PTOTSLPCohortArchives of physical medicine and rehabilitation2026

Clinically interpretable prediction models of stroke functional outcomes: A national cohort study of adults in inpatient rehabilitation facilities in the US.

Alison M Cogan, Yan Wen, Dingyi Nie and 3 others

PMID 42586400

WHAT IT FOUND

In US inpatient rehab, discharge self-care and mobility were best predicted by admission self-care and mobility, continence and length of stay.

Therapy minutes had limited influence.

Key findings

01Admission self-care score was the strongest predictor of discharge self-care, and admission mobility score was the strongest predictor of discharge mobility.

02Bladder and bowel incontinence and length of stay were among the strongest predictors, while therapy minutes had limited influence.

03On test data, the self-care model accounted for 58% of variance and the mobility model accounted for 62% of variance.

STILL TO COME

How it was doneWhat they foundWhat it means for PTsWhat it means for OTsWhat it means for SLPs

Read the rest of this summary

You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.

Already have one?

What it does not show

This is an observational prediction study, so it does not show that any predictor caused a better outcome. The data came from inpatient rehabilitation administrative records, so they do not capture patient preferences, goals or stroke-specific details. Therapy minutes had limited variability because facilities usually provide three hours per day, so the study cannot evaluate whether more therapy improves function. Section GG items are hierarchical, so patients with short stays may not have attempted harder items, which can affect scores. The sample is limited to adults admitted to US inpatient rehabilitation facilities that use eRehabData. The models were checked on a split of the same dataset, not on a separate group of facilities.

Declared interests

The paper reports individual author disclosures rather than one funding source. Disclosures included federal and foundation research support, book royalties, and consulting or advisory payments to device companies. One author also reported a journal editor role with no role in evaluation.

The easy way to misread this

Do not read these models as proof that therapy minutes or treatment components cause better function. The study is predictive and observational; length of stay and admission function drove predictions more than therapy minutes.

Read it on PubMed →