Predicting Post-ICU Functional Impairment During Early ICU Admission Using Real-world Electronic Health Record Data.
Anna Krupp, You Wang, Chao Wang and 3 others
PMID 40468706WHAT IT FOUND
Lower AMPAC, older age, reduced nurse-documented mobility, lower Braden, and prior walking device in the first 24 hours predicted functional impairment at discharge in adults with ICU sepsis.
Key findings
01The first AMPAC score was the most important predictor in both the top 5 and top 10 models.
02Among patients with an initial AMPAC score of 18 or less, 61.6% had functional impairment at hospital discharge, compared with 7.8% of patients with a score above 18.
03The top 5 feature model using 24-hour data achieved an AUC-ROC of 0.83 and an accuracy of 0.77.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for RNs
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.
What it does not show
The model was internally validated only and has not been tested externally. It was developed from adults with sepsis in one healthcare setting, so it may not apply to other ICU populations. The model was trained with 1,598 synthetic records, and all synthetic records were placed in the training set. A small percentage of the sample, 14.6%, had only one documented AMPAC score. The model was not tested prospectively or evaluated for effect on patient outcomes. The association between absence of sedative use and impairment was unexpected and was not seen in the 48-hour model. Misclassified cases were more common when patients had lower initial AMPAC scores, dependence on assistive walking devices, and older age, or higher mobility and higher initial AMPAC scores.
Declared interests
The authors declared no potential conflicts of interest.
The easy way to misread this
Do not treat the model as a tested clinical tool or as proof that any single predictor causes impairment. It was developed retrospectively in one sepsis population, used synthetic data, and has not been externally validated or tested for impact on outcomes.