Translation and Psychometric Evaluation in Cancer Care of the German Version of collaboRATETM-a 3-item Patient-reported Measure of Shared Decision-Making.
Background: The collaboRATETM measure assesses the shared decision-making process from patients' perspective with three items. Because of its shortness, it is especially feasible in routine care. It was developed in English and has been translated into several languages. This study aimed to translate collaboRATE into German, test its comprehensibility and evaluate its psychometric properties.
Methods: Translation followed the TRAPD protocol. Comprehensibility was tested in cognitive interviews with lay people (N = 18). Psychometric properties were evaluated in a secondary analysis of a sample of 1703 patients with cancer. They rated the collaboRATE items to assess their care experience in general at the respective department of one large university medical centre. We calculated collaboRATE sum and top scores and assessed item characteristics (i.e., acceptance and ceiling effects), convergent validity with the 9-item Shared Decision-Making Questionnaire (SDM-Q-9 for one specific medical encounter) and satisfaction with care (single item), and divergent validity with psychosocial distress (NCCN distress thermometer). Completion rates, percentages of highest score, skewness, item endorsability and different correlation coefficients informed the evaluation of these psychometric properties.
Results: During translation and cognitive interviewing, the necessity to simplify sentence structures to enhance comprehensibility became apparent. Adaptations led to good comprehensibility. The mean collaboRATE sum score was 82.9 (SD = 19.3), with 466 participants (28.9%) indicating the top score. Item characteristics suggested good acceptability and ceiling effects. Correlations with SDM-Q-9 were lower than expected (sum score: r = 0.47, p < 0.001; top score: pbr = 0.27, p < 0.001). Correlations were as expected for satisfaction with care (sum score: rs = 0.46, p < 0.001; top score: χ2 = 218.3, p < 0.001, Cramer's V = 0.37) and minimally higher than expected for distress (sum score: r = -0.11, p < 0.001; top score: pbr = -0.09, p < 0.001).
Conclusions: A well-comprehensible German version of collaboRATE is now available. However, ceiling effects were found and convergent validity could not be established in a secondary analysis of a sample from cancer care. Further evaluation is needed regarding the psychometric properties of German collaboRATE. Members of the public were involved in developing the original English collaboRATE and testing the comprehensibility of German collaboRATE. Patients with cancer provided data for psychometric testing.