Risk prediction models for post-intensive care syndrome of ICU discharged patients: A systematic review
Objectives: This systematic review aimed to assess the properties and feasibility of existing risk prediction models for post-intensive care syndrome outcomes in adult survivors of critical illness. Methods: As of November 1, 2023, Cochrane Library, PubMed, Embase, CINAHL, Web of Science, PsycInfo,...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-01-01
|
| Series: | International Journal of Nursing Sciences |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S235201322400111X |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Objectives: This systematic review aimed to assess the properties and feasibility of existing risk prediction models for post-intensive care syndrome outcomes in adult survivors of critical illness. Methods: As of November 1, 2023, Cochrane Library, PubMed, Embase, CINAHL, Web of Science, PsycInfo, China National Knowledge Infrastructure (CNKI), SinoMed, Wanfang database, and China Science and Technology Journal Database (VIP) were searched. Following the literature screening process, we extracted data encompassing participant sources, post-intensive care syndrome (PICS) outcomes, sample sizes, missing data, predictive factors, model development methodologies, and metrics for model performance and evaluation. We conducted a review and classification of the PICS domains and predictive factors identified in each study. The Prediction Model Risk of Bias Assessment Tool was employed to assess the quality and applicability of the studies. Results: This systematic review included a total of 16 studies, comprising two cognitive impairment studies, four psychological impairment studies, eight physiological impairment studies, and two studies on all three domains. The discriminative ability of prediction models measured by area under the receiver operating characteristic curve was 0.68–0.90. The predictive performance of most models was excellent, but most models were biased and overfitted. All predictive factors tend to encompass age, pre-ICU functional impairment, in-ICU experiences, and early-onset new symptoms. Conclusions: This review identified 16 prediction models and the predictive factors for PICS. Nonetheless, due to the numerous methodological and reporting shortcomings identified in the studies under review, clinicians should exercise caution when interpreting the predictions made by these models. To avert the development of PICS, it is imperative for clinicians to closely monitor prognostic factors, including the in-ICU experience and early-onset new symptoms. |
|---|---|
| ISSN: | 2352-0132 |