Improving the Validity of the Behavioral Risk Factor Surveillance System Alcohol Measures

Abstract

Background: Valid measurement of alcohol use can be difficult in surveys, which are subject to biases like underreporting and differential nonresponse. Still, monitoring trends, policy impacts, disparities, and related issues all require valid individual-and state-level drinking data collected over time.Here, we propose a double-adjustment approach for improving the validity of the Behavioral Risk Factor Surveillance System (BRFSS) alcohol measures.Methods: Validity analyses of the 1999 to 2016 BRFSS, a general population survey of U.S. adults. Measures are aggregated to state level for N = 918 observations, single-adjusted for BRFSS methodologic changes, and double-adjusted by per capita consumption. Fixed-effects models: (i) assess predictive validity using adjusted BRFSS drink volume to predict mortality outcomes and (ii) assess outcome validity using state-level alcohol taxes to predict adjusted BRFSS volume.Results: Neither the raw nor the single-adjusted BRFSS drinking measures were related to mortality in the expected direction, while double-adjusted BRFSS volume and 5+ days were significantly positively related to mortality, as expected. Spirits and beer taxes were not related to single-adjusted BRFSS drinking in the expected direction. However, spirits and beer taxes were both significantly related to double-adjusted BRFSS volume in the expected directions.Conclusions: Future studies should consider using the double-adjusted BRFSS measures to ensure the validity of drinking survey data in analyses where variation over time is considered.

Association of Recreational Cannabis Legalization With Alcohol Use Among Adults in the US, 2010 to 2019

“…However, prior literature finds that RCLs increased cannabis use during our study period . Second, BRFSS respondents may underestimate alcohol use owing to underreporting and non-response bias . This measurement error is only problematic for our DiD approach if it is systematically related to which states implemented RCLs, which is unlikely.…”

“… 47 , 48 Second, BRFSS respondents may underestimate alcohol use owing to underreporting and non-response bias. 49 , 50 This measurement error is only problematic for our DiD approach if it is systematically related to which states implemented RCLs, which is unlikely. Third, RCLs may have different effects across states owing to states’ particular approaches to cannabis legalization, as well as cultural perceptions of cannabis use.…”

Improved estimates for individual and population-level alcohol use in the United States, 1984-2020

“…However, it is based on several assumptions that require further discussion. First, it assumes that the APC is an ideal estimate for overall consumption level (Gmel & Rehm, 2004), an assumption which seems reasonable for the US and the individual states given extant literature (Poznyak et al, 2013;Subbaraman et al, 2020). Second, it may be reasonable to suspect that much of the difference between the survey's results and the APC is due to alcohol consumed by relatively small groups of heavy drinkers not covered by the usual representative surveys, such as the homeless, the institutionalized, or military personnel (Gmel & Rehm, 2004;Rehm, 1998).…”

“…By triangulating APC and survey data, we can address both of their limitations and achieve a more accurate and representative estimate of alcohol consumption including that of different population subgroups. Triangulation is the standard procedure used in aggregate-level modelling (Manthey et al, 2019), and has been used to adjust drinking data for the United States (US) at the aggregate level, whereby drinking variables are adjusted for demographic subgroups and not for individuals (Subbaraman et al, 2020).…”

The impact of COVID-19 on alcohol sales and consumption in the United States: A retrospective, observational analysis

“…We elected to use both sales and behavior data to provide multiple complementary data components. While the NIAAA sales data benefit from being objective figures, it is limited by being available for only 14 states; conversely, the BRFSS consumption data are survey data and potentially less reliable but available for all 50 states ( Subbaraman, Ye, Martinez, Mulia, & Kerr, 2020 ). In addition, large differences between purchasing and consumption may provide insight into consumer behaviors such as stockpiling.…”