Predictive modeling of health trajectories
Flagship Program: Changing Health Trajectories in Chronic Disease
Every person’s health and wellness trajectory is different, and influenced by biological, environmental and social factors. Many of us will suffer from chronic diseases and some of us will experience adverse events as part of the care we receive. If we are unfortunate enough to have an accident, the timing and degree of our recovery will be different from one another, along with our ability to return to normal work and life.
The ability to predict which of us is at risk of developing a chronic disease, experiencing an adverse event as part of our care, or recovering poorly from an injury has enormous future value for improving health and wellness, reducing cost and maximising quality of life and productivity.
Predictive individual risk is not only useful for the purpose of targeting specific populations, but also, to the extent that the risk cannot be avoided, for the purpose of financial planning and allocation of healthcare resources.
Under this project, researchers from both Western Sydney University (WSU) and Southern Methodist University (SMU) – and a PhD student from SMU – will work with one of Digital Health CRC’s US-based participants, HMS, to develop algorithms that can both measure and predict risk at an individual level.
HMS is particularly interested in the following types of risk – the risk of preventable hospital re-admissions, the risk of developing chronic conditions such as diabetes and cardiovascular disease, and the risk of a significant progression of chronic disease (eg. the development of chronic kidney disease in patients with diabetes).
HMS has already developed solutions for Population Health Management, which include several aspects of risk management. This research is expected to complement and augment the existing solutions and add to the state-of-the-art in health predictive and prescriptive modeling.
While prediction of the incidence of chronic conditions and preventable hospitalisation is not a new topic in itself, it has never been done on comprehensive data of this type – including inpatient, outpatient and pharmacy data.
HMS is particularly interested in two types of risks: the risk of preventable hospital readmissions and the risk of developing chronic conditions such as diabetes and cardiovascular disease. The precise list of which chronic conditions should be predicted depends on a number of factors, such as the quality and availability of precise diagnosis in the claims data. In addition, Digital Health CRC may develop predictive models for chronic conditions in the context of other projects.