Injuries/Violence
Losing sleep? The effects of exposure to gun violence near one’s residence on self-reported sleep quantity in Chicago, 2020-2021 Suzanne G. McLone* Suzanne G McLone Chad M. Coleman Hannah Matzke Nikhil Prachand Matthew P. Fox Jonathan S. Jay
Very few studies have investigated the association between exposure to gun violence and sleep. We used data from the Healthy Chicago Survey, an annual survey with an address-based sampling scheme of adults in Chicago, IL, to address this question. Self-reported survey responses included poverty level, BMI status, health status, and sleep duration. The dates and locations of firearm shootings were accessed via the Chicago Data Portal. Exposure to gun violence was measured as shootings occurring within 400 meters of respondent’s residence either 7 days before (exposed group) or 7 days after (unexposed group) taking the survey. Sleep was dichotomized into <7 hours and >=7 hours. We fitted log-linear regression models to generate RRs and 95% CIs of exposure to gun violence prior to the survey on <7 hours of sleep. We used probabilistic quantitative bias analysis to estimate the magnitude and direction of the impact of differential outcome misclassification on results; the error in classification depends on respondents’ BMI status and/or health status. Overall, 741 survey respondents were included in the study, 40% of whom reported getting <7 hours of sleep. The RR (95% CI) of the impact of gun violence on sleep duration of <7 hours was 1.14 (0.95, 1.37) in the fully adjusted model with age, race/ethnicity, gender, poverty level, BMI status, and health status as covariates. A modest, positive effect between exposure to gun violence and <7 hours of sleep was observed, although with some imprecision. Results of the quantitative bias analysis indicated that our conventional results were biased away from the null, and that there was no effect of exposure to gun violence on < 7 hours of sleep (RR=1.04, 95% SI: 0.89, 1.22). These results indicate that studies investigating the effects of various exposures on self-reported sleep quantity ought to perform quantitative bias analysis on the results to account for the effect of outcome misclassification due to self-report bias.