Surveying general practitioners: a new avenue

Abstract

Background: The validity of opinions expressed in questionnaire surveys diminishes with lower response rates. Recent research suggests that general practitioners (GPs) are more likely to respond to questionnaire surveys if they are motivated to do so because the subject of the survey is perceived as relevant to the practitioner at the time of the survey, and there is an opportunity to respond without encroaching on existing obligations or interests.Aim: To compare the results of a questionnaire survey administered by two routes: a postal survey with reminders and after an educational meeting with no reminders.Methods: A validated questionnaire with 19 questions using a Likert scale seeking views about the ideal contents of a referral letter to colorectal surgeons was administered to GPs attending a monthly educational meeting in one locality in the UK and by post to a different group in a second locality. The response rate and demographic characteristics of the GPs were compared for respondents, non-respondents, attenders and non-attenders to the educational meeting. The mean responses to the questions on the Likert scale were used to produce a schedule to score the contents of referral letters. The schedules produced by the two methods (postal/educational meeting) were used to score 551 referral letters at one district general hospital. Mean scores calculated using the two schedules derived by the postal and educational routes were compared. Cases diagnosed with 'significant organic pathology' as defined in a previous study were identified. The predictive value of the scores in identifying these cases was compared.Results: 88% (138/157) of GPs in the locality attended the educational meeting and were invited to participate in the survey. Non-attenders were more likely to be assistants, locums or part-time practitioners. Overall 71% (98) of the GPs responded to the survey. There were no significant demographic differences between the respondents and non-respondents to the survey. Response rates to the survey conducted by the two routes were similar. Referral letter 'quality' scores derived via the two alternative routes were equally valuable in predicting which patients had significant organic pathology.Conclusions: Responses to questionnaire surveys are equally valid when administered via an educational meeting or by post. The value of the educational meeting approach was the relative ease of administering the questionnaire and the rapid turn around of replies. However this method may exclude the views of part-time practitioners who were more likely to fail to attend. These practitioners may need to be targeted by alternative routes in questionnaire surveys.