The Effects of Adolescent Cannabis Use on Psychosocial Functioning

INTRODUCTION

Cannabis use in adolescence has been consistently linked to a range of negative psychosocial outcomes, including problematic peer, romantic, and parent–child relationships, worse educational outcomes, lower adult socioeconomic status, and legal consequences. In this narrative review, we present key findings from research published in the past 20 years on long-term psychosocial outcomes in adolescents who use cannabis. We first summarize current evidence from relevant longitudinal studies. Because we focus on long-term psychosocial outcomes, findings from these studies may reflect cumulative effects of adolescent cannabis use, as well as those of repeated cannabis exposure into adulthood—although in some cases, long-term outcomes (eg, justice system involvement) may be driven simply by possession of cannabis, the shorter-term consequences of use, or acute intoxication. We use the term “cannabis use” broadly to encompass the varying definitions in these studies, providing more precise descriptors (ie, in terms of cannabis use disorder status or cannabis use frequency) when possible. We then consider evidence for potential mechanisms linking adolescent cannabis use and long-term psychosocial outcomes, drawing largely from the smaller number of relevant causally informative studies, which include those conducting sibling/twin comparisons or taking advantage of “natural experiments” (eg, changes in the legal status of cannabis). We critically consider the current evidence base for concluding the causal effects of adolescent cannabis use on psychosocial functioning, given methodological and study design limitations. Finally, we offer suggestions for future investigations seeking to identify causal mechanisms and clinic care points for clinical providers and public policy makers concerned with adolescent cannabis use and its effects on later psychosocial outcomes.

Current Evidence

There is extensive evidence from research published over the past 20 years that adolescent cannabis use is associated with negative outcomes in important psychosocial domains, including functioning in peer and romantic relationships, the quality of the parent-child relationship, educational outcomes, adult socioeconomic status, and legal consequences (Box 1).

Box 1. Psychosocial outcomes.

Interpersonal relationships Social network size and support Relationship quality/satisfaction Attachment Delinquent/substance-using peers Sexual behavior Parental monitoring/knowledge

Socioeconomic status School performance and drop-out Educational attainment Employment and occupational prestige Income and debt

Legal Driving while intoxicated Car crashes Delinquency Justice system involvement

Peer and romantic relationships

Decades of research on substance use have sought to understand its associations with functioning in peer and romantic relationships, given the importance of these social bonds for well-being.^1,2 Cannabis use is associated with decreased social network size and, in turn, lower perceived social support.³ A number of studies have examined the role of peers in adolescent cannabis use, inferring a causal pathway from peer relationships to cannabis initiation and use in adolescence.^4–6 However, there is some evidence that the direction of the effect may be reversed, so that early-onset cannabis use, in adolescence, leads to later associations with delinquent and substance-using peers.^7,8 Regular cannabis use in adolescence and early adulthood has also been found to be associated with lower likelihood of being in a romantic relationship.⁹ For those who are in romantic relationships, adolescent cannabis use is associated with lower levels of relationship satisfaction, even after accounting for other potential confounders, such as comorbid substance use, family functioning, and psychopathology.¹⁰ There is also evidence of greater engagement in risky sexual behaviors among women who first initiated cannabis use in adolescence relative to those who first initiated in early adulthood or later.¹¹ Taken together, evidence from these longitudinal studies suggests that adolescent cannabis use is associated with smaller social networks, lower social support, fewer peer and romantic relationships, greater affiliation with deviant and substance-using peers, lower romantic relationship satisfaction, and more risky sexual behaviors.

Parent–child relationships

Parent–child relationships continue to serve as a key source of support and guidance for adolescents as they become increasingly independent in later adolescence and early adulthood.¹² Most studies examining associations between these relationships and cannabis use have examined how different aspects of the parent–child relationship and specific parenting behaviors relate to cannabis initiation and course of use in adolescence. Collectively, these studies suggest that the quality of the parent–child relationship, parental monitoring, household rules about cannabis and other substance use, and parents’ knowledge of their adolescents, their peers, and their activities may act as particularly important buffers against cannabis initiation, use, and negative consequences of use.^13,14 There is also evidence that parental factors can moderate the negative influences of deviant peers on adolescent cannabis use.¹⁵ Fewer studies have considered potential effects of adolescent cannabis use on parent–child relationship quality. Qualitative studies suggest adolescent cannabis use is associated with more problematic interactions and relationships with family members.^16,17 Similarly, research seeking to understand the bidirectional nature of adolescent cannabis use and family relationships has shown that externalizing behavior in early adolescence, which has been linked to substance use, is negatively associated with the quality of the parent–child relationship.¹⁸ One meta-analysis found that the association between earlier relationship experiences and later substance use was significantly stronger than that between earlier substance use and later attachment in adulthood.¹⁹ This suggests that, perhaps, parent–child relationships may affect later substance use more than substance use affects parent–child relationships. However, even longitudinal studies of this kind are not sufficient to determine a causal pathway in either direction. Taken together, evidence from the existing longitudinal studies considering adolescent cannabis use and functioning in the parent–child relationship suggests that lower parental monitoring and knowledge are associated with adolescent cannabis use, and that adolescent cannabis use may be associated with lower-quality parent–child relationships.

School performance and educational attainment

An extensive body of longitudinal studies has considered associations between adolescent cannabis use and school performance and educational attainment. Several systematic reviews conducted over the past several decades find that adolescent cannabis use is associated with increased risk of school drop-out and lower educational attainment.^20–22 However, as these reviews note, many (though not all) of the existing studies find that associations between adolescent cannabis use and educational outcomes are dramatically attenuated after accounting for potential confounders, including family demographics, child cognitive ability at baseline, and behavior problems.^21,23

Adult socioeconomic status

A substantial body of research has considered associations between cannabis use in adulthood and financial and job-related outcomes. In contrast, there are markedly fewer longitudinal studies considering longer-term associations between adolescent cannabis use and adult socioeconomic outcomes. Still, the existing studies paint a relatively consistent picture, indicating that greater cannabis use or more chronic patterns of use in adolescence are associated with greater risk of unemployment, lower occupational prestige, lower income, greater debt, and more difficulty paying for medical necessities in young or middle adulthood.^24–26 These findings are largely mirrored by longitudinal studies in adulthood, which indicate that greater cannabis use at baseline is prospectively associated with greater risk of unemployment, lower income, and job loss among people used at baseline.^27–30 Notably, although findings from these longitudinal studies are remarkably consistent in suggesting that both adolescent and adult cannabis use are associated with varied indicators of adult socioeconomic status, many of these studies also indicate that associations between cannabis use and adult socioeconomic outcomes are significantly moderated by several factors, including income, race/ethnicity, sex, and age of cannabis initiation. For example, associations between cannabis use and risk of job loss were found to be are highest at the lowest and highest extremes of the income distribution, though the mechanisms driving this finding are as yet unclear.²⁹ Adolescent cannabis use was also prospectively associated with lower occupational prestige and lower income among White participants, but not Black participants, in a United States sample.³¹ Another study of a nationally representative sample of United States high school seniors (from the classes of 1976–1988) reported that frequent cannabis use was prospectively associated with lower occupational attainment 10 years later among men, but not among women.³² Finally, a study of the 2015 National Survey of Drug Use and Health (NSDUH) found that adolescents who initiated cannabis use earlier had lower odds of graduating from high school and greater risk of unemployment relative to those who initiated cannabis later.³³ Taken together, however, the literature on adolescent and adult cannabis use paint a relatively consistent picture, indicating that greater use is associated with unemployment, lower occupational prestige, lower income, job loss, and greater debt in adulthood.

Legal consequences

There is considerable interest in the potential legal consequences of adolescent cannabis use. Most of the research in this area has focused on cross-sectional associations between cannabis use in adolescence and risky behavior (eg, driving while intoxicated, car crashes), delinquency, and justice system involvement, with markedly fewer longitudinal studies considering longer-term associations with legal consequences in adulthood. Adolescent cannabis use, and particularly early cannabis initiation, has been found to be associated with greater criminal involvement and delinquency in later adolescence and adulthood.^7,34,35 However, the direction of this association remains unclear, and there is evidence that adolescents higher in behavioral disinhibition and risk proneness are also more likely to initiate cannabis use and use more cannabis.^36,37 Of note, one rationale for more permissive cannabis laws (medical and recreational cannabis legalization) is to reduce racial/ethnic inequities in law enforcement and inequities in legal consequences for cannabis use (eg, disproportionate number of arrests for cannabis use and possession of people who are Black, though rates of use are similar among people who are Black and White).³⁸ Taken together, evidence from the relatively small number of longitudinal studies considering adolescent cannabis use and legal consequences suggests that adolescent cannabis use is associated with risky behavior, delinquency, and justice system involvement in adolescence and adulthood, though it remains unclear whether this reflects causal effects.

CONSIDERATIONS

Although links between adolescent cannabis use and psychosocial impairment in both adolescence and adulthood are well established, the mechanisms underlying these links remain largely unclear. Unfortunately, methodological limitations have seriously limited the inferences and conclusions that can be drawn regarding the mechanistic links between adolescent cannabis use and long-term psychosocial functioning outcomes in adulthood.

Limitations of Current Evidence

Methodological limitations to studies of adolescent cannabis use and psychosocial functioning outcomes include issues of sample ascertainment, measurement, and study design, which result in limited generalizability of findings to the broader population and serious risk of study confounding (Box 2). Sample ascertainment bias resulting from inadequate identification or enrollment of a representative sample and differential attrition due to factors relevant to cannabis use or psychosocial impairment, or both, have the potential to limit the generalizability of findings and substantially bias results. Measurement issues include unreliability of measurement, overreliance on the use of retrospective self-reports to assess cannabis use, and the difficulty of establishing a standard cannabis “dose” in a way similar to the measurement of alcohol (eg, a standard drink) or nicotine (eg, a cigarette), given different routes of cannabis administration and variation in THC concentrations across cannabis products. Studies also commonly use different definitions of “use,” ranging from quantity or frequency of cannabis use to a diagnosis of cannabis use disorder. Perhaps the most serious methodological limitation is that most study designs cannot adequately account for confounding factors. Many studies have attempted to “control” or statistically account for putative confounders, including demographic characteristics and socioeconomic factors (eg, sex, race/ethnicity, socioeconomic status), as well as other potentially important moderators (eg, parental substance use disorder, age of cannabis initiation, cohort effects such as cannabis legalization). However, attempts to statistically account for confounders by including additional variables in statistical models are necessarily limited to those putative confounders/moderators identified and measured by researchers, which may or may not adequately reflect or capture the entire universe of confounders, and are themselves subject to measurement issues, including unreliability in measurement and reliance on retrospective self-reports that may be distorted by well-known cognitive biases.³⁹

Box 2. Methodological limitations.

Sample ascertainment Nonrepresentative samples Differential attrition

Measurement Unreliability Retrospective self-reports Nonstandard cannabis “dose” Nonstandard definitions of cannabis “use”

Study designs Correlational/observational Unaccounted confounding factors

Alternative Mechanistic Explanations for Adolescent Cannabis Use and Psychosocial Functioning Links

The predominant explanation for links between adolescent cannabis use and psychosocial functioning outcomes posits causal effects of cannabis on brain integrity,^40,41 which results in poorer cognitive performance and difficulties with emotion regulation^20,42 or amotivation⁴³ and consequent impairment in important domains of psychosocial functioning. This explanation is consistent with findings from the experimental animal literature, which clearly indicates that heavy, chronic substance use causally and adversely affects the brain, including during the critical adolescent period of rapid brain development.⁴⁴ However, in experimental animal studies, it is possible to randomly assign animals to the administration of a defined quantity of cannabis exposure, thereby isolating the putative causal mechanism. This is simply not possible in human studies—it is not ethical to randomly assign adolescents to cannabis exposure. This means that any preexisting factors that might lead some adolescents to initiate cannabis use or to use more cannabis than other adolescents may confound study results, and it is not possible to isolate cannabis exposure, as opposed to these other confounding factors, as the causal factor for psychosocial impairment. That is, associations between adolescent cannabis use and negative psychosocial outcomes may be due primarily (or entirely) to confounding by shared vulnerability factors that account for both initiation and use of cannabis in adolescence and concurrent and subsequent psychosocial impairment. This alternative explanation warrants serious consideration, given converging streams of evidence suggesting that many forms of psychosocial adversity (eg, maltreatment or socioeconomic disadvantage) occur disproportionately among individuals who use cannabis.^45,46

Evidence from Causally Informative Studies

Fortunately, there are study designs that allow us to approximate an experiment, even when true random assignment is not possible—natural experiments or quasi-experimental study designs^47,48 (Box 3). Although these quasi-experimental approaches cannot yield findings that are as definitive as a true experiment, when multiple such study designs are used in a complementary manner and yield converging findings, our confidence in causal inferences is considerably increased. Relevant study designs include discordant sibling designs (and the even more informative variant, discordant twin designs), Mendelian randomization studies, and quasi-interventions, such as changes in the legal status of cannabis or age of legal use. Unfortunately, to our knowledge, these more causally informative approaches have so far been almost exclusively limited to the examination of educational outcomes and adult socioeconomic indicators, meaning evidence for a specific causal pathway for the effects of cannabis use in adolescence on functioning in peer, romantic, or parent–child relationship domains, or legal consequences, is seriously limited (see^49,50; but see also⁵¹). Nonetheless, although small in number, the relevant causally informative studies of adolescent cannabis use and educational and socioeconomic outcomes are informative. One study among adult male twins who served in the US military during the Vietnam era found that twin differences in retrospectively reported adolescent cannabis use and cannabis dependence in adolescence were not associated with educational attainment.⁵² A second study of a community-based sample of adult Australian twins found that although retrospectively-reported adolescent cannabis use was associated with risk of school dropout, this association was not observed at the within-pair level (ie, even among twin pairs differing in their levels of cannabis use, both twins were at increased risk of school dropout), suggesting shared vulnerability factors account for the association, rather than a causal effect of cannabis exposure.⁵³ In contrast to these earlier studies, a more recent, prospective study of 3 community-based samples of Minnesota same-sex twins found among monozygotic twin pairs differing in their levels of cannabis use, the twin who used more cannabis in adolescence also had lower educational attainment, occupational prestige, and income in young adulthood than their lesser-using cotwin, consistent with a potential causal effect of cannabis exposure on adult socioeconomic outcomes.⁵⁴ It is possible that these discrepant findings arise due to differences in statistical power, as the Schaefer and colleagues (2021) study is the only causally informative analysis we are aware of that used continuous measures of both cannabis use and educational outcomes (rather than dichotomous indicators such as early vs cannabis later use, or high school graduation vs school dropout). Another study took advantage of a change in the regulation of cannabis sales in Dutch coffee shops to assess the effects of cannabis use on university students’ school performance.⁵⁵ To reduce drug tourism, the city of Maastricht in the Netherlands enacted restrictions on the nationality of students who could purchase cannabis in coffee shops in the city—students from the Netherlands, Belgium, and Germany could access coffee shops while students from France and Luxembourg could not. Among the students who could no longer legally buy cannabis in coffee shops, grades improved by 0.10 of a standard deviation relative to the students who could still buy cannabis, and their course pass rates improved by 5%, with gains that were 3.5 times greater in courses that required numerical/mathematical skills. These findings are consistent with a causal effect of cannabis exposure on educational outcomes, though an important limitation is that, because many students were likely able to acquire cannabis from other students still able to buy cannabis, it is possible the study underestimated the effects of a true ban. Taken together, evidence from the as yet relatively small number of causally informative studies considering adolescent cannabis and psychosocial outcomes suggests that adolescent cannabis use may be causally associated with poorer school performance, lower educational attainment, lower occupational prestige, and lower income in adulthood, even accounting for other shared vulnerability factors.

Box 3. Causally informative study designs.

Approach	Definition	Strengths	Limitations
Discordant sibling/twin design	Compares outcomes when (twin) siblings are discordant in their cannabis use. If the effects of cannabis are causal, then the sibling reporting more cannabis use should experience the outcome at higher levels than their less-cannabis-using sibling.	Controls for confounding by shared genetics and environment.	Potential for confounding by factors NOT perfectly shared by siblings; affected more strongly by measurement error; limited power.
Mendelian randomization (MR)	MR uses one or more genetic variants associated with cannabis use as an instrumental variable to test the causal effect of cannabis use on an outcome.	Genetic variants are not subject to confounding from environmental factors and cannot be caused by the outcome, so this approach controls for confounding by shared genetics and environment, reverse causality, and selection bias. Genetic variants also do not change over time and are measured with high accuracy, reducing measurement error.	Generally low power, possibility of biased estimates due to horizontal pleiotropy, population stratification, assortative mating, or dynastic effects.
Quasi-interventions	Capitalizes on an external event or intervention occurring at a specific timepoint that affects cannabis use in one population but not another in a quasi-random fashion (eg, changes in cannabis legalization). These populations are then compared with see if the population with greater cannabis exposure or access experiences a higher proportion of negative outcomes.	Can include study settings that would be impractical or unethical to produce by researchers; addresses confounding by genetics and environment and reverse causality.	Possibility of selection bias if “exposed” and “unexposed” groups are not sufficiently comparable (ie, some unobserved confounding may remain).

SUMMARY

Adolescent cannabis use is associated with negative outcomes in a range of important psychosocial domains. Although the existing literature varies somewhat by psychosocial domain, there is consistent evidence that adolescent cannabis use is associated with impairment in peer, romantic, and parent–child relationship domains; poorer school performance and lower educational attainment; lower adult socioeconomic status; and legal consequences. Although these associations are often considered to reflect a causal effect of cannabis exposure on the developing brain and consequent impairment, alternative explanations, including shared vulnerability factors that account for both adolescent cannabis use and negative psychosocial outcomes, have been proposed. Unfortunately, the number of causally informative studies considering adolescent cannabis and long-term psychosocial outcomes in adulthood is critically lacking, meaning that the potential mechanisms underlying these associations remain to be explicated. The few existing causally informative studies do suggest that adolescent cannabis use may be causally associated with poorer school performance, lower educational attainment, lower occupational prestige, and lower income in adulthood, even accounting for other shared vulnerability factors.

FUTURE DIRECTIONS

An extensive body of literature documenting research conducted over the past several decades establishes associations between adolescent cannabis use and impairment in important psychosocial domains in both adolescence and adulthood. We must now turn our attention to the mechanisms underlying these associations. Although it is not ethical to randomly assign adolescents to cannabis exposure, as in an experiment, the existing causally informative studies have provided important initial insights into potentially causal effects of cannabis exposure on psychosocial functioning. Additional causally informative research using epidemiologic samples and quasi-experimental designs are critical to move the field forward. One exciting future direction lies with the Adolescent Brain Cognitive Development (ABCD) study,⁵⁶ a multisite investigation of more than 11,000 children first assessed at age 9 or 10 years, who will be followed up every 6 months for 10 years, through adolescence, and holds considerable promise for mapping adolescent cannabis use and the developing addiction cycle. Critically, the ABCD study also includes more than 800 twin pairs, which, particularly in the context of its prospective, longitudinal study design, increases causal inference considerably.⁵⁷ Of note, although adolescence and the transition into emerging adulthood is an optimal time for investigations of cannabis initiation and use and misuse, there is also great potential value in even earlier investigations, such as the recently launched Healthy Brain and Child Development (HBCD) study,⁵⁸ as the risk processes that lead some adolescents to initiate cannabis use earlier or misuse cannabis were in place years earlier, indexed by familial and contextual adversities and early-evident individual differences in behavioral disinhibition. Considered together, existing research suggests delineating the developmental origins of cannabis use to identify the earliest causal risk factors and maximize early prevention efforts will promote adaptive psychosocial functioning in adolescence and adulthood.

KEY POINTS.

• Observational studies have shown that cannabis use in adolescence is consistently associated with impairments in multiple domains of psychosocial functioning, including interpersonal relationships, academics, and socioeconomic attainment. Cannabis use is also associated with increased risk of legal consequences.

• The mechanisms responsible for these associations are largely unclear. Although it is reasonable to assume that associations reflect a causal effect of cannabis on functioning, it is also possible that they arise due to reverse causation or confounding by shared vulnerability factors.

• Causally informative studies are key to differentiating between these possibilities. However, these designs have focused primarily on educational and socioeconomic outcomes to date, leaving the causal status of associations between cannabis use and other important psychosocial outcomes uncertain.

CLINICS CARE POINTS.

• Adolescent cannabis use is associated with psychosocial impairment in a range of important domains of functioning, including interpersonal relationships, socioeconomic status, and legal domains

• To date, very few causally informative studies have been conducted, meaning mechanistic links between adolescent cannabis use and long-term psychosocial functioning in adulthood remain unclear

• Causally informative studies suggest a potential causal effect of cannabis exposure on adult socioeconomic outcomes

• Clinicians working with adolescents using cannabis should be aware of individual and contextual vulnerability factors that might lead to negative psychosocial functioning outcomes, even with the successful cessation of cannabis use