Design
We conducted a nationwide observational longitudinal prospective study. This study was approved by the Ethics Committee Research (EC Research) of the University Hospitals Leuven (UZ Leuven) (S63869) and the Belgian Information Security Committee, division social security and health (IVC/KSZG/20/218).
Participants
Belgium is a federal state comprised of three regions: the Flemish region (Flanders) in the north, the Walloon region (Wallonia) in the south and the Brussels-Capital region. Flanders and Wallonia each represent five provinces. In total, Belgium has 43 administrative arrondissements.
On 31 December 2019, there were 16,722 active GPs in Belgium and 2,209 GPs in training.12 Belgian GPs are organized into 153 different GP circles: 78 in Flanders13, 61 in Wallonia14 and 14 in Brussels. In total, 8 different electronic medical record (EMR) systems are used in general practice in Belgium.
Measuring instrument
eForms (HealthConnect, Corilus, Gent, Belgium) allow a uniform and secure electronic exchange of structured data (e.g., FHIR, CSV, a custom defined format) between health care workers and health care organizations in Belgium and are integrated into EMR systems in general practice. eForms are typically used to exchange structured data on individual patients for a specific workflow. However, for the current study, an eForm to exchange aggregated data on the level of a GP centre was developed and made available for all Belgian GPs. The eForm was developed between the 17th and 19th of March 2020. At the beginning of the study, the eForms were not available in all EMR systems. Therefore, an alternative method for data collection using a web form was developed by Healthdata (www.healthdata.be), a federal platform organized by Sciensano (www.sciensano.be) to store health care data in Belgium. During the study, the eForms were implemented in all EMR systems.
The eForm and web form consisted of three parts: a part on the burden on the GP centre, an epidemiological part and a part on the availability of personal protection equipment (PPE). The first part contained five questions about the need for support, created by Vioras (Vioras, Tielt-Winge, Belgium, www.vioras.be): 1) were you able to perform your critical tasks today? 2) were you able to perform your tasks with a sufficient level of quality (according the guidelines)? 3) Do you need help? 4) Do you have enough resources (personnel and material) to execute your duties tomorrow? and 5) Do you have enough resources (personnel and material) to execute your duties next week? Questions were answered with yes or no, and extra information could be added in free text. The other questions concerned the presence and absence (due to an illness) of GPs in the GP centre and the workload in the GP centre (increased, normal or decreased). The epidemiological part questioned: 1) the number of telephone and physical consultations, 2) the percentage of consultations related to respiratory problems, and 3) the number of patients referred to a triage centre or the emergency department due to COVID-19 suspicion. The last part contained questions on the availability of PPE, namely, surgical masks, FFP2 masks, disinfectant hand gel, gloves and protective suits. The respondents indicated whether they had a significant shortage, a small shortage, a small surplus or a significant surplus for each category.
GPs were informed by their EMR developer (email, website) that the eForm (and web form) was available and they were invited to participate. A member (GP) of each GP centre was asked to send an eForm or fill out the web form each day, reporting the situation at that centre. A reminder to fill out the eForm (or web form) was sent each day (except weekends and holidays) to all GPs that participated at least once. From the beginning of June 2020, GP centres were asked to send the eForm once a week instead of each day. Data from the 19th of March until the 17th of August 2020 were included.
The eForms were sent to the principal investigator of this project (BV), who is also a general practitioner, who integrated the received eForms automatically into a .csv file. This file was shared with Healthdata (www.healthdata.be). Healthdata integrated all data from the eForms and web forms and shared the data with Vioras (www.vioras.be), which created daily reports based on the five questions about the need for support.
Reporting
Each day, a report was prepared for the chairmen of the GP circles with the situation in their circle (for example, see www.vioras.be). This report consisted of a graphical presentation showing all GP centres in their circle and a colour code representing the status of each centre (red: urgent action needed, orange: upcoming problems in the following day(s), light green: possible problems in the following week, dark green: everything under control, grey: no response). In addition to the graphical presentation, they also received a detailed report with all of the specific comments the GP centres registered. This report directed the actions of the circles to support the GP centres in their region.
Because the GP centres and circles were identified based on their geographical location, a hierarchical structure could be made, and overarching reports were generated on the level of primary care zones (approximately 100,000 inhabitants, only in Flanders), arrondissements, provinces, federal regions and country to give policy makers tools to direct their actions.
Data analysis
Descriptive statistics were used. The denominator to calculate the epidemiological incidences was based on the number of GPs per practice and the estimated number of patients per GP. This number was estimated separately for each district (arrondissement) in Belgium based on 2018 data on population statistics (www.statbel.fgov.be) and the number of active GPs, defined as GPs having \(\ge\) 500 contacts/year (www.riziv.fgov.be). SAS 9.4 was used to perform the data analyses.