Determinants of survival time among breast cancer patients in Uganda: a survival analysis.

Abstract

Breast cancer remains a leading cause of cancer-related mortality among women in Uganda, with survival outcomes substantially lower than those reported in high-income countries. Despite the increasing burden, limited context-specific evidence exists quantifying the independent socio demographic and economic, clinical, and treatment-related determinants of survival time. This study sought to identify and quantify determinants of survival time among breast cancer patients treated at the Uganda Cancer Institute between 2015 and October 2025. A retrospective cohort design was employed using registry data for 400 patients with confirmed primary breast cancer. The primary outcome was survival time (in days) from diagnosis to death, with right-censoring for patients alive at last follow-up. Survival functions were estimated using the Kaplan–Meier method and compared using log-rank tests. Multivariable analysis was conducted using the Stratified Cox Proportional Hazards regression model. Model adequacy was assessed through Schoenfeld residuals, deviance residual diagnostics, Cox–Snell residuals, and Harrell’s concordance statistic. The overall 365 days (1-year), 1,095 days (3-year), and 1,825 days (5-year) survival probabilities were 75.7%, 45.3%, and 21.6%, respectively, with a median survival time of 966 days. Earlier stage at diagnosis was strongly protective, with patients diagnosed at Stage I experiencing an 85.8% lower hazard of death compared to Stage IV patients (aHR = 0.142; 95% CI: 0.02–0.97). HIV comorbidity significantly increased mortality risk (aHR = 1.45; 95% CI: 1.01–2.11). Treatment-related factors were also critical determinants of survival. Receipt of surgery, radiotherapy, and hormonal therapy significantly reduced mortality risk, while failure to complete treatment increased the hazard of death by more than twofold (aHR = 2.532; 95% CI: 1.729–3.71). Longer delays between diagnosis and treatment initiation were also associated with poorer survival outcomes. The global Schoenfeld test showed no significant violation of the proportional hazards assumption, confirming the validity of the model. In conclusion, advanced stage at diagnosis, HIV comorbidity, and incomplete treatment were found to be key contributors to poor survival outcomes in Uganda. Strengthening early detection strategies, integrating HIV and oncology care, and improving treatment adherence may substantially enhance survival.