Psychometric Validation of a Nonproprietary Survey of Patient Satisfaction for Use in Outpatient Radiology Centers

Abstract Objective Demonstrate the psychometric evaluation process for and results from our radiology-specific patient experience measure. Methods We developed a survey to measure five dimensions of patient experience: (1) appointment, (2) reception, (3) registration, (4) procedure, and (5) facility. Each dimension included three to five questions. Each question was answered using a Likert scale (very dissatisfied to very satisfied). Data on procedure type, facility, radiologist interaction, health rating, survey length, and demographics were collected. The survey was implemented at 12 radiology offices. Analyses were conducted using responses from March 2018 to April 2019. Construct validation of the five dimensions was accomplished using confirmatory factor analysis (CFA). Internal consistency was examined using Cronbach and Guttman analysis. Results The sample included 20,736 subjects. There was strong evidence for construct validity of the five dimensions of patient experience. The CFA achieved the best fit with the five-factor model relative to other models (comparative fit index: 0.98, standardized root mean square error residual: 0.0307, root mean square error of approximation: 0.0371). There was high internal consistency (Cronbach’s α 0.94, Guttman coefficient 0.93). Item analysis showed that no questions were consistently skipped. Eighty-two percent of participants said the survey was not too long. Patients reported high satisfaction on all dimensions of satisfaction across modalities and office sites. Discussion The CFA and internal consistency analyses provide evidence for this survey having good psychometric properties: construct validity for five dimensions of patient experience and high internal consistency among the items. This survey is intended to be used by, and to benefit, radiology practices and their patients.