Quarterly figures on health, lifestyle and care use

Conducting the Health Survey during the coronavirus crisis presented a challenge to CBS. Data on (self-perceived) health, lifestyle and care use are obtained by conducting surveys among the population. The survey participants are selected in such a way as to ensure a representative coverage of the entire population living in private households. CBS conducts the survey both online and by means of face-to-face interviews with the aim of obtaining the highest possible response rate. Due to the coronavirus measures, CBS had to discontinue interview surveys temporarily as of mid-March. On a positive note, more people participated online. Although CBS attempted to persuade people who would ordinarily be interviewed at home to take part in the online survey, there was only limited response.
Usually, CBS only publishes the annual figures from the Health Survey. In this case, the need for fast survey results incorporating the possible impact of the coronavirus crisis on health, lifestyle and care use called for a deviation from this method by producing quarterly figures. The quarterly figures are not only over 2020 but also over the quarters of previous years. This helps put the 2020 figures into perspective.
The estimation method for these quarterly figures is based on so-called structural time series models. In this estimation method, data from previous time periods is used to produce reliable quarterly estimates. This model consists of a trend modelling the gradual change in the series and a seasonal component which models the seasonal pattern, if any, over a period of twelve months. These models also take into account the temporary suspension of interviews as a surveytaking method by looking at the differences between online and interview response rates observed in the past.
The coronavirus crisis has led to relatively substantial changes in the survey results over 2020 on a number of variables. As a result, it has become necessary to produce a more flexible trend for this period. The model estimates for these periods are therefore less dependent on patterns from the past. Consequently, some of the error margins around the quarterly figures over 2020 are wider than in previous quarters.

Due to the exceptional circumstances and accompanying uncertainties and methodological challenges, the quarterly figures from the Health Survey are marked provisional.
Back to article