In 2006, the National Heart Forum (NHF) was commissioned by the U.K. Government to predict and analyse future trends in disease and societal costs attributable to the rise in obesity. In order to achieve this, the NHF built two computer programs one to model future distributions which feeds into a second, a micro-simulation, to model future impacts of changes in the risk factor. This work was published in the Foresight report, "Tackling Obesities: Future Choices-Modelling Future Trends in Obesity and Their Impact on Health".
Subsequently, the NHF has secured a contract with Department of Health (DH, England) to produce an annual update, undertaken work for the U.S. Government Center for Disease Control (CDC) and has undertaken analysis of individual interventions and their long term impact at a local level, among other things.
What is the micro-simulation model?
The NHF micro-simulation is capable of dynamically modelling demographically specified populations of people, on an individual basis. Provided the necessary defining statistics exist the populations can be representative of whole countries (England and the U.S. are working examples); or geographical regions or simply collections of individuals. The populations can be stratified by any number of variables such as sex, class, ethnicity, income, family, geographical region, etc. Various features of the populations can be predicted into the future - until now the emphasis has been on predicting future health outcomes dependent on predicted body mass index (BMI) growth and tobacco consumption (i.e. the number of cigarettes smoked by an individual).
How it works
The micro-simulation operates by simulating representative individuals and deriving population statistics from Monte Carlo methods. Members of any selected population are modelled individually from birth; they age, give birth, their BMI or the number of cigarettes smoked change as they age, contract diseases from which they die or survive, and so on. The individual growth patterns are arranged so that there is agreement with known population statistics at the present day and there is agreement with projected population statistics at selected times in the future. For example, within the context of obesity, the distribution of BMI within the English or U.S. population is forecast to rise in a way that can be quantified. The NHF micro-simulation models individual BMI growth so that the resulting population has the desired distributions at the various check points - every few years, for example. In other words, the individuals are organised so as longitudinally to interpolate the population's known cross sectional characteristics. Diseases that have incidence, survival and mortality risks that are dependent on an individual's BMI, age, sex, etc., are modelled at the individual level; health statistics for the population are derived from aggregating individuals' experiences.
The micro-simulation is capable of modelling the financial and health consequences of making interventions that alter disease risk factors. Risk factors come in many different sorts and are, more or less, easy to model. These include BMI, tobacco consumption and alcohol intake.
