Probabilistic models for heart disease and related conditions

The topic of this thesis is mortality and the prevalence in the UK of ischaemic heart disease (IHD) and stroke. In particular, we consider changes in each of these between 1981 and 2000 and quantify the extent to which these changes are due to known risk factors for heart disease: body mass index, diabetes, hypertension, hypercholesterolaemia. Chatterjee et al. (2008a) presented a multiple state Markov model for the development of heart disease and stroke. We develop two parameterisations for this model, one consistent with the 1981 prevalences in the UK of all the risk factors, and IHD, stroke and mortality, and the other consistent with the 2000 prevalences. By taking the 2000-consistent model and then changing the parameters for a given risk factor back to those in the 1981-consistent model, we can quantify the effect on IHD, stroke and mortality if this risk factor had not changed. We also carry out this exercise changing combinations of risk factors.