Agreement of decision analyses and subsequent clinical studies in infectious diseases

Abstract

Purpose: Decision analysis techniques can compare management strategies when there are insufficient data from clinical studies to guide decision making. We compared the outcomes of decision analyses and subsequent clinical studies in the infectious disease literature to assess the validity of the conclusions of the decision analyses.

Methods: A search strategy to identify decision analyses in infectious disease topics published from 1990 to 2005 was developed and performed using PubMed. Abstracts of all identified articles were reviewed, and infectious disease-related decision analyses were retained. Subsequent clinical trials and observational studies that corresponded to these decision analyses were identified using prespecified search strategies. Clinical studies were considered a match for the decision analysis if they assessed the same patient population, intervention, and outcome. Agreement or disagreement between the conclusions of the decision analysis and clinical study were determined by author review.

Results: The initial PubMed search yielded 318 references. Forty decision analyses pertaining to 29 infectious disease topics were identified. Of the 40, 16 (40%) from 13 infectious disease topics had matching clinical studies. In 12 of 16 (75%), conclusions of at least 1 clinical study agreed with those of the decision analysis. Three of the 4 decision analyses in which conclusions disagreed were from the same topic (management of febrile children).

Conclusions: There was substantial agreement between the conclusions of decision analyses and clinical studies in infectious diseases, supporting the validity of decision analysis and its utility in guiding management decisions.

Figure 1

Sample decision tree.

Figure 2

Flow chart of search results and articles reviewed for the study.