Error budgets: a system to characterize error source in health care experimentation.

Abstract

OBJECTIVE: In the design of an experiment, it is important that all significant sources of error be identified and characterized accurately before the experiment is executed. For many experiments, the experiment designer is less inclined to ignore significant error factors when he adopts the use of an error budget in his experiment design process. A plan for generating an accurate error budget is presented along with a discussion concerning the disposition of significant error sources. METHODS: Specific methods discussed to generate the error budget are the fishbone diagram, factor weighing and Pareto analysis. Specific dispositions suggested are to eliminate, attenuate, characterize or to simply acknowledge the error with no remedy. These procedures, adhered to by most successful experiment design teams in industry, have evolved over recent decades and seem to be of use as they are applied to health care research. RESULTS: Use of the outlined method forces the experimenter to become familiarized with the most significant variables of the experiment he or she is designing through the deliberate examination of experimental noise contributors and through the comparison of the experimental noise to the magnitudes of the measurements in the experiment. CONCLUSION: The design of health care research experiments may be streamlined by incorporating the methods outlined in the generation of an error budget and by properly using the error budget to dispose of each source of error properly.