Choice of surveys influences results of health indicators

Knowing the variations in the number of patients or the average life expectancy by age and country are just some of many health indicators used to describe and investigate public health situation and its issues. Measuring health is a necessary task to develop proper policies and strategies, but it is also a difficult one. And the choice of the tools used to perform it may influence its results.

A study published on the International Journal of Public Health by members of the LIFEPATH project revealed that survey choice strongly impacts estimates of educational inequalities and disability using one of the main health indicators used in Europe.

The GALI (Global Activity Limitation Indicator) is one of the three questions of a module developed to be used in EU social surveys, statistical analysis and the Eurobarometer. It has been shown to be a reliable indicator with good predictive validity but, being a self-reported indicator, it is subject to variations in the tendency to report health problems. The GALI is provided through surveys that can vary in terms of sampling design, method of data collection, response rate, whether or not proxy respondents are allowed, and the phrasing of the GALI question.

LIFEPATH researchers wanted to assess whether different surveys that measure the GALI indicator may lead to similar conclusions regarding prevalence and educational inequalities. They analysed three widely used nationally representative European surveys, in order to see whether they provided similar or different estimates of prevalence of GALI disability and of educational inequalities in GALI disability in Europe. 

What they found is that there are important differences in the prevalence and the educational inequalities of GALI disability between the surveys included in the analysis. Which means that is possible to draw different interpretations of the health status of a country, its relative position to other countries and the size of educational inequalities, depending on what survey is used for the measurement. 

These findings have relevant implications for population health monitoring and, in turn, for the development of proper health policies. «At the national level, it is difficult to make reliable assessments of the prevalence of disability since the agreement between different surveys is lacking and there is no gold standard among the three surveys», said Jose Rubio Valverde, junior researcher in public health at the Erasmus MC in Rotterdam and leading author of the study. «We think that further investigations are required to understand what drives these discrepancies».

As long as there is no way of knowing which survey represents reality, the only option is to combine all available data sources and search for patterns that are consistent between surveys. If the surveys agree in the trends, this would mean that a country is consistently improving or worsening. However, comparisons between different country are even harder to perform reliably. LIFEPATH experts thus recommend to be very cautious in using these surveys for cross-country comparisons of (inequalities in) GALI disability.

PDF icon lifepath_press_20181129.pdf47.43 KB
Publication date: 
Thursday, November 29, 2018