Evidence-based research: understanding the best estimate

摘要

Introduction: Best estimates of intervention outcomes are used when uncertainties in decision making are evidenced. Best estimates are often, out of necessity, from a context of less than quality evidence or needing more evidence to provide accuracy. Purpose: The purpose of this article is to understand the best estimate behavior, so that clinicians and patients may have confidence in its quantification and validation. Methods: To discover best estimates and quantify uncertainty, critical appraisals of the literature, gray literature and its resources, or both are accomplished. Best estimates of pairwise comparisons are calculated using meta-analytic methods; multiple comparisons use network meta-analysis. Manufacturers provide margins of performance of proprietary material(s). Lower margin performance thresholds or requirements (functional failure) of materials are determined by a distribution of tests to quantify performance or clinical competency. The same is done for the high margin performance thresholds (estimated true value of success) and clinician-derived critical values (material failure to function clinically). This quantification of margins and uncertainties assists clinicians in determining if reported best estimates are progressing toward true value as new knowledge is reported. Analysis: The best estimate of outcomes focuses on evidence-centered care. In stochastic environments, we are not able to observe all events in all situations to know without uncertainty the best estimates of predictable outcomes. Point-in-time analyses of best estimates using quantification of margins and uncertainties do this. Conclusion: While study design and methodology are variables known to validate the quality of evidence from which best estimates are acquired, missing are tolerance margins, or upper and lower performance requirements and clinician critical values, within which best estimates behave and are validated. Understanding the best estimate behavior toward true value may provide clinicians and patients confidence in decision making under uncertainty.