Risk prediction models for depression in cancer survivors: A systematic review and meta-analysis

Background: To systematically evaluate the risk prediction models for depression in cancer survivors, so as to provide guidance for establishing and improving models. Methods: CNKI, Wanfang Database, Sinmed, PubMed, Web of Science, The Cochrane Library, and Embase were searched for studies on cancer survivors published before July 1, 2024. The prediction model risk of bias assessment tool was used to evaluate the quality of the studies on the prediction model, and the Stata 15 software was used to conduct a meta-analysis of the predictive variables for the establishment of the model. Results: Seven articles were included, and 7 prediction models from 7 studies were included in this review. The results of the PROBAST evaluation showed that all the 7 studies were at high risk of bias, but the applicability was good. The area under the curve (AUC) = 0.685 to 0.885. All the included studies were validated internally, using the Bootstrap test, Hosmer–Lemeshow test, and temporal validation, respectively. The random effects model was used to perform the meta-analysis on AUC and C-index. The meta-analysis result of AUC was 0.705 (95% CI: 0.691–0.718), I 2 = 92.6% ( P < .001), P < .0001. The meta-analysis of C-index was 0.833 (95% CI: 0.814–0.852), I 2 = 66.5% ( P = .030), P < .0001. Conclusions: At present, risk prediction models for depression in cancer survivors are still in the research and development stage, and the overall quality of research needs to be further improved.