Relationships of policy and place with substance use, alcohol misuse, and other risk behaviors
Ranker, Lynsie Renee
MetadataShow full item record
Substance use has been linked to adverse health and social costs, including morbidity and mortality. Substance use patterns are driven not only by individual-level determinants, but also by social, political and ecological factors. Given this inter-connected web, further exploration into the interactions between individuals and their environments may help inform strategies to intervene and reduce substance use and related risk behaviors. The purpose of this dissertation was to examine the dimensions of place in determining an individual’s substance use behaviors. In the first two studies, we explored the policy dimension of an individual’s environment, specifically how the state-level policy environment shapes frequency of alcohol use. We used data from the National Longitudinal Survey of Youth 1997 (NLSY97) cohort, a nationally representative, longitudinal sample of emerging adults. In the first study, we examined associations between stringency of an individual’s state-level policy environment (as measured via the Alcohol Policy Scale, APS) and past 30-day drinking and binge drinking. We found that 10-unit higher (more stringent) APS score at interview was associated with small reductions in past 30-day drinking (-0.03 drinking days, 95% Confidence Interval (CI) -0.17, 0.11) while associations were null for past 30-day binge drinking (0.004, 95% CI -0.06, 0.06), after adjusting for individual and contextual confounders. Given interest in impacts of policy changes on drinking, we also assessed associations for cross-interview change in APS. The mean interview-to-interview (roughly 1 year) change in APS score was 0.4 (+/- 2.6, range -38.1 to 39.7). In confounder-adjusted models, a 10-unit greater increase in APS change score across an interview interval was associated with -0.15 reduced drinking days (95% CI -0.30, 0.01) while reductions in binge drinking days were smaller (-0.05, 95% CI -0.14, 0.05). For the change in APS analyses, stratification by prior (starting) APS score showed the largest reduction among those in the highest quartile of prior APS (-0.25 drinking days, 95% CI -0.57, 0.07) while those in the lowest quartile of prior APS had the largest reductions in binge drinking (-0.10, 95% CI -0.13, -0.07). Age-based stratification showed the impacts were primarily driven by years when individuals were under age 21 (under legal drinking age). The second study of this dissertation also used data from the NLSY97 cohort, but focused on alcohol use over time rather than drinking at single time points. Specifically, our second study identified sub-groups of longitudinal drinking patterns—or trajectories—across 18 years of follow-up using latent class growth analysis. We then estimated the association between state-level policy score (APS) at baseline and trajectory group membership. For both the past 30-day drinking and past 30-day binge drinking outcomes, distinct trajectories were identified and described. For drinking days, a 5-group solution was the best fitting model. The 5 drinking-days groups were: late escalating (10.8%), normative (19.0%), high frequency (5.6%), low frequency (47.0%) and no/infrequent use (17.7%). A 10-unit higher baseline APS score was associated with slightly higher odds of membership in the late escalating group (adjusted Odd Ratio, aOR = 1.13, 95% CI 0.98, 1.31) and reduced odds of membership in the normative group (aOR = 0.94, 95% CI 0.81, 1.10) compared with membership in the no/infrequent group. In a separate trajectory model, we identified 5 binge drinking-days groups: later onset (10.5%), high frequency (4.4%), once-a-month (34.8%), earlier onset (11.3%), and no/infrequent (39.1%). A 10-unit higher baseline APS score was not associated odds of binge drinking group membership compared to the no/infrequent comparator outcome. For study 2 overall, we found weak associations between state-level APS at approximately age 14 and drinking or binge drinking trajectory membership. However, we also found a consistent, slight decreased odds of membership in the normative drinking days group (compared to both the no/infrequent and low frequency drinking groups) with a more stringent policy environment. Furthermore, sensitivity analyses showed APS score may decrease odds of membership in drinking day trajectory groups typified by earlier initiation of alcohol use. In study 3, we examined a more granular environmental exposure—the neighborhood environment. Using baseline data from participants in a human immunodeficiency virus (HIV) prevention study who lived in Baltimore, we evaluated the association between neighborhood disorder and substance use behaviors. We compared neighborhood-level disorder at residential address (live/sleep neighborhood) to a more dynamic definition of exposure to neighborhood disorder based on where individuals reported spending their time engaged in specific activities (activity spaces). Using data from a previously collected Neighborhood Inventory for Environmental Typology (NIfETy) instrument assessment, we calculated objective neighborhood disorder scores for participants (live/sleep scores and activity-weighted scores). While higher live/sleep neighborhood disorder was associated with higher prevalence of self-reported harmful alcohol use, results were imprecise (adjusted prevalence ratio, PR: 1.14, 95% CI 0.94, 1.37). We found a similar directional relationship between live/sleep disorder and injection drug use in the past 6 months (adjusted prevalence ratio: 1.03, 95% CI 0.88, 1.22). The relationship of live/sleep disorder to injection-related risk behaviors (among those who injected drugs in the past 6 months) was inconsistent. Higher live/sleep neighborhood disorder was associated with reduced prevalence of non-alcohol related treatment (among those with a history of drug use; adjusted prevalence ratio: 0.97, 95% CI 0.89, 1.07). Contrary to our hypothesis, there was little variation in results or model fit between the live/sleep and activity-weighted exposures. However, time spent in live/sleep location modified observed relationships for harmful alcohol use and injection drug use. Understanding exposure to factors such as disorder at a micro-space level (both location and time spent at a location) may help explain risk behaviors and identify priorities for intervention. These three studies highlight state policy and neighborhood as two of the many environmental forces that influence individual substance use behaviors. A better understanding of how state-level policy and neighborhood environments influence behavior is critical to initiatives focused on substance use and related harm reduction. These findings are therefore relevant to policy makers and community advocates at both the state and local-level. These dissertation studies also highlight the complexity inherent in quantifying and assessing the impact of both policy and neighborhood exposures on individuals. Further research is needed to examine critical time periods for the influence of neighborhood and state-level policies on substance use across the life-course.