- Split View
-
Views
-
Cite
Cite
Joan GM Deckers, W John Paget, François G Schellevis, Douglas M Fleming, European primary care surveillance networks: their structure and operation, Family Practice, Volume 23, Issue 2, April 2006, Pages 151–158, https://doi.org/10.1093/fampra/cmi118
- Share Icon Share
Abstract
Background. In many European countries, primary care surveillance networks play a role in public health surveillance.
Objectives. To update an inventory of surveillance networks, to describe them and to report on their organization and function in a standardized way. To investigate whether and under what conditions their information can contribute to surveillance at a European level.
Methods. Surveillance networks were defined as ‘A network of practices or community based primary care physicians who monitor one or more specific illness problems on a regular and continuing basis’. For the inventory questionnaires were sent out, followed by site visits to seven networks using a standardized audit checklist.
Results. We sent out 75 questionnaires and received 57 back (73% response rate), with 33 (58% of responders) fitting our selection criteria. National surveillance networks were identified in 11 countries. Many had an infectious disease surveillance component, particularly for influenza. Most were funded by the Ministry of Health, some by research funds. The median number of general practitioners was 120, comprising a stable group of general practitioners and covering a representative sample of the general population. The frequency of reporting varied from daily to annually, depending on the purpose of the network.
Conclusions. A large number of primary care surveillance networks exist in Europe. Their value has been shown with the surveillance of influenza, but the challenge is now to extend their use to other diseases. When fulfilling identical minimal criteria they can provide comparable estimates of morbidity, ultimately leading to improved national and European surveillance.
Deckers JGM, Paget WJ, Schellevis FG and Fleming DM. European primary care surveillance networks: their structure and operation. Family Practice 2006.23: 151–158.
Introduction
The health of the population is a major concern for national governments and for supranational authorities. Data on the health of populations can be obtained from various sources like mortality statistics, hospital discharge registers, disease-specific registers, notification registers or prescription databases.1 However, no data source is totally comprehensive and all have their specific limitations. Many conditions do not lead to hospital admission or death but have a great impact on the health of the population. Data from primary care and population-based surveys are needed to inform on these.1,2
Primary care is—sometimes the only—the entry-point into the health care system and the majority of health problems are managed exclusively in primary care.3 Therefore primary care based networks are potentially an important source of public health information.4 Moreover, primary care is the first source of professional health information next to the general population. In comparison with information obtained in population-based health surveys, data from primary care have the additional advantage of professional assessment by a health care professional. Moreover, although sometimes problematic, the epidemiological denominator can be defined more easily in primary than in secondary health care facilities. In countries where fixed patient lists exist, characteristics of the denominator population are readily available.5
Primary care based surveillance networks typically include a sample of GPs who report about the occurrence of diseases—based on information from daily patient care—to a central office. The networks frequently provide data on infectious diseases, particularly vaccine preventable infections and are involved in public health surveillance.4,6,7 Participating GPs voluntarily provide information from their practice for surveillance purposes.8 These practices are sometimes called sentinel practices, indicating that the information from these practices is used to monitor the health of the entire population by monitoring the morbidity observed in a sample of practices.
Inventories of primary care based surveillance networks in Europe were carried out in 1987 and 1990 by the Belgian Scientific Institute of Public Health in Brussels and by the European General Practice Research Network in 1998.4,9
The objective of the present inventory was to update the previous inventories, to describe the different networks in Europe and to compare the findings with the previous inventories. From this overview well-established surveillance networks were selected to provide more in-depth information during a site visit. The aim of these site visits was to provide a consistent description of each network and to clarify details of their organization and functioning. These visits were also meant to establish the quality of these networks and hence the validity of the information provided. Ultimately we tried to formulate criteria for a common approach and methodology for surveillance networks so that others can learn from this experience. Eventually this may lead to an increased comparability of the information provided by these networks on the European level, and in that way contribute to the Public Health Information activities of the European Commission (EC).
Methods
The activities described in this paper were part of two successive projects funded by the Public Health Programme of the EC. The inventory was carried out in 1999 and 2000 and the site visits were performed in 2002 and 2003.
Inventory
For the inventory the following definition of a primary care network was used: ‘A network of practices or community based primary care physicians who monitor one or more specific illness problems on a regular and continuing basis.’ Networks were contacted by (i) using previous inventories; (ii) contacting national general practitioner associations in Europe; and (iii) informal contacts with colleagues, the project's steering committee, persons met at conferences and so on. When no information was available for a country, letters were sent to organizations we expected to be able to provide this information like the National Ministry of Health.
Seventy-five letters were sent out with a short one-page questionnaire containing questions about the objectives of the network, some organizational characteristics, how the network operates and how it is funded. The first questionnaire was sent off in 1999 and a reminder was posted to non-responders in 2000.
Site visits
Site visits were made to seven networks using a standardized audit checklist to investigate aspects of organization, function and recording quality. We selected networks that responded to the 1999/2000 inventory, had been in place for >10 years, and monitored a variety of diseases and were not limited to one (often infectious) disease. Lastly networks had to be located in a EU Member State during the study period. Networks completed a standardized questionnaire before the site visit and a standardized audit checklist was used for the actual visit. The pre-visit questionnaire was used to collect background information on the network. During the site visit the entire process from data provision by the participating GPs, to data entry, storage and analysis by the central organization was recorded. Details of the seven networks that were visited are shown in the Appendix. These networks are still existing and functioning to date.
Results
Inventory of networks
We received 57 responses to the inventory, giving an overall response rate of 73%. A total of 33 networks (58%) fitted in our definition for a primary care surveillance network. Our inventory indicated that there were no such surveillance networks in four of the older EU Member States (Finland, Italy, Luxembourg and Sweden) and most countries that have joined the EU in May 2004.
The three countries with the highest number of networks (Table 1) were The Netherlands (n = 9), France (n = 6) and Spain (n = 5). Many of the networks in The Netherlands were regional, based at a university and used for general practice research via a medical school (Tables 1 and 2). In France and Spain the networks were mostly regional surveillance networks and used for the surveillance of infectious diseases (Tables 1 and 2). Somewhat less than half of the networks had a national coverage (Belgium, France, Denmark, England & Wales, Germany, Ireland, The Netherlands, Portugal, Slovenia and Switzerland). The median number of primary care physicians involved in all 33 networks was 120 (range 8–1300) and their main objective was usually disease surveillance (76%; Table 2). Most networks collected data on infectious diseases (influenza was frequently cited) although non-infectious diseases were also reported. Funding generally came from government and university budgets (73%; Tables 2 and 3), a few networks (12%) were funded by the private sector (pharmaceutical companies and/or health insurance companies). The networks were usually based in governmental institutions (regional or central government) or universities (medical schools).
Country . | Networks contacted . | Networks responding . | Networks fitting definitiona . | Coverage . | . | |
---|---|---|---|---|---|---|
. | . | . | . | National . | Regional . | |
Austria | 1 | 1 | 1 | 1 | ||
Belgium | 2 | 2 | 1 | 1 | ||
Denmark | 2 | 1 | 1 | 1 | ||
Finland | 2 | 2 | 0 | |||
France | 8 | 6 | 6 | 1 | 5 | |
Germany | 3 | 3 | 3 | 2 | 1 | |
Greece | 3 | 2 | 1 | 1 | ||
Ireland | 1 | 1 | 1 | 1 | ||
Italy | 1 | 1 | 0 | |||
The Netherlands | 17 | 14 | 9 | 3 | 6 | |
Norway | 3 | 1 | 0 | |||
Portugal | 1 | 1 | 1 | 1 | ||
Slovenia | 1 | 1 | 1 | 1 | ||
Spain | 9 | 5 | 5 | 1 | 4 | |
Switzerland | 2 | 1 | 1 | 1 | ||
United Kingdomb | 22 | 15 | 2 | 1 | 1 | |
Total (%) | 78 | 57 (73c) | 33 (58d) | 14 (42e) | 19 (58e) |
Country . | Networks contacted . | Networks responding . | Networks fitting definitiona . | Coverage . | . | |
---|---|---|---|---|---|---|
. | . | . | . | National . | Regional . | |
Austria | 1 | 1 | 1 | 1 | ||
Belgium | 2 | 2 | 1 | 1 | ||
Denmark | 2 | 1 | 1 | 1 | ||
Finland | 2 | 2 | 0 | |||
France | 8 | 6 | 6 | 1 | 5 | |
Germany | 3 | 3 | 3 | 2 | 1 | |
Greece | 3 | 2 | 1 | 1 | ||
Ireland | 1 | 1 | 1 | 1 | ||
Italy | 1 | 1 | 0 | |||
The Netherlands | 17 | 14 | 9 | 3 | 6 | |
Norway | 3 | 1 | 0 | |||
Portugal | 1 | 1 | 1 | 1 | ||
Slovenia | 1 | 1 | 1 | 1 | ||
Spain | 9 | 5 | 5 | 1 | 4 | |
Switzerland | 2 | 1 | 1 | 1 | ||
United Kingdomb | 22 | 15 | 2 | 1 | 1 | |
Total (%) | 78 | 57 (73c) | 33 (58d) | 14 (42e) | 19 (58e) |
‘A network of practices or community based primary care physicians monitoring one or more specific illness problems on a regular and continuous basis.’
Two networks fitted the definition; one had coverage for England and the other for Wales.
Percentage is based on the number of networks contacted.
Percentage is based on the number of networks responding.
Percentages are based on the number of networks fitting the definition.
Country . | Networks contacted . | Networks responding . | Networks fitting definitiona . | Coverage . | . | |
---|---|---|---|---|---|---|
. | . | . | . | National . | Regional . | |
Austria | 1 | 1 | 1 | 1 | ||
Belgium | 2 | 2 | 1 | 1 | ||
Denmark | 2 | 1 | 1 | 1 | ||
Finland | 2 | 2 | 0 | |||
France | 8 | 6 | 6 | 1 | 5 | |
Germany | 3 | 3 | 3 | 2 | 1 | |
Greece | 3 | 2 | 1 | 1 | ||
Ireland | 1 | 1 | 1 | 1 | ||
Italy | 1 | 1 | 0 | |||
The Netherlands | 17 | 14 | 9 | 3 | 6 | |
Norway | 3 | 1 | 0 | |||
Portugal | 1 | 1 | 1 | 1 | ||
Slovenia | 1 | 1 | 1 | 1 | ||
Spain | 9 | 5 | 5 | 1 | 4 | |
Switzerland | 2 | 1 | 1 | 1 | ||
United Kingdomb | 22 | 15 | 2 | 1 | 1 | |
Total (%) | 78 | 57 (73c) | 33 (58d) | 14 (42e) | 19 (58e) |
Country . | Networks contacted . | Networks responding . | Networks fitting definitiona . | Coverage . | . | |
---|---|---|---|---|---|---|
. | . | . | . | National . | Regional . | |
Austria | 1 | 1 | 1 | 1 | ||
Belgium | 2 | 2 | 1 | 1 | ||
Denmark | 2 | 1 | 1 | 1 | ||
Finland | 2 | 2 | 0 | |||
France | 8 | 6 | 6 | 1 | 5 | |
Germany | 3 | 3 | 3 | 2 | 1 | |
Greece | 3 | 2 | 1 | 1 | ||
Ireland | 1 | 1 | 1 | 1 | ||
Italy | 1 | 1 | 0 | |||
The Netherlands | 17 | 14 | 9 | 3 | 6 | |
Norway | 3 | 1 | 0 | |||
Portugal | 1 | 1 | 1 | 1 | ||
Slovenia | 1 | 1 | 1 | 1 | ||
Spain | 9 | 5 | 5 | 1 | 4 | |
Switzerland | 2 | 1 | 1 | 1 | ||
United Kingdomb | 22 | 15 | 2 | 1 | 1 | |
Total (%) | 78 | 57 (73c) | 33 (58d) | 14 (42e) | 19 (58e) |
‘A network of practices or community based primary care physicians monitoring one or more specific illness problems on a regular and continuous basis.’
Two networks fitted the definition; one had coverage for England and the other for Wales.
Percentage is based on the number of networks contacted.
Percentage is based on the number of networks responding.
Percentages are based on the number of networks fitting the definition.
Country . | Number of networks . | Number of GPs per network . | Main objective . | . | Funding . | . | . | Based . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Median (range, if applicable) . | Surveillance (%) . | Research (%) . | Public Gov/univ . | Private . | Both . | Univ. . | Gov. . | Other . | |||||
Austria | 1 | 15 | 1 | 1 | 1 | ||||||||||
Belgium | 1 | 140 | 1 | 1 | 1 | ||||||||||
Denmark | 1 | 110 | 1 | 1 | 1 | ||||||||||
France | 6 | 120 (40–1300) | 6 | 3 | 1 | 2 | 4 | 2 | |||||||
Germany | 3 | 120 (29–600) | 2 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | |||||
Greece | 1 | 9a | 1 | 1 | 1 | ||||||||||
Ireland | 1 | 20 | 1 | 1 | 1 | ||||||||||
The Netherlands | 9 | 35 (8–160) | 3 | 6 | 6 | 2 | 2 | 6 | 1 | 2 | |||||
Portugal | 1 | 200 | 1 | 1 | 1 | ||||||||||
Slovenia | 1 | 808 | 1 | 1 | 1 | ||||||||||
Spain | 5 | 110 (75–167) | 4b | 4b | 4 | 1 | |||||||||
Switzerland | 1 | 250 | 1 | 1 | 1 | ||||||||||
United Kingdom | 2 | 182 (31–333) | 2 | 2 | 1 | 1 | |||||||||
Total (%c) | 33 | 120 (8–1300) | 25 (76) | 7 (21) | 24 (73) | 4 (12) | 4 (12) | 16 (48) | 11 (33) | 6 (18) |
Country . | Number of networks . | Number of GPs per network . | Main objective . | . | Funding . | . | . | Based . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Median (range, if applicable) . | Surveillance (%) . | Research (%) . | Public Gov/univ . | Private . | Both . | Univ. . | Gov. . | Other . | |||||
Austria | 1 | 15 | 1 | 1 | 1 | ||||||||||
Belgium | 1 | 140 | 1 | 1 | 1 | ||||||||||
Denmark | 1 | 110 | 1 | 1 | 1 | ||||||||||
France | 6 | 120 (40–1300) | 6 | 3 | 1 | 2 | 4 | 2 | |||||||
Germany | 3 | 120 (29–600) | 2 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | |||||
Greece | 1 | 9a | 1 | 1 | 1 | ||||||||||
Ireland | 1 | 20 | 1 | 1 | 1 | ||||||||||
The Netherlands | 9 | 35 (8–160) | 3 | 6 | 6 | 2 | 2 | 6 | 1 | 2 | |||||
Portugal | 1 | 200 | 1 | 1 | 1 | ||||||||||
Slovenia | 1 | 808 | 1 | 1 | 1 | ||||||||||
Spain | 5 | 110 (75–167) | 4b | 4b | 4 | 1 | |||||||||
Switzerland | 1 | 250 | 1 | 1 | 1 | ||||||||||
United Kingdom | 2 | 182 (31–333) | 2 | 2 | 1 | 1 | |||||||||
Total (%c) | 33 | 120 (8–1300) | 25 (76) | 7 (21) | 24 (73) | 4 (12) | 4 (12) | 16 (48) | 11 (33) | 6 (18) |
Primary health care centres.
Main objective and funding status of one Spanish network unknown.
Percentages are based on the number of networks fitting the definition.
Country . | Number of networks . | Number of GPs per network . | Main objective . | . | Funding . | . | . | Based . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Median (range, if applicable) . | Surveillance (%) . | Research (%) . | Public Gov/univ . | Private . | Both . | Univ. . | Gov. . | Other . | |||||
Austria | 1 | 15 | 1 | 1 | 1 | ||||||||||
Belgium | 1 | 140 | 1 | 1 | 1 | ||||||||||
Denmark | 1 | 110 | 1 | 1 | 1 | ||||||||||
France | 6 | 120 (40–1300) | 6 | 3 | 1 | 2 | 4 | 2 | |||||||
Germany | 3 | 120 (29–600) | 2 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | |||||
Greece | 1 | 9a | 1 | 1 | 1 | ||||||||||
Ireland | 1 | 20 | 1 | 1 | 1 | ||||||||||
The Netherlands | 9 | 35 (8–160) | 3 | 6 | 6 | 2 | 2 | 6 | 1 | 2 | |||||
Portugal | 1 | 200 | 1 | 1 | 1 | ||||||||||
Slovenia | 1 | 808 | 1 | 1 | 1 | ||||||||||
Spain | 5 | 110 (75–167) | 4b | 4b | 4 | 1 | |||||||||
Switzerland | 1 | 250 | 1 | 1 | 1 | ||||||||||
United Kingdom | 2 | 182 (31–333) | 2 | 2 | 1 | 1 | |||||||||
Total (%c) | 33 | 120 (8–1300) | 25 (76) | 7 (21) | 24 (73) | 4 (12) | 4 (12) | 16 (48) | 11 (33) | 6 (18) |
Country . | Number of networks . | Number of GPs per network . | Main objective . | . | Funding . | . | . | Based . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Median (range, if applicable) . | Surveillance (%) . | Research (%) . | Public Gov/univ . | Private . | Both . | Univ. . | Gov. . | Other . | |||||
Austria | 1 | 15 | 1 | 1 | 1 | ||||||||||
Belgium | 1 | 140 | 1 | 1 | 1 | ||||||||||
Denmark | 1 | 110 | 1 | 1 | 1 | ||||||||||
France | 6 | 120 (40–1300) | 6 | 3 | 1 | 2 | 4 | 2 | |||||||
Germany | 3 | 120 (29–600) | 2 | 1 | 1 | 1 | 2 | 1 | 1 | 1 | |||||
Greece | 1 | 9a | 1 | 1 | 1 | ||||||||||
Ireland | 1 | 20 | 1 | 1 | 1 | ||||||||||
The Netherlands | 9 | 35 (8–160) | 3 | 6 | 6 | 2 | 2 | 6 | 1 | 2 | |||||
Portugal | 1 | 200 | 1 | 1 | 1 | ||||||||||
Slovenia | 1 | 808 | 1 | 1 | 1 | ||||||||||
Spain | 5 | 110 (75–167) | 4b | 4b | 4 | 1 | |||||||||
Switzerland | 1 | 250 | 1 | 1 | 1 | ||||||||||
United Kingdom | 2 | 182 (31–333) | 2 | 2 | 1 | 1 | |||||||||
Total (%c) | 33 | 120 (8–1300) | 25 (76) | 7 (21) | 24 (73) | 4 (12) | 4 (12) | 16 (48) | 11 (33) | 6 (18) |
Primary health care centres.
Main objective and funding status of one Spanish network unknown.
Percentages are based on the number of networks fitting the definition.
Operational information on the networks
All seven networks that were visited monitor a proportion of the national or regional population which varies from 0.5 to 5%. Most networks covered an age, gender and geographically representative population. In Belgium and in both French networks representativeness was assumed without supporting evidence. In health care systems where ambulatory specialists are directly accessible the completeness of the data is questionable. Therefore the Spanish network also includes paediatricians.
Where valid comparisons could be made (e.g. by age, gender and size of practice) participating GPs were often not representative of the national GP population (Table 3). The motivation of the GPs to participate in the networks was reflected in the low annual turn-over rate of the GPs, which was limited to 5–10%.
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Established in | 1979 | 1964 | 1988 | 1984 | 1992 | 1970 | 1989 |
Size | 1.5% | 1.2% | 5% | 2% | 0.25–0.5% | 0.9% | 2.6–3% |
Representativeness patient population | Assumed | - Age | - Age | Assumed | Assumed | - Gender | - Age |
- Gender | - Gender | - Geographic distribution | - Gender | ||||
- Geographic distribution | - Urbanization | ||||||
- Industrialization | - Urbanization | ||||||
- Population Density | |||||||
Approximate number GPs | 158 GPs | 362 GPs | 142 GPs 27 Paediatricians | 1300 (∼150 report weekly) | 133 GPs | 65 GPs | 171 GPs |
Representativeness GPs | Older | - Younger | Younger | Older | Older | Yes | Older |
- Larger practices | Smaller practices | More males | More males | ||||
- More females | |||||||
Annual turn-over network | 10% | <5% | 2–5% | 5–10% | <5% | <5% | <5% |
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Established in | 1979 | 1964 | 1988 | 1984 | 1992 | 1970 | 1989 |
Size | 1.5% | 1.2% | 5% | 2% | 0.25–0.5% | 0.9% | 2.6–3% |
Representativeness patient population | Assumed | - Age | - Age | Assumed | Assumed | - Gender | - Age |
- Gender | - Gender | - Geographic distribution | - Gender | ||||
- Geographic distribution | - Urbanization | ||||||
- Industrialization | - Urbanization | ||||||
- Population Density | |||||||
Approximate number GPs | 158 GPs | 362 GPs | 142 GPs 27 Paediatricians | 1300 (∼150 report weekly) | 133 GPs | 65 GPs | 171 GPs |
Representativeness GPs | Older | - Younger | Younger | Older | Older | Yes | Older |
- Larger practices | Smaller practices | More males | More males | ||||
- More females | |||||||
Annual turn-over network | 10% | <5% | 2–5% | 5–10% | <5% | <5% | <5% |
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Established in | 1979 | 1964 | 1988 | 1984 | 1992 | 1970 | 1989 |
Size | 1.5% | 1.2% | 5% | 2% | 0.25–0.5% | 0.9% | 2.6–3% |
Representativeness patient population | Assumed | - Age | - Age | Assumed | Assumed | - Gender | - Age |
- Gender | - Gender | - Geographic distribution | - Gender | ||||
- Geographic distribution | - Urbanization | ||||||
- Industrialization | - Urbanization | ||||||
- Population Density | |||||||
Approximate number GPs | 158 GPs | 362 GPs | 142 GPs 27 Paediatricians | 1300 (∼150 report weekly) | 133 GPs | 65 GPs | 171 GPs |
Representativeness GPs | Older | - Younger | Younger | Older | Older | Yes | Older |
- Larger practices | Smaller practices | More males | More males | ||||
- More females | |||||||
Annual turn-over network | 10% | <5% | 2–5% | 5–10% | <5% | <5% | <5% |
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Established in | 1979 | 1964 | 1988 | 1984 | 1992 | 1970 | 1989 |
Size | 1.5% | 1.2% | 5% | 2% | 0.25–0.5% | 0.9% | 2.6–3% |
Representativeness patient population | Assumed | - Age | - Age | Assumed | Assumed | - Gender | - Age |
- Gender | - Gender | - Geographic distribution | - Gender | ||||
- Geographic distribution | - Urbanization | ||||||
- Industrialization | - Urbanization | ||||||
- Population Density | |||||||
Approximate number GPs | 158 GPs | 362 GPs | 142 GPs 27 Paediatricians | 1300 (∼150 report weekly) | 133 GPs | 65 GPs | 171 GPs |
Representativeness GPs | Older | - Younger | Younger | Older | Older | Yes | Older |
- Larger practices | Smaller practices | More males | More males | ||||
- More females | |||||||
Annual turn-over network | 10% | <5% | 2–5% | 5–10% | <5% | <5% | <5% |
Data collection and analysis
Data collection was performed either by standardized paper forms or electronically (Table 4). For electronic data collection, standardized forms on the Internet or standardized extractions from electronic medical records were used. The data collected in the networks in England & Wales and France-SFMG cover all morbidity presented to GPs (‘comprehensive recording’). All 5 others recorded the occurrence of a selection of diseases, usually between 5 and 10. The minimum dataset in all countries consisted of GP and/or patient ID, diagnosis, age and gender of patient (Table 4). The frequency of reporting to the central office was weekly, except for France-SFMG (quarterly) and England & Wales (twice weekly). There were a number of quality control measures in place in the various networks; however, their amount varied. Paper forms are visually inspected, consistency checks are incorporated in data entry programmes and manual assessment of data after data entry is applied (Table 4). Preliminary analyses are performed to compare data to expected figures or national averages. External information sources, such as mortality statistics, disease registers or laboratory reports, are being used for validation of the aggregated data.
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Data collection | Paper form | Electronic data extraction | Paper form | Standard internet form | Electronic data extraction | Paper form | Paper form |
Diseases recorded | 6–8/year | All | 3–5/year | ∼10/year | All | Maximum 15/year | 4–12/year |
Minimum data set | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID |
- Week | - Date | - Date | - Date | - Patient ID | - Week | - Patient ID | |
- No. Pt contacts | - Age | - Age | - Age | - Date | - Number of days recorded | - Date | |
- Age(group) | - Gender | - Gender | - Gender | - Age | - Gender | ||
- Diagnosis | - Diagnosis | - Diagnosis | - Diagnosis | - Gender | - Age group | - Age | |
- Episode type | - Diagnosis | - Gender | - Diagnosis | ||||
- Diagnosis | |||||||
Frequency of reporting by GP | Weekly | Twice weekly | Weekly | Weekly | Preferably monthly or quarterly | Weekly | Weekly |
Frequency of reporting by Organization | Annually | Twice weekly and annually | Quarterly and annually | Weekly and annually | Annually | Annually | Annually |
Quality control measures | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data follow-up | - Missing data follow-up |
- Consistency check during data entry | - Ad hoc validation | - Visual inspection | - Consistency check | - Improbable combinations | - Visual inspection | - Improbable combinations | |
- External validation disease registers, national statistics | - External validation with laboratory reports | - Consistency check during data entry | - Logically expected cases | - Logically expected cases | - Manual assessment | - Consistency check during data entry | |
- Preliminary analysis | - Logically expected incidence figures | - Sampling | |||||
- External validation disease registers, national statistics | |||||||
Analysis | Incidence/prevalence/f requency of consultations by age | Mean weekly incidence rates per 100 000 persons by age and gender | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 100 000 inhabitants per region | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 10 000 patients by age, gender, urbanization, geographic location | Incidence by age and gender |
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Data collection | Paper form | Electronic data extraction | Paper form | Standard internet form | Electronic data extraction | Paper form | Paper form |
Diseases recorded | 6–8/year | All | 3–5/year | ∼10/year | All | Maximum 15/year | 4–12/year |
Minimum data set | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID |
- Week | - Date | - Date | - Date | - Patient ID | - Week | - Patient ID | |
- No. Pt contacts | - Age | - Age | - Age | - Date | - Number of days recorded | - Date | |
- Age(group) | - Gender | - Gender | - Gender | - Age | - Gender | ||
- Diagnosis | - Diagnosis | - Diagnosis | - Diagnosis | - Gender | - Age group | - Age | |
- Episode type | - Diagnosis | - Gender | - Diagnosis | ||||
- Diagnosis | |||||||
Frequency of reporting by GP | Weekly | Twice weekly | Weekly | Weekly | Preferably monthly or quarterly | Weekly | Weekly |
Frequency of reporting by Organization | Annually | Twice weekly and annually | Quarterly and annually | Weekly and annually | Annually | Annually | Annually |
Quality control measures | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data follow-up | - Missing data follow-up |
- Consistency check during data entry | - Ad hoc validation | - Visual inspection | - Consistency check | - Improbable combinations | - Visual inspection | - Improbable combinations | |
- External validation disease registers, national statistics | - External validation with laboratory reports | - Consistency check during data entry | - Logically expected cases | - Logically expected cases | - Manual assessment | - Consistency check during data entry | |
- Preliminary analysis | - Logically expected incidence figures | - Sampling | |||||
- External validation disease registers, national statistics | |||||||
Analysis | Incidence/prevalence/f requency of consultations by age | Mean weekly incidence rates per 100 000 persons by age and gender | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 100 000 inhabitants per region | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 10 000 patients by age, gender, urbanization, geographic location | Incidence by age and gender |
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Data collection | Paper form | Electronic data extraction | Paper form | Standard internet form | Electronic data extraction | Paper form | Paper form |
Diseases recorded | 6–8/year | All | 3–5/year | ∼10/year | All | Maximum 15/year | 4–12/year |
Minimum data set | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID |
- Week | - Date | - Date | - Date | - Patient ID | - Week | - Patient ID | |
- No. Pt contacts | - Age | - Age | - Age | - Date | - Number of days recorded | - Date | |
- Age(group) | - Gender | - Gender | - Gender | - Age | - Gender | ||
- Diagnosis | - Diagnosis | - Diagnosis | - Diagnosis | - Gender | - Age group | - Age | |
- Episode type | - Diagnosis | - Gender | - Diagnosis | ||||
- Diagnosis | |||||||
Frequency of reporting by GP | Weekly | Twice weekly | Weekly | Weekly | Preferably monthly or quarterly | Weekly | Weekly |
Frequency of reporting by Organization | Annually | Twice weekly and annually | Quarterly and annually | Weekly and annually | Annually | Annually | Annually |
Quality control measures | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data follow-up | - Missing data follow-up |
- Consistency check during data entry | - Ad hoc validation | - Visual inspection | - Consistency check | - Improbable combinations | - Visual inspection | - Improbable combinations | |
- External validation disease registers, national statistics | - External validation with laboratory reports | - Consistency check during data entry | - Logically expected cases | - Logically expected cases | - Manual assessment | - Consistency check during data entry | |
- Preliminary analysis | - Logically expected incidence figures | - Sampling | |||||
- External validation disease registers, national statistics | |||||||
Analysis | Incidence/prevalence/f requency of consultations by age | Mean weekly incidence rates per 100 000 persons by age and gender | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 100 000 inhabitants per region | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 10 000 patients by age, gender, urbanization, geographic location | Incidence by age and gender |
. | Belgian Sentinel Network of GPs Belgium . | Weekly Returns Service England & Wales . | Sentinel Network Castilla y Léon Spain . | Sentinel Network Inserm France . | Open Rome/French College of GPs France . | Continuous morbidity registration The Netherlands . | Sentinel network Médicos-Sentinela Portugal . |
---|---|---|---|---|---|---|---|
Data collection | Paper form | Electronic data extraction | Paper form | Standard internet form | Electronic data extraction | Paper form | Paper form |
Diseases recorded | 6–8/year | All | 3–5/year | ∼10/year | All | Maximum 15/year | 4–12/year |
Minimum data set | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID | - GP ID |
- Week | - Date | - Date | - Date | - Patient ID | - Week | - Patient ID | |
- No. Pt contacts | - Age | - Age | - Age | - Date | - Number of days recorded | - Date | |
- Age(group) | - Gender | - Gender | - Gender | - Age | - Gender | ||
- Diagnosis | - Diagnosis | - Diagnosis | - Diagnosis | - Gender | - Age group | - Age | |
- Episode type | - Diagnosis | - Gender | - Diagnosis | ||||
- Diagnosis | |||||||
Frequency of reporting by GP | Weekly | Twice weekly | Weekly | Weekly | Preferably monthly or quarterly | Weekly | Weekly |
Frequency of reporting by Organization | Annually | Twice weekly and annually | Quarterly and annually | Weekly and annually | Annually | Annually | Annually |
Quality control measures | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data not entered | - Missing data follow-up | - Missing data follow-up | - Missing data follow-up |
- Consistency check during data entry | - Ad hoc validation | - Visual inspection | - Consistency check | - Improbable combinations | - Visual inspection | - Improbable combinations | |
- External validation disease registers, national statistics | - External validation with laboratory reports | - Consistency check during data entry | - Logically expected cases | - Logically expected cases | - Manual assessment | - Consistency check during data entry | |
- Preliminary analysis | - Logically expected incidence figures | - Sampling | |||||
- External validation disease registers, national statistics | |||||||
Analysis | Incidence/prevalence/f requency of consultations by age | Mean weekly incidence rates per 100 000 persons by age and gender | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 100 000 inhabitants per region | Incidence/prevalence/frequency of consultations by age and gender | Incidence per 10 000 patients by age, gender, urbanization, geographic location | Incidence by age and gender |
As a routine, scientific and technical staff determined the incidence or prevalence rates by age and gender (Table 4). In Belgium, The Netherlands and Portugal standardization for age and sex to the national population was performed. Those networks that record a selection of diseases often collect additional information about specific diseases on an ad hoc basis; e.g. detailed description of symptoms, patient follow-up, referral to other health care professionals, prescribing information or taking samples. Annual analyses are performed routinely in all networks. Some analyse more regularly, varying from twice weekly (England & Wales) to weekly (France-Inserm) and quarterly (Spain). The only exception concerns the surveillance of influenza-like-illness which during winter is analysed on a weekly basis in all seven networks.
Discussion
In summary, this inventory shows that according to our definition in the year 2000 there were at least 33 primary care based surveillance networks in Europe. Such networks exist in 13 countries and national primary care surveillance networks exist in 11 countries. With the exception of Switzerland, all are current EU Member States. The national networks usually have an infectious disease component, e.g. for the surveillance of influenza.
There were some weaknesses in our study. Firstly, some respondents did not report surveillance systems that we are aware of (e.g. in Sweden and Italy10) and this indicates that the inventory was heavily dependent on the knowledge of the respondents. Secondly, we used a broad definition for a primary care surveillance network. This resulted in a wide range of responses, varying from an institutionalized working group of GPs interested in research to an established computerized network collecting morbidity data on a continuous basis. Thirdly, more networks were contacted in The Netherlands and UK than in other countries. This is partly due to the strong position of primary care in these countries. Yet these two countries were not overrepresented in the final selection of networks. Ultimately we did get a response from the core, mature surveillance networks in Western Europe.
The Eurosentinel survey identified 33 sentinel networks with GPs collecting data on a permanent or periodic basis.9 On the whole, there appears to be relative stability in the networks over the 10 year period. Nine out of eleven national sentinel networks identified in the 1990 inventory were also found 10 years later in the 2000 inventory. A number of regional networks reported in the first inventory were not reported in the second one, particularly in UK. Two national networks were established in the period between the first and the second inventory in Denmark and Germany.
A pan-European primary care based surveillance system already exists for influenza, the European Influenza Surveillance Scheme (www.EISS.org).11 Five of the networks visited participate in EISS. The added value of EISS in influenza surveillance is that the cooperation on the international level provides a critical mass resulting in an exchange of expertise and methodologies, and that national surveillance systems benefit from it, e.g. via harmonization efforts. In a growing EU, such a collaboration results in an improved infrastructure particularly in newly accessed Member States. The challenge now is to identify other conditions that require monitoring by primary care surveillance systems and from there, to create and develop other primary care based disease surveillance networks across the whole of Europe.
Recommendations
Based upon the information we obtained, we formulated the following minimal criteria for a primary care based surveillance network. This concept of a minimum standard could be seen as a pre-requisite when the data are used to describe trends and to make (inter)national comparisons within the context of public health. These include The first three and the fifth recommendations were in place in all seven networks. As stated above, determination of the denominator is dependent on the patient list system, therefore four out of seven networks could easily do so. Quality control varied from network to network, we consider this an issue that needs further attention by all seven networks. Electronic data collection was performed in three networks.
Continuous surveillance of a representative sample of the population. By continuous surveillance throughout the year seasonal effects are taken into account. Furthermore surveillance for several years in succession makes it possible to follow trends over time. The sample size is advised to be ∼1% of the population, which allows the study of common diseases. Increasing the sample size much more will lead to an increased workload, but is not expected to result in additional information;
Representativeness of the population by age and sex. The sampling method has a large impact on the representativeness of the data.12 Representativeness for age and sex can be assumed in most countries, but information on other health-determining factors like socio-economic status (SES) or ethnic origin is not readily available. In England & Wales the SES distribution of the network has been shown to differ from the general population, with an overrepresentation of the least deprived population.13 SES is known to influence consulting rates,14 and consequently biases information obtained in primary care networks. Where networks report on GP-interventions there is also a need to consider the representativeness of the GPs;
Use of standard age groups with break points at 0–5–15–45–65 years. Additional cut-off points at 25 and 75 years are highly recommended. Collection by age group allows a better identification of affected persons and potential risk groups in the community;1
A registered patient list is preferred for determination of the epidemiological denominator. This information is only readily available in health care systems with patients listed to GPs. In health care systems where fixed patient lists do not exist, methods can be developed to calculate estimates.5 Recent developments in Belgium and France where patients can choose their own GP will facilitate the availability of denominator data;
Ability to deliver information on a timely basis. For surveillance of infectious diseases and diseases with a high seasonal influence reporting by the network is preferred to be on a weekly basis. For other conditions annual reporting suffices;
Quality control measures, although time-consuming and therefore costly, must be in place to ensure reliable and valid data. The basic and determining quality of a primary care network is the quality of the diagnosis made by the GPs. It has been shown that the quality of morbidity coding for conditions with clear diagnostic features is higher compared with conditions with more subjective criteria.15 On the level of the entire network many efforts can be made to validate the information, e.g. by comparing this with external information sources;
Electronic data collection obtained from electronic medical records. Networks involved in surveillance currently use either paper forms or automated recording to collect their data. By switching to automated data collection the frequency of reporting can be increased to weekly or even daily and comprehensive recording can be introduced.
In conclusion, primary care based surveillance networks are potentially a valuable tool for public health surveillance in Europe. Although some of the existing networks are limited to the surveillance of infectious diseases, primary care based networks can be a useful source of information on health and health-related problems. However to guarantee adequate monitoring, continuous financial support is essential. Ultimately primary care surveillance networks may prove to be a valuable source of information for the European Public Health Programme.
Declaration
Opinions expressed in this paper are exclusively those of the authors.
Appendix
Contact details of the seven primary care networks that were visited (all websites were visited at 12 November 2005)
Belgium, Brussels, the Scientific Institute of Public Health, Belgian Sentinel Network of GPs: http://www.iph.fgov.be.
Spain, Valladolid, General Directorate of Public Health, Sentinel Network Castilla y Léon: http://www.jcyl.es.
France, Paris, Open Rome/French College of General Practitioners (SFMG): http://omg.sfmg.org.
France, Paris, Inserm, Sentinel Network Inserm: http://www.sentiweb.org.
The Netherlands, Utrecht, Netherlands Institute for Health Services Research, Continuous Morbidity Registration The Netherlands: http://www.nivel.nl.
Portugal, Lisbon, National Institute of Public Health, Sentinel network Médicos–Sentinela: http://www.onsa.pt.
United Kingdom, Birmingham, Birmingham Research Unit of Royal College of General Practitioners, Weekly Returns Service: http://www.rcgp.org.uk/bru/index.asp.
The authors would like to express their gratitude to all coordinators of the participating primary care networks and their staff. This study was funded by the European Commission, Directorate General Health and Consumer Protection as part of the Health Monitoring Programme, and the Health Information Strand, respectively, of the Public Health Action Programme (Project numbers 1998/IND/1021 and 2001/IND/2096).
References
Declich S, Carter AO. Public health surveillance: historical origins, methods and evaluation.
Randall T. National morbidity data have their beginnings in offices of private practice physicians.
Schlaud M, Brenner MH, Hoopmann M, Schwartz FW. Approaches to the denominator in practice-based epidemiology: a critical overview.
Thomas P, Griffiths F, Kai J, O'Dwyer A. Networks for research in primary health care.
Paget WJ, Zimmermann H, Vorkauf H, and the Sentinella Working Group. A national measles epidemic in Switzerland in 1997: consequences for the elimination of measles by the year 2000.
Chauvin P, Valleron AJ. Monitoring the compliance of sentinel general practitioners in public health surveillance: which GPs persevere?
van Casteren V, Leurquin P. Eurosentinel: development of an international sentinel network of general practitioners.
Manuguerra JC, Mosnier A, on behalf of EISS (European Influenza Surveillance Scheme). Surveillance of influenza in Europe from October 1999 to February 2000.
Fleming DM, van der Velden J, Paget WJ. The evolution of influenza surveillance in Europe and prospects for the next 10 years.
Byass P. Empirical modelling of population sampling: lessons for designing sentinel surveillance.
Harcourt SE, Edwards DE, Fleming DM, Smith RL, Smith GE. How representative is the population covered by the RCGP spotter practice scheme? Using Geographical Information Systems to assess.
Carr-Hill RA, Rice N, Roland M. Socioeconomic determinants of rates of consultation in general practice based on fourth national morbidity survey of general practices.
Author notes
aNetherlands Institute for Health Services Research, Utrecht,, The Netherlands and bRoyal College of General Practitioners, Birmingham Research Unit, Birmingham, UK