Improving decision making in the management of hospital readmissions using modern survival analysis techniques

Prediction of 30-day unplanned hospital readmission through survival analysis

“…Previous works using survival models have used random survival forests and neural networks achieving 0.7 of concordance index (C-index) [10] . In [11] , a survival Bayesian additive regression kernel model is proposed for modelling 30-day hospital readmission data, reporting 1.74 root-mean-square error (RMSE) between the observed and posterior survival predicted outcome.…”

“… [7] , [8] , [9] Survival analysis (ML) 0.7 C-index, 1.74 RMSE Not as widely used in conjunction with ML. [10] , [11] …”