This analysis evaluated the relationship between recent cannabis use and self-rated health in young and middle-aged adults using National Health and Nutrition Examination Survey (NHANES) data from 2009–2018. The authors aimed to address prior methodological limitations by applying propensity score weighting to balance measured covariates between cannabis users and non-users before outcome modeling.
The study analyzed NHANES respondents aged 20–59 who completed the survey cannabis questionnaire. The final analytic sample included 22,055 adults. Within this sample, 14.4% reported cannabis use in the 30 days prior to survey response.
The primary exposure was self-reported cannabis use in the 30 days before the survey. The primary outcome was self-rated health measured on a five-level ordinal scale. Secondary outcomes were counts of days in the past month with: (i) poor physical health, (ii) poor mental health, and (iii) activity limitations related to poor health.
To reduce confounding, the investigators used propensity score weighting to balance cannabis users and non-users across sociodemographic, medical, and lifestyle characteristics available in NHANES. Outcome models were then estimated using survey-weighted regression methods appropriate to each outcome.
The primary analysis used a weighted proportional odds regression model to estimate the association between recent cannabis use and ordered levels of self-rated health. After reweighting the sample for covariate balance, recent cannabis use was not statistically significantly associated with higher or lower self-rated health (odds ratio 0.90, 95% CI 0.80–1.01).
For the three count-based secondary outcomes, the investigators used weighted zero-inflated negative binomial models to account for excess zeros and overdispersion in counts of days with poor health or activity limitations.
Results indicated that recent cannabis use was associated with increased reports of poor mental health and with greater activity limitations in the past month. There was no observed association between recent cannabis use and counts of days with poor physical health.
The study found that, after adjustment using propensity score weighting, recent cannabis use was not associated with worse or better overall self-rated health in adults aged 20–59. However, recent use was linked to greater reports of poor mental health and to more activity limitations in the prior month. Based on these findings, the authors recommend that cannabis users who are at risk of poor mental health be connected with clinicians to help guide assessment and therapy.
Clinicians should interpret these results within the context of cross-sectional survey data and the measured covariates available in NHANES. The analysis used robust weighting and outcome-specific regression models, but the source reports no causal inference beyond observed associations.
The investigation used publicly available NHANES data. The authors stated they followed relevant ethical guidelines and noted that NHANES oversight applies. They declared no competing interests. The NHANES dataset and documentation are available from the CDC website; the source reports that data are publicly accessible.