Does maternal education moderate the relationship between adolescent cannabis use and mental health in early adulthood?
MRC Integrative Epidemiology Unit, Bristol Medical School, University of Bristol, Bristol, UK
Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK
Medical Research Council Centre for Neuropsychiatric Genetics and Genomics, School of Medicine, Cardiff University, Cardiff, UK
Department of Psychology, University of Bath, Bath, UK
Abstract
Introduction
Socioeconomic disadvantage has been associated with cannabis use and poor mental health. It is therefore hypothesised that lower maternal education, a proxy for socioeconomic disadvantage, may increase the risk of cannabis-related mental health and substance use consequences.
Methods
A total of 5099 participants from the Avon Longitudinal Study of Parents and Children reported cannabis use via questionnaires at 16 or 18. Logistic regression was used to examine the relationship between any and regular (weekly or more) adolescent cannabis use with depression, anxiety, psychotic experiences, and problematic cannabis use at age 24. Maternal education was included as an effect modifier. Missing data were addressed through multiple imputation using chained equations.
Results
In total, 36.5% of participants reported adolescent cannabis use and, of these, 14% reported regular use. Adolescent cannabis use was associated with an increased likelihood of anxiety and problematic cannabis use; however, there was little evidence of moderation by maternal education. Regular cannabis use was associated with an increased likelihood of problematic cannabis use, with little evidence of moderation by maternal education. There was weak evidence that the association between regular cannabis use and depression (interaction p-value = 0.024) and anxiety (interaction p-value = 0.056) was stronger in people with high maternal education.
Discussion and Conclusions
Adolescent cannabis use is associated with increased risk of anxiety and cannabis use disorder, but there was insufficient evidence that childhood socioeconomic position (proxied by maternal education) modifies this relationship. Improved public health messages for all adolescents about these risks may be warranted.
Untitled section
Keywords: ALSPAC, cannabis, mental health, socioeconomic position
Article notes
Untitled section
Issue date 2024 Nov 1.
1.Introduction
Cannabis is the most commonly used illicit drug world-wide [1]. Adolescent cannabis use is associated with mental health disorders, including psychosis [2], depression, and anxiety [1], particularly with earlier and more frequent cannabis use [2]. Understanding this relationship is challenging due to complex and overlapping risk factors, such as lower socioeconomic position (SEP) [3]. SEP may confound the relationship between substance use and mental health, or it could moderate the relationship through a clustering of risk factors that impact resilience to substance effects.
Understanding the joint effects of cannabis and SEP on mental health may help inform targeted public health campaigns and interventions. Putative moderating effects of various SEP aspects, including education and income, have been observed for alcohol-related mortality, and for tobacco use and depression [4, 5], though have not been robustly replicated. Studies of moderation by SEP for cannabis-related outcomes are scarce [5] and weakened by cross-sectional design and focus on adults who tend to experience lesser harms than adolescents [2]. Moreover, there is need for further evidence to confirm that frequent cannabis use increases the risk of adverse consequences [1, 2], particularly regarding mental health consequences of adolescent use.
The current research aims to address gaps in existing literature by investigating the effect of adolescent cannabis use (and regular use) on adult mental health in a UK birth cohort, where cannabis is illegal. We also aim to explore whether this relationship is moderated, on a multiplicative scale, by maternal education as a proxy of childhood SEP using prospective, longitudinal data to minimise recall bias and reverse causality. We used maternal education as our proxy of childhood SEP because it captures knowledge-related assets and correlated aspects, such as income and occupation, and has been linked with offspring health [6]. Moreover, maternal education is available for the largest number of participants and is not influenced by changes to mother’s labour market participation during early years of parenting.
2.Methods
2.1.Study population
The Avon Longitudinal Study of Parents and Children (ALSPAC) is a UK population-based birth cohort which recruited pregnant women residing in the former Avon region with expected delivery dates between 1 April 1991 and 31 December 1992. Full details have been previously outlined [7–9]. The study website contains details of data available through a data dictionary and variable search tool (http://www.bristol.ac.uk/alspac/researchers/our-data/). Study data were collected and managed using Research Electronic Data Capture (REDCap) tools hosted at the University of Bristol. REDCap is a secure, web-based software platform designed to support data capture for research studies [10]. Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committees. Informed consent for the questionnaire and clinic data was obtained from participants following recommendations at the time.
Of the 14,541 pregnancies enrolled, 13,988 children were alive at age 1. A total of 5099 participants reported on cannabis use at 16/18 and a subsample of 1859 who reported using cannabis were eligible for analysis of cannabis use frequency (Figure 1).
2.2.Measures
2.2.1.Exposures
At 16 and 18, participants were asked “Have you ever tried cannabis?” Individuals who reported ever using cannabis at 16 or 18 are classified as experiencing adolescent cannabis use. Those who reported cannabis use were asked to report frequency of use; this was dichotomised to occasional (monthly or less) and regular (weekly or more) (Supporting information S1).
2.2.2.Outcomes
At 24, moderate/severe depressive episodes or generalised anxiety disorder (past week) were assessed with the Clinical Interview Schedule Revised and clinically significant criteria were applied [11]. Psychotic experiences (PE) were assessed with the Psychotic-Like Symptoms interview [12]. Participants were classified as having PEs if they reported a suspected/definite PE in the past 12 months, which was frequent or distressing. Problematic cannabis use in the past 6 months was measured with the Cannabis Abuse Screening Test (CAST; Supporting information S1) [13]. Responses were dichotomised to no/non-problematic (<2 CAST items) and problematic cannabis use (≥2 CAST items).
2.2.3.Moderator
As a proxy for SEP, mothers self-reported their highest level of education during pregnancy, which we dichotomised to O (Ordinary) level (compulsory qualifications at 16) or below and A (Advanced) level (post-16 qualifications) or above.
2.2.4.Confounders
Mother-reported confounders were parity, age at participant birth, and tobacco use during pregnancy. Participant confounders were sex assigned at birth, ethnicity, intelligence quotient at 8, depressive symptoms and PEs at 13, and tobacco and alcohol use at 16 (Supporting information S1).
2.3.Statistical analysis
Analyses were conducted in Stata, version 16. Logistic regression was used to estimate the association between maternal education and cannabis use at 16/18, and between cannabis use and mental health at 24, adjusting for confounders, followed by multivariable regression with maternal education as an effect modifier. Odds ratios (OR) and 95% confidence intervals (CI) are presented. This analysis was not pre-registered and should be considered exploratory.
2.4.Missing data
The sample was limited to those who had complete data on ethnicity and cannabis use at 16 or 18 (N = 5099). Missing data in outcomes and covariates were addressed through multiple imputation using chained equations (Supporting information S1).
3.Results
Of 5099 participants, 36.46% reported cannabis use at 16/18 and, of these, 14% reported using cannabis regularly. Table 1 presents the proportions of outcomes and covariates according to cannabis use.
| Characteristic | Adolescent cannabis use (N = 5099) | Frequency of cannabis use (N = 1859) | |||||
|---|---|---|---|---|---|---|---|
| Yes (36.46%) | No (63.54%) | p-value | Weekly or more (14%) | Monthly or less (86%) | p-value | ||
| Outcomes | |||||||
| Depression (moderate or severe symptoms) at 24 | 10.38% | 6.90% | <0.001 | 12.18% | 10.34% | <0.001 | |
| Generalised anxiety disorder at 24 | 12.40% | 8.27% | <0.001 | 13.37% | 12.43% | <0.001 | |
| Psychotic-like experiences at 24 | 6.76% | 4.68% | <0.001 | 11.68% | 6.47% | <0.001 | |
| Problematic cannabis use at 24 | 7.13% | 2.71% | <0.001 | 29.94% | 3.39% | <0.001 | |
| Demographic variables | |||||||
| Low maternal education | 48.90% | 54.10% | <0.001 | 51.90% | 48.41% | <0.001 | |
| Maternal smoker during pregnancy | 51.35% | 37.17% | <0.001 | 64.20% | 49.24% | <0.001 | |
| Maternal age, years (mean [SD]) | 29.64 (0.108) | 29.23 (0.078) | NA | 29.49 (0.326) | 29.67 (0.114) | <0.001 | |
| Maternal parity (mean [(SD]) | 0.79 (0.021) | 0.70 (0.015) | NA | 0.87 (0.062) | 0.78 (0.023) | <0.001 | |
| Male sex | 39.05% | 42.13% | <0.001 | 55.31% | 36.41% | <0.001 | |
| Black or minority ethnic group | 5.33% | 3.33% | <0.001 | 5.47% | 5.30% | 0.471 | |
| Below average IQ at 8 | 11.74% | 16.08% | <0.001 | 13.70% | 11.63% | <0.001 | |
| Adolescent substance use | |||||||
| Alcohol use at 16 | 99.28% | 90.93% | <0.001 | 99.04% | 99.44% | <0.001 | |
| Tobacco use at 16 | 83.20% | 26.68% | <0.001 | 87.82% | 82.23% | <0.001 | |
| Adolescent mental health | |||||||
| MFQ Score at 13 (mean [SD]) | 5.77 (0.136) | 4.54 (0.084) | NA | 6.46 (0.386) | 6.62 (0.133) | NA | |
| No. of PEs at 13 (mean [SD]) | 0.20 (0.015) | 0.14 (0.010) | NA | 0.27 (0.049) | 0.19 (0.017) | NA | |
3.1.Maternal education and cannabis use
Lower maternal education was associated with reduced odds of adolescent cannabis use (OR 0.81; 95% CI 0.72–0.91; Table 2), but there was little evidence of an association with regular use (OR 1.15; 95% CI 0.87–1.51).
| Higher maternal education | Lower maternal education | ||||||
|---|---|---|---|---|---|---|---|
| N | OR (95% CI) | p-value | N | OR (95% CI) | p-value | ||
| Adolescent cannabis use | 2437 | 1 (ref) | — | 2662 | 0.81 (0.72-0.91) | <0.001 | |
| Regular cannabis use | 950 | 1 (ref) | — | 909 | 1.15 (0.87-1.51) | 0.316 | |
3.2.Mental health and substance use
Those who reported adolescent cannabis use had increased odds of adulthood anxiety (OR 1.70; 95% CI 1.26–2.29) and problematic cannabis use (OR 3.17; 95% CI 1.40–7.17) after adjusting for confounders. Adolescent cannabis was not associated with depression (OR 1.30; 95% CI 0.89–1.88) or PEs (OR 1.16; 95% CI 0.77–1.74) following adjustment (Table 3).
| Unstratified analysis | Effect modification | |||||
|---|---|---|---|---|---|---|
| Outcome | OR (95% CI) | p-value | OR for high maternal education (95% CI) | OR for low maternal education (95% CI) | p-value for interactiona | |
| Depression (moderate or severe symptoms)b | Unadjusted | 1.56 (1.17–2.08) | 0.003 | 1.45 (0.98–2.17) | 1.69 (1.15–2.48) | 0.586 |
| Adjusted | 1.30 (0.89–1.88) | 0.173 | 1.19 (0.74–1.92) | 1.38 (0.88–2.17) | 0.594 | |
| Generalised anxiety disorder | Unadjusted | 1.57 (1.23–2.00) | <0.001 | 1.37 (0.94–1.99) | 1.79 (1.28–2.48) | 0.300 |
| Adjusted | 1.70 (1.26–2.29) | 0.001 | 1.41 (0.94–2.12) | 1.94 (1.32–2.85) | 0.240 | |
| Psychotic-like experiencesc | Unadjusted | 1.48 (1.05–2.09) | 0.027 | 1.43 (0.86–2.40) | 1.57 (1.00–2.47) | 0.790 |
| Adjusted | 1.16 (0.77–1.74) | 0.486 | 1.15 (0.66–2.02) | 1.16 (0.69–1.93) | 0.994 | |
| Problematic cannabis use | Unadjusted | 2.91 (1.39–6.07) | 0.005 | 2.89 (0.64–12.98) | 3.36 (1.47–7.67) | 0.863 |
| Adjusted | 3.17 (1.40–7.17) | 0.006 | 2.94 (0.64–13.51) | 3.68 (1.36–9.96) | 0.809 | |
Amongst those who used cannabis, regular use was associated with an 11-fold increase in the odds of problematic cannabis use (OR 11.30; 95% CI 5.92–21.57) at 24, after confounder adjustment. The association between regular cannabis and PEs was attenuated after adjusting for confounders (OR 1.80; 95% CI 0.94–3.44). There was little evidence regular cannabis use was associated with depression (OR 1.03; 95% CI 0.53–2.00) or anxiety (OR 1.11; 95% CI 0.65–1.90) at 24 (Table 4).
| Unstratified analysis | Effect modification | |||||
|---|---|---|---|---|---|---|
| Outcome | OR (95% CI) | p-value | OR for high maternal educationa (95% CI) | OR for low maternal educationb (95% CI) | p-value for interactionc | |
| Depression (moderate or severe symptoms)d | Unadjusted | 1.19 (0.67–2.12) | 0.548 | 1.90 (0.91–3.98) | 0.73 (0.32–1.67) | 0.079 |
| Adjusted | 1.03 (0.53–2.00) | 0.936 | 2.05 (0.90–4.66) | 0.53 (0.21–1.35) | 0.024 | |
| Generalised anxiety disorder | Unadjusted | 1.08 (0.65–1.80) | 0.768 | 1.76 (0.91–3.41) | 0.65 (0.28–1.51) | 0.070 |
| Adjusted | 1.11 (0.65–1.90) | 0.707 | 1.89 (0.96–3.72) | 0.67 (0.28–1.58) | 0.056 | |
| Psychotic-like experiencese | Unadjusted | 1.89 (1.04–3.45) | 0.037 | 1.15 (0.39–3.36) | 2.36 (1.15–4.85) | 0.265 |
| Adjusted | 1.80 (0.94–3.44) | 0.075 | 0.98 (0.30–3.15) | 2.42 (1.13–5.18) | 0.196 | |
| Problematic cannabis usef | Unadjusted | 12.41 (6.55–23.49) | <0.001 | 12.37 (5.45–28.07) | 12.91 (5.18–32.16) | 0.948 |
| Adjusted | 11.30 (5.92–21.57) | <0.001 | 9.89 (4.17–23.48) | 12.50 (4.89–31.92) | 0.729 | |
3.3.Effect modification by maternal education
For adolescent cannabis use, there was no evidence for effect modification, but estimates were slightly higher for low maternal education groups for depression (interaction p = 0.594), anxiety (interaction p = 0.240), PEs (interaction p = 0.994) and problematic cannabis use (interaction p = 0.809) at 24 (Table 3).
For frequent cannabis use, there was no evidence for effect modification for problematic cannabis use (interaction p = 0.729) or PEs (interaction p = 0.196), although the effect estimates were slightly higher for low maternal education groups (Table 4). Individuals with higher maternal education were slightly more likely to experience depression (interaction p = 0.024) and anxiety (interaction p = 0.056) following frequent cannabis use after adjustment for confounders.
The complete case results are in Tables S3–S6.
4.Discussion
This study found people with lower childhood SEP (proxied by maternal education) are less likely to use cannabis in adolescence, but no evidence that frequency of use differs. Adolescent cannabis use was associated with increased odds of anxiety, and both any and regular cannabis use were associated with increased odds of problematic use in adulthood. There was little statistical evidence of effect modification under a multiplicative model, meaning the relative odds of mental health outcomes in cannabis users or regular users compared with non-users or occasional users was similar regardless of maternal education, as a proxy of SEP.
This study builds upon previous literature reporting relationships between adolescent cannabis use and mental health by suggesting they may be accounted for by confounders [1, 14], notably earlier mental health, for which measures are often absent [1]. Previous work from this cohort found the cross-sectional relationship between higher-potency cannabis and PEs to be robust to adjustment for earlier PEs [15], whereas the present work indicates that longitudinal relationships are less robust to this adjustment. Future research would benefit from considering the temporality in onset of cannabis use and mental health.
To our knowledge, this is the first study examining the moderating effect of SEP on the relationship between adolescent cannabis use and later mental health. Our results support previous findings that low SEP groups are less likely to experiment with cannabis but more likely to use cannabis daily [16], as well as associations between cannabis and adverse mental health [1, 2]. When exploring regular cannabis use, there was weak evidence that the odds of depression and anxiety were greater in groups with high, compared with low, maternal education; however, the imprecise effect estimates make it difficult to identify the effect of cannabis in each group. This contrasts evidence relating to the relationship between SEP and mental health outcomes when cannabis exposure is not incorporated [17], and warrants exploration if robustly replicated. Otherwise, there was little evidence that SEP moderated the associations on a multiplicative scale, which is consistent with patterns of risk when there is co-exposure to cannabis use for psychiatric outcomes [18].
4.1.Strengths and limitations
This study is strengthened by use of prospective longitudinal data and adjustment for many confounders; however, residual confounding cannot be excluded. The reliance on self-reported cannabis use may result in misclassification, particularly given its illegality. Attrition prior to 16/18 may have resulted in underrepresentation of low SEP groups. Such factors may impact the nature of the observed associations, meaning replication in more representative cohorts would be useful. Whilst maternal education is a viable proxy for SEP, it does not encapsulate all SEP aspects or SEP later in childhood/adolescence. Further research to address these limitations would enable more robust conclusions, which could better aid in the development of public health interventions. Finally, the interaction results should be considered cautiously due to difficult interpretation with regards to the choice of statistical model, inferences from findings, and frequent lack of robust replication.
5.Conclusion
The findings suggest that adolescent cannabis use increases the odds of anxiety and problematic cannabis use in early adulthood. There was little evidence of moderation by childhood SEP (proxied by maternal education), supporting the need for improved public health messages for all adolescents about the risks of cannabis use.
Supplementary Material
Untitled section
Boxed Text
- Adolescent cannabis use was less likely amongst those with lower maternal education.
- Anxiety and problematic cannabis use were more likely in those who used cannabis in adolescence.
- Frequent cannabis use in adolescence was associated with a further increase in likelihood of problematic cannabis use in adulthood.
- There was little evidence that such relationships were altered by socioeconomic position (proxied by maternal education).
Acknowledgements
We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses.
Funding information
UK Medical Research Council and Wellcome Trust, Grant/Award Number: 217065/Z/19/Z; UK Medical Research Council, Grant/Award Numbers: MR/ L022206/1, MR/M020894/1; Wellcome Trust, Grant/Award Numbers: 218495/Z/ 19/Z, 209158/Z/17/Z; NIHR Bristol Biomedical Research Centre, Grant/Award Number: NIHR203315
The UK Medical Research Council and Wellcome Trust (Grant ref: 217065/Z/19/Z) and the University of Bristol provide core support for ALSPAC. A comprehensive list of grants funding is available on the ALSPAC website (http://www.bristol.ac.uk/alspac/external/documents/grant-acknowledgements.pdf); this work was specifically funded by the MRC (Grant ref: MR/L022206/1). This publication is the work of the authors and Gemma Sawyer, Laura Howe, and Lindsey Hines will serve as guarantors for the contents of this article. Gemma Sawyer is supported by a Wellcome Trust PhD studentship in Molecular, Genetic and Lifecourse Epidemiology (Grant ref: 218495/Z/19/Z). Lindsey Hines is funded by the Wellcome Trust (Grant ref: 209158/Z/17/Z). Laura Howe is supported by a career development award from the UK Medical Research Council (Grant ref: MR/M020894/1). Stanley Zammit is supported by the NIHR Bristol Biomedical Research Centre (Grant ref: NIHR203315). Matthew Hickman acknowledges support from the NIHR Health Protection Research Unit in Behavioural Science and Evaluation at the University of Bristol. This research was funded in whole, or in part, by the Wellcome Trust. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.
Footnotes
Footnote Group
References
Untitled section
References
- 1.Gobbi G, Atkin T, Zytynski T, Wang S, Askari S, Boruff J, et al. Association of cannabis use in adolescence and risk of depression, anxiety, and suicidality in young adulthood: a systematic review and meta-analysis. JAMA Psychiatry. 2019;76:426–34. doi: 10.1001/jamapsychiatry.2018.4500.
- 2.Hall W, Stjepanović D, Caulkins J, Lynskey M, Leung J, Campbell G, et al. Public health implications of legalising the production and sale of cannabis for medicinal and recreational use. Lancet. 2019;394:1580–90. doi: 10.1016/S0140-6736(19)31789-1.
- 3.Hakulinen C, Mok PLH, Horsdal HT, Pedersen CB, Mortensen PB, Agerbo E, et al. Parental income as a marker for socioeconomic position during childhood and later risk of developing a secondary care-diagnosed mental disorder examined across the full diagnostic spectrum: a national cohort study. BMC Med. 2020;18:323. doi: 10.1186/s12916-020-01794-5.
- 4.Probst C, Kilian C, Sanchez S, Lange S, Rehm J. The role of alcohol use and drinking patterns in socioeconomic inequalities in mortality: a systematic review. Lancet Public Health. 2020;5:e324–32. doi: 10.1016/S2468-2667(20)30052-9.
- 5.Matta J, Hoertel N, Airagnes G, Wiernik E, Limosin F, Goldberg M, et al. Does substance use explain social differences in terms of depression? Findings from the Constances cohort. Compr Psychiatry. 2020;102:152203. doi: 10.1016/j.comppsych.2020.152203.
- 6.Howe LD, Lawlor DA, Propper C. Trajectories of socioeconomic inequalities in health, behaviours and academic achievement across childhood and adolescence. J Epidemiol Community Health. 2013;67:358–64. doi: 10.1136/jech-2012-201892.
- 7.Boyd A, Golding J, Macleod J, Lawlor DA, Fraser A, Henderson J, et al. Cohort profile: the ‘Children of the 90s’-the index offspring of the Avon longitudinal study of parents and children. Int J Epidemiol. 2013;42:111–27. doi: 10.1093/ije/dys064.
- 8.Fraser A, Macdonald-wallis C, Tilling K, Boyd A, Golding J, Davey smith G, et al. Cohort profile: the Avon longitudinal study of parents and children: ALSPAC mothers cohort. Int J Epidemiol. 2013;42:97–110. doi: 10.1093/ije/dys066.
- 9.Northstone K, Lewcock M, Groom A, Boyd A, Macleod J, Timpson N, et al. The Avon longitudinal Study of Parents and Children (ALSPAC): an update on the enrolled sample of index children in 2019 [version 1; peer review: 2 approved] Wellcome Open Res. 2019;4:51. doi: 10.12688/wellcomeopenres.15132.1.
- 10.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)-a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–81. doi: 10.1016/j.jbi.2008.08.010.
- 11.Lewis G, Pelosi AJ, Araya R, Dunn G. Measuring psychiatric disorder in the community: a standardized assessment for use by lay interviewers. Psychol Med. 1992;22:465–86. doi: 10.1017/s0033291700030415.
- 12.Zammit S, Horwood J, Thompson A, Thomas K, Menezes P, Gunnell D, et al. Investigating if psychosis-like symptoms (PLIKS) are associated with family history of schizophrenia or paternal age in the ALSPAC birth cohort. Schizophr Res. 2008;104:279–86. doi: 10.1016/j.schres.2008.04.036.
- 13.Legleye S, Karila L, Beck F, Reynaud M. Validation of the, CAST., a general population cannabis abuse screening test. J Subst Use. 2007;12:233–42.
- 14.Moore TH, Zammit S, Lingford-Hughes A, Barnes TR, Jones PB, Burke M, et al. Cannabis use and risk of psychotic or affective mental health outcomes: a systematic review. Lancet. 2007;370:319–28. doi: 10.1016/S0140-6736(07)61162-3.
- 15.Hines LA, Freeman TP, Gage SH, Zammit S, Hickman M, Cannon M, et al. Association of high-potency cannabis use with mental health and substance use in adolescence. JAMA Psychiatry. 2020;77:1044–51. doi: 10.1001/jamapsychiatry.2020.1035.
- 16.Legleye S, Janssen E, Beck F, Chau N, Khlat M. Social gradient in initiation and transition to daily use of tobacco and cannabis during adolescence: a retrospective cohort study. Addiction. 2011;106:1520–31. doi: 10.1111/j.1360-0443.2011.03447.x.
- 17.Lemstra M, Neudorf C, D’arcy C, Kunst A, Warren LM, Bennett NR. A systematic review of depressed mood and anxiety by SES in youth aged 10-15 years. Can J Public Health. 2008;99:125–9. doi: 10.1007/BF03405459.
- 18.Zammit S, Lewis G, Dalman C, Allebeck P. Examining interactions between risk factors for psychosis. Br J Psychiatry. 2010;197:207–11. doi: 10.1192/bjp.bp.109.070904.