Nonresponse Bias Against the Underserved in Emergency Department Patient Experience Surveys: A Cross-Sectional Analysis

Background

Patient experience scores are used to assess emergency physicians, but only a small fraction of patients actually respond to patient experience surveys. No prior studies have determined patient characteristics that predict emergency department (ED) patient experience survey response.

Objectives

To determine which patient characteristics are associated with ED patient experience survey response.

Methods

This was a cross-sectional analysis of a random sample of 2500 patients from our hospital system who were discharged from an ED between January 1 and June 30, 2022. Our hospital system has one hospital-based ED and two freestanding EDs, one of which is in an economically disadvantaged area. For each randomly selected patient, we used chart review to gather the following data points: ED visited, patient age, gender, race, preferred language, primary residence, time of arrival, marital status, health insurance, and emergency severity index for that visit. We used multivariable logistic regression to determine which of those variables were associated with patient experience survey response.

Results

Among 2500 patients sent surveys, 207 (8.3%) responded. Those with the lowest response rates were from a freestanding ED in an economically disadvantaged area (5.9%), were uninsured (4.9%), or were homeless (2.1%). On multivariable analysis, adjusted odds ratios for survey response for those characteristics were as follows: 0.54 (95% CI 0.33–0.84), 0.56 (95% CI 0.34–0.90), and 0.30 (95% CI 0.02–1.44), respectively.

Conclusion

Multiple markers of lower socioeconomic status were associated with decreased patient experience survey response rates, suggestive of nonresponse bias against these individuals.