Evaluating psychometric properties of an instrument addressing comprehensiveness of care among dentists

Abstract

This study aimed to evaluate the psychometric properties of an instrument to assess comprehensiveness of care from dentists using a combination of classical test theory and item response theory. A 46-item instrument was developed and tested by a panel of experts, followed by a pilot test and administration to 187 primary care dentists in a large Brazilian city. The 46 items were evaluated using the following criteria: acceptability, internal consistency, temporal stability, inter-item correlation, and tetrachoric correlation. This evaluation led to a shortened version consisting of 11 items that met all the criteria previously described. The temporal stability was measured using Cohen’s kappa, and all 11 items presented values greater than 0.5. The Cronbach’s alpha value was 0.72. None of the 11 items had missing data on the distribution of responses, and the model considering the discrimination as varying fit the data better than the model considering discrimination as a constant parameter (p<0.001). Item characteristic curves showed that 54.5% of items could be considered difficult, i.e., only dentists with a good understanding of comprehensiveness responded favorably. The 11-item instrument to assess comprehensiveness of care by dentists is considered to have good psychometric properties.