Mountford Timothy, MSc

PhD Thesis: Harnessing Healthcare Data for Targeted Mental Health Interventions: Temporo-Geographical Cohort Characterization, Risk Assessment, and Predictive Modelling for Early Identification in the Western Cape province of South Africa. (Supervisor: PD Dr Andreas Haas)

His current research aims to improve mental health care outcomes in the Western Cape Province of South Africa by leveraging electronic health records and machine learning to predict high-risk patients. Given the significant burden of neuropsychiatric disorders in South Africa, particularly anxiety and substance use disorders, the study seeks to address gaps in service accessibility and high mortality rates among psychiatric patients.

The project consists of three interconnected studies:

1. Spatio-Temporal Analysis of Mental Health Service Utilization – This study characterizes the sociodemographic and diagnostic profiles of mental health service users, examines comorbidities, and maps geographical and temporal trends in service utilization.
2. Risk Factors for Readmission and Premature Mortality – This study identifies key predictors of psychiatric readmission and early mortality post-discharge, including demographic factors, comorbid conditions, treatment adherence, and service utilization patterns.
3. Machine Learning for Predictive Risk Modelling – Using insights from the first two studies, this study develops and validates a predictive model to identify patients at high risk of readmission or premature mortality. The model will be integrated into the Provincial Health Data Centre (PHDC) to support clinical decision-making and targeted interventions.

By employing statistical and machine learning techniques, this research aims to enhance resource allocation in the public healthcare system, facilitate early interventions, and ultimately reduce adverse health outcomes for mental health patients in South Africa.