ABSTRACT
OBJECTIVES: This study aimed to evaluate competing risks and functional ability measures among patients who had a stroke. DESIGN: A joint model comprising two related submodels was applied: a cause-specific hazard submodel for competing drop-out and stroke-related death risks, and a partial proportional odd submodel for longitudinal functional ability. SETTING: Felege Hiwot Referral Hospital, Ethiopia. PARTICIPANTS: The study included 400 patients who had a stroke from the medical ward outpatient stroke unit at Felege Hiwot Referral Hospital, who were treated from September 2018 to August 2021. RESULTS: Among the 400 patients who had a stroke, 146 (36.5%) died and 88 (22%) dropped out. At baseline, 14% of patients had no symptoms and/or disability while 24% had slight disability, and 25% had severe disability. Most patients (37.04%) exhibited moderate functional ability. The presence of diabetes increased the cause-specific hazard of death by 3.95 times (95% CI 2.16 to 7.24) but decreased the cause-specific hazard of drop-out by 95% (aHR 0.05; 95% CI 0.01 to 0.46) compared with non-diabetic patients who had a stroke. CONCLUSION: A substantial proportion of patients who had a stroke experienced mortality and drop-out during the study period, highlighting the importance of considering competing risks in stroke research. Age, diabetes, white cell count and stroke complications were significant covariates affecting both longitudinal and survival submodels. Compared with stand-alone models, the joint competing risk modelling technique offers comprehensive insights into the disease's transition pattern.
