Development and Psychometric Evaluation of Healthcare Access Measures among Women with Ovarian Cancer

Abstract: Introduction: There are no validated instruments for measuring healthcare access (HCA). This study aimed to develop a cohesive HCA instrument for cancer survivors and evaluate the factor structure, reliability, and psychometric properties of two HCA domains—Acceptability and Accommodation—that require patient-reported survey data. Methods: This study reports data from three general methodological approaches: (1) concept elicitation using focus groups with 32 cancer survivors (63% Black, 18% Hispanic) to inform the development of new HCA survey items; (2) refining the new survey items using cognitive interviews with seven ovarian cancer survivors (n = 3 Black, n = 1 Hispanic) and pilot testing with 54 ovarian cancer survivors (74% White, 14% Black); and (3) psychometric evaluation of the Acceptability and Accommodation items among 333 ovarian cancer survivors (82% White, 13% Black). Multiple model structures were assessed for each HCA dimension using confirmatory factor analysis methods, and composite reliability was estimated for selected models. Results: In focus group discussions, cancer survivors expressed challenges navigating cancer treatment across multiple HCA domains, with the Acceptability domain (quality of patient–provider interaction) being the most salient across all racial groups. Lack of empathy, compassion, and poor communication overshadowed positive aspects of providers’ specialty, experience, or reputation. Cognitive interviews and pilot testing of new HCA survey items helped to clarify the language of specific items and refine the recruitment and consent process for implementation of the survey. In psychometric evaluation, the Accommodation domain (convenience and organization of services) was best accounted for by a two-factor structure: satisfaction with care and access to support services (comparative fit index (CFI) = 0.99). For the Acceptability domain, all fit indices supported the retention of a five-factor higher-order model (CFI = 0.96). Composite reliability estimates were >0.80 for all latent factors in the two-factor Accommodation model and the higher-order Acceptability model. Conclusions: Empirical evidence supports the utility of standardized measures of Accommodation and Acceptability using self-reported survey items, which will contribute to the better characterization of HCA dimensions among diverse cancer survivors.