Socioeconomic, treatment, and sociodemographic predictors of health related quality of life among cancer patients in Bangladesh

Cancer patients in low- and middle-income countries face substantial physical, psychological, and financial challenges that compromise quality of life (QoL). This study assessed socioeconomic barriers, clinical treatment, and sociodemographic predictors of Health-related Quality of Life (HRQoL) among cancer patients in Bangladesh. It was a cross-sectional study that randomly enrolled 278 cancer inpatients from three tertiary hospitals in Dhaka and Chattogram (July–September 2025). Data on sociodemographic status, clinical profiles, and socioeconomic barriers were collected via interviews and medical records. HRQoL was assessed using the EORTC QLQ-C30 and analyzed with a median split of the Global health status score of QoL. Following Boruta feature selection, ten machine learning classifiers were developed using stratified 10-fold cross-validation, with model interpretability achieved via SHapley Additive Explanations (SHAP). Mean QoL global health score was severely compromised (23.73 ± 13.59), with high symptom burden for fatigue (79.94 ± 26.01), pain (71.35 ± 27.86), and financial difficulties (89.41 ± 19.76). Bivariate analyses showed significant QoL associations with income, education, marital status, family type, mode of diagnosis, caregiver support, and treatment financing. In the primary median-split analysis, Boruta confirmed 27 predictors, and Gradient Boosting Machine and CatBoost performed best (accuracy 0.914, 95% CI 0.874–0.944; ROC-AUC 0.939; PR-AUC 0.954), well above the no-information rate of 0.622. SHAP analysis of CatBoost model identified mode of cancer diagnosis, family type, monthly family income, primary caregiver identity, and treatment-cost financing as the most influential predictors. Machine learning found key sociodemographic and socioeconomic determinants of HRQoL, highlighting the need for targeted financial protection, caregiver-centered interventions, patient education, and enhanced palliative care access to reduce inequities in inpatient oncology services.