Rates, Delays, and Completeness of General Practitioners' Responses to a Postal Versus Web-Based Survey: A Randomized Trial

Background Web-based surveys have become a new and popular method for collecting data, but only a few studies have directly compared postal and Web-based surveys among physicians, and none to our knowledge among general practitioners (GPs). Objective Our aim is to compare two modes of survey delivery (postal and Web-based) in terms of participation rates, response times, and completeness of questionnaires in a study assessing GPs' preventive practices. Methods This randomized study was conducted in Western Switzerland (Geneva and Vaud) and in France (Alsace and Pays de la Loire) in 2015. A random selection of community-based GPs (1000 GPs in Switzerland and 2400 GPs in France) were randomly allocated to receive a questionnaire about preventive care activities either by post (n=700 in Switzerland, n=400 in France) or by email (n= 300 in Switzerland, n=2000 in France). Reminder messages were sent once in the postal group and twice in the Web-based group. Any GPs practicing only complementary and alternative medicine were excluded from the study. Results Among the 3400 contacted GPs, 764 (22.47%, 95% CI 21.07%-23.87%) returned the questionnaire. Compared to the postal group, the participation rate in the Web-based group was more than four times lower (246/2300, 10.70% vs 518/1100, 47.09%, P<.001), but median response time was much shorter (1 day vs 1-3 weeks, P<.001) and the number of GPs having fully completed the questionnaire was almost twice as high (157/246, 63.8% vs 179/518, 34.6%, P<.001). Conclusions Web-based surveys offer many advantages such as reduced response time, higher completeness of data, and large cost savings, but our findings suggest that postal surveys can be still considered for GP research. The use of mixed-mode approaches is probably a good strategy to increase GPs' participation in surveys while reducing costs.