Methods for Addressing Missingness in Electronic Health Record Data for Clinical Prediction Models: Comparative Evaluation
Traditional imputation methods for inferential statistics, such as multiple imputation, may not be optimal for prediction models. The amount of missingness influenced performance more than the missingness mechanism. In datasets with frequent measurements, LOCF and native support for missing values in machine learning models offer reasonable performance for handling missingness at minimal computational cost in predictive analyses.
Authors
- Jean Digitale (ORCID: https://orcid.org/0000-0002-6909-8478)
- Charles E. McCulloch (ORCID: https://orcid.org/0000-0002-1279-6179)
- Mark J. Pletcher (ORCID: https://orcid.org/0000-0002-6966-1312)
- Efstathios D. Gennatas (ORCID: https://orcid.org/0000-0001-9280-3609)
- Deborah Franzon (ORCID: https://orcid.org/0009-0008-9905-6448)
Institutions
- University of California, San Francisco (US)
- UCSF Benioff Children's Hospital (US)
Publication Details
- Journal
- JMIR Medical Informatics
- Published
- 2025-11-14
- DOI
- https://doi.org/10.2196/79307
- Citations
- 2
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 4.96