Australian Epidemiology Association ASM 2018

Using big data to improve vascular risk prediction and better targeted treatment (#114)

Rodney Jackson 1
  1. School of Population Health, University of Auckland, Auckland, New Zealand

Readily available treatments can halve the risk of premature vascular disease but under- and over-treatment is common and there are substantial ethnicity- and deprivation-related inequities in vascular disease burden. The effectiveness of most treatments depends on patients’ risks of developing vascular disease but estimating risk is difficult without risk prediction algorithms and few valid algorithms have been developed.

We have established three overlapping ‘big-data’ cohort studies, a primary care cohort, a hospital cohort and a national cohort. These cohorts are electronically linked to the same routine national health datasets of laboratory investigations, drug treatment, hospitalisations and deaths. Using these linked data we are: i. developing new risk prediction algorithms to assist clinicians estimate vascular risk in multiple high-risk populations; ii. investigating in whom, where and why, under- and over-treatment and inequities in vascular risk and risk management occur; iii. developing and implementing a multi-algorithm risk prediction engine and a ‘big-data’ vascular health information platform to support initiatives to increase appropriate treatment, reduce inequities in vascular disease outcomes and improve overall vascular health.