Risk prediction models for post-intensive care syndrome of ICU discharged patients: A systematic review.
Pengfei Yang, Fu Yang, Qi Wang and 3 others
PMID 39990991WHAT IT FOUND
Sixteen prediction models for post-ICU syndrome could tell which ICU survivors might develop problems, but most were biased and overfitted, so they should be used with caution.
Key findings
01Sixteen studies reported 16 prediction models for post-intensive care syndrome in adult ICU survivors.
02Only one study was at low risk of bias, and most models were biased and overfitted.
03The models commonly used age, pre-ICU function, ICU experiences such as delirium or agitation, and early symptoms such as sleep disorder or early psychological symptoms.
STILL TO COME
How it was doneWhat they foundWhat 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
Most included studies had high risk of bias, and only one was low risk. The models were often biased and overfitted because of missing external validation, missing data reporting, insufficient sample size, incomplete analysis, and binarization of continuous variables. Only three studies had external validation, and one study used temporal validation. PICS was measured with many different tools and at different times, so the review could not combine the studies numerically. The search was limited to English and Chinese, so relevant studies in other languages may have been missed. The review did not find enough evidence to say which prediction model was better than another.
Declared interests
This work was supported by Shanghai health and hospital research projects. The funders had no role in design, implementation, or analysis. The authors declared no conflict of interest.
The easy way to misread this
Do not read the reported validation scores as proof these models are ready for use. Only three studies had external validation, and most models were biased and overfitted.