Multistate Models in Public Health Review and Application to the Framingham Heart Study.

Thesis Report.

ABSTRACT: Three major issues are important to understand the overall health of population: health status, classification of population into different health states and the representation of population health in a single index. At the population level, the widely acceptable indicator is ' healthy life expectancy' which is increasingly the best known index for summarizing health status of the population and addressing the discussion of 'compression' and 'expansion' of morbidity. The Sullivan method is used to construct the single index using prevalence data through life table techniques and multistate life tables are used for incidence data. From the methodological and empirical perspective, we have investigated the difference between Sullivan method and multistate method and emphasized the potential utility of the multistate models in the public health. Explicit description of the input data construction from panel data and the representation of output from the multistate life table have been illustrated. We have reviewed some important articles from different international journals that discussed the multistate and Sullivan method that contributed to the methodological development in the public health literature. For the application of life table, we have used Framingham Heart Study (FHS)- a 40 years longitudinal panel study in the field of epidemiology and medicine. A nested series of life table has been constructed. Using the FHS, we have investigated the history of cardiovascular disease. Eventually in the paper two, issues are emphasized throughout (i) the concept of prevalence data and the incidence data, (ii) the measurement of age (age in completed years versus exact age). Key words: Age, Incidence, Prevalence, Health expectancy, Sullivan method, Multistate life table, Matrioska model, Panel study, Cardiovascular disease

Eng.