Background: General Practitioners (GPs) respond poorly to postal surveys. Consequently there is potential for reduced data quality and bias in the findings. In general population surveys, response to postal questionnaires may be improved by reducing their length and offering incentives. The aim of this study was to investigate whether questionnaire length and/or the offer of an incentive improves the response of GPs to a postal questionnaire survey.
Methods: A postal questionnaire survey was sent to 800 UK GPs randomly selected from Binley's database; a database containing contact details of professionals working in UK general practices. The random sample of GPs was assigned to one of four groups of 200, each receiving a different questionnaire, either a standard (eight sides of A4) or an abbreviated (four sides of A4) questionnaire, with or without the offer of an incentive (a prize draw entry for a £100 voucher) for completion. The effects of questionnaire length and offer of incentive on response were calculated.
Results: Of 800 mailed questionnaires, 19 GPs did not meet inclusion criteria and 172 (adjusted response 22.0%) completed questionnaires were received. Among the four groups, response ranged from 20.1% (standard questionnaire with no incentive and abbreviated questionnaire with incentive) through 21.8% (standard questionnaire with incentive), to 26.0% (abbreviated questionnaire with no incentive). There were no significant differences in response between the four groups (p = 0.447), between the groups receiving the standard versus the abbreviated questionnaire (% difference -2.1% (95% confidence interval (CI) -7.9, 3.7)) or the groups offered an incentive versus no incentive (% difference -2.1% (95% CI -7.9, 3.7).
Conclusions: Strategies known to improve response to postal questionnaire surveys in the general population do not significantly improve the response to postal questionnaire surveys among GPs. Further refinements to these strategies, or more novel strategies, aimed at increasing response specifically among GPs need to be identified in order to maximise data quality and generalisability of research results.