Commentary on Sacks-Davis <i>et al.</i> (2016): Quantifying the risk environment-effect modification and precision population health

Substance use epidemiology focuses increasingly on the risk environment; however, few studies quantify the differential impacts of physical, social and structural environments on population health risk factors and outcomes. To do so, researchers studying the effects of the risk environment on drug-related harms should consider effect modification analyses. During the last two decades, substance use epidemiology has focused on the role of the ‘risk environment’ and the ways in which contextual factors shape substance use and substance use-related harms. As a conceptual framework, the risk environment refers to the social, economic, structural and political forces that influence substance use and related risk behaviors 1. To date, studies that have attempted to operationalize and quantify the effects of various risk environments have relied primarily on geospatial or multi-level methods. For example, some research has used census tracts or zip codes as administratively defined boundaries to define one's geographic location and environmental exposures 2. Other studies have operationalized an individual's risk environment context via measures of neighborhood poverty, segregation, perceived neighborhood disorder, residential mobility, etc. 3, 4. In this issue of Addiction, Sacks-Davis and colleagues investigate whether the risk environment—operationalized as ‘living context’—modifies the relationship between prescription opioid injection (POI) and the incidence of hepatitis C virus (HCV) in Montréal, Canada 5. They define ‘living context’ as living in the inner city versus the surrounding areas of Montréal, noting differences in total population, average income, and crime rates between these two regions. Both the overall rates of POI and the hazard of HCV infection associated with POI were elevated in the inner city, compared to the surrounding neighborhoods of Montréal. Importantly, these findings emphasize that the effect of prescription opioid injection on HCV infection is different among those residents living in the inner city compared to that among residents living in the surrounding neighborhoods, even after adjusting for established HCV-related risk factors and local socio-economic disadvantage. These results suggest that novel harm reduction strategies focused on POI, as well as increased uptake of opioid agonist therapies (e.g. methadone, buprenorphine/naloxone), may be most impactful among people who inject drugs in the inner city. This work highlights the importance of quantifying effect modification by place of residence to elucidate how risk factors operate differently across contexts, and to identify how public health interventions can be targeted most effectively. While there has been increasing recognition of the overall impact of the risk environment on substance use-related harms (e.g. overdose, HIV disease) 6, 7, there is a paucity of literature that evaluates quantitatively how the risk environment shapes population health risk factors. The findings presented by Sacks-Davis et al. underscore the importance of formally testing for effect modification, and should serve as a call for other researchers to employ these methods. Because aspects of the risk environment are frequently hypothesized to influence substance use and related harms, a strong rationale exists to test for effect modification by factors such as place of residence, drug use context, etc. Reporting guidelines also support the increased adoption of effect modification analyses in substance use epidemiology. For example, the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) recommendations state that effect modification should be conducted and presented with enough information for the reader to calculate effect modification on an additive and multiplicative scale 8. Knol & VanderWeele provide clear guidelines to do so, and recommend presenting effect measures and confidence intervals of the main exposure of interest for each stratum of the potential effect modifier, as well as measures of effect modification on both additive and multiplicative scales, with corresponding confidence intervals and P-values 9. We support these recommendations, and encourage researchers who study the effects of the risk environment to consider, conduct and present the results of effect modification analyses. By quantifying how much the risk environment influences the effect of particular risk factors, research can inform whether and the extent to which public health interventions ought to be targeted. In this manner, effect modification represents one method that might be referred to a ‘precision population health’ approach—that is, the use of social epidemiology to inform context-specific, geographically tailored and population-relevant interventions. We note that effect modification analyses are not without challenges: they require larger sample sizes to detect significant differences and precise measurement of the effect modifier (the value of which can change over time) 10. None the less, the study by Sacks-Davis et al. represents an excellent and demonstrative example of how and why effect modification should be conducted in studies of the risk environment and drug-related harm. B.D.L.M. is supported by the National Institute on Drug Abuse (DP2-DA040236) and by a Henry Merrit Wriston Fellowship from Brown University. None.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.