Tools for Assessing Potential Significance of Pharmacist Interventions: A Systematic Review

Abstract

Introduction: Assessing the significance of pharmacist interventions (PIs) is essential to demonstrate the added value of pharmacists. Methods and tools for assessing the potential significance of PIs are diverse and their properties are questionable. Objectives: We aimed to systematically review the tools available to assess the potential significance of PIs. Methods: We conducted a systematic search for English- or French-language publications from 1986 to 2013 in PubMed, PsycINFO, PASCAL, and CINAHL. Studies were screened by two independent reviewers based on inclusion/exclusion criteria and were abstracted for content, structure of tools, and validation process. Results: Of 873 citations screened, 82 distinct tools were identified from 133 studies. While clinical aspects were often defined quite clearly, terminology regarding humanistic, economic, and process-related aspects of PIs was omitted, incomplete, or ambiguous in most tools. The probabilities of consequences of PIs/drug-related problems were evaluated in 20/82 tools. Few tools simultaneously measured economic, clinical, humanistic, and process-related variables. Structure of the tools varied from an implicit, mono-dimensional tool to an explicit, multi-dimensional algorithm. Validation processes were diverse in terms of quantification and number of raters, rating method, and psychometric parameters. Of 133 identified studies, there was limited evidence of validity (8/133, 6.0 %), inter-rater reliability (49/133, 36.8 %), and intra-rater reliability (2/133, 1.5 %). Conclusions: The majority of tools focused primarily on assessing clinical aspects and failed to detect comprehensive impacts. The heterogeneity of tools and assessment processes hindered our ability to synthesize the results of evaluations. Limited results for their validity and reliability cast doubt on the credibility of this methodology for justification of the value of PIs. Recommendations for development of tools with optimal theoretical, pragmatic, and psychometric properties are proposed.