JMIR Publications

ABSTRACT

Background:

Electronic health records (EHRs) are increasingly used for epidemiologic research to advance public health practice. However, key variables are susceptible to missing data or misclassification within EHRs, including demographic information or disease status, which could affect estimation of disease prevalence or risk factor associations.

Objective:

In this article, we applied methods from literature on missing data and causal inference to assess whether we could mitigate information biases when estimating measures of association between potential risk factors and diabetes among a patient population of New York City (NYC) young adults.

Methods:

We estimated odds ratios (OR) for diabetes by race/ethnicity and asthma status using EHR data from NYU Langone Health. Methods from the missing data and causal inference literature were then applied to assess the ability to control for misclassification of health outcomes in the EHR data. We compared EHR-based associations with associations observed from two national health surveys, the Behavioral Risk Factor Surveillance System and National Health and Nutrition Examination Survey, representing traditional public health surveillance systems.

Results:

Observed EHR-based associations between race/ethnicity and diabetes were comparable to health survey-based estimates, but the association between asthma and diabetes was significantly overestimated (OR EHR=3.01 vs. OR BRFSS=1.23). Missing data and causal inference methods reduced information biases in these estimates, yielding relative differences from traditional estimates below 50% (OR Missing Data=1.79, OR Causal=1.42).

Conclusions:

Findings suggest that without bias adjustment, EHR analyses may yield biased measures of association, driven in part by subgroup differences in healthcare utilization patterns. However, applying missing data or causal inference frameworks can help control for and, importantly, characterize residual information biases in these estimates.