Globorisk uses a novel approach for risk prediction that makes the model more applicable to diverse populations. It does so by using local data on risk factors (i.e. smoking, diabetes, blood pressure and cholesterol) as well as data on average death rates of heart disease and stroke from the World Health Organization. In contrast, most other cardiovascular disease models are only applicable to the population(s) in which they were developed and need to be recalibrated for other populations, which is often not done due to technical difficulties.
To estimate the rates of fatal and non-fatal heart disease or stroke, Globorisk uses estimates of case fatality rates for heart disease and stroke by age, sex, and country. For details of how these rates are estimated see the Lancet Diabetes and Endocrinology publication in 2015. The estimated case fatality rates incorporate the fact that these rates increase by age and that they are generally higher in low- and middle-income countries.
To allow for different age patterns of cardiovascular disease, Globorisk uses age as the time scale of the (Cox proportional hazard) model and therefore allows any age-pattern of cardiovascular disease by taking this factor out of the estimation process and using local data on age-specific cardiovascular diseases as explained above. This novel approach removes the need to have age in the prediction model which often leads to underestimating the risk of heart disease and stroke in younger ages and therefore lower use of lifestyle change or medication in these patients.