CFE 2020: Start Registration
View Submission - CMStatistics
Title: Informative presence bias in electronic health records Authors:  Glen McGee - University of Waterloo (Canada) [presenting]
Sebastien Haneuse - Harvard TH Chan School of Public Health (United States)
Brent Coull - Harvard University (United States)
Ran Rotem - Harvard University (United States)
Abstract: Electronic health records (EHRs) offer unprecedented opportunities to answer epidemiological questions. However, unlike in ordinary cohort studies or randomized trials, EHR data are collected somewhat idiosyncratically. In particular, patients who have more contact with the medical system have more opportunities to receive diagnoses, which are then recorded in their EHRs. The goal is to clarify the nature and scope of this phenomenon, known as informative presence, which can cause bias if not accounted for. Whereas previous work has introduced this as an instance of confounding, we instead frame it in the context of misclassification. As a consequence, we show that informative presence bias can occur more broadly than previously thought and that covariate adjustment may not be fully correct for bias. Motivated by a study of autism spectrum disorder, we report on a comprehensive series of simulations to shed light on when to expect informative presence bias and how to mitigate it.