Back

Comparing Missing Data Imputation Methods for Patient-Reported Outcomes in Esophageal Cancer Research

Kweon, Y. J.; Salman, Y.; Dhillon, S.; Dehghani, M.; Mohammed, E. A.; Crump, R. T.

2025-09-12 health informatics
10.1101/2025.09.10.25335531 medRxiv
Show abstract

Missing data is common in patient-reported outcomes (PRO) research, particularly in oncology settings. We evaluated common methods for handling missing data in esophageal cancer quality of life measurements, namely: multiple imputation by chained equations, variational autoencoder, denoising autoencoder, Bayesian principal component analysis, a deep autoen-coder method with patient-specific embeddings and temporal pattern modeling, SoftImpute, and K-nearest neighbors. Using data from McGill Universitys Esophageal and Gastric Data- and Bio-Bank, we compared these imputation methods for 44 variables of the Functional Assessment of Cancer Therapy-Esophageal patient-reported outcome measure on execution time, distribution preservation, correlation maintenance, imputation accuracy, and clinical classification performance. Our comprehensive validation framework provides evidence-based recommendations for selecting appropriate imputation methods for esophageal cancer PRO research, which may improve the validity and reliability of research findings in this domain.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.