Causal effect modeling with naturalistic rather than experimental data is challenging. In observational studies participants in different treatment conditions may also differ on pretreatment characteristics that influence outcomes. Propensity score methods can theoretically eliminate these confounds for all observed covariates, but accurate estimation of propensity scores is impeded by large numbers of covariates, uncertain functional forms for their associations with treatment selection, and other problems. This article demonstrates that boosting, a modern statistical technique, can overcome many of these obstacles. The authors illustrate this approach with a study of adolescent probationers in substance abuse treatment programs. Propensity score weights estimated using boosting eliminate most pretreatment group differences and substantially alter the apparent relative effects of adolescent substance abuse treatment.
BackgroundAn association between testosterone therapy (TT) and cardiovascular disease has been reported and TT use is increasing rapidly.MethodsWe conducted a cohort study of the risk of acute non-fatal myocardial infarction (MI) following an initial TT prescription (N = 55,593) in a large health-care database. We compared the incidence rate of MI in the 90 days following the initial prescription (post-prescription interval) with the rate in the one year prior to the initial prescription (pre-prescription interval) (post/pre). We also compared post/pre rates in a cohort of men prescribed phosphodiesterase type 5 inhibitors (PDE5I; sildenafil or tadalafil, N = 167,279), and compared TT prescription post/pre rates with the PDE5I post/pre rates, adjusting for potential confounders using doubly robust estimation.ResultsIn all subjects, the post/pre-prescription rate ratio (RR) for TT prescription was 1.36 (1.03, 1.81). In men aged 65 years and older, the RR was 2.19 (1.27, 3.77) for TT prescription and 1.15 (0.83, 1.59) for PDE5I, and the ratio of the rate ratios (RRR) for TT prescription relative to PDE5I was 1.90 (1.04, 3.49). The RR for TT prescription increased with age from 0.95 (0.54, 1.67) for men under age 55 years to 3.43 (1.54, 7.56) for those aged ≥75 years (ptrend = 0.03), while no trend was seen for PDE5I (ptrend = 0.18). In men under age 65 years, excess risk was confined to those with a prior history of heart disease, with RRs of 2.90 (1.49, 5.62) for TT prescription and 1.40 (0.91, 2.14) for PDE5I, and a RRR of 2.07 (1.05, 4.11).DiscussionIn older men, and in younger men with pre-existing diagnosed heart disease, the risk of MI following initiation of TT prescription is substantially increased.
Background The built environment can constrain or facilitate physical activity. Most studies of the health consequences of the built environment suffer from problems of selection bias associated with confounding effects of residential choice and transportation decisions. Purpose To examine the cross-sectional associations between objective and perceived measures of the built environment, BMI, obesity (BMI>30 kg/m2), and meeting weekly recommended physical activity (RPA) levels through walking and vigorous exercise. To assess effect of using light rail transit system (LRT) on changes in BMI, obesity, and meeting weekly RPA levels. Methods Data were collected on individuals before (July 2006–February of 2007) and after (March 2008–July 2008) completion of a light rail system in Charlotte, NC. BMI, obesity, and physical activity levels were calculated for a comparison of these factors pre- and post-LRT construction. A propensity score weighting approach adjusted for differences in baseline characteristics among LRT and non-LRT users. Data were analyzed in 2009. Results More positive perceptions of one’s neighborhood at baseline were associated with a −0.36 (p<.05) lower BMI, 15% lower odds (95% CI=0.77, 0.94) of obesity, 9% higher odds (95% CI = 0.99, 1.20) of meeting weekly RPA through walking, and 11% higher odds (95% CI= 1.01, 1.22) of meeting RPA levels of vigorous exercise. The use of light rail transit to commute to work was associated with an average −1.18 reduction in BMI (p<0.05) and an 81% reduced odds (95% CI= 0.04, 0.92) of becoming obese over time. Conclusions The results of this study suggest that improving neighborhood environments and increasing the public’s use of LRT systems could provide improvements in health outcomes for millions of individuals.
The key problem in testing for racial profiling in traffic stops is estimating the risk set, or "benchmark," against which to compare the race distribution of stopped drivers. To date, the two most common approaches have been to use residential population data or to conduct traffic surveys in which observers tally the race distribution of drivers at a certain location. It is widely recognized that residential population data provide poor estimates of the population at risk of a traffic stop; at the same time, traffic surveys have limitations and are more costly to carry out than the alternative that we propose herein. In this article we propose a test for racial profiling that does not require explicit, external estimates of the risk set. Rather, our approach makes use of what we call the "veil of darkness" hypothesis, which asserts that police are less likely to know the race of a motorist before making a stop after dark than they are during daylight. If we assume that racial differences in traffic patterns, driving behavior, and exposure to law enforcement do not vary between daylight and darkness, then we can test for racial profiling by comparing the race distribution of stops made during daylight to the race distribution of stops made after dark. We propose a means of weakening this assumption by restricting the sample to stops made during the evening hours and controlling for clock time while estimating daylight/darkness contrasts in the race distribution of stopped drivers. We provide conditions under which our estimates are robust to a substantial nonreporting problem present in our data and in many other studies of racial profiling. We propose an approach to assess the sensitivity of our results to departures from our maintained assumptions. Finally, we apply our method to data from Oakland, California and find that in this example the data yield little evidence of racial profiling in traffic stops.
Propensity score analysis (PSA) is a common method for estimating treatment effects, but researchers dealing with data from survey designs are generally not properly accounting for the sampling weights in their analyses. Moreover, recommendations given in the few existing methodological articles on this subject are susceptible to bias. We show in this article through derivation, simulation, and a real data example that using sampling weights in the propensity score estimation stage and the outcome model stage results in an estimator that is robust to a variety of conditions that lead to bias for estimators currently recommended in the statistical literature. We highly recommend researchers use the more robust approach described here. This article provides much needed rigorous statistical guidance for researchers working with survey designs involving sampling weights and using PSAs.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.