Internet and ivr survey: Choice, engagement, data equivalency

Objectives: Telephonic interactive voice response (IVR) systems and Internet-based surveys have provided new methods of collecting longitudinal data for injury researchers. Providing a choice of multiple survey modes has been shown to increase response rate, but may lead to different responses to the same questions. We examined data equivalency and loss to follow-up rate from these two survey modes. Methods: 475 workers from 36 limited-service restaurants participated in a 12-week prospective-cohort study to investigate risk factors for slipping. Participants were given a choice of reporting their weekly slip experience by telephone using an IVR system, Internet-based survey, or by completing and mailing questionnaire forms. All survey materials were made available in English, Spanish and Portuguese. Demographic differences, loss to follow-up and reported rate of slipping were compared among those who chose internet-based survey and IVR. Results: Out of 475 participants, 315 chose IVR, and 154 chose internet-based survey to complete their weekly surveys. Younger participants, college graduates, and English and Portuguese speaker were more likely to choose the Internet-based survey mode over the IVR mode. Loss to follow-up rate was slightly higher for the IVR mode. Rate of slipping was not significantly different by survey mode (RR 0.84 95% CI 0.63-1.13). Conclusions: Surveys relying solely on the Internet may lead to selective exclusion of certain populations. Results provided no evidence of non-equivalency of data by survey mode. It should be possible to combine Internet-based survey and IVR to collect longitudinal information about injury causing events and near misses.