Jones CD, Falvey J, Hess E, et al. - Researchers constructed a parsimonious predictive model to foretell hospital readmissions from home healthcare (HHC) in Medicare fee-for-service beneficiaries. They developed this model by using patient-level clinical variables. They also sought validation for this model in this retrospective analysis using multivariable logistic regression and gradient boosting machine (GBM) learning. Elixhauser comorbidity index, index hospital length of stay, urinary catheter presence, patient status (ie, fragile health with high risk of complications or serious progressive condition), two or more hospitalizations in prior year, pressure injury risk or presence, and surgical wound presence were the seven variables that were incorporated in the final model as individually predictive of outcome. Fewer readmissions were reported in relation to surgical wounds, either from a total hip or total knee arthroplasty procedure or another surgical procedure. For the full model and parsimonious model, the optimism-corrected c-statistics were 0.67 and 0.66, respectively. This suggests fair discrimination. A good calibration was also reported. The GBM model found similar predictive variables. Experts concluded that the risk of readmission can be predicted by variables accessible to HHC clinicians at the first postdischarge HHC visit, and these can inform care plans in HHC. A better prediction of this outcome might be achieved by including measures of social determinants of health, such as housing instability or social support, in future analyses.