menu ☰
menu ˟

Prevalence modelling methods

What prevalence data does IPH produce?

Prevalence is the number of people who have a condition during a time period. Prevalence data are useful because they tell us how many people have a condition and can help us develop services where and when they are needed.

IPH systematically estimates and forecasts the prevalence of a number of chronic conditions nationally and sub-nationally on the island of Ireland. The data show how prevalence varies across the island and what changes are expected between 2010, 2015 and 2020.  Data is available by age and sex for each Local Health Office of the Health Service Executive (HSE) in the Republic of Ireland and each Local Government District in Northern Ireland.

How are the prevalence data produced?

IPH uses a systematic and consistent method to produce prevalence data for all conditions.

This section provides a broad description of our approach. For full details on a particular condition please go to the conditions page by clicking the link on the right of the page.

 

Estimating and forecasting population prevalence at national and sub-national levels involves four steps. The steps are implemented separately in the Republic of Ireland and Northern Ireland using data specific to the particular country.

 

Step I: Estimating risk in the population

A reference study is used to build the best predictive model of risk in the national population. The best predictive model includes a number of explanatory variables for the condition. The model: 

 

  • Divides the population into risk groups defined by the categories of the explanatory variables
  • Provides an estimate of the risk (at national level) of having the condition in each of the risk groups.

Step II: Modify the initial model and identify the final model

The initial model is the best predictive model of risk based on the reference study. However, it may not be the best model based on other criteria. The initial model is assessed for bias and the feasibility of estimating/projecting the number of people in the population in each of its risk groups. In Step II a final model – a possibly simplified form of the initial model – is identified that:

  • Is as close as possible to the initial model
  • Provides sub-national estimates of population prevalence that are as unbiased and precise as possible
  • Allows the population size of the risk groups in each local area to be estimated as satisfactorily as possible 

Step III: Estimate and forecast the number of people in each risk group in the population

Population-based data (for age and sex) and data from the reference studies (for the other explanatory variables in the final model) are used to estimate and forecast the number of people in each risk group in the population.

Step IV: Estimate and forecast national and sub-national population prevalence

The final model’s national group risk estimates (Step II) are multiplied by the corresponding group population count estimates and forecasts (Step III) to estimate and forecast the number of people with the condition.

Comparison of IPH prevalence estimates with other survey-based estimates

IPH has compared its prevalence estimates with prevalence estimates from other health surveys on the island. The comparison found that the various estimates are broadly consistent. The comparison report includes prevalence estimates from IPH and other health surveys,  highlights the methodological issues in comparing estimates from different surveys, and explains why IPH estimates may be marginally different to the estimated prevalence per cents taken directly from the references studies.

Share this: