Press Ganey Survey in Ambulatory Upper-Extremity Care: Assessing Nonresponse Bias
Patient surveys are now being widely used by hospital systems to monitor patient satisfaction with the process of inpatient and outpatient musculoskeletal care. While data from the surveys can help guide quality-improvement efforts, many clinicians have some concerns with the survey results in that the patients who respond may not be representative of all patients, and patient-care experiences may differ between survey “responders” and “nonresponders.”
In the September 1, 2021 issue of JBJS, Weir et al. delve further into this topic in their report on the response rate and factors associated with the completion of the Press Ganey Ambulatory Surgery Survey (PGAS) among patients treated with upper-extremity procedures in their outpatient surgical center. Of the 1,489 included patients, only 13.5% (201 patients) responded to the survey. The authors found significant differences between the responder and nonresponder groups with respect to baseline characteristics, including race (72% vs 57% White in the 2 groups, respectively), education (49% vs 40% with a college degree), employment status (88% vs 79% employed), income (49% vs 34% with income ≥$70,000), and marital status (54% vs 43% currently married). The responders also had better pre-intervention PROMIS scores across multiple domains, although the authors note that these differences were not clinically meaningful.
While emphasizing that factors influencing response rates are multifactorial and complex, the authors state that “The existence of substantial differences between responders and nonresponders raises concern for potential nonresponse bias for the PGAS.” They further point out that “surgical centers may be disproportionately missing the experiences of minority groups with lower socioeconomic status, and more focused efforts may be needed to ensure that these patients have equitable care experiences.”
It seems to me that avenues toward increasing the collection of patient responses might include improved processes for following up with nonresponders using personalized phone calls or emails, or potentially other incentives to collect these data. Survey vendors themselves have a role to play, working with hospital systems to enhance the credibility of these commonly utilized tools. With more inclusive response, providers are likely to be more confident in applying survey feedback to the practice environment, thereby improving the process of care for our patients.
Marc Swiontkowski, MD