Abstract
In the Netherlands, several general practice registrations exist. Groups of general practitioners register elements of patient care according to agreed-upon criteria, and these data are collected in a central database. By means of a questionnaire the authors interviewed the managers of all nine computerized registration networks extensively about the possibilities and limitations of their registration. In addition, respondents answered some questions with data from the central database of their network. Various items are collected by nearly all the registration networks, while other items are collected by only one network. Answering questions with data from the central database turned out to be difficult. Organization and manpower are the main obstacles.
In the Netherlands general practitioners hold a key position in health care. They are the first health care professionals people see when they are ill, and they act as gatekeepers to secondary care. Most cases that present to a general practice are taken care of by the general practitioner alone.
The vast majority of general practitioners in the Netherlands register their patient data by computer.1 They use their computer not only to collect and store patient data but also for practice organization and financial administration. Several software systems exist side by side, each with its own possibilities and obstacles.
In the Netherlands several general practice registration networks have been established in the past 25 years. A registration network is a group of general practitioners from different practices who collect elements of patient care according to agreed-on criteria.2,3,4 The collected items are often coded. At regular intervals these data are collected in a central database. We made a systematic inventory of these registrations.
Methods
By means of a questionnaire we interviewed the managers of all 14 registration networks. Data collection took place between summer 1995 and spring 1996 and all the managers participated. The questionnaire addressed objectives, methods, size, and content of the registrations, availability of morbidity items such as symptoms, diagnoses, side effects of therapy, and items referring to the management of the general practitioners and whether the registrations would enable further analyses in their databases and could be used as a sampling frame for follow-up studies. In addition, we asked the managers to answer some specific research questions with data from their central database, to determine whether obstacles would arise when answering research questions.
Results
Each registration network consists of a group of general practitioners from different practices. The size varies from 3 to 103 practices with 7 to 161 general practitioners. These general practitioners collect data daily from 10,000 to 200,000 patients. Mostly the general practitioners code the data themselves, using the International Classification of Primary Care.5 Nine of the registrations are computer based; two others are in a phase of transition from paper forms to computer registration. We consider only the nine computer-based registration networks, working with four different commercially available computer programs. They are all connected to a university.
It is a challenge for the managers of the networks to get the registering general practitioners on a par. The managers of five of these networks made a manual available to the participating physicians. Furthermore, all nine managers organize consensus meetings several times a year in which they discuss and create firm criteria for labeling the presented morbidity. This standardization is important for minimizing interdoctor variation with respect to labeling morbidity and interventions. The managers are responsible for the final quality of the registration and an optimal structure of the database; these are regularly tested through cross-checking for impossible codes and combinations.
Five registrations have developed a central database in which patients are not identifiable, for research purposes. Three of them have set up their database primarily as a sampling frame, allowing researchers to select patients with particular health problems for subsequent recruitment by the patient's general practitioner. Other objectives are education (one network), postmarketing surveillance (one), description and explanation of the “transition” from reasons for encounter to diagnosis (one), mapping the management of the general practitioner (three), improving such management (one), and policy development (one).
The content of the registrations concerns mostly morbidity, connected to the management of the general practitioner in a specific case. Five registrations collect data of all morbidity, whereas four restrict themselves, for instance, to morbidity of a few chapters of the ICPC-classification or to chronic diseases only. ▶ gives an overview of the different registered morbidity items and the management of the general practitioner.
Table 1.
Number of Computerized Networks in the Netherlands Registering Morbidity and Management Items
| Coded | Free Text | Restricted to Selected Diseases | Not Registered | |
|---|---|---|---|---|
| Symptoms | 1 | 2 | 2 | 4 |
| Acute diseases | 4 | None | 4 | 1 |
| Chronic diseases | 5 | None | 4 | None |
| Side effects | 3 | None | None | 6 |
| Reasons for encounter | 1 | 2 | 1 | 5 |
| Physical examination | 1 | 3 | 1 | 4 |
| Diagnostic tests | 2 | 3 | 2 | 2 |
| Watchful waiting | None | 3 | None | 6 |
| Patient education | 1 | 2 | 1 | 5 |
| Follow-up policy | None | 3 | 1 | 5 |
| Referrals | 8 | None | None | 1 |
| Prescriptions | 7 | None | 1 | 1 |
Answering research questions with data from the central database turned out to be difficult. The managers answered fewer questions than they ought to have been able to answer on the basis of their registration data. Various reasons have been advanced for these disappointing results. A few registration networks collect items for specific disease groups only; a specific question can concern another disease. Organization and lack of manpower seemed to be the main obstacles.
Conclusion
When starting research with an existing database it is important from a methodologic viewpoint to be well informed about the method of registration of the specific network. Researchers should be quite specific in what they want from networks and choose the network that best fits their question. For this purpose a brochure has been written.6 On the other hand it would be advisable for managers of the registration networks to make their networks better accessible for external researchers. They should develop and lay down procedures for situations in which outsiders ask for information. Good logistics are important. An adequate organization of the privacy of patient data asks for attention.
References
- 1.Van der Lei J, Duisterhout JS, Westerhof HP, et al. The introduction of computer-based patient records in The Netherlands. Ann Intern Med. 1993;119:1036-41. [DOI] [PubMed] [Google Scholar]
- 2.Meyboom-de Jong B. Morbidity registration in general practice in the Netherlands. Huisarts Wet. 1993;36(suppl):49-68. [Google Scholar]
- 3.Metsemakers JF, Höppener P, Knottnerus JA, Kocken RJ, Limonard CB. Computerized health information in the Netherlands: a registration network of family practices. Br J Gen Pract. 1992;42:102-6. [PMC free article] [PubMed] [Google Scholar]
- 4.Middelkoop BJC, Bohnen AM, Duisterhout JS, Hoes AW, Pleumeekers HJCM, Prins A. Rotterdam general practitioners report (ROHAPRO): a computerised network of general practices in Rotterdam, The Netherlands. J Epidemiol Commun Health. 1995;49:231-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lamberts H, Wood M (eds). International Classification of Primary Care. New York: Oxford University Press, 1987.
- 6.Hart HE, van der Wouden JC, Höppener P, van Schendel GJ, Metsemakers JFM, Knottnerus JA. Huisartsgeneeskundige registraties in Nederland. Rotterdam/Maastricht: Instituut Huisartsgeneeskunde, 1996.