What question did this study set out to answer?

To demonstrate the application of generalized raking and multiple imputation to control for unmeasured confounding when key covariates are available only in a subset of linked electronic health records.

August 16, 2026American Journal of Epidemiology

Empirical application of missing data methods to address confounding in health insurance claims-based analyses using electronic health records available on a subset: an example from the Food and Drug Administration’s sentinel system.

View Full Paper

Ask AI

Bookmark

To demonstrate the application of generalized raking and multiple imputation to control for unmeasured confounding when key covariates are available only in a subset of linked electronic health records.
Linked claims and electronic health record (EHR) data for 912 hospitalized patients (449 with COVID-19 and 463 with influenza) from Kaiser Permanente Washington.
Evaluated 90-day arterial thromboembolism (ATE) risk, applying generalized raking (GR) and multiple imputation (MI) to adjust for EHR-ascertained body mass index (BMI) available in 31.0% of COVID-19 and 38.5% of influenza patients.
Adjusted hazard ratios for 90-day arterial thromboembolism risk were qualitatively similar with and without adjustment for EHR-ascertained BMI.
Both generalized raking and multiple imputation produced consistent effect estimates when accounting for partially missing confounder data.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Observational study shows consistent thromboembolism risks after adjusting for missing BMI data in COVID-19 and flu patients, indicating robust missing data methods.

Cite This Study

Pocobelli et al. (2026) studied this question.

synapsesocial.com/papers/6a817a4cf2fb91fc834ae026 https://doi.org/10.1093/aje/kwag198

View Full Paper

Ask AI

Bookmark

Implication

Observational study shows consistent thromboembolism risks after adjusting for missing BMI data in COVID-19 and flu patients, indicating robust missing data methods.

Empirical application of missing data methods to address confounding in health insurance claims-based analyses using electronic health records available on a subset: an example from the Food and Drug Administration’s sentinel system.

Key Points

Discussion

Implication

Cite This Study

Empirical application of missing data methods to address confounding in health insurance claims-based analyses using electronic health records available on a subset: an example from the Food and Drug Administration’s sentinel system.

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider