Statistical Application and Cost Saving in a Dental Survey.

Abstract

To effectively achieve a robust survey response rate in a timely manner, an alternative approach to survey distribution, informed by statistical modeling, was applied to efficiently and cost-effectively achieve the targeted rate of return. A prospective environmental scan surveying adoption of health information technology utilization within their practices was undertaken in a national pool of dental professionals (N=8000) using an alternative method of sampling. The piloted approach to rate of cohort sampling targeted a response rate of 400 completed surveys from among randomly targeted eligible providers who were contacted using replicated subsampling leveraging mailed surveys. Two replicated subsample mailings (n=1000 surveys/mailings) were undertaken to project the true response rate and estimate the total number of surveys required to achieve the final target. Cost effectiveness and non-response bias analyses were performed. The final mailing required approximately 24% fewer mailings compared to targeting of the entire cohort, with a final survey capture exceeding the expected target. An estimated $5000 in cost savings was projected by applying the alternative approach. Non-response analyses found no evidence of bias relative to demographics, practice demographics, or topically-related survey questions. The outcome of this pilot study suggests that this approach to survey studies will accomplish targeted enrollment in a cost effective manner. Future studies are needed to validate this approach in the context of other survey studies.