From community to validated measure: Development and validation of the patient-informed cancer communication scale (PICCS).

Abstract

159 Background: Assessing patient-centered communication (PCC) is critical to improving quality patient-centered cancer care. While shown to be reliable and valid, most PCC measures were created by researchers, likely neglecting concerns of patients, families, and community members. PCC measures should be driven, in part, by the patient population of interest, especially given disparities in clinical communication. Our objective was to develop and validate a patient-informed PCC scale to assess cancer treatment discussions. Methods: As part of a larger study on communication and minority accrual to trials, we convened a panel of black and white cancer survivors, caregivers, and advocates. Panel members (n = 11) included 5 black and 3 white men and 1 black and 2 white women. Among them were 6 survivors and 5 caregivers. Panel members met regularly over six months to observe and discuss video-recorded treatment discussions between black and white men with prostate cancer and their physicians. Through an iterative process of generating and refining a list of physician communication behaviors they considered to be critical to PCC in a diverse patient population, they produced a list of 22 items, titled Patient-Informed Cancer Communication Scale (PICCS). We then applied the list as an observational scale to a set of video-recorded treatment discussions (n = 61) from the larger study. Trained raters applied the list and had acceptable inter-rater reliability. We used findings to determine constructs using scale development and factor analysis, then validated the scale through correlation with established scales. Results: We evaluated each item for content validity and feasibility. We divided some items that were assessing multiple attributes. The result was a 28-item scale. Using Classical Test Theory, we reduced the scale to 22 items. Using factor analysis, we identified five factors, including: 1. Treatment options (10 items a =.92); 2. Clinical relationship (6-items a =.92); 3. Prognosis and goals of treatment (2-items a =.79); 4. Explanations (2-items a =.43); and 5. Context (1-item). To validate, we correlated factors with two validated scales, one measuring physician PCC and the other patient active participation. Factor 1 was positively correlated with patient active participation ( r=.46; p=.003); Factor 2 with PCC ( r=.54, p. <.001); Factor 3 with patient active participation ( r=.48; p=.002); and Factor 5 with PCC ( r=.47, p=.002). The full PICCS scale was positively correlated with patient active participation ( r=.37, p=.02). PICCS Factor 4 was not correlated to the scales. Conclusions: This community-engaged research produced a reliable and valid patient-informed scale to assess PCC during cancer clinical interactions. Next steps include translating the findings by using PICCS to train physicians to communicate effectively in a diverse patient population.