OBJECTIVES: The Evidence and Value: Impact on DEcisionMaking (EVIDEM) is a multicriteria decision analysis-based framework developed to support the decision-making process in healthcare. In an application of the EVIDEM core model, the Likert scale is employed for the scoring scales. However, categorical scales do not necessarily display interval properties, such that equal increments on a scoring scale represent equal increments of value. This study investigated the relationship between scores and partial values for each criterion.

METHODS: We elicited partial value functions for the 13 criteria of the EVIDEM through direct value rating by the public and by healthcare professionals. In this study, the scoring scales was based on a 5-point Likert scale (6- or 11-point scales in the EVIDEM) with anchors at the two ends, where score 5 (e.g., “very severe” for the criterion “disease severity”) represented the highest level of fulfillment of the criterion and 1 (e.g., “not severe”) the lowest. Respondents were asked to rate scores 2-4 for the 13 criteria on a scale from 0 to 100. Scores 1 and 5 were predefined as 0 and 100.

RESULTS: The survey was completed by 1,141 members of the public and 1,066 healthcare professionals. Only 2.1%-2.8% of the public and 1.6%-2.7% of healthcare professionals considered that scores have interval properties. The means of differences by each criterion between scores 1 and 2 were the largest (range of the means: 36.2-40.0 for the public vs. 34.5-41.5 for professionals; between 2 and 3, 15.4-17.0 vs. 15.7-19.3; between 3 and 4, 16.4-17.9 vs. 16.8-19.9; between 4 and 5, 27.2-30.0 vs. 25.5-28.9).

CONCLUSIONS: Most partial value functions displayed non-linearities and were similar in shape, indicating that the scoring scale of the EVIDEM should not automatically be used as interval scales. Scores need to be adjusted by their partial value functions before calculating the aggregate score.