Prediction models for post-traumatic stress disorder in family members of ICU patients: A systematic review

Background: Several predictive models have been developed for post-traumatic stress disorder (PTSD) in intensive care unit (ICU) family members. However, significant differences persist across related studies in terms of literature quality, model performance, predictor variables and scope of applicability.

Aim: This study aimed to systematically review risk prediction models for PTSD in family members of ICU patients, to make recommendations for health care professionals in selecting appropriate predictive models.

Study design: China National Knowledge Infrastructure, VIP database, Wanfang database, SinoMed, PubMed, Web of Science, Cumulative Index to Nursing and Allied Health Literature, The Cochrane Library, Embase and OVID were searched from inception to 1 May 2024. Two independent researchers conducted literature screening, data extraction and applied a risk of bias assessment tool for predictive models to evaluate included studies. The systematic review was registered on PROSPERO (registration number: CRD42024560815).

Results: Seventeen studies were included, with sample sizes ranging from 32 to 2734. Incidence rates of outcomes ranged from 1.6% to 63.6%. The most frequently used predictors were relative's female sex, longer duration of ICU stay, patient's death in the ICU and type of relationship with the patient. Two models reported area under the receiver operating characteristic curve (AUC) values ranging from 0.73 to 0.74; only three models reported calibration, and one study conducted internal validation. Overall, the 17 studies showed good applicability but exhibited a high risk of bias, particularly in data analysis.

Conclusions: Research on predictive models for PTSD risk in family members of ICU patients is in the developmental stage. Future studies should validate existing models or develop high-performance localized predictive models.

Relevance to clinical practice: PTSD can have a significant impact on the families of ICU patients, making early identification of high-risk populations essential for health care professionals to implement timely interventions.