Advertisement

A charitable donation incentive did not increase physician survey response rates in a randomized trial

      Strategies to improve physician survey response rates are needed. Promised charitable donations might appeal to the respondent's sense of altruism, by allowing them to support an organization with which they share common goals and values [
      • Robertson D.H.
      • Bellenger D.N.
      A new method of increasing mail survey responses: contributions to charity.
      ].
      We conducted a randomized trial comparing the effect of a $40 charitable donation to the Kidney Foundation of Canada with no incentive on response rates in a survey of nephrologists' attitudes toward home dialysis. Our sampling frame consisted of all practicing nephrologists in Canada, drawn from provincial registers, and listservs. We used a computerized random number generator and stratified by province to randomize nephrologists in a 1:1 ratio to receive the incentive vs. standard notifications with no incentive. We used a modified Tailored Design Method [
      • Dilman D.A.
      • Smyth J.D.
      • Christian L.M.
      Internet, mail, and mixed-mode surveys: the tailored design method.
      ], with prenotification, invitation, first reminder, and final reminder letters sent by e-mail to both groups simultaneously over a 4-week period. Letters were identical for both groups with the exception of the following phrase: “By completing the survey, and in compensation for your time, a charitable donation of $40 CAD will be made, on your behalf, to the Kidney Foundation of Canada, by our study sponsors.” We used a commercial web-based survey platform (Fluid Surveys).
      Characteristics of study participants are presented in Table 1. Among 484 eligible participants, 81 of 241 (33.6%) vs. 93 of 243 (38.3%) responded in the intervention and control groups, respectively (relative risk=1.10, 95% confidence interval [CI]: 0.85–1.42, P=0.46). In a time-to-event analysis (Cox regression), the hazard ratio was 1.15 (95% CI: 0.83–1.48, P=0.51); we found no significant interaction between respondent characteristics and the intervention.
      Table 1Characteristics of study participants
      CharacteristicsIntervention (N=241)Control (N=241)
      n (%)n (%)
      Female gender75 (31.1)73 (30.0)
      Practice setting
       University78 (32.4)79 (32.5)
       Community149 (61.8)147 (60.5)
       Missing14 (5.8)17 (7.0)
      Province
       Alberta27 (11.2)26 (10.7)
       British Columbia29 (12.0)31 (12.8)
       Manitoba10 (4.1)8 (3.3)
       Maritimes18 (7.5)19 (7.8)
       Ontario86 (35.7)84 (34.6)
       Quebec61 (25.3)66 (27.2)
       Saskatchewan5 (2.1)6 (2.5)
       Missing5 (2.1)3 (1.2)
      Our findings are consistent with those of at least one other study. Gattellari and Ward [
      • Gattellari M.
      • Ward J.E.
      Will donations to their learned college increase surgeons' participation in surveys? A randomized trial.
      ], in a study of Australian surgeons, found that a promised charitable donation of $10 AUS to their professional college decreased the response rate. In the Australian study, the control group response rate was 70%, providing little room for improvement. Moreover, unlike our study, the Australian study provided monthly reminders over a 4-month period. Multiple successive reminder prompts may have impacted negatively on participants' enthusiasm for supporting the study (or the charity) over time.
      In contrast, we followed a published method that involved four points of contact over a 4-week period [
      • Dilman D.A.
      • Smyth J.D.
      • Christian L.M.
      Internet, mail, and mixed-mode surveys: the tailored design method.
      ]. We deliberately restricted our study period and number of prompts to avoid antagonizing participants in our fairly small national community. A more persistent approach may have yielded a higher response rate. Survey fatigue, a lack of interest in the topic, and a perception among participants that they support the Kidney Foundation through other means may explain the ineffectiveness of the incentive.
      Strengths of this study include its randomized design, balanced respondent characteristics between groups, and the use of a comprehensive (and hence, representative) sample. Limitations include unknown generalizability beyond the Canadian nephrology community and suboptimal sample size.
      This study adds to a small but important body of evidence suggesting that promised charitable donations might not be effective in improving the response rates in e-mail surveys of physicians. With the growing demands placed on physicians' time, better strategies to engage them in survey research are needed and worthy of further study.

      References

        • Robertson D.H.
        • Bellenger D.N.
        A new method of increasing mail survey responses: contributions to charity.
        J Market Res. 1978; 15: 632-633
        • Dilman D.A.
        • Smyth J.D.
        • Christian L.M.
        Internet, mail, and mixed-mode surveys: the tailored design method.
        John Wiley & Sons, Inc., Hoboken, NJ2009
        • Gattellari M.
        • Ward J.E.
        Will donations to their learned college increase surgeons' participation in surveys? A randomized trial.
        J Clin Epidemiol. 2001; 54: 645-649