On February 27, 2020, the first patient who tested positive for the coronavirus (COVID-19) was reported in the Netherlands, followed by the first death from the same virus on March 6. The social consequences of corona were major. It was the task of Statistics Netherlands to map out these consequences. The new reality required CBS researchers to use all their knowledge and expertise in the field of statistics to show the impact of the virus on society.
Last month, the website published an article explaining how deeply the corona crisis has affected the lives of the inhabitants of the Netherlands. Many spent more time behind a desk, they exercised less and more people (especially young people) reported psychological complaints. But because more people were at home, there were also fewer burglaries in homes. In the first corona year, the higher mortality and the decline in immigration already led to a halving of the population growth compared to 2019.
What did corona do to statistics?
On April 3, 2020, the first CBS report was published reporting that higher mortality in the preceding weeks coincided with the start of the corona crisis. A month later, the first publication appeared in which the term excess mortality was used. It is the statistical term used in research to get a handle on the number of extra deaths, for example during a heat wave, flu epidemic and now the corona pandemic. On May 22, 2020, the first custom table was published with an estimate of the expected number of deaths according to various characteristics if there had been no corona pandemic. This table would appear weekly from then on.
Acceleration of statistics
Many statistics have been published more quickly, such as the development of the number of bankruptcies pronounced by courts and the absenteeism statistics. The publication frequency of part of the health statistics has been increased from once a year to once a quarter. New publications have appeared to evaluate corona policies, such as the use of emergency financial arrangements. Consider, for example, the use of the Tozo (Temporary Bridging Scheme for Self-Employed Entrepreneurs), the number of users of the TOGS (Entrepreneurs Affected Sectors COVID-19 Allowance) and the NOW (Emergency Measure Bridging for the Preservation of Employment).
Problems with perception
Due to the lockdown and the contact restrictions, problems quickly arose with the observation by Statistics Netherlands when conducting interviews at home. For that reason, statistics programs have been adapted. Because the hairdressers had to close their doors for a while, the price of a haircut could no longer be observed for the purposes of compiling the consumer price index (CPI). This gradually applied to more goods and services whose sales were temporarily halted. As a result, estimated prices were used for the CPI calculation for these goods and services; incidentally, a way of working that other statistical offices have also used and has been coordinated with the European statistical office Eurostat. By doing so, the CPI could still be published.
The health survey as an example
To show how the statistics had to change under the influence of corona and the corona measures taken, CBS researcher Jan Willem Bruggink was asked to evaluate how the statistics program based on the Health Survey was adjusted and how the publication policy was shaped during the corona years. Bruggink emphasizes that it is very important for reliable statistics to make adjustments to the research process as quickly as possible in the extreme situation of a pandemic, which you can reasonably expect to influence observations. ‘The Health Survey was about three things: how do you deal with the loss of part of the observation? Is it possible to publish more frequently? And should the questions in the survey be adjusted, and if so, in what way?’
Decline response addressed with time series analysis
With all the restrictions on contact, surveying at home was sometimes no longer possible between March 2020 and January 2022. As a result, an important part of the regular survey flow was canceled. People were given extra opportunity to participate in the surveys via the Internet, but this yielded little additional response. All in all, the statistic lost about ten percent response on an annual basis due to the partial disappearance of the at-home surveys. An additional problem was that in-home interviews often reach people with different characteristics than internet observations. Statistics Netherlands was able to rely on experience in using (time series) model analysis to correctly correct this missed response. Bruggink: ‘Surveys from previous years revealed how diverse internet respondents and home-based respondents responded to the questions from the Health Survey. This knowledge was used to correct the results for the period without home monitoring. This was done for eight variables and on a quarterly basis. Those quarterly figures were then included in the weighting model for 2020 and 2021. Variables other than the eight for which the time series models were made also benefited from the correction. The use of (time series) model analysis was possible in this case because it concerned a relatively short period without home observations and sufficient data from the past were available.’
Additional questions about the impact of corona
In order to better monitor the impact of corona, questions were added to the Health Survey from August to December 2020. People were asked if they had (had) corona and how sick it made them. Questions have also been added that can clarify the impact of corona on loneliness, anxiety, gloom and stress. And on people’s behaviour: on their work situation, on their lifestyle (exercise, sports, substance use, diet) and the use of care. These questions have been coordinated with the Municipal Medical Services (GGDs) and the National Institute for Public Health and the Environment (RIVM), who also conducted corona research themselves and with whom Statistics Netherlands collaborated in the Health Monitor. In this way it was possible for the three institutes to publish consistent information. The Health Monitor is published every four years. To gain more insight into the possible impact of corona on health, lifestyle and perceived well-being more quickly, an extra edition was published in 2022, two years ahead of schedule. The corona questions that Statistics Netherlands added to the Health Survey in the August-December 2020 period could already be published in 2021. The 2021 questionnaire has been adjusted with questions about vaccination readiness. Those questions were quickly removed because the vaccination campaign was already in full swing and figures became available about the actual numbers of people who were vaccinated.
The wish to publish health figures more frequently has mainly been expressed by external parties, such as the Social and Cultural Planning Office (SCP), RIVM and the Ministry of Health, Welfare and Sport.
Read also: Pandemic increases the need for sustainable and flexible mobility solutions: survey among 250 Dutch organisations
Also read: Dutch SMEs risk lower credit scores and delayed bankruptcy due to COVID-19