Introduction: Methods now exist to detect residual confounding. One requiresan "indicator" with two key properties: conditional independence of the outcome(given exposure and measured covariates) absent confounding and other modelmiss-specification; and an association with unmeasured confounders (like theexposure). We now present a new method for correcting for residual confoundingin time-series and other epidemiological studies. We argue that estimators frommodels that include an indicator with these key properties should have lessbias than those from models without the indicator. Methods: Using causal reasoning and basic regression theory we presenttheoretical arguments to support our claims. In simulations, we empiricallyevaluate our approach using a time-series study of ozone effects on emergencydepartment visits for asthma (AV). We base simulations on observed data forozone, meteorological factors and asthma. Results: In simulations, results from models that included ozoneconcentrations one day after the AV yielded effect estimators with slightly ormodestly less residual confounding. Conclusion: Theory and simulations show that including the indicator based onfuture air pollution levels can reduce residual confounding. Our method differsfrom available methods because it uses a regression approach involving anexposure-based indicator rather than a negative outcome control.
展开▼