Association between cannabis use and physical activity in the United States based on legalization and health status | Journal of Cannabis Research

Data

This is a cross-sectional study that uses a standardized questionnaire to obtain data from U.S. adults 2016 through 2022. Data were obtained from the Behavior Risk Factor Surveillance System (BRFSS), which is a national system of health-related telephone surveys that collect state and territory data about U.S. residents regarding their health-related risks behaviors, chronic health conditions, and use of preventive services. The BRFSS completes over 400,000 adult surveys each year. The survey design uses random probability samples of the adult (ages 18 and older) population. The questionnaire consists of three parts: (1) core questions on demographics, current health-related conditions, and behaviors adopted by all states and U.S. territories; (2) optional modules on specific topics (e.g., marijuana or cannabis use) that states may choose to use; and (3) state-added questions developed by states for their own use (Healthy People 2030). Overall median response rates for participating areas were 47.1% in 2016, 45.9% in 2017, 49.9% in 2018, 49.4% in 2019, 47.9% in 2020, 44.0% in 2021, and 45.1% in 2022 (CDC BRFSS 2016, 2017, 2018, 2019, 2020, 2021, 2022).

Survey questions on cannabis use were added to the BRFSS in 2016 as an optional module and have been used by U.S. states and territories since. This study includes participants who were administered the cannabis use module in the 2016 through 2022 BRFSS surveys. The number of participating areas during these years are 10 states (n = 106,820), 10 states and 1 territory (n = 63,451), 13 states and 2 territories (n = 113,543), 12 states and 1 territory (n = 89,007), 20 states and 1 territory (n = 80,188), 20 states and 1 territory (n = 137,560), and 17 states and 1 territory (n = 94,919), respectively.

All participants provided informed consent prior to the interview. Information about the BRFSS survey design, questionnaires, and data collection is available elsewhere (CDC BRFSS 2024). This study was determined to be exempt from human subject research review by the author’s institutional review board because the BRFSS provides publicly available deidentified data.

Measures

The primary dependent variable was adult physical activity. This variable was based on the question: “Did you participate in physical activity or exercise during the past 30 days apart from your regular job?” (CDC BRFSS 2022). The primary independent variable was current cannabis use. This variable was based on the question “During the past 30 days, on how many days did you use marijuana or cannabis?” (Azofeifa et al. 2016). Participants responded “yes” or “no” to both these questions.

Note that BRFSS treats marijuana and cannabis as synonymous. Although the cannabis plant contains about 540 chemical substances, the word marijuana typically refers to the part of or products from the plant that contain substantial amounts of tetrahydrocannabinol (THC) (NCCIH 2019; Steinmetz 2017). This study will also treat marijuana and cannabis synonymously.

Morbidity was assessed by the question, “Has a doctor, nurse, or other health professional ever told you that you had any of the following?” with answers including stroke, heart attack, coronary heart disease (CHD), asthma, chronic obstructive pulmonary disease (COPD), diabetes, arthritis, kidney disease, skin cancer, other types of cancer, and depressive disorder (CDC BRFSS 2022). A variable was created to indicate whether they had any of these chronic medical conditions.

A variable that was only available in 2018 through 2021 surveys identified the primary purpose for using cannabis, based on the question “When you used marijuana or cannabis during the past 30 days, was it usually for (1) medical reasons, (2) non-medical reasons, or (3) for both medical and non-medical reasons” (CDC BRFSS 2021). This study combines reasons 2 and 3.

Other variables included were age (18–34, 35–54, ≥ 55), sex (men, women), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and other), education level (< high school, high school, some college, and college), employment status (employed, not employed, student, homemaker/retired), smoked > 100 cigarettes in lifetime (yes, no), body mass index (BMI), and legal status (recreational, medical, not legal). All areas that legalized recreational cannabis had also legalized it for medical purposes. This variable was determined by identifying for each year whether the areas participating in the cannabis module had legalized cannabis for recreational use, medical use, or neither (Forbes Health 2024). Four categories of BMI provided by the BRFSS are: underweight (BMI < 18.5 kg/m²), normal weight (18.5 ≤ BMI < 25.0 kg/m²), overweight (25.5 ≤ BMI < 30.0 kg/m²), and obese (≥ 30 kg/m²) (National Heart, Lung, and Blood Institute 2008; World Health Organization 1995).

Statistical analysis

Prevalence of physical activity and current cannabis use were estimated by taking the survey stratum and sampling weights into consideration. Multiple logistic regression on sample survey data was used to identify whether there were associations after adjusting for age, sex, marital status, race/ethnicity, education, employment, smoking, BMI, legal status, and chronic medical conditions. Adjusted odds ratios were combined across years by taking their sample size weighted average. Variable effects were assessed for statistical significance in the model using the t test. Interactions were assessed for statistical significance using the F test. The modifying effect of legal cannabis status on cannabis use and physical activity and of cannabis use on chronic medical conditions and physical activity were assessed by comparing whether stratified ORs significantly differed from unity. Odds ratios were estimated to measure the association between variables, with corresponding 95% confidence intervals. Confidence intervals that do not overlap 1 indicate statistical significance of the odds ratio. Statistical significance was based on the 0.05 level. Statistical analyses were conducted using Statistical Analysis System (SAS) software, version 9.4 (SAS Institute Inc., Cary, NC, USA, 2014).