Electronic health records (EHR) provide valuable resources for longitudinal studies and understanding risk factors associated with poor clinical outcomes. However, they may not contain complete follow-ups, and the missing data may not be at random since hospital discharge may depend in part on expected but unrecorded clinical outcomes that occur after patient discharge. These non-ignorable missing data requires appropriate analysis methods. Here, we are interested in measuring and analyzing individual treatment beneflts of medical treatments in patients recorded in EHR databases. We present a method for predicting individual beneflts that handles non-ignorable missingness due to hospital discharge. The longitudinal clinical outcome of interest is modeled simultaneously with the hospital length of stay using a joint mixed-effects model, and individual beneflts are predicted through a frequentist approach: the empirical Bayesian approach. We illustrate our approach by assessing individual pain management beneflts to patients who underwent spinal fusion surgery. By calculating sample percentiles of empirical Bayes predictors of individual beneflts, we examine the evolution of individual beneflts over time. We additionally compare these percentiles with percentiles calculated with a Monte Carlo approach. We showed that empirical Bayes predictors of individual beneflts do not only allow examining beneflts in speciflc patients but also reflect overall population trends reliably.
|Translated title of the contribution||Measuring Individual Benefits of Medical Treatments Using Longitudinal Hospital Data with Non-Ignorable Missing Responses Caused by Patient Discharge: Application to the Study of Benefits of Pain Management Post Spinal Fusion|
|Number of pages||26|
|Journal||Revista Colombiana de Estadistica|
|State||Published - Jul 14 2022|
Bibliographical noteFunding Information:
The authors thank Lorraine Maw, M.A., from the Mental Health Research Center at Eastern State Hospital, Lexington, KY, USA, for editorial assistance. They also thank Drs. Jonathan Mahnken, Jo Wick, and Milind Phadnis from the University of Kansas Medical Center, KS, USA, for their comments and suggestions. The authors thank two anonymous reviewers for comments and suggestions that contributed to improve the quality and contents of the paper. No commercial organizations had any role in writing this paper for publication. Dr. Zhang is a consultant at Boston Strategic Partners, Inc. (BSP), Boston, MA, but received no financial support from BSP for this study. None of the authors had conflict of interest in the last 36 months. None of the authors is or has been an employee or a paid contractor of Cerner Corporation, the owner of the pain database used in this article, and none of them was involved in the collection of the data contained in the database.
© 2022, Universidad Nacional de Colombia. All rights reserved.
- Electronic health records
- Empirical Bayesian prediction
- Joint mixed models
- Non-ignorable missing data
- Observational data
- Random effects
ASJC Scopus subject areas
- Statistics and Probability