Methods for Improving Response Rates in an Emergency Department Setting – A Randomized Feasibility Study

Abstract

Standard single survey mode approaches, such as mail-only or telephone-only, generally produce lower response rates than sequential mixed mode (SMM) approaches. SMM approaches may be particularly important when lower response propensities and/or poor contact information exist, as with emergency department (ED) patients. Results were examined from a randomized feasibility study testing five survey administration modes among ED patients who are discharged-to-community. A total of 4,017 discharged patients from eight EDs during February 2016 were randomized in equal numbers to five modes: on-site distribution of a paper survey; mail notification of a web survey; email notification of a web survey; two-stage SMM (mailed survey with telephone follow-up); and three-stage SMM (email notification of a web survey with mail follow-up and then telephone follow-up). Sampled patients received a 43-item instrument focusing on ED patient experience. The two-stage SMM, commonly used for other patient experience surveys, was the reference mode. Three-stage SMM had a higher response rate (30.7%) than two-stage SMM (25.3%); the other three experimental modes had significantly lower response rates (0.8–9.6%; p<0.001 for all comparisons). On-site distribution was difficult to operationalize, possibly promoting selection bias. Valid email capture rates ranged from 4.0% to 48.3% by hospital and 30.1% overall, even though participating hospitals reported capture rates of at least 20%; the odds of having a valid email address were significantly lower for older and male patients. Findings highlight the limitations of web-only and on-site approaches, and identify SMM with a web-based component as a promising approach for survey administration in the ED setting.