Longitudinal Investigation of Rehospitalization Patterns in Spinal Cord Injury and Traumatic Brain Injury Among Medicare Beneficiaries.

The sample was limited to older Medicare fee-for-service beneficiaries who received inpatient rehabilitation after acute hospitalization for spinal cord injury or traumatic brain injury, so it may not apply to younger patients, Medicare Advantage enrollees, or people who did not need inpatient rehabilitation. Patients who did not survive 90 days after discharge, had missing key data, enrolled in Medicare Advantage, or had a later rehabilitation stay were excluded. The study describes claims patterns and does not prove that any variable caused readmission or that treatment changed risk. The authors intended to describe the whole dataset, not to build a validated predictive model, so the interactive tools should not be treated as proven prediction tools. Some factors that may affect readmission, such as traumatic brain injury severity, were not available in the claims data and were not modeled. Emergency department visits without admission were not counted as readmissions. Monthly rates could include more than one readmission for the same person. The models used only months 1, 2, 4, 6, 8, 10, and 12 because of memory limitations.

The authors report no conflicts of interest and state that the manuscript does not contain information about medical devices.

Do not treat the interactive estimates as a validated prediction tool or as evidence that therapy prevents readmission. The study describes associations in older Medicare beneficiaries with spinal cord injury or traumatic brain injury, and the authors say the analysis was descriptive rather than predictive.