RESUMO
Mendelian randomization (MR) is a widely-used method to estimate the causal relationship between a risk factor and disease. A fundamental part of any MR analysis is to choose appropriate genetic variants as instrumental variables. Genome-wide association studies often reveal that hundreds of genetic variants may be robustly associated with a risk factor, but in some situations investigators may have greater confidence in the instrument validity of only a smaller subset of variants. Nevertheless, the use of additional instruments may be optimal from the perspective of mean squared error even if they are slightly invalid; a small bias in estimation may be a price worth paying for a larger reduction in variance. For this purpose, we consider a method for "focused" instrument selection whereby genetic variants are selected to minimise the estimated asymptotic mean squared error of causal effect estimates. In a setting of many weak and locally invalid instruments, we propose a novel strategy to construct confidence intervals for post-selection focused estimators that guards against the worst case loss in asymptotic coverage. In empirical applications to: (i) validate lipid drug targets; and (ii) investigate vitamin D effects on a wide range of outcomes, our findings suggest that the optimal selection of instruments does not involve only a small number of biologically-justified instruments, but also many potentially invalid instruments.
RESUMO
Lead exposure still threatens children's health despite policies aiming to identify lead exposure sources. Some US states require de jure universal screening while others target screening, but little research examines the relative benefits of these approaches. We link lead tests for children born in Illinois between 2010 and 2014 to geocoded birth records and potential exposure sources. We train a random forest regression model that predicts children's blood lead levels (BLLs) to estimate the geographic distribution of undetected lead poisoning. We use these estimates to compare de jure universal screening against targeted screening. Because no policy achieves perfect compliance, we analyze different incremental screening expansions. We estimate that 5,819 untested children had a BLL ≥5µg/dL, in addition to the 18,101 detected cases. 80% of these undetected cases should have been screened under the current policy. Model-based targeted screening can improve upon both the status quo and expanded universal screening.
