Assessing the performance of QLQ-C30 in predicting all-cause mortality in community cancer patients

Background: To evaluate the ability of QLQ-C30 in predicting mortality in community-based cancer patients in China. Methods: A whole-cluster sampling method was adopted to enroll cancer patients in four communities in Shanghai from 2018 to 2019. The patients were surveyed using a questionnaire enquiring demographic information, cancer types, and QLQ-C30 scale. Death information of participants was collected and updated from community health care centers. Cox regression models were used to assess the relationship between various QLQ-C30 scores (i.e., total score, five dimension scores, and utility score) and all-cause mortality. Results: A total of 3,304 participants were enrolled with a mean age of 63.9 years. Among them, 2,710 patients survived while 594 died by 2023. The mean total QLQ-C30 score in living patients was statistically significantly higher than that in deceased patients (92.96 vs. 85.21, p < 0.001); and the mean values of the five dimension scores and utility score were also significantly higher for the living patients. Cox regression models with the adjustment of covariates also confirmed that higher QLQ-C30 scores were associated with lower risk of death, with the hazard ratio value being 0.81 for the total score, 0.83–0.89 for the dimension scores, and 0.83 for the utility score, respectively (p < 0.001 for all). Conclusions: QLQ-C30 could accurately predict all-cause mortality in Chinese community cancer patients.