Understanding trajectories of cannabis use frequency across the lifespan based on routine screening in a large outpatient population - SUMMARY Cannabis use is prevalent and increasing in the US, with growth in older adult use outpacing increases in all other age groups. Cannabis use increases risk of cannabis use disorder (CUD) and other adverse health outcomes, with up to 33% of people who use cannabis having a CUD. In the context of legalized, more frequent and higher potency cannabis use, longitudinal studies examining the long-term risks of cannabis use are critically needed. Latent class trajectory modeling is an established approach to modeling cannabis use patterns among adolescent and young adult research participants. Few studies have included middle-aged or older adults, for whom cannabis morbidity may be most concerning due to risks of drug interactions, diminished health, falls and injuries, and impaired cognition. The proposed study responds to NIDA’s call for research on the health effects of cannabis, particularly among older adults, and trajectories of substance use and adverse health outcomes. This study will use 10 years of electronic health record data from a large health system that screens patients annually with a valid practical, single-item cannabis screen. The sample includes more than 331,000 adult patients (≥ 18 years)—including large samples of middle aged and older adults—who have completed the cannabis screen on 3 or more occasions as part of routine clinical care (2016 – 2025). The screen asks about the frequency of cannabis use (none to daily), with frequency of use being the most important predictor of CUD and adverse health conditions, even when accounting for heterogeneity of cannabis products. Specific aims are to conduct developmental research to inform a future R01 that will assess the extent to which different longitudinal cannabis use patterns predict subsequent adverse health outcomes. Aim 1 is to conduct preliminary analyses to understand sample biases, cohort effects (e.g., COVID- 19) and data missing. Aim 2 is to apply multistep trajectory modeling to identify groups of patients, separately for 4 age groups (18-34, 35-49, 50-64, ≥ 65), who follow similar trajectories of cannabis use and characterize patients in each trajectory group by demographics, health conditions, medication use, health care utilization and diagnosed CUD. Aim 3 is to describe, for each age-based trajectory group, the year-by-year prevalence of concurrent adverse health outcomes associate with cannabis use (i.e., depression, psychotic disorder, chronic pain, polysubstance use, cognitive impairment, diagnosed CUD) over the study period. Secondarily, by age group, we will repeat Aims 2 and 3 separately in women and men and in 4 subgroups defined by race and ethnicity: Black, Hispanic, Asian and White. Public Health Impact: More than 59 million US adults use cannabis, yet little is known about the long-term patterns of cannabis use and associated adverse health outcomes, particularly for middle-aged and older adults. Results will have direct clinical implications for care of patients whose cannabis use is associated with adverse outcomes and build the foundation for future research to predict subsequent adverse health outcomes by trajectory group.
