AI and Machine learning: Risk prediction models for stroke outcomes

Stroke is a major healthcare burden, but wide variation exists in the uptake of effective interventions. Efforts to improve the quality of stroke management rely on rigorous outcome data to avoid misleading comparisons being made between hospitals. Various stroke indexes are available (e.g. Charlson, Elixahauser, Stroke severity scales and Frailty) to predict poor outcome. Limited data are available that compares these indexes for stroke using the same cohort. The student will be working within the Big Data, Epidemiology and Prevention Division (Stroke and Ageing Research) from Monash University to reach the following research aims: • evaluate published current evidence related to stroke indexes e.g. Charlson, Elixahauser, Stroke severity scales and Frailty for patients with stroke (phase 1) • identify and compare stroke indexes that may have a predictive value for a poor outcome (phase 2) After obtaining ethics approval, the student will access large, linked databases and will be expected to complete a systematic review and conduct data analysis. The student is expected to summarise the work completed and their contributions to the research in a thesis by publication. With supportive supervision, the student will develop the following skills and experience during the research program: • A taste of working within Public Health Research • Basic knowledge of statistical analysis using STATA • Ability to undertake a systematic literature review • Publication in high level scientific journals