Analysis of self-transcendence status and influencing factors in gastric cancer patients undergoing chemotherapy: a random forest model-based study.

Abstract

To investigate the current status and determinants of self-transcendence in gastric cancer patients undergoing chemotherapy and to establish a foundation for clinical development of interventions. A convenience sampling method was employed to select 507 gastric cancer patients undergoing chemotherapy in the oncology department of a tertiary hospital in Wuxi City, Jiangsu Province, from October 2024 to May 2025. Questionnaire surveys were carried out using the Demographic Characteristics Questionnaire, Brief Illness Perception Questionnaire, Herth Hope Index, Medical Coping Modes Questionnaire, and Self-Transcendence Scale. A random forest model and LASSO regression were used to rank the importance of influencing factors and select characteristic variables. Then, multiple linear regression analyses were combined to determine the main influencing factors. The mean value of the self-transcendence score of gastric cancer patients undergoing chemotherapy was 44.08. The random forest model results demonstrated that prioritizing the key variables was most effective at a lambda (λ) value of 0.048, which aligned with five influencing factors. The independent variables with the top five importance rankings were the degree of confrontation, the self-rated health status, the degree of resignation, the level of hope, and the degree of avoidance. Multiple linear regression analysis revealed that self-transcendence in gastric cancer patients undergoing chemotherapy was predominantly affected by confrontation degree, self-rated health status, resignation degree, hope level, and avoidance degree (P<0.05). The self-transcendence level of gastric cancer patients undergoing chemotherapy is observed to be in the moderate to low spectrum. Healthcare personnel can improve this level by targeting the relevant influencing factors, thereby enhancing quality of life during chemotherapy.