, N. (2018), ‘It is premature to

expand access to medicinal cannabis in hopes of solving the US opioid crisis’,

Addiction

113,

pp. 987-988. https://doi.org/10.1111/add.14139.

Hansen, B., Miller, K. and Weber, C. (2017a),

The taxation of recreational marijuana: evidence from

Washington state,

National Bureau of Economic Research, Cambridge, MA.

Hansen, B., Miller, K. and Weber, C. (2017b),

The grass is greener on the other side: how extensive

is the interstate trafficking of recreational marijuana?,

(Working Paper No. 23762), National

Bureau of Economic Research, Cambridge, MA. https://doi.org/10.3386/w23762.

Hansen, B., Miller, K. S. and Weber, C. (2018),

Early evidence on recreational marijuana legalization

and traffic fatalities,

(Working Paper No. 24417), National Bureau of Economic Research,

Cambridge, MA. https://doi.org/10.3386/w24417.

Hao, Z. and Cowan, B. (2017),

The cross-border spillover effects of recreational marijuana

legalization,

(Working Paper No. 23426), National Bureau of Economic Research, Cambridge,

MA. https://doi.org/10.3386/w23426.

Harrison, L. D., Martin, S. S., Enev, T. and Harrington, D. (2007),

Comparing drug testing and self-

report of drug use among youths and young adults in the general population,

Substance

Abuse and Mental Health Services Administration, Office of Applied Studies, Rockville, MD.

Hasin, D. S., Wall, M., Keyes, K. M., Cerdá, M., Schulenberg, J., O’Malley, P. M., Galea, S., Pacula,

R. and Feng, T. (2015), ‘Medical marijuana laws and adolescent marijuana use in the USA

from 1991 to 2014: results from annual, repeated cross-sectional surveys’,

Lancet Psychiatry

2, pp. 601-608. https://doi.org/10.1016/S2215-0366(15)00217-5.

Highway Loss Data Institute (2017), ‘Recreational marijuana and collision claim frequencies’, Vol. 34,

No. 14, Highway Loss Data Institute, Arlington, VA.

Hughes, B., Quigley, E., Ballotta, D. and Griffiths, P. (2017) ‘European observations on cannabis

legalization’,

Addiction

112, pp. 1136-1137.

Hunt, P. and Pacula, R. L. (2017), ‘Early impacts of marijuana legalization: an evaluation of prices in

Colorado and Washington’,

Journal of Primary Prevention

38, pp. 221-248.

https://doi.org/10.1007/s10935-017-0471-x.

Hunt, P., Pacula, R. L. and Weinberger, G. (2018),

High on crime? Exploring the effects of marijuana

dispensary laws on crime in California counties,

IZA Discussion Papers, Institute of Labor

Economics (IZA), Bonn.

Jikomes, N. and Zoorob, M. (2018), ‘The cannabinoid content of legal cannabis in Washington state

varies systematically across testing facilities and popular consumer products’,

Scientific

Reports

8, 4519. https://doi.org/10.1038/s41598-018-22755-2.

Johnson, L. (2015),

Marijuana dispensary regulations approved in Vancouver,

CBC News,

https://www.cbc.ca/news/canada/british-columbia/marijuana-dispensary-regulations-

approved-in-vancouver-1.3126111, accessed on 4 December 2019.

Jones, J., Jones, K. and Peil, J. (2018), ‘The impact of the legalization of recreational marijuana on

college students’,

Addictive Behaviors

77, pp. 255-259.

https://doi.org/10.1016/j.addbeh.2017.08.015.

Jones, C. M., Underwood, J. M. and Volkow, N. D. (2019), ‘Challenging the association of marijuana

laws with teen marijuana use’,

JAMA Pediatrics,

doi:10.1001/jamapediatrics.2019.4235.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Kerr, D., Bae, H., Phibbs, S. and Kern, A. C. (2017), ‘Changes in undergraduates’ marijuana, heavy

alcohol and cigarette use following legalization of recreational marijuana use in Oregon’,

Addiction

112, pp. 1992-2001. https://doi.org/10.1111/add.13906.

Kerr, D. C., Bae, H. and Koval, A. L. (2018), ‘Oregon recreational marijuana legalization: Changes in

undergraduates’ marijuana use rates from 2008 to 2016’,

Psychology of Addictive

Behaviors

32(6), p. 670.

Kerr, W. C., Lui, C. and Ye, Y. (2017), ‘Trends and age, period and cohort effects for marijuana use

prevalence in the 1984-2015 US National Alcohol Surveys’,

Addiction

113, pp. 473-481.

https://doi.org/10.1111/add.14031.

Kerr, W. C., Ye, Y., Subbaraman, M. S., Williams, E. and Greenfield, T. K. (2018), ‘Changes in

marijuana use across the 2012 Washington state recreational legalization: is retrospective

assessment of use before legalization more accurate?’,

Journal of Studies on Alcohol and

Drugs

79, pp. 495-502. https://doi.org/10.15288/jsad.2018.79.495.

Kilmer, B. (2018),

Should Canada ‘start low and go slow’ when it comes to cannabis potency?.

Testimony presented before the Standing Committee on Social Affairs, Science, and

Technology Senate of Canada on May 7, 2018. Available at:

https://www.rand.org/pubs/testimonies/CT492.html.

Kilmer, B. and MacCoun, R. J. (2017), ‘How medical marijuana smoothed the transition to marijuana

legalization in the United States’,

Annual Review of Law and Social Science

13, pp. 181-202.

https://doi.org/10.1146/annurev-lawsocsci-110615-084851.

Kilmer, B. and Pacula, R. L. (2017), ‘Understanding and learning from the diversification of cannabis

supply laws’,

Addiction

112, pp. 1128-1135. https://doi.org/10.1111/add.13623

Kilmer, B., Caulkins, J., Pacula, R. L., MacCoun, R. J. and Reuter, P. H. (2010),

Altered State?,

RAND Corporation, Santa Monica, CA,

http://www.rand.org/pubs/occasional_papers/OP315.html, accessed 31 March 2016.

Kilmer, B., Caulkins, J. P., Midgette, G., Dahlkemper, L., MacCoun, R. J. and Pacula, R. L. (2013),

Before the grand opening,

RAND Corporation, Santa Monica, CA.

Kilmer, B., Caulkins, J. P. and Everingham, S. S. (2014),

What America’s users spend on illegal

drugs, 2000-2010,

RAND Corporation, Santa Monica, CA.

Kilmer, B., Reuter, P.and Giommoni, L. (2015), ‘What can be learned from cross-national

comparisons of data on illegal drugs?’,

Crime and Justice

44, pp. 227-296.

https://doi.org/10.1086/681552.

Kim, H. S. and Monte, A. A. (2016), ‘Colorado cannabis legalization and its effect on emergency care’,

Annals of Emergency Medicine

68, pp. 71-75.

https://doi.org/10.1016/j.annemergmed.2016.01.004.

Kim, H. S., Hall, K. E., Genco, E. K., Van Dyke, M., Barker, E. and Monte, A. A. (2016), ‘Marijuana

tourism and emergency department visits in Colorado’,

New England Journal of Medicine

374, pp. 797-798. https://doi.org/10.1056/NEJMc1515009.

Lane, T. J. and Hall, W. (2019), ‘Traffic fatalities within US states that have legalized recreational

cannabis sales and their neighbours’,

Addiction

114(5), pp. 847-856.

Lankenau, S. E., Fedorova, E. V., Reed, M., Schrager, S. M., Iverson, E. and Wong, C. F. (2017),

‘Marijuana practices and patterns of use among young adult medical marijuana patients and

non-patient marijuana users’,

Drug and Alcohol Dependence

170, pp. 181-188.

Livingston, M. D., Barnett, T. E., Delcher, C. and Wagenaar, A. C. (2017), ‘Recreational cannabis

legalization and opioid-related deaths in Colorado, 2000-2015’,

American Journal of Public

Health

107, pp. 1827-1829. https://doi.org/10.2105/AJPH.2017.304059.

Lopez, G. (2018),

Mexico just took a big step toward marijuana legalization,

Vox,

https://www.vox.com/world/2018/11/1/18051592/mexico-marijuana-legalization-supreme-

court, accessed on 29 May 2019.

MacCoun, R. J. (1993), ‘Drugs and the law: a psychological analysis of drug prohibition’,

Psychological bulletin

113, p. 497.

Mair, C., Freisthler, B., Ponicki and W. R., Gaidus, A. (2015), ‘The impacts of marijuana dispensary

density and neighborhood ecology on marijuana abuse and dependence’,

Drug and Alcohol

Dependence

154, pp. 111-116.

Makin, D. A., Willits, D. W., Wu, G., DuBois, K. O., Lu, R., Stohr, M. K., Koslicki, W., Stanton, D.,

Hemmens, C., Snyder, J. and Lovrich, N. P. (2018), ‘Marijuana legalization and crime

clearance rates: testing proponent assertions in Colorado and Washington state’,

Police

Quarterly

22, pp. 31-55. https://doi.org/10.1177/1098611118786255.

Mason, W. A., Fleming, C. B., Ringle, J. L., Hanson, K., Gross, T. J. and Haggerty, K. P. (2016),

‘Prevalence of marijuana and other substance use before and after Washington state’s

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

change from legal medical marijuana to legal medical and non-medical marijuana: cohort

comparisons in a sample of adolescents’,

Substance Abuse

37, pp. 330-335.

https://doi.org/10.1080/08897077.2015.1071723.

Matias, J., Kalamara, E., Mathis, F., Skarupova, K., Noor, A., Singleton, N. and the European Web

Survey on Drugs Group (2019), ‘The use of multi-national web surveys for comparative

analysis: lessons from the European Web Survey on Drugs’,

International Journal of Drug

Policy

73, pp. 235-244. https://doi.org/10.1016/j.drugpo.2019.03.014.

McCarthy, J. (2018),

Two in three Americans now support legalizing marijuana,

Gallup

https://news.gallup.com/poll/243908/two-three-americans-support-legalizing-marijuana.aspx

McGinty, E. E., Niederdeppe, J., Heley, K. and Barry, C. L. (2017), ‘Public perceptions of arguments

supporting and opposing recreational marijuana legalization’,

Preventive Medicine

99, pp. 80-

86. https://doi.org/10.1016/j.ypmed.2017.01.024.

Miller, A. M., Rosenman, R. and Cowan, B. W. (2017), ‘Recreational marijuana legalization and

college student use: Early evidence’,

SSM — Population Health

3, pp. 649-657.

https://doi.org/10.1016/j.ssmph.2017.08.001.

Miroff, N. (2017), ‘In Uruguay’s marijuana experiment, the government is your pot dealer’,

The

Washington Post,

https://www.washingtonpost.com/world/the_americas/in-uruguays-

marijuana-experiment-the-government-is-your-pot-dealer/2017/07/07/6212360c-5a88-11e7-

aa69-3964a7d55207_story.html, accessed on 5 December 2019.

National Conference of State Legislatures (2019),

State medical marijuana laws,

National Conference

of State Legislatures, http://www.ncsl.org/research/health/state-medical-marijuana-laws.aspx,

accessed on 17 September 2019.

Newsom, G., Humphreys, K. and Soltani, A. (2015),

Pathways report: policy options for regulating

marijuana in California,

Blue Ribbon Commission on Marijuana Policy,

www.safeandsmartpolicy.org https://www.safeandsmartpolicy.org/wp-

content/uploads/2015/07/BRCPathwaysReport.pdf.

Orens, A., Light, M., Lewandowski, B., Rowberry, J. and Saloga, C. (2018),

Market size and demand

for marijuana in Colorado: 2017 market update,

Marijuana Policy Group, Denver, CO.

Ouellet, M. L., Macdonald, M., Bouchard, M., Morselli, C. and Frank, R. (2017),

The price of cannabis

in Canada,

Public Safety Canada (Sécurité publique Canada), Ottawa.

Pacula, R. L. and Smart, R. (2017), ‘Medical marijuana and marijuana legalization’,

Annual Review of

Clinical Psychology

13, pp. 397-419.

Pacula, R. L., Jacobson, M. and Maksabedian, E. J. (2016), ‘In the weeds: a baseline view of

cannabis use among legalizing states and their neighbours’,

Addiction

111, pp. 973-980.

https://doi.org/10.1111/add.13282.

Pacula, R. L., Powell, D., Heaton, P. and Sevigny, E. L. (2015), ‘Assessing the effects of medical

marijuana laws on marijuana use: the devil is in the details’,

Journal of Policy Analysis and

Management

34, pp. 7-31. https://doi.org/10.1002/pam.21804.

Pardal, M. (2018),

A study of cannabis social clubs in Belgium. Insights into the supply model, the

movement, and its key actors,

Ghent University, Ghent.

Parnes, J. E., Smith, J. K. and Conner, B. T. (2018), ‘Reefer madness or much ado about nothing?

Cannabis legalization outcomes among young adults in the United States’,

International

Journal of Drug Policy

56, pp. 116-120. https://doi.org/10.1016/j.drugpo.2018.03.011.

Rapoport, E., Keim, S. A. and Adesman, A. (2019), ‘Challenging the association of marijuana laws

with teen marijuana use’,

JAMA Pediatrics

174(1) pp. 98-99.

Rusby, J. C., Westling, E., Crowley, R. and Light, J. M. (2018), ‘Legalization of recreational marijuana

and community sales policy in Oregon: impact on adolescent willingness and intent to use,

parent use, and adolescent use’,

Psychology of Addictive Behaviors

32, pp. 84-92.

https://doi.org/10.1037/adb0000327.

SAMHSA (2018),

Key substance use and mental health indicators in the United States: results from

the 2017 National Survey on Drug Use and Health,

Center for Behavioral Health and

Statistics and Quality, Substance Abuse and Mental Health Services Administration,

Rockville, MD.

Shi, Y., Meseck, K. and Jankowska, M. M. (2016), ‘Availability of medical and recreational marijuana

stores and neighborhood characteristics in Colorado’,

Journal of Addiction,

doi:10.1155/2016/7193740.

Smart, R. (2016),

Essays on the effects of medical marijuana laws,

University of California, Los

Angeles.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Smart, R. and Pacula, R. L. (2019), ‘Early evidence of the impact of cannabis legalization on cannabis

use, cannabis use disorder, and the use of other substances: Findings from state policy

evaluations’,

The American Journal of Drug and Alcohol Abuse

45(6), pp. 644-663.

Smart, R., Caulkins, J. P., Kilmer, B., Davenport, S. and Midgette, G. (2017), ‘Variation in cannabis

potency and prices in a newly legal market: evidence from 30 million cannabis sales in

Washington state’,

Addiction

112, pp. 2167-2177. https://doi.org/10.1111/add.13886.

Sokoya, M., Eagles, J., Okland, T., Coughlin, D., Dauber, H., Greenlee, C. and Winkler, A. A. (2018),

‘Patterns of facial trauma before and after legalization of marijuana in Denver, Colorado: A

joint study between two Denver hospitals’,

The American Journal of Emergency Medicine

36,

pp. 780-783. https://doi.org/10.1016/j.ajem.2017.10.014.

Subbaraman, M. S. and Kerr, W. C. (2016), ‘Marijuana policy opinions in Washington state since

legalization: would voters vote the same way?’,

Contemporary Drug Problems

43, pp. 369-

380. https://doi.org/10.1177/0091450916667081.

Subramaniam, V. (2019),

‘Absolutely horrible’: Cannabis industry expectations sliding after slow start

to legal era,

Financial Post, https://business.financialpost.com/cannabis/cannabis-

business/cannabis-investing/absolutely-horrible-cannabis-industry-expectations-being-

tempered-after-slow-start-to-legal-era, accessed on 5 December 2019.

Subritzky, T., Pettigrew, S. and Lenton, S. (2016), ‘Issues in the implementation and evolution of the

commercial recreational cannabis market in Colorado’,

International Journal of Drug Policy

27, pp. 1-12. https://doi.org/10.1016/j.drugpo.2015.12.001.

van Laar, M., Frijns, T., Trautmann, F. and Lombi, L. (2013), ‘Sizing the cannabis market: a demand-

side and user-specific approach in seven European countries’,

Current Drug Abuse Reviews

6, pp. 152-164.

Volkow, N. D., Swanson, J. M., Evins, A. E., DeLisi, L. E., Meier, M. H., Gonzalez, R., Bloomfield,

M. A. P., Curran, H. V. and Baler, R. (2016), ‘Effects of cannabis use on human behavior,

including cognition, motivation, and psychosis: a review’,

JAMA Psychiatry

73, pp. 292-297.

https://doi.org/10.1001/jamapsychiatry.2015.3278.

Wang, G. S., Lait, M.-C. L., Deakyne, S. J., Bronstein, A. C., Bajaj, L. and Roosevelt, G. (2016),

‘Unintentional pediatric exposures to marijuana in Colorado, 2009-2015’,

JAMA Pediatrics

170, e160971-e160971. https://doi.org/10.1001/jamapediatrics.2016.0971.

Wang, G. S., Hall, K., Vigil, D., Banerji, S., Monte, A. and VanDyke, M. (2017), ‘Marijuana and acute

health care contacts in Colorado’,

Preventive Medicine

104, pp. 24-30.

https://doi.org/10.1016/j.ypmed.2017.03.022.

Wang, G. S., Davies, S. D., Halmo, L. S., Sass, A. and Mistry, R. D. (2018), ‘Impact of marijuana

legalisation in Colorado on adolescent emergency and urgent care visits’,

Journal of

Adolescent Health

63, pp. 239-241. https://doi.org/10.1016/j.jadohealth.2017.12.010.

Wen, H. and Hockenberry, J. M. (2018), ‘Association of medical and adult-use marijuana laws with

opioid prescribing for Medicaid enrollees’,

JAMA Internal Medicine

178, pp. 673-679.

https://doi.org/10.1001/jamainternmed.2018.1007.

Wen, H., Hockenberry, J. M. and Cummings, J. R. (2015), ‘The effect of medical marijuana laws on

adolescent and adult use of marijuana, alcohol, and other substances’,

Journal of Health

Economics

42, pp. 64-80.

Wilkins, C. (2016), ‘After the legalisation of cannabis: the Cannabis Incorporated Society (CIS)

regulatory model for recreational cannabis in New Zealand’,

New Zealand Medical Journal

129, pp. 17-20.

Zuccato, E., Chiabrando, C., Castiglioni, S., Bagnati, R. and Fanelli, R. (2008), ‘Estimating community

drug abuse by wastewater analysis’,

Environmental Health Perspectives

116, pp. 1027-1032.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Appendix A: Identification and commentary on ongoing/planned

studies of legalisation

Here, we identify some ongoing or planned studies from the US, Uruguay and Canada that examine

the impacts of cannabis policy changes. This is not a comprehensive list; the goal is to highlight the

variety of studies being conducted, to learn more about cannabis and the consequences of various

policy changes.

US studies

Studies funded by the National Institutes of Health (NIH)

The NIH, largely through the National Institute on Drug Abuse (NIDA), has supported a broad portfolio

of research on cannabinoids and their use in the endocannabinoid system (

). This includes funding

ongoing research into the use of phytocannabinoids, purified cannabinoids, such as CBD and THC,

and synthetic cannabinoids. In FY 2017, the NIH funded research into cannabinoids to the value of

USD 140 million through various research arms, including NIDA (USD 88 million) and the National

Institute on Alcohol Abuse and Alcoholism (NIAAA) (USD 21 million). In addition, the NIH provided an

additional USD 36 million for research into therapeutic cannabinoids, allocating USD 16 million to

NIDA and USD 6.5 million to the National Institute on Alcohol Abuse and Alcoholism (NIAAA). The

NIH also provided an additional USD 15 million for research into CBD, USD 11 million of which went

to NIDA.

NIDA is currently undertaking several studies that are specific to cannabis use and the impacts of

legalisation. The most ambitious study is the Longitudinal Study of Adolescent Brain Cognitive

Development (the ABCD study), which aims to enrol almost 12 000 healthy children aged 9 and 10

years (singletons and twins) across the US and follow them into early adulthood. According to NIDA,

this is the largest long-term study of brain development and child health in the US.

The study, which is being carried out in partnership with the NIAAA, will examine how biology and

environmental factors relate to developmental and physical outcomes, including mental health and life

achievements. Researchers will employ brain imaging technology to evaluate changes over time.

Specific to cannabis legalisation, the study will examine the impact of the changing state and local

policies on youth drug use and related development. For example, it will examine the extent to which

casual or regular use of cannabis during adolescence has an impact on neurodevelopment in

adulthood.

The study is still currently enrolling participants. Since September 2016, researchers have enrolled

just over 10 000 participants. Baseline data obtained from the first 4 500 participants, suggest that

drug use among participants is minimal.

One active study by NIDA is looking at medical cannabis use among primary care patients in states

with legal medical and recreational cannabis laws. The aims of the Medical Cannabis Use among

Primary Care Patients study are to better understand the medical use of cannabis in a single, large

health system that asks patients about their cannabis use and to describe their cannabis use

behaviours and compare them with others who use cannabis without medical recommendations.

Studies funded by the National Institute of Justice (NIJ)

The NIJ is currently funding multiple studies that examine the impacts of cannabis in the US, including

( ) See

https://www.drugabuse.gov/drugs-abuse/marijuana/nih-research-marijuana-cannabinoids

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

one study examining the effects of Colorado’s REC law on DUIs and crime (

). The NIJ has also

funded a USD 1 million study on the impacts of marijuana legislation on law enforcement and crime in

Washington state.

Studies funded by US states

Colorado (

) state has granted USD 9 million of public funds to cannabis research, dispersing the

funds across nine studies. The studies focus on the health benefits of using medical cannabis to treat

inflammatory bowel syndrome, Parkinson’s disease, epilepsy (two studies), post-traumatic stress

disorder (PTSD) (two studies), palliative care, insomnia and pain relief. The state has also funded

USD 2.4 million of research into the public health effects of REC, focusing on driving impairment, the

acute effects of dabbing (

), the concentration of cannabinoids in breast milk, use in older

populations, the adverse effects of edibles, cardiovascular effects and a general pre/post analysis of

broad public health impacts.

Under Washington (

) state’s initiative, a portion of cannabis tax revenue is earmarked for research.

For the period 2015-17, the Alcohol and Drug Abuse Institute of the University of Washington

received USD 454 000 from the state’s Dedicated Marijuana Fund. Current and ongoing research

focuses on areas of epidemiology, chronic pain management, impacts on driving and the efficacy of

innovative prevention programmes.

Uruguay

There are two ongoing academic monitoring and evaluation groups. The first is the Monitor Cannabis

Uruguay group, affiliated with the School of Social Science at the University of the Republic of

Uruguay. Monitor Cannabis Uruguay researches ongoing developments related to cannabis in the

country. Current ongoing research focuses on the impact of the law on security. The second

academic group is housed at the Catholic University of Uruguay, which maintains the Latin American

Marijuana Research Initiative (LAMRI) (

). LAMRI is funded by the Open Society Institute (a non-

governmental organisation that advocates drug policy reform) and evaluates ongoing cannabis policy

trends in Uruguay and the rest of the region. LAMRI-affiliated researchers have published papers on

consumption patterns, cannabis user opinions and cannabis club design in Uruguay.

The Uruguayan Drug Observatory published a preliminary study (

) in 2015 on drug use among

university students to evaluate their opinions, prevalence and source of cannabis. Though the law had

not been fully implemented, researchers included questions on home cultivation.

Canada

The Canadian government (

), through the Canadian Institute of Health Research, has allocated

CAD 1.4 million to study the effects of cannabis legalisation on certain groups and evaluate existing

regulatory models. Funds are to be distributed across 14 studies led by universities and hospitals,

looking at youth prevention, exposure in pregnancy, impaired driving, trajectories of cannabis use,

outcomes related to opioid use, cannabis use in the workplace, cannabis use in secondary schools,

monitoring mental health outcomes and provincial responses to federal legalisation. More recently,

Health Canada released tenders that focused on studies of cryptomarket cannabis sales and public

attitudes.

( ) See

https://www.nij.gov/topics/drugs/Pages/research-projects.aspx

( ) See

https://www.colorado.gov/pacific/cdphe/marijuana-research

( ) Dabbing is the action or practice of inhaling small quantities of a concentrated and vaporized drug, typically cannabis oil or

resin.

( ) See

http://learnaboutmarijuanawa.org/research.htm

( ) See

https://ucu.edu.uy/es/lamri

( ) See

http://www.cicad.oas.org/oid/pubs/UniversityStudyReport_Uruguay_SPA.pdf

( )

https://www.canada.ca/en/institutes-health-research/news/2018/01/cannabis_and_populationhealthresearch.html

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Appendix B: Research papers on the impacts of recreational cannabis laws

Shading indicates presence in text.

Reference

Location and

years evaluated

Control measures

Outcome

Independent

variable(s)

REC-specific

policy change

Data and methods

Findings

Peer-reviewed articles

Aydelotte et

al. (2017)

Washington and

Colorado

2009-15

States with similar

traffic and roadways:

Alabama, Indiana,

Kentucky, Missouri,

South Carolina,

Tennessee, Texas

and Wisconsin

Motor vehicle

crash fatality

REC

REC adoption

Annual number of motor

vehicle fatalities reported in

the Fatality Analysis

Reporting System;

difference-in-difference

approach; random effects

Pre recreational marijuana legalisation,

annual changes in motor vehicle crash

fatality rates for Washington and Colorado

were similar to those for the control states.

Post recreational marijuana legalisation,

changes in motor vehicle crash fatality

rates for Washington and Colorado also did

not significantly differ from those for the

control states (adjusted difference-in-

difference coefficient =

+0.2

fatalities/billion

vehicle miles travelled; 95 % CI

−0.4

+0.9).

In the five years after legalisation, fatal

crash rates increased more in Colorado

and Washington than would be expected

had they continued to parallel crash rates in

the control states (+1.2 crashes/billion

vehicle miles travelled, CI: -0.6 to 2.1,

p = 0.087),

but not significantly so. The

effect was more pronounced and

statistically significant after the opening of

commercial dispensaries (+1.8

crashes/billion vehicle miles travelled, CI:

+0.4 to +3.7,

p = 0.020).

MED laws not associated with any reported

changes in cannabis use. REC laws were

associated with an 8 % decrease (OR,

0.92; 95 % CI, 0.87-0.96).

Aydelotte et

al. (2019)

Washington and

Colorado

2007-17

Hawaii, Montana,

New Mexico, Rhode

Island, Vermont,

Idaho, Kansas,

Nebraska, South

Dakota

Monthly motor

vehicle crash

fatality

REC

implementation

(i.e., when

stores opened)

Annual number of motor

vehicle fatalities reported in

the Fatality Analysis

Reporting System;

difference-in-difference

approach; random effects

Anderson et

al. (2019)

States with MED

or REC laws

States without MED

or REC laws

Past-month

prevalence

rates

MED or REC

adoption

Past-month prevalence rates

for US high school students

from the Youth Risk

Behavioral Survey.

Multivariate logistic

regression for reporting

past-month use.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Bell et al.

(2015)

Location and

years evaluated

Colorado

January 2008-

August 2014

Control measures

Pre/post

Outcome

Burns

Independent

variable(s)

MED

REC

REC-specific

policy change

Pre REC to

MED/REC

enactment and

implementation

Data and methods

Cross-sectional analysis of

American Burn Association’s

National Burn Repository for

hydrocarbon burns

Findings

Twenty-nine patients with butane hash oil

burns were admitted to the local burn

centre during the study period. No cases

presented prior to medical liberalisation, 19

(61.3 %) presented during medical

liberalisation (October 2009 to December

2013) and 12 presented (38.7 %) in 2014,

after legalisation. The majority of patients

were Caucasian (72.4 %) and male

(89.7 %). The median age was 26 years

(range 15-58 years). The median total body

surface area covered by burns was 10 %

(range 1-90 %). The median length of

hospital admission was 10 days.

Many retailers had no security measure to

determine age (41 % in Colorado; 35 % in

Washington). Approximately 61 % of

retailers in Colorado and 44 % in

Washington made health claims about the

benefits of marijuana, including reduction in

anxiety, depression, insomnia and

pain/inflammation. Inferred demographic

characteristics of followers of Weedmaps

on Twitter and Instagram revealed that over

60 % were male and nearly 70 % or more

were aged 20-29 years, yet some (15-

18 %) were under the age of 20.

Longer REC duration (Odds ratio (OR)

vaping: 2.82, 95 % CI 2.24 to 3.55; OR

edibles: 3.82, 95 % CI 2.96 to 4.94) and a

higher dispensary density (OR vaping:

2.68, 95 % CI 2.12 to 3.38; OR edibles:

3.31, 95 % CI 2.56 to 4.26) were related to

a higher likelihood of trying vaping and

edibles. Permitting home cultivation was

related to a higher likelihood (OR 1.93,

95 % CI 1.50 to 2.48) and younger age at

onset (β

−0.30,

95 % CI

−0.45

−0.15)

consuming edibles.

Bierut et al.

(2017)

Washington and

Colorado

June-July 2015

None

Cannabis

advertising

online

REC

None

Cross-sectional sample of

licensed retailers that

advertise through

Weedmaps (n = 146; 89

from Colorado, 57 from

Washington); social media

followers of Weedmaps

(Twitter = 57 752;

Instagram = 2 249);

descriptive analysis and

non-parametric hypothesis

testing (chi-squared test)

Cross-sectional sample of

cannabis-using young

people (14 to 18 years;

= 2 630) surveyed online;

logistic and linear regression

Borodovsky et

al. (2017)

REC states

29 April 2016 to

18 May 2016

Non-REC states

Youth use of

cannabis,

including ever

vaping or ever

using edibles

REC,

operational

dispensary,

home

cultivation

Point in time

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Cerdá et al.

(2017)

Location and

years evaluated

Washington and

Colorado

2010-12 and

2013-15

Control measures

Non-REC states

Outcome

Adolescent

past-month

marijuana

prevalence of

use and

perception

Independent

variable(s)

REC

REC-specific

policy change

Pre REC to REC

enactment and

implementation

Data and methods

Repeated cross-sectional

MTF surveys (n = 253 902);

difference-in-difference

approach to estimate

perception/use

Findings

In Washington, perceived harmfulness

declined by 14.2 % and 16.1 % among 8th

and 10th graders, respectively, while

marijuana use increased by 2.0 % and

4.1 % during the periods 2010-12 and

2013-15, respectively. In contrast, among

states that did not legalise REC, perceived

harmfulness decreased by 4.9 % and 7.2 %

among 8th and 10th graders, respectively,

and marijuana use decreased by 1.3 % and

0.9 % over the same periods. Difference-in-

difference estimates comparing

Washington with states that did not legalise

REC indicated that these differences were

significant for perceived harmfulness (8th

graders

−9.3

% (SD 3.5 %),

= 0.01; 10th

graders

−9.0

% (SD 3.8 %),

= 0.02) and

marijuana use (8th graders: % [SD], 5.0

[1.9];

= 0.03; 10th graders 3.2 % (SD

1.5 %),

= 0.007). No significant

differences were found regarding perceived

harmfulness or marijuana use among 12th

graders in Washington or for any of the

three grades in Colorado.

Among respondents aged 12 to 17 years,

past-year CUD increased from 2.18 % to

2.72 % after REC enactment, a 25 %

higher increase than that for the same age

group in states that did not enact REC

(odds ratio [OR], 1.25; 95 % CI, 1.01-1.55).

Unmeasured confounders would need to

be more prevalent in REC states and

increase the risk of cannabis use by 1.08 to

1.11 times to explain observed results,

indicating results that are sensitive to

omitted variables. No associations were

found among the respondents aged 18 to

25 years. Among respondents 26 years or

older, past-month marijuana use after REC

enactment increased from 5.65 % to

7.10 % (OR, 1.28; 95 % CI, 1.16-1.40),

past-month frequent use from 2.13 % to

2.62 % (OR, 1.24; 95 % CI, 1.08-1.41), and

past-year CUD from 0.90 % to 1.23 % (OR,

1.36; 95 % CI, 1.08-1.71).

Cerdá et al.

(2019)

Washington,

Colorado, Oregon

and Alaska

2008 to 2016

Non-REC states

Adolescent

past-month

marijuana

prevalence of

use and self-

reported

cannabis use

disorder

REC

3-level variable

determining

never passed

REC; before

REC enactment,

after REC

enactment

Repeated cross-sectional

NSDUH surveys; difference-

in-difference approach

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Cheng et al.

(2018)

Location and

years evaluated

Colorado localities

January 2010-

August 2015

Control measures

Outcome

Housing

values

Independent

variable(s)

REC

REC-specific

policy change

Data and methods

Difference-in-difference

ordinary least squares

estimation

Findings

Legalisation leads to an average 6 %

increase in housing values, indicating that

the capitalised benefits outweigh the costs.

In addition, we find suggestive evidence

that this relatively large housing value

appreciation is probably because REC has

induced strong housing demand while

having no discernible effect on housing

supply. Finally, authors show that the effect

of REC is heterogeneous across locations

and property types.

Weighted estimates showed that 39 % of

the full sample and 9 % of non-marijuana

users supported marijuana legalisation.

Multivariable models showed that lower

marijuana harm perceptions and lifetime

and past 30-day tobacco use were

common predictors of support for marijuana

legalisation and intentions to use marijuana

among non-users of marijuana. State-level

marijuana policy was not associated with

the level of agreement for marijuana

legalisation.

Prevalence of past-year driving while under

the influence of marijuana was 43.6 %

among respondents. The prevalence of

driving within 1 hour of using marijuana at

least five times in the past month was

23.9 %. Increased perception that driving

high is unsafe was associated with lower

odds of past-year marijuana DUIs

(OR = 0.31,

< 0.01) and lower past-month

odds of driving five or more times within 1

hour of using marijuana (OR = 0.26,

< 0.01).

Cohn et al.

(2017)

October 2014

None

Young adult

opinion for

REC

Demographic

Point in time

Nationally representative

cross-sectional sample of

adults aged 18-34 years

from the Truth Initiative

Young Adult Cohort survey

(n = 3 532); multinomial

logistic regression

Davis et al.

(2016)

Colorado and

Washington

September 2014

None

Self-reported

driving while

intoxicated or

within 1 hour

of consuming

marijuana

Demographic

variables

Point in time

Online survey of past-month

users (n = 865; 399 from

Colorado, 446 from

Washington);

logistic regression

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Dilley et al.

(2017)

Location and

years evaluated

Washington

June 2016

Control measures

None

Outcome

Local-level

regulations

and public

opinion

Independent

variable(s)

REC

REC-specific

policy change

Point in time

Data and methods

Descriptive assessment of

marijuana-related municipal

and county ordinances

(n = 181)

Findings

A total of 125 cities and 30 counties had

passed local ordinances to address

recreational marijuana retail sales. Multiple

communities implemented retail market

bans, including some temporary bans

(moratoria), while considering whether or

not to pursue other policy options. As of 30

June 2016, 30 % of the state population

lived in places that had temporarily or

permanently banned retail sales.

Communities frequently enacted zoning

policies that explicitly regulated where

marijuana businesses could be established.

Other policies included in ordinances

placed limits on business hours and

distance requirements (buffers) between

marijuana businesses and youth-related

land use types or other sensitive areas.

HYS shows statistically significant declines

in prevalence from 2010-2012 to 2014-

2016 among both 8th graders (from 9.8 %

[95 % CI, 9.1 %-10.5 %] to 7.3 % [95 % CI,

6.6 %-8.0 %];

P < .001)

and 10th graders

(from 19.8 % [95 % CI, 18.6 %-21.0 %] to

17.8 % [95 % CI, 16.7 %-18.9 %];

P = .01).

Findings indicated a significantly positive

correlation between marijuana-related

consequences and perceived risk post

legalisation. Despite relatively equal use in

both groups, adolescents in the legalisation

group experienced higher levels of

perceived risk and increased negative

consequences.

Current use increased among adults living

in areas within 18 miles of a retailer and,

especially, within 0.8 miles (odds ratio

[OR] = 1.45; 95 % confidence interval

[CI] = 1.24, 1.69). Frequent use increased

among adults living within 0.8 miles of a

retailer (OR = 1.43; 95 % CI = 1.15, 1.77).

Dilley et al.

(2019)

Washington

2010-12 and

2014-16

Pre/post REC

Past month

prevalence

Washington Healthy Youth

Survey of 8th, 10th, and

12th graders

Estoup et al.

(2016)

Washington

2010-15

Pre-REC cohort

Perceived risk

and frequency

of use of

marijuana in

adolescents

REC

No date or year

specified

Self-referred students with

problematic drug use,

enrolled in school-base

substance use intervention

(n = 262; 144 pre-

legalisation); mediation

model with non-parametric

hypothesis testing

Washington Behavioral Risk

Factor Surveillance System

(BRFSS)

Everson et al.

(2019)

Washington

2009-16

None. Proximity to

dispensary

Past month

prevalence

REC

REC enactment

and proximity to

nearest retail

outlet to

respondent ZIP

code

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Fiala et al.

(2017)

Location and

years evaluated

Oregon

November 2015

and April-May

2016

Control measures

Pre/post

Outcome

Viewing

marijuana

promotion

Independent

variable(s)

Demographic

variable and

county of

residence

REC-specific

policy change

None

Data and methods

Repeated cross-sectional

online survey (n = 4 001) of

adults (18+ years); non-

parametric hypothesis

testing (chi-squared test).

Findings

More than half of adults (54.8 %) state-wide

reported seeing marijuana advertising in

the past month. These adults reported that

they most frequently saw store-front

(74.5 %), street-side (66.5 %) and billboard

(55.8 %) advertising. Exposure did not

significantly differ by participant’s age or

marijuana use but was higher among those

living in counties with retail sales (56.5 %)

than among those living in counties without

(32.5 %).

Independent of the effects of covariates,

densities of marijuana outlets were

unrelated to property and violent crimes in

local areas. However, the density of

marijuana outlets in spatially adjacent

areas was positively related to property

crime in spatially adjacent areas over time.

Furthermore, the density of marijuana

outlets in local and spatially adjacent block

groups was related to higher rates of

marijuana-specific crime. This study

suggests that the effects of the availability

of marijuana outlets on crime do not

necessarily materialise within the specific

areas where these outlets are located but

may materialise in adjacent areas.

Women who completed the intervention

after marijuana legalisation were

significantly more likely (OR = 2.1,

< 0.0001) to report marijuana use on

exiting the programme than women who

completed the intervention before

marijuana legalisation. Across both cohorts

(pre and post legalisation), the authors

found a positive link between marijuana

use on exit and alcohol, illegal methadone,

other opioids, amphetamines and cocaine

use.

Freisthler et

al. (2017)

Denver census

tracts

January 2013-

October 2015

None

Violent,

property and

marijuana-

specific crime

(crimes that

involved

marijuana and

licensed

facilities)

REC

MED/REC

dispensary

density over

time

Denver Police data of 481

census blocs over 34

months (n = 16 354);

Bayesian Poisson space-

time model

Grant et al.

(2017)

Washington

February 2001 to

July 2015

Pre-REC cohort

Self-reported

past-month

use of

marijuana in

pregnant

women

REC

Pre REC to REC

enactment and

implementation

Women enrolled in the

Parent-Child Assistance

Program (n = 1 359; pre

REC = 997, post

REC = 362); non-parametric

hypothesis testing (chi-

squared test)

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Hall, K. E. et

al. (2018)

Location and

years evaluated

Colorado

2012-14

Control measures

Pre/post

Outcome

ED episodes

Independent

variable(s)

REC

REC-specific

policy change

Pre-REC to REC

implementation

Data and methods

Colorado Hospital

Association data

Findings

State-wide data demonstrated a fivefold

rise in the prevalence of mental health

diagnoses among cannabis-associated ED

visits (prevalence rate = 5.35, 95 % CI 5.27

to 5.43), compared with ED visits not

related to cannabis. The hospital

subpopulation supported this finding with a

fourfold rise in the prevalence of psychiatric

complaints among ED visits attributable to

cannabis (prevalence rate = 4.87, 95 %

CI 4.36 to 5.44), compared with ED visits

that were not attributable to cannabis.

State-wide rates of ED visits associated

with both cannabis and mental health

significantly increased from 224.5 per

100 000 in 2012 to 268.4 per 100 000 in

2014 (p < 0.0001).

Results indicate that there were no impacts

on the prices paid for medical or

recreational marijuana by state

representative residents within the short 4-

to 5-month window following legalisation.

However, there were differences in how

much people paid for marijuana for

recreational purposes from a recreational

store.

The prevalence of marijuana use in

Colorado college students is much higher

than the national average (71 % v 39 %,

respectively;

< 0.001), especially the

percentage of daily or almost daily users

(25 % v 2 %, respectively;

< 0.001).

There were significant differences found

between non-users of marijuana and the

marijuana users that use once a week or

more but not daily in regard to the grade

point average (F

(6,227)

= 2.935,

< 0.001).

In addition, it seems that the relationship

between alcohol and marijuana use in

general has decreased since Amendment

64 was passed but not among binge

drinkers.

Rates of Oregon college students’

marijuana use increased (relative to that of

students in other states) following

recreational marijuana legislation in 2015

but only for those who reported recent

heavy use of alcohol.

Hunt and

Pacula (2017)

Colorado and

Washington

October 2013,

May 2014,

October 2014

Pre-REC cohort

Reported retail

price

REC/MED

REC enactment

to REC

implementation

Longitudinal survey of three

waves (RAND Marijuana

Use in West Coast States

Survey) (n = 317 past-month

users);

Weedmap data on price

(n = 3 802); difference-in-

difference approach

Jones et al.

(2018)

Colorado

October 2013 to

March 2015

US (whether REC

was in place or not)

Frequency of

marijuana use

in college

students

REC

REC enactment

to REC

implementation

Repeated cross-sectional

surveys (n = 1 413; pre

REC = 424; post

REC = 989), compared with

responses from National

College Health Assessment

for 2015; parametric

hypothesis testing, ANOVA

Kerr, D. et al.

(2017)

Students in one

large public

university in

Oregon

2012-16

Students from six

universities in non-

REC states

Changes in

marijuana,

alcohol and

cigarette use

in college

students

REC

Pre REC to REC

enactment and

implementation

in July 2015

Repeated cross-sectional

survey (n = 10 924) of

undergraduates; mixed-

effects logistic regression

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Kerr, D. et al.

(2018)

Location and

years evaluated

University

students in

Oregon

2008-16

Control measures

Students from 123

other universities in

non-REC states

Outcome

Changes in

cannabis use

prevalence

Independent

variable(s)

REC

REC-specific

policy change

REC

implementation

Data and methods

Repeated cross-sectional

survey in Oregon (n = 7 412)

and 123 institutions (n =

274 340)

Findings

Following REC, Oregon students

(compared to non-REC-state students)

showed relative increases in rates of

marijuana use (odds ratio [OR] = 1.29,

95 % confidence interval [CI: 1.13, 1.48],

= .0002, and decreases in tobacco use

rates (OR = .71, 95 % CI [.60, .85],

< .0001).

A small increase of 1.2 percentage points

in past-year use prevalence, from 24.3 %

(22.3-26.5 %) to 25.6 % (23.6-27.6 %),

which is not statistically significant, was

found when combining the surveys. No

statistically significant change was found in

the prevalence of the simultaneous use of

cannabis and alcohol — it decreased from

12.9 % (11.3-14.7 %) to 12.6 % (11.0-

14.4 %).

Period effects were the main driver of rising

marijuana use prevalence. Models

including indicators of medical and

recreational marijuana policies did not find

any significant positive impacts.

The prevalence of hospitalisations for

marijuana exposure in patients aged 9

years and more than doubled after the

legalisation of medical marijuana (from 15

per 100 000 hospitalisations during the

period 2001-09 to 28 per 100 000

hospitalisations during the period 2010-13;

< 0.001), and ED visits nearly doubled

after the legalisation of recreational

marijuana (from 22 per 100 000 ED visits

during the period 2010-13 to 38 per

100 000 ED visits during January to June

2014;

< 0.001).

Kerr, W. C. et

al. (2018)

Washington

2012, 2014-15

None

Self-reported

past-year use

in survey year

and in 2012 in

adults

REC

Pre REC to REC

enactment and

implementation

Cross-sectional

representative phone

surveys (n = 3 451);

multinomial logistic

regression

Kerr, W. C. et

al. (2017)

Respondents in

states with

REC/MED policy

1984-2015

Respondents in

states without

REC/MED policy

Past-year

marijuana use

MED, REC,

provision for

dispensary,

home

cultivation

MED

REC

REC adoption

Repeated cross-sectional

quinquennial National

Alcohol Survey of adults

(18+) (n = 37 359);

fixed-effects approach for

age-period-cohort effects

Retrospective data from the

Colorado Hospital

Association, a consortium of

more than 100 hospitals in

the state; parametric

hypothesis testing

Kim and

Monte (2016)

Colorado

2001-14

Pre/post

ED episodes

for cannabis

Pre REC to

MED/REC

enactment and

implementation

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Kim et al.

(2016)

Location and

years evaluated

Colorado

2011-14

Control measures

Pre/post

Outcome

ED episodes

for cannabis

Independent

variable(s)

REC

REC-specific

policy change

Pre REC to

MED/REC

enactment and

implementation

Data and methods

Retrospective data from

Colorado Hospital

Association

Findings

The data from the Colorado Hospital

Association did not show a significant

change from 2011 to 2012 in the rate of ED

visits with ICD-9 codes of cannabis use

among out-of-state residents; however,

from 2012 to 2014, the state-wide rate

among out-of-state residents rose from 78

per 10 000 visits in 2012 to 112 per 10 000

visits in 2013 and 163 per 10 000 visits in

2014 (rate ratios, 1.44 (2012-13) and 1.46

(2013-14);

< 0.001 for both comparisons).

Among Colorado residents, from 2011 to

2014, the rate of ED visits possibly related

to cannabis use increased from 61 to 70,

86 and 101, respectively, per 10 000 visits

(rate ratios, 1.14 (2011-12), 1.24 (2012-13)

and 1.17 (2013-14);

< 0.001 for all

comparisons).

There was a pooled step increase of 1.08

traffic fatalities per million residents

followed by a trend reduction of

−0.06

per

month (both

< 0.001). The results

suggest that legalizing the sale of cannabis

for recreational use can lead to a temporary

increase in traffic fatalities in legalizing

states that can spill over into neighbouring

jurisdictions.

Lane and Hall

(2019)

Colorado,

Washington and

Oregon

2009-16

Kansas, Nebraska,

New Mexico,

Oklahoma and Utah

(Colorado

neighbours); British

Columbia and

Oregon (Washington

neighbours); and

California and

Nevada (Oregon

neighbours)

Pre/post

Monthly motor

vehicle

fatalities

REC

implementation

in each state

Interrupted time series using

CDC WONDER data on

vehicle fatalities

Livingston et

al. (2017)

Colorado

2000-15

Monthly

opioid-related

fatalities

REC

implementation

January 2014

Interrupted time series

Colorado’s legalisation of recreational

cannabis sales and use resulted in a 0.7

deaths-per-month (b =

−0.68;

95 % CI =

−1.34

−0.03)

reduction in opioid-related

deaths. This reduction represents a

reversal of the upwards trend in opioid-

related deaths in Colorado.

Findings suggest there are no negative

effects of legalisation on crime clearance

rates. Moreover, evidence suggests that

some crime clearance rates have

improved. Our findings suggest legalisation

has resulted in improvements in some

clearance rates.

Makin et al.

(2018)

Colorado and

Washington

2010-15

The rest of the US

as one unit

Violent and

property crime

clearance

rates

REC

Pre REC to REC

enactment

Monthly counts of crime

from FBI Uniform Crime

Reporting data; interrupted

time-series analysis

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Mason et al.

(2016)

Location and

years evaluated

Washington

2010/11 to

2013/14

Control measures

Pre-REC cohort

Outcome

Self-reported

past-month

use in

adolescents

Independent

variable(s)

Second survey

wave (just

prior to REC)

REC-specific

policy change

Pre REC to REC

enactment

Data and methods

Two wave cohort study of

8th/9th graders from

Tacoma, Washington

(n = 238); multivariate,

multilevel modelling

Nationally representative

sample of US adults

(n = 979; 334 from Colorado,

Washington, Arkansas,

Oregon, DC);

ordered logit and non-

parametric hypothesis

testing (chi square)

Findings

Marijuana use was higher for the second

cohort than for the first cohort, but this

difference was not statistically significant.

Rates of cigarette and alcohol use were

slightly lower in the second cohort than in

the first cohort.

Respondents rated pro-legalisation

arguments highlighting beneficial economic

and criminal justice consequences as more

persuasive than anti-legalisation

arguments, emphasising adverse public

health effects. Respondents were more

likely to agree with arguments highlighting

legalisation’s potential to increase tax

revenue (63.9 %) and reduce prison

overcrowding (62.8 %) than arguments

emphasising the negative consequences of

motor vehicle crashes (51.8 %) and youth

health (49.6 %). The highest rated anti-

legalisation arguments highlighted the

conflict between state and federal

marijuana laws (63.0 %) and asserted that

legalisation will fail to eliminate the illicit

market (57.2 %). Respondents who

endorsed pro-legalisation economic and

criminal justice arguments were more likely

than other respondents to support

legalisation. Respondents living in

Arkansas, Colorado, DC, Oregon and

Washington were significantly (p < 0.05)

more likely to agree with 11 of the 13 pro-

legalisation arguments and significantly

less likely to agree with 10 of the 13 anti-

legalisation arguments than respondents

living in non-legalisation states.

Students at Washington State University

experienced a significant increase in

marijuana use after legalisation. This

increase is larger than the increase that

would be predicted by national trends. The

change is strongest among females, Black

students and Hispanic students. The

increase for underage students is as much

as that for legal-age students. We find no

corresponding changes in the consumption

of tobacco, alcohol or other drugs.

McGinty et al.

(2017)

Respondents in

states that passed

REC

April 2016

Respondents in

states that did not

pass REC

Public opinion

and

arguments

for/against

legalisation

REC

Point in time

Miller et al.

(2017)

Washington

2005-15

Pre-REC cohort

Past-month

marijuana use

in college

students

REC

Pre REC to REC

enactment and

implementation

Repeated cross-sectional

surveys of undergraduates

at Washington State

University who participated

in the National College

Health Assessment

(n = 13 335); logit and OLS

regression

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Pacula et al.

(2016)

Location and

years evaluated

Colorado and

Washington

October 2013

Control measures

Oregon and New

Mexico

Outcome

Prevalence

rates and use

behaviours

between MED

and REC

users

Independent

variable(s)

REC

REC-specific

policy change

Point-in-time

measure during

REC enactment

Data and methods

Cross-sectional sample of

representative phone survey

(n = 1 994;

Washington = 787,

Oregon = 506;

Colorado = 503; New

Mexico = 213)

Findings

Recreational use is considerably higher

than medical use across all states (41 %),

but it is the highest in Oregon and

Washington. Approximately 86 % of people

who report ever using cannabis for

medicinal purposes also use it

recreationally. Fewer than one in five

recreational users report simultaneous use

of alcohol and cannabis most or all of the

time, and fewer than 3 % of medicinal

users report frequent simultaneous use of

alcohol and cannabis. In the US, the

degree of overlap between medicinal and

recreational cannabis users is 86 %.

Medicinal and recreational cannabis users

favour different modes and amounts of

consumption.

Cannabis use has increased for all

students since recreational legalisation, but

more so for those over 21 years. No

differences in past-month use frequency

were found between pre- and post-

legalisation. The influence of cannabis laws

on non-resident students’ decision to attend

a Colorado college predicted lifetime and

past 30-day use. In addition, out-of-state

students reported higher past 30-day use

than in-state students.

In communities opting out of sales, the

prior-to-legalisation cohort was less likely to

increase their willingness and intent to use

marijuana, and the legalisation cohort was

more likely to increase their intent to use.

For young people who used marijuana,

legalisation was associated with increased

use, and those in communities opting out of

sales experienced a larger growth in

marijuana use. Community policy appears

to impact young people’s attitudes towards

and use of marijuana. The results suggest

that the legalisation of recreational

marijuana did not increase marijuana use

among young people who did not use

marijuana, but it did increase use among

young people who were already using.

Parnes et al.

(2018)

Colorado

Spring 2013 to

autumn 2015

Pre-REC cohort

Self-reported

cannabis use

in college

students

REC

REC enactment

implementation

Undergraduate survey

(n = 5 241); non-parametric

hypothesis test (chi-squared

test)

Rusby et al.

(2018)

Oregon

Spring 2014 and

2015

Pre-REC cohort

Self-reported

cannabis use

in adolescents

REC

REC enactment

implementation

Two cohorts of 8th/9th

graders (n = 444);

multivariate linear regression

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Smart et al.

(2017)

Location and

years evaluated

Washington

July 2014-

September 2016

Control measures

None

Outcome

Changes in

price, potency

and product

Independent

variable(s)

REC

REC-specific

policy change

Post REC

implementation

only

Data and methods

Transaction sale data from

regulators

Findings

Traditional cannabis ﬂowers still account for

the majority of spending (66.6 %), but the

market share of extracts for inhalation

increased by 145.8 % between October

2014 and September 2016, now comprising

21.2 % of sales. The average THC-level for

cannabis extracts is more than triple that

for cannabis ﬂowers (68.7 % compared to

20.6 %). For ﬂower products, there is a

statistically signiﬁcant relationship between

price per gram and both THC [coefﬁcient =

0.012; 95 % conﬁdence interval (CI) =

0.011–0.013] and CBD (coefﬁcient = 0.017;

CI = 0.015–0.019).

The estimated discount elasticity is +0.06

(CI = +0.07 to +0.05). Traditional cannabis

flowers still account for the majority of

spending (66.6 %), but the market share of

extracts for inhalation increased by

145.8 % between October 2014 and

September 2016, now accounting for

21.2 % of sales. The average THC level for

cannabis extracts is more than triple that

for cannabis flowers (68.7 % compared

with 20.6 %). For flower products, there is a

statistically significant relationship between

price per gram and both THC

(coefficient = 0.012; 95 % CI 0.011 to

0.013) and CBD (coefficient = 0.017; 95 %

CI 0.015 to 0.019). The estimated discount

elasticity is –0.06 (95 % CI –0.07 to –0.05).

Subbaraman

and Kerr

(2016)

Washington

January-October

2014

Pre/post

Support for

legalisation

Demographic

variable

Post REC

implementation

Random digit dialling of

residents (n = 2007);

bivariate tests and

multivariate regressions

Less than 5 % of those who voted for

marijuana legalisation would change their

vote, whereas 14 % of those who voted

against legalisation would change their

vote. In multivariable models controlling for

demographics, substance use and

marijuana-related opinions, those who

voted for legalisation had half the odds of

changing their vote than those who voted

against it. Among past-year non-marijuana

users, almost 10 % were somewhat/very

likely to use marijuana if they could buy it

from a legal store.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Sokoya et al.

(2018)

Location and

years evaluated

Denver, Colorado

2012-15

Control measures

Pre/post

Outcome

ED visits

Independent

variable(s)

REC

REC-specific

policy change

Pre REC to

MED/REC

enactment and

implementation

Data and methods

Hospital data from the

University of Colorado

Hospital and Denver Health

Medical Centre;

non-parametric hypothesis

test

Wang et al.

(2016)

Colorado

2009-15

Pre/post

Paediatric

exposures to

marijuana

MED

REC

Pre REC to

MED/REC

enactment and

implementation

Retrospective cohort study

of hospital child admissions

at Children’s Hospital

Colorado and cannabis

exposure incidents reported

to poison control;

Poisson regression

Findings

Maxillary and skull base fracture

proportions significantly increased following

legalisation (p < 0.001 and

< 0.001,

respectively). No significant differences

were seen in the proportion of patients who

lived in urban and rural counties before and

after legalisation (p > 0.05).

The mean rate of marijuana-related visits to

the children’s hospital increased from 1.2

per 100 000 population 2 years prior to

legalisation to 2.3 per 100 000 population 2

years after legalisation (p = 0.02). Annual

poison control paediatric marijuana cases

increased more than fivefold from 2009 (9)

to 2015 (47). Colorado had an average

increase in poison control cases of 34 %

(p < 0 .001) per year, while the remainder

of the US had an increase of 19 %

< 0.001).

From 2000 to 2015, hospitalisation rates

with marijuana-related billing codes

increased from 274 per 100 000

hospitalisations to 593. Overall, the

prevalence of mental illness among ED

visits with marijuana-related codes was five

times higher (5.07, 95 % CI 5.0 to 5.1) than

the prevalence of mental illness among ED

visits without marijuana-related codes.

Poison control calls remained constant

from 2000 to 2009. However, in 2010, after

local medical marijuana policy

liberalisation, the number of marijuana

exposure calls significantly increased, from

42 to 93; in 2014, after recreational

legalisation, calls significantly increased by

79.7 %, from 123 to 221 (p < 0.0001).

Marijuana-related visits increased from 1.8

per 1 000 visits in 2009 to 4.9 in 2015

(p = < 0.0001).

Wang et al.

(2017)

Colorado

2000-15

Pre/post

Annual rates

hospitalisation

s, ED events

and poison

control calls

for marijuana

MED

REC

Pre REC to

MED/REC

enactment and

implementation

Univariate Poisson

regression

Wang et al.

(2018)

Colorado

2005-15

Pre/post

Children ED

visits for

cannabis

REC

MED

Pre REC to

MED/REC

enactment and

implementation

Retrospective review of

annual number of marijuana-

related visits to ED

(n = 4 202)

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Wen and

Hockenberry

(2018)

Location and

years evaluated

Alaska, Colorado,

Oregon and

Washington

Q1 2011 to Q2

2016

Control measures

Non-REC states

Outcome

State-level

opioid

prescribing

rates covered

by Medicaid

Independent

variable(s)

MED

REC

REC-specific

policy change

REC

implementation

Data and methods

State drug utilisation data

from the Centres for

Medicare and Medicaid

Services;

difference-in-difference

approach, two-way fixed

effects

Findings

The state implementation of medical

marijuana laws was associated with a

5.88 % lower rate of opioid prescribing

(95 % CI

−11.55

% to approximately

−0.21

%). Moreover, the implementation of

adult-use marijuana laws, which all

occurred in states with existing medical

marijuana laws, was associated with a

6.38 % lower rate of opioid prescribing

(95 % CI

−12.20

% to approximately

−0.56

%).

Regardless of store type, marijuana stores

were more likely to be located in

neighbourhoods that had a lower proportion

of young people, a larger racial and ethnic

minority population, a lower household

income, a higher crime rate or a greater

density of on-premise alcohol outlets. The

availability of medical and recreational

marijuana stores was differentially

correlated with household income and

racial and ethnic composition.

Shi et al.

(2016)

Census tracts in

Colorado

2015

REC and MED

store density

MED

REC

Cross-sectional ecological

study of all census tracts in

the state (n = 1 249);

parametric hypothesis

testing

Working papers

Hansen et al.

(2017b)

Washington

2015

Pre/post

Inter-state

trafficking after

REC

implementatio

n in Oregon

REC (opening

of retail stores

in Oregon)

REC

implementation

in Oregon

Washington retail

transaction data;

regression discontinuity

design

Washington retailers situated along the

Oregon border experienced a 41 % decline

in sales immediately following Oregon’s

cannabis market opening. In counties that

are the closest crossing point for the

majority of the neighbouring population, the

estimated decline in sales has grown to

58 % and is the largest for the biggest

transactions.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Reference

Hansen et al.

(2018)

Location and

years evaluated

Washington and

Colorado

2000-16

Control measures

Non-REC states in

the US

Outcome

Motor vehicle

crash fatality

Independent

variable(s)

REC

REC-specific

policy change

REC

implementation

in 2014

Data and methods

Annual number of motor

vehicle fatalities reported in

the Fatality Analysis

Reporting System; synthetic

control approach

Findings

Between 2013 and 2016, the number of

drivers who tested positive for THC

increased in Colorado and Washington, by

92 % and 28 % respectively. However,

identifying a causal effect is difficult

because of the presence of significant

confounding factors. Hansen et al. found

that ‘the synthetic control groups saw

similar changes in marijuana-related,

alcohol-related and overall traffic fatality

rates despite not legalising recreational

marijuana’.

There is no conclusive evidence that

marijuana sale/manufacture arrests, DUI

arrests or opium/cocaine possession

arrests in border states are affected by

REC. The NSDUH data show that self-

reported marijuana use in states that

border REC states increased after REC,

compared with those states that do not

share borders with REC states.

Results from single-state analyses as well

as the combination of the three states

indicate that collision claim frequencies

increased significantly when retail sales

commenced. When states are examined

individually, the frequency of collision

claims increases by between 4.5 % and

13.9 %. A single analysis that combined the

three states with legal recreational use

found a smaller yet significant increase of

2.7 %.

Hao and

Cowan (2017)

Counties in

Washington and

Colorado

2009-14

Counties in

neighbouring states

just over the border

Arrests and

self-reported

marijuana use

REC

Pre REC to REC

enactment

FBI Uniform Crime

Reporting Data at county

level; NSDUH state-level

prevalence data; difference-

in-difference approach,

synthetic controls for

robustness check

Highway Loss

Data Institute

(2017)

Colorado,

Washington and

Oregon

January 2012 to

October 2016

Neighbouring states:

Nebraska, Utah,

Wyoming, Montana,

Idaho and Nevada

Auto

insurance

collision claim

rates

REC

implementation

Monthly insurance collision

claims; Poisson regression

and non-parametric

hypothesis testing

REC, recreational marijuana law; MED, medical marijuana law; SD, standard deviation; ANOVA, analysis of variance; OLS, ordinary least squares.

REC adoption: when an initiative was passed by voters but was not in effect.

REC enactment: when laws were put into place to effect a legal change.

REC implementation: when the laws were fully operational and stores were open. Note: advocate-produced reports are not included.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Appendix C: Additional trend data

Prevalence

In the figures below, we plot prevalence rates reported by the NSDUH for various age cohorts and states.

To help the reader, we also include a reference table of legal MED/REC dates of adoption and

implementation, the latter referring to when qualified patients or adults were able to access and use

cannabis. Keep in mind that many early MED states did not have any dispensaries or commercial store

fronts for several years after adopting or implementing MED.

TABLE AC1

Dates of MED/REC adoption and implementation

Medical cannabis

State

Alaska

Colorado

District of Columbia

Oregon

Washington

Adoption

November 1998

November 2000

November 1998

Year that stores

legally opened

2005

2010

2009

Adoption

November 2014

November 2012

November 2014

November 2012

Recreational cannabis

Implementation

October 2016

January 2014

February 2015

October 2015

July 2014

The series of charts below plot the past-month prevalence rate of use of cannabis among the general

population (aged 12 and over) in Alaska, Colorado, DC, Oregon, Washington and the US as a whole

between 2003 and 2017. State estimates were taken from the NSDUH’s 2-year pooled estimates,

assigning the latter year for each of the year (e.g. pooled estimates for 2016 to 2017 represent 2017). To

show the change in prevalence rates relative to policy changes, we have plotted vertical dashed/dotted

lines to indicate the year in which MED dispensaries opened (a dotted line, where applicable), when REC

was adopted (i.e. voted into law; this is represented by a light-grey dashed line) and when REC was

implemented (i.e. when stores started opening; this is represented by a dark-grey dashed line). Because

we are using annual data, these vertical lines roughly approximate the timing of a policy change. For

example, voters of a state may have adopted REC in November 2014, even though the year was nearly

over after the law passed.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

FIGURE AC1

Past-month cannabis use prevalence (%) in Alaska among those aged 12 years and older

2003

2004

2005

2006

Alaska

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

REC Adoption

REC Implement

Source: NSDUH

FIGURE AC2

Past-month cannabis use prevalence (%) in Colorado among those aged 12 years and older

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Colorado

REC Adoption

REC Implement

MED store

Source: NSDUH

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

FIGURE AC3

Past-month cannabis use prevalence (%) in DC among those aged 12 years and older

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

REC Adopt

REC Implement

MED store

Source: NSDUH

FIGURE AC4

Past-month cannabis use prevalence (%) in Oregon among those aged 12 years and older

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Oregon

REC Adopt

REC Implement

MED store

Source: NSDUH

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

FIGURE AC5

Past-month cannabis use prevalence (%) in Washington state among those aged 12 years and

older

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Washington

REC Adopt

REC Implement

Source: NSDUH

Data from the MTF survey show little change in past-month cannabis use among 8th, 10th and 12th

graders (pooled) between 2009 and 2017. In 2009, 13.8 % of respondents reported using cannabis at

least once in the past month. This rate peaked in 2013, at 15.6 %, declining to 14.5 % in 2017. However,

past-month use rates have steadily increased among college students. In 2009, 18.5 % of respondents

reported using cannabis at least once in the past month. By 2016, the rate had climbed to 22.2 %.

Data from biennial state surveys add additional information to the changes in prevalence rates among

high school students post REC. These state surveys randomly sample students from selected middle

schools and high schools. The HKCS sampled about 17 000 students from 157 schools across the state

in 2015. The Washington state’s HYS surveyed over 230 000 students in over 1 000 schools in 2016. In

Table AC2 we report available prevalence estimates for high school students and adults reported in state

surveys in recent years, along with estimates from the Behaviour Risk Factor Surveillance System

(BRFSS) of the Centers for Disease Control and Prevention. However, past-month prevalence rates

among high school students or 12th graders have remained flat or declined in both states.

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

TABLE AC2

Surveys and reported past-month cannabis prevalence rates for Colorado and Washington

Survey

Colorado, past-month (18+) BRFSS

Colorado, daily/near-daily (18+) BRFSS

Colorado, past-month (high school) HKCS*

Washington, past-month (12th grade) HYS

Washington, heavy past-month (12th grade) HYS**

*Estimates are for odd years; otherwise they are even.

**Heavy use is defined as using on 10+ days in the last 30.

2009/10

2011/12

2013/14

13.6

2015/16

13.5

6.35

Hospital and poison control centres

Data from hospitals and poison control centres allow researchers to gauge the initial public health

incidents associated with post-REC implementation. Below we plot the rate of ED events and

hospitalisations involving cannabis. ED episodes have increased since 2011 but not as rapidly as

hospitalisations. Rates remained relatively flat until 2013, when they started to rise just prior to REC

enactment. Data for 2015 are an estimated annual rate, as only counts from January to September are

publicly available. We were unable to find ED or hospitalisation data from Washington state.

FIGURE AC6

Adverse events involving cannabis per 100 000 population in Colorado

4000

3000

2000

1000

2009

2010

2011

2012

Hospitalizations

2013

2014

2015

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

The figure below shows the reported calls to poison control centres involving cannabis per 100 000

residents for both Washington and Colorado. In Colorado since 2011, calls involving cannabis between

2011 and 2014, when they levelled off in 2014 at about 4 per 100 000. Washington saw a similar

increase, with rates rising in 2014, then levelling off for 2 years before rising again in 2016/17 to similar

per capita rates as in Colorado.

FIGURE AC7

Poison control calls per 100 000 population

2009

2010

2011

2012

Colorado

2013

2014

Washington

2015

2016

2017

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Traffic fatalities

The National Highway Traffic Safety Administration’s Fatality Analysis Reporting System (FARS) reports

data on the number of traffic fatalities per 100 million vehicle miles travelled. The rate of traffic fatalities in

most states analysed was fairly linear between 2009 and 2017 (the obvious exceptions are DC, which is

mostly urban, and Alaska, which is mostly rural; both may suffer from extreme year-to-year variability).

However, starting around 2013, Oregon, Colorado and Washington show a marked increase compared

with the rest of the US.

FIGURE AC8

Traffic fatality rate per 100 million vehicle miles travelled

2.0

1.5

1.0

0.5

0.0

2009

2010

Alaska

2011

Colorado

2012

2013

2014

Oregon

2015

Washington

2016

2017

Source: FARS

TECHNICAL REPORT

Monitoring and evaluating changes in cannabis policies: insights from the Americas

Arrests for cannabis possession and supply offences

The figures below show rates of arrest for cannabis offences over time (where available). In all three

states, arrest rates for cannabis offences have declined over the period.

FIGURE AC9

Cannabis possession arrests per 100 000 population

350

300

250

200

150

100

2012

2013

Alaska

2014

Washington

2015

Colorado

2016

FIGURE AC10

Cannabis supply arrests per 100 000 population

2012

2013

Alaska

2014

Washington

2015

Colorado

2016