138 – Modeling Applications for Backcasting, Nowcasting and Forecasting
Mode Effect, Patient-Mix Adjustment, and Nonresponse Analysis in the Outpatient and Ambulatory Surgery Consumer Assessment of Healthcare Providers and Systems (OAS CAHPS) Survey
Patrick Chen
RTI International
Marjorie Hinsdale-Shouse
RTI International, Research Triangle Park, NC
Shampa Saha
RTI International
A randomized mode experiment was conducted for the Outpatient Ambulatory Surgery Consumer Assessment of Healthcare Providers and Systems (OAS CAHPS) Survey to test the effect on survey responses of using three data collection modes: mail only, telephone only, and mixed mode (mail with telephone follow-up to nonrespondents). Eligible patients selected from Medicare-certified hospital outpatient departments and ambulatory surgery centers were randomly assigned to one of the three data collection modes. In addition to the mode effect analysis, the data were analyzed to determine if patient-mix or nonresponse adjustments were necessary. The results of multivariate linear regression models indicated that there were no significant mode effects; however, six characteristics were identified as significant patient-mix adjustors: age group, overall health, overall mental health, education, English-language proficiency, and surgery category. Patient-mix adjustment was needed to produce survey estimates for each facility. Nonresponse analysis results suggested that no nonresponse-adjusted weights were needed to produce survey estimates for OAS CAHPS.