CANN2021 survey and registry-linkages cohort on cannabis involvement among Norwegian high school students: design, measures and sample characteristics – cohort profile

Abstract

Purpose

CANN2021 is a nationwide cohort of Norwegian high school students created with the aim of addressing emerging issues in epidemiology of cannabis use through the initial surveillance and examination of its correlates, causes and consequences.

Participants

Between 25 February 2021 and 10 April 2021, a core baseline sample of 3490 students (48% boys; 11th grade 35.5%, 12th grade 31.2%, 13th grade 33.3%) from 34 high schools in Norway anonymously completed comprehensive e-surveys assessing their cannabis-related involvement and experiences. A total of 1510 (43.3%) participants (45.8% boys; 11th grade 28.9%, 12th grade 31.1%, 13th grade 40.0%) provided identifying information and consented to administrative contact entailing individual-level linkages of their survey responses to their health and census data as recorded in various national registries since 2010, thus establishing the CANN2021 registry-linkages cohort.

Findings to date

The core baseline sample (N=3490) was largely representative of the Norwegian high school youth between the ages of 17 and 19 years, and as such of relevance to national surveillance needs. One in five (20.3%) reported having used cannabis at least once during their lifetime; of these, 40.9% consented to registry linkages.

Future plans

E-survey data from the registry cohort will be linked at the individual level to health and administrative registries such as the Norwegian Patient Registry, Education, Crime, Income and Population Registry in 2025, 2029 and 2031. The retrospective and prospective linkages of baseline e-surveys with registry data can thus be used to address a range of epidemiological and public health questions, including examination of temporal associations between various types of early cannabis involvement and putative risk and protective factors, and subsequent health and social outcomes.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A sizeable baseline sample largely representative of the Norwegian high school-aged youth, and thus of relevance to public health surveillance of emerging issues in early cannabis involvement.

• Baseline e-survey extended beyond the basic assessments of prevalence and frequency of cannabis use and included multiple modules assessing various dimensions of early cannabis involvement such as cannabis-related attitudes, beliefs, motivations, perceptions and knowledge.

• Individual-level, prospective and retrospective registry-based follow-up of the consented participants, enabling examination of the associated health, education, crime and labour characteristics over time.

• Limitations include a relatively small number of cannabis users in the registry-linkages cohort and its slight departures from national representativeness.

Introduction

Cannabis use during adolescence is associated with increased risk of repeated and high-risk use, use disorder and adverse health and social outcomes that may persist throughout a lifetime.¹,⁸ Consequently, understanding the risk and protective factors, as well as the consequences of early cannabis involvement, remains one of the key public health undertakings. Even though alcohol and tobacco use seem to be on the decline among adolescents from European countries,^{9 10} there are persisting concerns about their cannabis use in the context of rapid shifts in cannabis legislative modes, products and practices.⁹,¹¹ In addition, the extant cohort studies appear to be disproportionately based on the North American and Australian adolescent cohorts,^{2 6} and predominantly on the standard epidemiological indicators (such as the prevalence or frequency of use only) obtained from self-reports despite recent changes in cannabis types, modes and potency.^{2 12 13} Cannabis epidemiology is thus in urgent need of more nuanced investigation of both use and the related phenomena¹⁴,¹⁸ while the relevant longitudinal research can benefit from incorporation of alternative data sources such as the administrative and health registries readily available in Nordic counties.¹⁹,²¹ The hereby presented CANN2021 thus aims to advance epidemiology of cannabis use by establishing the first cohort study focusing primarily on cannabis involvement among high school youth outside the dominant North American/Australian social, cultural and policy contexts.

The persisting concerns about youth cannabis involvement are juxtaposed against the global cannabis liberalisation trends,²² which are challenging and reshaping the traditional cannabis markets, public opinions, individual perceptions, use patterns and (sub)cultures.^{15 23 24} This is first and foremost evident in the young people’s evolving cannabis-narratives, legalisation support and ostensible perceptions of cannabis use as an ordinary activity in many Western societies.²⁵,²⁸ These societal and cultural shifts indicative of cannabis normalisation may translate into its increased availability²⁹ and concrete use opportunities, the first and necessary factor in drug use trajectories.³⁰,³² The settings and reasons of young people’s cannabis use seem to be shifting as well, as American adolescents are increasingly engaging in solitary use,³³ and as a means of coping with negative moods.³⁴ Given these trends, it may not be surprising that many adolescents seem to plan or expect to use cannabis in the future—especially if doing so is no longer illegal.^{35 36} Yet, the nature and extent of cannabis-related attitudes, perceptions, use exposures and opportunities, and (non)use motivations among European youth, including Norway, remain underexamined despite direct relevance of these putative risk and protective factors to public health.

Further, recent changes in cannabis regulatory frameworks are accompanied by the surges of novel products such as edibles or high potency cannabis extracts and novel intake modes such as digestion or vaping.³⁷ These changes are not without consequences for young people’s cannabis use patterns and modes. For example, while smoking remained the primary intake mode, both the ingestion of edible products and vaping were commonly used methods in a national sample of American adolescents.³⁸ Prevalence of past-month cannabis vaping increased by sevenfold between 2013 and 2020³⁹ and by roughly twofold between 2017 and 2019 alone⁴⁰ among adolescents from Canada and the USA, but not as steeply among adolescents from England where cannabis sales remain illegal.⁴⁰ More than 8 out of 10 current users from California consumed edibles in their lifetime, and such users were characterised by more severe risk profiles and use histories.⁴¹ How such new intake modes and ever more potent products⁴² may affect adolescents’ health and development remains a concern,⁴³ especially given the concordance between the increases in delta-9-tetrahydrocannabinol content and the first-time treatment admissions for cannabis-related disorders in general population.⁴⁴ Yet, the extent of adolescents’ involvement with and knowledge of these novel cannabis markets and products is not fully known, underscoring the needs for timely and comprehensive investigation.

In sum, our knowledge of the evolving cannabis involvement among adolescents from Europe, including Norway,⁴⁵ remains limited. While underage cannabis use has been comparably low in Norway,⁴⁶ recent findings also suggest changes in cannabis-related experiences among Norwegian adolescents.³²⁴⁷,⁴⁹ Despite these trends, the national and international monitoring tools such as the UngData and the European School Survey Project on Alcohol and Other Drugs (ESPAD),^{46 48 50} for example, only gather basic epidemiological indicators of cannabis use, information on only a handful of risk factors, and—in case of ESPAD—only among younger, 16-year-old middle school students. These projects, together with a handful of relevant large-scale adolescent cohorts in Norway,^{51 52} were also designed to investigate youth development, health and substance use more generally. In other words, the extant studies are not cannabis-specific and thus cannot effectively capture the highly relevant issues in cannabis epidemiology¹⁴,¹⁶ such as the various types of cannabis exposure and use opportunities,³⁰,³² usage of novel products with potentially differential risk associations with health and social outcomes,^{33 37 43} and evolving attitudes, perceptions and knowledge of cannabis.²⁵,²⁸ Most importantly, while valuable in providing national-level and community-level estimates of cannabis use, the extant monitoring tools^{46 48 50} based on anonymous repeated cross-sectional assessments cannot directly engage or follow up the participants and, therefore, offer only limited inferences concerning causes and consequences of early cannabis involvement.

Against this backdrop, we initiated CANN2021⁵³: a cohort study on cannabis involvement among high school-aged adolescents in Norway within the context of rapidly changing cannabis landscape. The current study aims to advance monitoring of and the research on underage cannabis involvement in two primary ways: first, by providing more complex cannabis-relevant indicators beyond the simple prevalence and frequency of use estimations, and second, by advancing our understanding of their health and social correlates, causes and consequences. These aims also underscore the need for deeper investigation of cannabis involvement during sensitive periods, both at the individual level (ie, developmentally sensitive periods such as adolescence) and at the societal level (ie, historically sensitive periods such as policy and cultural shifts). To the best of our knowledge, this is the first cohort study to focus primarily yet broadly on cannabis involvement and related experiences among youth. Additional study features include:

1. A large baseline sample of high school students facilitating national-level estimations of early cannabis involvement and providing recruitment pool for longitudinal registry cohort.

2. Comprehensive baseline assessment of early cannabis involvement, including cannabis-related experiences, opinions, beliefs, motivations, perceptions, and knowledge.

3. Prospective and retrospective individual-level linkages of the consenting participants’ survey responses with their administrative records across multiple national registries containing health, education, crime and labour market indicators.

Cohort description

Setting

Norway is a Nordic country characterised by high commitment to welfare and well-being of its residents, currently ranking at the top of the Human Development Index. In terms of drug policy, Norway allows only limited and highly regulated sales of a prescription cannabis product for medical purposes.⁵⁴ Any other cannabis sales, production, possession and use remain illegal in Norway despite recent decriminalisation initiatives.⁵⁵,⁵⁷

Norway also provides tuition-free public education from elementary school throughout university, with the initial ten years (grades 1–10) being compulsory. Nevertheless, nearly all adolescents (97.7%) transition into high school at the age of 16 (grade 11) in the form of either 3 years of general studies programme or 4 years of vocational training.⁵⁸

Recruitment

We aimed to obtain a representative sample of high school-aged youth (ie, of 17–19-year-old adolescents) and a sample size comparable to other national youth samples⁴⁶ in order to facilitate national-level estimations and subsequent recruitment for administrative, registry-based follow-up. To achieve this, we collaborated with the Norwegian Gallup (https://kantar.no/om/norsk-gallup/) to develop appropriate sampling strategy and weighting procedures, if needed. In the context of near-complete high school enrolment in Norway, a simple random sampling strategy without stratification or special consideration of socio-geographic characteristics was ultimately selected, with high schools serving as both the sampling frame and the recruitment platform.⁵⁹

Aiming to recruit approximately 30 schools and 2000 students, 56 high schools were randomly selected from 415 high schools registered by the Department of Education in Norway in 2021. To lessen the school burden, CANN2021 targeted the entire grade 11, 12 or 13 cohorts per participating school. These 56 schools with an estimated targeted enrolment of 6800 students across the selected grade cohorts were approached in late 2020 with information letters detailing study design and purpose. A total of 36 schools consented to study participation. However, during the planning phase, two schools withdrew citing the lack of resources, leaving 34 high schools with 4122 eligible students scheduled for baseline data collection. All 34 schools executed data collection during the Spring semester 2021, thus surpassing the 30-school target. Out of 4122 initially eligible students from 236 classrooms, a total of 3490 students from 210 classrooms completed e-questionnaires (89% classroom, and 84.7% individual student response rate) well surpassing the target of 2000 student participants. These students constituted the core CAN2021 baseline sample (ie, completion of e-surveys) and provided a recruitment pool for the longitudinal study arm (ie, consent for future administrative contact through registry linkages and/or additional direct contact).

Recruitment and data collection stages are shown as part of the study flow (figure 1).

Figure 1. CANN2021 recruitment and data collection flow.

Baseline data collection procedures

Two information and training sessions were conducted with school representatives before the data collection started. Data collection took place between 25 February 2021 and 10 April 2021, and it entailed administration of e-surveys to students during regular school hours under teachers’ coordination. Specifically, teacher coordinators described the study aims and methods; addressed anonymity and confidentiality issues while clarifying consent procedures; provided the unique school/classroom link to e-questionnaires and answered student questions, if any. Student participants were not individually reimbursed for their time, but the participating classrooms received group reimbursements to their classroom account amounting to NOK4000/NOK2000 (approximately US$480/US$240) depending on the class size. All teacher coordinators received reimbursement of NOK1000 (approximately US$120) per supervised classroom.

Baseline e-survey participation was anonymous and voluntary, and students were informed that financial reimbursements were not contingent on classroom participation rates. After they electronically submitted their responses, participants were asked via web-form whether they would consent for: (a) administrative contact in the form of individual-level linkages of their survey responses to their data from various national registries and (b) future direct contact, including possible invitations to follow-up e-surveys and/or qualitative interviews. In addition to the explanations as to why such follow-ups and linkages may be needed and how they will be processed, this e-consent also clarified that study participation would no longer be anonymous because personally identifying information such as the email address or Social Security Number would also need to be provided.

The final step in the creation of the CANN2021 registry-cohort involves individual-level prospective and retrospective linkages of the consenting participants’ survey responses to their health, education, labour, etc administrative data as recorded in various population-based registries since early 2000s.¹⁹,²¹ Administrative and ethical approvals are currently obtained for retrospective linkages (starting in 2010) and for the 4, 8 and 10-year follow-ups; that is, for merging of the consenting participants’ survey responses with their various administrative records in 2025 (4-year follow-up), in 2029 (8-year follow-up) and in 2031 (10-year follow-up), with extension possibilities.

Recruitment and data collection stages are shown as part of the study flow (figure 1).

What has been measured?

Baseline assessment explored multiple dimensions of cannabis involvement among Norwegian high school students.

The e-survey instrument included items measuring basic sociodemographic characteristics, various individual characteristics ranging from perceived health to academic achievement and substance use histories (ie, alcohol and tobacco) of the participants. The instrument assessed cannabis involvement (ie, lifetime, past year and past month frequency of use) and in case of any use, follow-up questions about age of onset, use contexts and settings, type of products consumed and intake modes, and the means of obtaining the drug. This standard epidemiological module was expanded to capture additional dimensions of cannabis involvement, such as the experienced opportunities to use cannabis or negative consequences following use, for example.

Finally, the instrument included three additional modules broadly capturing cannabis-related beliefs, attitudes and opinions (such as use and non-use motives; political views, etc), cannabis-related perceptions (such as injunctive and descriptive norms; risk perceptions; etc) and cannabis-related knowledge (such as the knowledge of cannabis products, cultures and policies; access to information sources, etc). These modules were developed for this study to probe and explore a range of cannabis-relevant themes under-researched among high school-aged adolescents outside the North American cultural and policy settings. In addition, the response options included ‘don’t know’ or ‘not sure’ response categories whenever feasible. In addition to minimising the number of missing responses and various response biases, inclusion of such response categories allowed us to assess often lost dimensions of youth’s cannabis involvement such as the lack of knowledge or lack of certainty.^{60 61}

Table 1 summarises the key themes reflected in the e-survey, including the instruments and scales used across the five survey modules described above.

Table 1. CANN2021 baseline e-survey content.

Assessed domains	E-survey items	Sources
Module 1 Background characteristics	Gender; year of birth; resident family structure; parental marital status; parental education; own and parental country of birth; self-rated physical and mental health; academic achievement (grades) in Norwegian, English and math; sports and leisure; disposable income; employment	In line with other established surveys and monitoring tools such as ESPAD46 and Monitoring the Future70 or Developed for this study
	Perceived family socioeconomic status	MacArthur Scale of Subjective Social Status71
	Temperament	Zuckerman Sensation Seeking Scale72 Barratt Impulsiveness Scale-Brief73
	Future goal orientation	Time-income decision task74 75 College attendance plans76
	Alcohol, tobacco (cigarette and snus) and other drugs use	In line with ESPAD46 and Monitoring the Future70
	High risk alcohol use (eg, bingeing, blackout)	In line with ESPAD46 and Monitoring the Future70
	Use of e-cigarettes, vaping pens, JUUL, etc.	In line with ESPAD46 and Monitoring the Future70
Module 2 Cannabis-related experiences	Encountered cannabis exposures and use opportunities	Substantively expanded/modified known indicators31 46
	Any lifetime use	In line with ESPAD46 and Monitoring the Future70
	Users only (filter):
	Age of onset; past-year and past-month use	In line with ESPAD46 and Monitoring the Future70
	Use settings and contexts (eg, alone, at school, at parties)	Developed for this study, motivated by prior research77
	Sources and methods of obtaining cannabis	Developed for this study, motivated by prior research78
	Intake modes and used products (eg, smoking, edibles)	Developed for this study, motivated by prior research79
	Concurrent use of alcohol and cannabis	Substantively expanded/modified known indicators80 81
	Negative consequences checklist (eg, health, legal)	Substantively expanded/modified known indicators82 83
Module 3 Cannabis-related beliefs, attitudes, expectations, motivations and opinions	Expectancies/motivations for cannabis use	Substantively modified known indicators82 84
	Reasons for non-use (non-users) or non-increased use (users)	Substantively expanded/modified known indicators85
	Cannabis legalisation and decriminalisation attitudes	Developed for this study, motivated by prior research86
	Cannabis policy recommendations	Developed for this study, motivated by prior research86
	Expected cannabis use under non-prohibition	Developed for this study, motivated by prior research36 87
Module 4 Cannabis-related perceptions	Injunctive and descriptive norms (vs classroom peers)	Substantively modified/expanded known indicators88
	Perceived cannabis use by parents, siblings, friends, boy/girlfriends	Developed for this study, motivated by prior research89
	Perceived risk (health, legal, educational, etc) of cannabis use patterns	Substantively modified/expanded known indicators46 70 90
	Social comparison (perceptions of a typical cannabis user/non-user)	Substantively modified/expanded known indicators91
Module 5 Cannabis-related knowledge	Cannabis-related knowledge (health, product, cultural, etc)	Developed for this study, motivated by prior research92,94
	Knowledge of legal responses in case of own cannabis use/possession	Developed for this study, motivated by prior research92,94
	Sources for received or sought cannabis-related information	Developed for this study, motivated by prior research95
	Assessment of cannabis-related information sources	Developed for this study, motivated by prior research95

ESPAD, European School Survey Project on Alcohol and Other Drugs.

Because CANN2021 investigates both sensitive populations (ie, adolescents) and topics (ie, involvement with illegal drugs), security precautions were paramount. The Services for Sensitive Data at the University of Oslo (Tjenester for Sensitive Data, TSD) were responsible for the collection and storage of all e-survey data. Participants were informed that their responses would be confidentially stored at the TSD secure server accessible only to the key project researchers using two-step factor identification and that all findings will be published only in the aggregate format ensuring anonymity of all participants.

Patient and public involvement

None.

Findings to date

Participant characteristics

The primary recruitment goal was to obtain a baseline sample of high school students which would enable national-level estimations and provide a recruitment platform for the longitudinal study arm. Table 2 shows the distribution of participants from both the CANN2021 baseline sample (N=3490) and the registry-linkages cohort (n=1510, 43.3%) across the core sampling and individual characteristics, as well as their statistical comparisons with the general population of Norwegian aged 17–19 years as reported in the Statistics Norway’s publicly available census data. There were no divergences of the baseline sample from the corresponding general population in terms of structural sampling frame characteristics (ie, school/grade) and there were only slight discrepancies in terms of geographic region representation such that the South was somewhat over-represented and Mid/West region was slightly under-represented. Additional divergences (under 5% absolute error) were observed for gender and immigrant background, likely reflecting individual participation rates, non-responses and/or different operationalisations (eg, ‘other’ response option for gender checked by 28 participants in our e-survey, vs only male/female census categories for sex).

Table 2. CANN2021 sample characteristics.

Demographic characteristics	Norwegian population 17–19 year-olds	CANN2021
		Core baseline sample	Registry-linkages sample
	N=189 337	N=3490	n=1510 (43.3%)
	%	N (%)	n (%)
1. Geographic distribution*
East	41.2	1400 (40.1)	518 (34.3)†‡
Mid/West	35.8	1147 (32.9)†	555 (36.8)‡
South	14.0	644 (18.4)†	275 (18.2)†
North	9.0	299 (8.6)	162 (10.7)†‡
2. High school
Academic	66.4	2315 (66.3)	1081 (71.6)†‡
Vocational	33.6	1174 (33.7)	427 (28.4)†‡
3. Grade in high school/age§
Grade 11 (17-year-old)	33.4	1146 (32.8)	436 (28.9)†‡
Grade 12 (18-year-old)	32.9	1149 (32.9)	470 (31.1)
Grade 13 (19-year-old)	33.7	1195 (34.3)	604 (40.0)†‡
4. Gender
Girl	48.7	1785 (51.2)†	897 (53.5)†
Boy	51.3	1675 (48.0)†	691 (45.8)†
Other/No info	--	30 (.8)	11 (.7)
5. Immigrant background¶
No	82.9	2740 (78.5)†	1265 (83.8)‡
Yes	17.1	711 (20.4)†	238 (15.7)‡
No info	--	39 (1.1)	7 (.5)

Variables 1–3 reflect random sampling procedures and regional/school study participation. Variables 4–5 reflect individual student participation and response rates. All population-level statistics for 2021 were obtained from the publicly available census data managed by the Statistics Norway (https://www.ssb.no).

^*Reflects the traditional division of regions in Norway, encompassing all 11 counties.

^†Denotes significant differences at p<0.05 or lower for comparisons of proportions between the Norwegian population of 17–19 years in 2021 and (1) CANN2021 Baseline sample and (2) Registry-linkages cohort.

^‡Denotes significant differences at p<0.05 or lower for comparisons of proportions between the CANN2021 Baseline sample and the Registry-linkages cohort.

^§High school programme in Norway lasts 3 years (ie, grades 11–13) and enrolment is primarily based on birth year cohort. Consequently, there is little variation within school grades in terms of students’ chronological age, as students commonly start high school during the calendar year they turn 16.

^¶Statistics Norway uses the statistical standard for classification of persons by immigration background, including Born in Norway to Norwegian-born parents; Immigrants; Norwegian-born to immigrant parents; Foreign-born with one Norwegian-born parent; Norwegian-born with one foreign-born parent; Foreign-born to Norwegian-born parents (including intercountry adoptions). Accordingly, all CANN2021 participants who reported not being born in Norway to two Norwegian-born parents were classified to be of ‘immigrant background’.

Overall trends for the comparisons between the participants who consented for administrative contact/registry linkages and the general population of Norwegian aged 17–19 years suggest that boys, immigrants and younger students were less likely to consent for registry, but the CANN2021 registry-linkages cohort remained robust. If required for investigation of specific research questions, both the core baseline sample and the registry-linkages cohort can be weighted to ensure generalisability to the population of Norwegian aged 17–19 years.^{59 62}

Other findings

Roughly one in five participants (20.3%) reported having used cannabis at least once during their lifetime whereas roughly one in seven (15%) reported having used cannabis at least once in the past year.⁵³ Boys; older students; those perceiving their families to be of below-average wealth; and those reporting poorer mental and physical health were all more likely to report having used cannabis.⁵³ Of those participants who reported having used cannabis at least once during lifetime, 40.9% consented to registry linkages. These rates matched the overall consent rates for these planned study arms and ensured sufficient representation in the exposed groups (ie, cannabis users) in planned follow-ups.

Missing data

As noted, most e-survey items included ‘don’t know’ or ‘not sure’ response categories. We believe that for this reason the proportion of missing responses in the core sample was rather low on the substance use and cannabis-related indicators of substantive interest: ranging from the maximum of 2.2% (or 77 participants who did not respond to the question asking if they think their boyfriend/girlfriend has tried cannabis at least once) to the minimum of 0.14% (or 5 participants who did not respond to the question asking if they ever used an illegal drug other than cannabis). A total of 65 participants (or 1.8% of the core baseline sample) did not respond to the question asking if they ever used cannabis.

Missingness was greater on module 1 ‘background variables’, to the maximum of 8.7% (or 302 participants who did not respond to the question asking about their high school grades. However, this information can be obtained from the Education registry for the participants in the registry-linkages cohort. No advanced procedures for handling missing values from e-surveys are thus anticipated.

Future plans

Current plans involve the analyses of the CANN2021 baseline sample to address the emerging issues in underage cannabis involvement of relevance to public health and national monitoring needs. These analyses will generate a set of standard population-level parameters (ie, prevalence) of early cannabis involvement, as well as of other dimensions of cannabis involvement not typically assessed in the Norwegian cultural and policy setting (ie, use of cannabis edibles; variations in use settings and use motivations; use intentions under various policy scenarios, etc). Next, regression-based models will investigate the associations between these cannabis-involvement indicators and a range of sociodemographic (ie, gender, immigrant background, etc) and individual (ie, perceived health, risk perceptions, cannabis-related knowledge, etc) characteristics as self-reported in baseline e-surveys.

Future plans involve individual-level linkages of the consenting participants’ survey responses to their health, education, labor, etc administrative data as recorded in various national registries since early 2000s,¹⁹,²¹ and the corresponding longitudinal analyses of the CANN2021 registry cohort. Administrative and ethical approvals are currently obtained for the retrospective linkages with annual registry data starting from 2010, and for the prospective linkages with the annual registry data for 4, 8 and 10 years after the baseline survey, with the possibility of extensions. The relevant registries and key indicators of interest are shown in table 3. The planned cross-linkages of the survey self-reports and registry data either retrospectively (such as with the health diagnoses, criminal records or school performance data prior to CANN2021 survey participation) or prospectively (such as with the health diagnoses, criminal records or school performance data following CANN2021 survey participation) thus provide unique research opportunities. These include examination of both the putative risk factors for (eg, early school failure) and consequences of (eg, subsequent school failure) early cannabis involvement.

Table 3. Summary of key registries and variables to be used in CANN2021 registry-linkages cohort.

Registry	Key variables	Period
1. Norwegian Registry for Primary Health Care	All registered health contacts related substance use and mental health in primary health services.	Annual data 2010–2031
2. Norwegian Patient Registry	All treatment received for mental health problems through specialist health services. Facility and treatment type (outpatient/inpatient).	Annual data 2010–2031
3. Statistics Norway/National Education Database	Middle school results (grade 10), national tests for math (grades 5, 8 and 9). High school results (grades 11, 12, and 13), days of high school absence, highest completed educational level, and current school/college enrolment status.	Annual data 2010–2031
4. Statistics Norway/National Insurance Administration	NEET-status (a composite variable provided by Statistics Norway reflecting ‘not in education, employment or training’ status). SFP-status (a composite variable provided by Statistics Norway reflecting labour, educational activity, and welfare schemes intended to compensate for lack of earned income). Labour market status, unemployment and other welfare and disability benefits.	Annual data 2010–2031
5. Statistics Norway/Labor Market and Earnings Database	Individual and household income.	Annual data 2010–2031
6. Statistics Norway/Crime Registry	Charges and types of offences recorded, as well as adjudications and sentences.	Annual data 2010–2031
7. Statistics Norway/Population Registry	Immigrant status; residence community urbanity (in 2021); parental marital status and household type (2010–2031); mortality status (2021–2031).	Annual data as indicated

Shown is the list of registries with the obtained administrative approvals and planned coverage periods. (1) https://helsedata.no/en/forvaltere/norwegian-institute-of-public-health/norwegian-registry-for-primary-health-care-kpr/, (2) https://helsedata.no/en/forvaltere/norwegian-institute-of-public-health/norwegian-patient-registry-npr/, (3-7) https://www.ssb.no/en/data-til-forskning/utlan-av-data-til-forskere/variabellister.

Indeed, this cohort, predicated on registry-based longitudinal data, is poised to capitalise on the strengths of Nordic registry-based research.¹⁹,²¹ This study feature allows for unique research opportunities owing to the availability of high-quality health registries that contain information on all primary and specialist healthcare contacts within these systems of universal, tax-funded healthcare provision.^{19 20} In addition, detailed information on all residents is available on a range of social outcomes (eg, education, employment, crime) from other national administrations which are managed and coordinated by Statistics Norway (the Norwegian Bureau of Statistics). Cross-linkages of these data sources at the individual level are possible thanks to the unique national personal numbers (ie, social security numbers) assigned to each resident and used for referencing across all administrative databases and registries since their inception.

Registry-based epidemiology thus provides a distinctive platform for a multitude of longitudinal cohort studies. Examples of Nordic registry-based studies include a cohort of patients with gastro-oesophageal reflux disease from Denmark, Finland, Iceland, Norway and Sweden;⁶³ a cohort of prescription medication users from Sweden;⁶⁴ or a cohort of military recruits from Finland.⁶⁵ However, these studies, despite the impressive sample sizes and comprehensive registry data, lacked nuanced information obtainable directly from participants through surveys. Thus, combining the survey and registry data at the individual level offers significant advantages: specifically, surveys provide rich self-reported information on the participants, something that registry-only studies lack, while the continuously updated comprehensive records from the registries with negligible attrition uniquely supplement and extend survey self-reports. Indeed, several Nordic projects have combined individual-level survey and registry data to inform our understanding of substance use and associated issues broadly defined either in general^{66 67} or in vulnerable populations,^{68 69} but none focused on youth and cannabis extensively and exclusively.

Retrospective and prospective registry linkages with survey responses as planned in CANN2021 thus offer specific measures of relevant risk and protective factors, as well as of the health and social indicators (ie, substance use disorders and other health diagnosis, educational attainment, arrests, etc) associated with adolescent cannabis involvement. These study features can empower multiwave follow-ups, thus facilitating original multidisciplinary longitudinal research on the correlates, causes, consequences, patterns and mechanisms of early cannabis involvement well beyond high school years. Therefore, this study has the potential to critically inform our understanding of the emerging issues in cannabis epidemiology.

Collaboration

No collaboration efforts are currently planned.

Data availability statement

No data are available.