Measuring the Healing Context in CAM: Progress on Developing Patient-Reported Measures of Nonspecific Contributors to Treatment Outcome

Abstract

OA08.03 Purpose: In its 3rd strategic plan, NCCAM identified goals for future research, including “better understanding of the contributions of both specific and nonspecific effects influencing outcomes.” We present progress to date on our NCCAM-funded project, The Healing Context in CAM: Instrument Development and Initial Validation, which will measure nonspecific factors from the patient's perspective. Methods: We used PROMIS® methodology to develop patient-report measures of nonspecific factors that may influence healing, such as perceptions of the provider and healthcare environment, treatment expectations, and spirituality and optimism. We conducted 6 focus groups of CAM and conventional medicine patients, interviewed 22 CAM experts, and evaluated existing questionnaires. Through an iterative approach involving writing, rewriting, and revising items based on feedback from patients, we developed a set of item banks containing a total of 296 items. The items were tested on an Internet sample of 1400 adults who had experienced CAM or conventional treatments, and on a Pittsburgh sample of 257 patients who had started a CAM or conventional treatment for any non-acute condition within the previous 4 months. Data analysis included exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and item response theory (IRT). Additionally, 221/257 of the Pittsburgh sample rated their Clinical Global Impression (CGI) of improvement 6 weeks after completing the item banks. Results: EFA and CFA retained 250 items and identified 5 factors: 1) Patient-Provider Relationship+Healthcare Environment, 2) Optimism, 3) Spirituality, 4) Pro-CAM Attitudes, and 5) Treatment Expectations. IRT analyses resulted in retaining Healthcare Environment as a separate construct. Patients' IRT theta scores for all factors except Spirituality were significantly but modestly correlated with CGI, with Treatment Expectations showing the highest correlation with CGI (Spearman's Rho=.358, p<0.01). Conclusion: The project has successfully created item banks to measure nonspecific factors in healing, with most item banks exhibiting significant associations with patient reports of improvement. Further validity studies are forthcoming. Contact: Carol Greco, ude.cmpu@mcocerg