Intended for healthcare professionals

General Practice

Commentary: an opposing view from the Netherlands

BMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7059.734 (Published 21 September 1996) Cite this as: BMJ 1996;313:734
  1. Machiel P Springer, professor of general practicea,
  2. Harm W J van Marwijk, lecturer in general practicea
  1. a Department of Primary Care, University of Leiden, PO 2008, 2301 CB Leiden, Netherlands

    Dutch general practitioners receive similar excessive amounts of mail and are as overworked as their British colleagues—we like to think more so. But we are much more willing to answer questionnaires from our colleagues. Why should this be? Are we nicer, more obsessional, more anxious to express an option? Are we more supportive of research in general practice? Do we get paid for responding?

    The most important factors that determine the response rate to a questionnaire in any country are its subject, length, and quality, together with the “face” it presents. The Midlands study referred to by McAvoy and Kaner dealt with sensitive or perhaps even irritating subjects: general practitioners' attitudes to preventive medicine and to interventions about patients' alcohol use. Questionnaires about issues such as alcohol, ethics, yearly earnings, working hours, attitudes to genitourinary disease, or any subject that makes a general practitioner feel the need to make major changes in the way he or she works tread the same dangerous ground. A survey by our department of 494 Dutch general practitioners' attitudes to alcohol issues had the same low initial response rate, and three written reminders (compared with none in the Midlands study)1 improved the response rate to only 51%.2

    The need for a legitimising face

    However, we are confident that a relevant, short, good quality questionnaire will be filled in by most Dutch general practitioners. It helps the response if the body from which it stems is readily identifiable and presents a respected “face.” Groups that go outside such legitimising channels find it difficult to get good responses in Holland. A national body such as the Dutch institute for primary care research (NIVEL) gets a 100% response in its survey of employment among all Dutch general practitioners. This, however, may also be the result of its policy of constant reminders.

    For a more sensitive subject, it helps to make the questionnaires as local, personal, or interactive as possible—for example, by asking general practitioners over the phone whether they want to receive a particular questionnaire3 and to promise them a summary of the results after the study is finished.

    What else do we do to help improve the response rate? The practical and scientific organisation of Dutch general practitioners is increasingly focused on at the regional level, and most requests for cooperation with studies are dealt with there. Approval is then formally sought from our regional university department (one of eight in Holland), where all research involving general practitioners is also monitored. This approval process is supported strongly by the regional postgraduate training coordinator, and it greatly reduces the number of questionnaires sent out to general practitioners and avoids overuse of particular general practitioners. It is also well known in our local academic hospital and the local ethics committee that involvement of general practitioners has to go through our department. A national ethics committee for research in general practice (METOH) has been unsuccessful and has recently been wound up.

    So our advice from the Netherlands is not to resort to threats and bribery but to improve the “face” of research in British general practice. This is a question of funding, organisation, and ownership. We may have lost to England at football this summer, but we are far ahead in integrating research in general practice into postgraduate education and training. Perhaps achievingthis integration would be the best way to increase questionnaire response rates.

    References

    1. 1.
    2. 2.
    3. 3.