Abstract:

Objective

To systematically review the risk prediction models for incontinenceassociated dermatitis （IAD） in ICU patients.

Methods

The databases of CNKI，Wanfang，VIP，CBM，PubMed，Embase，Web of Science and the Cochrane Library were searched for the literature on risk prediction models for IAD in ICU patients published up to July 23，2023.Two researchers independently screened the literature，extracted relevant data，and evaluated the risk of bias and applicability of involved literature.Meta-analysis of the predictive value of predictors in the model was performed using Revman5.4.

Results

A total of eight studies were involved，including six model development studies and two external validation studies for the perineal assessment tool （PAT）.Six studies used logistic regression models，two used external validation，and three used internal validation.The area under the receiver operating characteristic curve of the involved models ranged from 0.795 to 0.974.Meta-analysis showed that serum albumin ［odds ratio （OR） = 2.45，95% confidence interval （CI） （1.68，3.59），Z = 4.62，P ＜0.001］，fever （body temperature ≥38 ℃） ［OR=2.29，95%CI （1.39，3.79），Z=3.23，P=0.001］，loose stools ［OR=6.93，95%CI （4.50，10.65），Z=8.81，P ＜0.001］ and PAT ［OR=2.56，95%CI （1.23，5.30），Z = 2.53，P = 0.01］ were influential factors for the risk of IAD in ICU patients.

Conclusions

The existing predictive models for IAD in ICU patients are not ideal，as they exhibit bias in the development，design，statistical analysis and reporting.In the future，model validation and updates should be conducted to improve their performance through large-scale and multi-center studies.

Key words: Intensive care unit, Incontinence-associated dermatitis, Prediction model, System evaluation