Monitoring COVID-19 in Belgian general practice: A tool for syndromic surveillance based on electronic health records

Figures & data

Table 1. Epidemiological data – diagnostics and codes.

Diagnostics	Codes
Suspected COVID-19	ICPC-2 R80 ICD-10 J11.1
Confirmed COVID-19	ICPC-2 A77 ICD-10 B34.2
ARI	ICPC-2 H71, ICPC-2 R74, ICPC-2 R75, ICPC-2 R76, ICPC-2 R77, ICPC-2 R78, ICPC-2 R81
ILI	ICPC-2 R80

ARI: acute respiratory infection; ICD: international classification of diseases; ICPC: international classification of primary care; ILI: influenza-like illness.

Table 2. Criteria to include electronic eForms in the study.

Validation criteria
Number of ARI, ILI, or COVID-19 (suspected/confirmed) cases [each] < number of active patients Percentage of coded diagnoses ≥ 70% and ≤ 100% Number of suspected COVID-19 cases ≤ number of influenza syndromes Daily rates of ARI, ILI, confirmed COVID-19 cases [each] ≤ 2000/100,000 inhabitants

ARI: acute respiratory infection; ILI: influenza-like illness.

Figure 1. Comparison between the Barometer’s data on the number of contacts with a GP for 100,000 inhabitants for suspected or confirmed COVID-19 or acute respiratory infection (ARI) and the official Belgian data on new cases of COVID-19.

Table 3. Description of the variables from the Barometer and the Belgian official data sources.

	Variables	N	Mean (SD)	Min-Max
Barometer	ARI	157	47,33 (17,26)	27,08–88,61
	Suspected COVID-19	157	106,84 (39,66)	64,41–242,91
	Confirmed COVID-19	157	28,18 (20,92)	14,18–97,23
Official sources	New cases of COVID-19	157	3588,89 (3156,92)	1590,29–16157,14
	Hospital admissions COVID-19	157	232,67 (160,42)	118,29–713,00

SD: Standard deviation.

Figure 2. Comparison between the Barometer’s data on the number of contacts with a GP for 100,000 inhabitants for suspected or confirmed COVID-19 or acute respiratory infection (ARI) and the official Belgian data on hospitalisations for COVID-19.

Table 4. Positive components and areas of improvement.

Positive points	Areas of improvement
Technical details
Positive feedback from users of the eForm-based computer syndromic surveillance tool eForms: easy and quick to implement Rapidly available data (day-to-day data collection)	Dependence on EMR software company’s schedule ^aValidation of technical developments to ensure standardisation among EMR software companies ^aFull automation of data transfer between audit and eForm (to avoid manual transcription and mistyped data) ^aConfirmation of reception of the forms by the Healthdata platform ^aUnique identifier for practices (not only GPs)
Project acceptability and participation
High level of participation maintained during the study period Voluntary participation (note: impact of financial incentive unassessed)	Support motivation (especially for long-term data collection)
Data and data quality
Tool effective in identifying health changes (good match with official data) Visualisation of practice outcomes/interest in feedback tool	Strengthen training in the use of diagnostic codes in patients EMR Limiting inter-personal variability in coding ^aStandardised calculation of the percentage of coded diagnoses (among software companies) External validity of data/comparability between different data sources
Project management
Set of communication tools (daily email reminders to participating practices to send data, website/FAQ, tutorials) Collaboration with GPs or GPs associations (meetings, advice and discussions to improve the project)	Financial aspects: integration with other funding sources/systems or based on a representative sample of GP practices should be assessed Implementation of a more formal, multidisciplinary scientific monitoring committee from the early stages of the project

^a Should be fully implemented in a future project.