Validity and reliability of the 2-item version compared with the 10-item version of the Japanese consultation and relational empathy (CARE) measure: primary analysis of a cross-sectional survey.

BACKGROUND: The Consultation and Relational Empathy Measure is an internationally used five point each and 10-item questionnaire that assesses a physician's empathy from the patient's perspective. A 2-item version that uses item 6 "Showing care and compassion" and item 9 "Helping you to take control" from the Japanese 10-item version without changing the text was preliminarily developed through secondary data analysis. OBJECTIVE: To examine the validity and reliability of the Japanese 2-item version. METHODS: Selectively sampled patients who visited 11 collaborating general practitioners working in primary care clinics in both urban and rural areas of the Tokai region of Japan completed the 2-item and 10-item versions, and a patient background questionnaire. Face validity, criterion validity, construct validity, and internal consistency were examined. The 10-item version was used to assess the criterion validity of the 2-item version. Inter-rater reliability was examined using generalizability theory. The correlation between each item in the 2-item version was identified. Missing data analysis was performed. RESULTS: Among all 349 participating patients, questionnaires completed by 347 patients who gave clear consent to participate in the study were analyzed. The 2-item version showed high face validity, with few missing and "does not apply" values. Criterion validity was identified with strong correlations between the 2-item and 10-item scores (Pearson's correlation coefficient 0.805, P < 0.001), along with construct validity with correlations between the 2-item score and patient satisfaction (Spearman's rho 0.583, P < 0.001). The two items showed acceptable internal consistency (Cronbach's α 0.919). Eighty-five patients were required to ascertain inter-rater reliability of the 2-item measure. A mean score of 7.5 and a score range of 7-8 was estimated in the 2-item total. A strong positive correlation was identified between each item in the 2-item version (Pearson's correlation coefficient 0.852, P < 0.001). Missing data analysis revealed that the 2-item version, 10-item version, and consultation characteristics were missing completely at random, except for patient characteristics. CONCLUSIONS: Although the high number of patients required per doctor to reliably discriminate between doctors' empathy may limit its feasibility in practice and careful interpretation is warranted, the Japanese 2-item version has high validity and internal reliability.