Abstract: Objective: To analyze the incidence and risk factors of post intensive care syndrome in family members of patients with hemorrhagic stroke and to develop the risk prediction model and evalute its prediction effect. Methods: From August 2020 to August 2021, 230 family members of patients with hemorrhagic stroke from ICU of a tertiary grade A hospital in Shiyan city were selected as the research objects by convenience sampling method. A cross-sectional survey was conducted by using the general information questionnaire, Pittsburgh Sleep Quality Index, Fatigue Assessment Instrument, Hospital Anxiety and Depression Scale, and Impact of Event Scale-Revised. Chi-squared Automatic Interaction Detector (CHAID) decision tree model and Logistic regression model were established respectively, and the differences between the results of the two models were analyzed and compared. Results: Logistic regression analysis showed that gender, age, family income and APACHE II score were independent influencing factors of post intensive care syndrome in family members (P<0.05). CHAID decision tree model analysis showed that APACHE II score was the main factor affecting the occurrence of post intensive care syndrome in the family members, followed by ICU hospitalization time, family income and family member's age. There was no significant difference between Logistic regression model and decision tree model (P>0.05). Conclusion: Logistic regression and decision tree model were combined to screen the risk factors of post intensive care syndrome in family members of patients with hemorrhagic stroke earlier and more comprehensively, and to provide reference for clinical intervention.

Key words: Post Intensive Care Syndrome-Family; hemorrhagic stroke; Logistic regression model; decision tree model; influencing factor