Longitudinal Data Analysis of Secondary Prevention of Stroke in Northern Ghana

Abstract

The purpose of the study was to apply an illness-to-death model that will enable us to observe the transition intensities of patients during rehabilitation at some discrete points in time. Finally, discover some risk factors of stroke and estimate the average length of stay of patients at different levels of disease states. A typical review of literature on stroke studies in many research works revealed that none of the articles estimated the possible probabilities of transiting from one disease state to another. To fill this gap, we employed Continuous Time Markov Model (CTMC) in Multi-state Models (MSM) to observe the transition rates of the patients at two monthly intervals for two years. Patient variables are age, sex, marital status, religion, educational status, occupation, location of the patient, comorbidity type, local treatment, smoking, alcohol intake, and hemiparesis. Results from our study reveal that the male sex, local treatment, and patients free from comorbidity contributed to early recovery in all states than the female sex, with no local treatment and one or more comorbidity. Patients with one or more comorbidity and alcohol intake decline in recovery. The outcome of the transition analysis indicated that patients with mild stroke remain in this state for about ten (10) months before recovery and will never become severe if the patient adheres to treatment. While old and older age groups have some chance of transiting to a less severe state and similar rates of transitioning from mild to a more severe state, the youth have better conditions in these states than these two groups. Further research should be carried out to investigate the role of traditional therapy in stroke rehabilitation.