O0110 Oct.3: 3:10pm - Sociodemographic Characteristics and Patient and Family Experience Survey Response Biases

Background: Enhancing Patient and Family Experience (PFE) is a priority in health care. Sociodemographic differences have been shown to impact health outcomes and experiences, but there is limited research on biases introduced from differential rates of PFE survey completion across various populations. This study’s objective was to assess survey completion rates across race/ethnicity, English proficiency, and socioeconomic status.

Methods: This retrospective study analyzed Cincinnati Children’s Hospital Medical Center (CCHMC)’s PFE ambulatory survey data collected January 1 – July 31, 2019. National Research Corporation (NRC) Real-Time administered the surveys. Outcomes of interest were 1) survey response (answer ≥1 question), 2) completion (answer all multiple-choice questions), and 3) presence of free response comment rates (present/absent). Predictor variables were sociodemographic characteristics, including race, ethnicity, spoken language, and address. Each variable was extracted from the electronic medical record. Addresses were geocoded, using HIPAA-compliant, CCHMC-based software, and connected to the corresponding census tract geography. Socioeconomic deprivation, measured at the census tract level, was defined using the Deprivation Index (DPI), calculated from ecological factors like poverty, educational attainment, and access to health insurance. DPI score was used to group respondents into quartiles. Quartile 1 (Q1) contained those with the lowest DPI scores and Quartile 4 represented the most deprivation. Associations between sociodemographic predictor variables and PFE survey outcome variables were then assessed using the Chi-square test.

Results: Surveys were attempted for 140,994 encounters, representing 82,294 patients; 77,627 encounters were included in the analysis, including one randomly selected visit per patient. Patients were white (76%), black (12%), multiracial (5%), and other (7%); 5% were Hispanic; 1% were Spanish speaking. Geocoding was not possible for 7,895 encounters (5.6%) and thus not connected to a DPI. Across the entire included sample, the response, completion, and comment rates were 20.1%, 17.6%, and 4.1% respectively (Table A). There were significant differences among all groups for response rates and comment rates. White respondents were more likely to complete the survey and leave a comment compared to Black respondents, despite similar response rates. Hispanic respondents (30.3%) and Spanish-speaking respondents (40.1%) had relatively high response rates, yet both groups were significantly less likely to leave a comment. There were significant differences among response and comment rates across deprivation quartiles; however, completion rates were similar for all groups. Those living in more deprived areas were slightly more likely to respond/complete the survey but less likely to comment.

Conclusions: PFE survey response patterns demonstrate potential sociodemographic biases. Minority families and those with higher DPI leave fewer comments, thus interventions and improvements might be biased towards those who provide qualitative feedback. Next steps include evaluating various methods and best practices of soliciting feedback that represents the diverse perspectives and backgrounds of our patients and families.