A Review of Epidemiological Approaches to Forecasting Mortality and Morbidity

This chapter discusses epidemiological models that take disease processes and related risk factors as the basis for modelling mortality and morbidity. These models can be roughly divided into two groups: statistical regression models and dynamic multistate models. For each group of models, examples are given for infectious and chronic diseases. Strong and weak aspects of each group of models are summarised in the context of their aims and data requirements. The chapter consists of four sections. Section 2.1 is an introduction. In Section 2.2, devoted to the regression models, the Alderson and Ashwood model for prediction of lung cancer mortality in England and Wales and the Murray and Lopez study of the global burden of disease are discussed. Section 2.3 focuses on dynamic multistate models. In Section 2.3.1 the (macrosimulation) method of back-calculation of HIV/AIDS-related mortality, and a (microsimulation) model of the spread of two sexually transmitted diseases, gonorrhoea and chlamydia, are given as examples of the multistate models for infectious diseases. The Dutch model PREVENT and the risk factor intervention models of Manton and colleagues are reported in Section 2.3.2 as examples of the multistate models of chronic diseases. Finally, the models MISCAN (Erasmus University, Rotterdam) and POHEM (Statistics Canada) illustrate the microsimulation approach in multistate modelling of chronic diseases. The final section (2.4) discusses the use of epidemiological models for research and health policy purposes.