Abstract: Objective To systematically evaluate the risk prediction model of acute altitude sickness,understand the bias risk of and application scope of the model. Methods The CNKI,VIP,WanFang,Web of Science,Ovid MEDLINE,PubMed,and Embase databases were electronically searched to collect studies related to the risk prediction model of acute mountain sickness,and the time frame of the search was from the establishment of the database to March 2024. Two researchers independently screened literature and extracted data,and the risk of bias assessment tool for prediction models（ PROBAST） was used to evaluate the risk of bias and applicability of the models. Results A total of 20 papers were included,including 27 prediction models for acute altitude illness. The AUCs of the predictive models included in the studies were 0.593-0.986,the sample sizes of the studies were 32-4 369 cases,and the number of predictors was 1-22.61.9% of the studies were modeled only without internal validation,and 90.5% of the studies were not externally validated. The reported model sensitivity ranged from 0.611 to 0.998. Conclusion The overall predictive performance of the acute altitude sickness risk prediction model is poor,and all included studies have a high risk of bias,and half of the studies have a high risk of applicability.

Key words: Acute mountain sickness, Prediction model, Systematic review, Prediction performance, Prediction model risk of bias assessment tool