BACKGROUND: Studies using decision analysis appear increasingly often
in the surgical literature. Because readers and reviewers may be unfam
iliar with decision analysis, we review the technique and provide guid
ance on its interpretation. STUDY DESIGN: Review article. RESULTS: Dec
ision analysis is a systematic approach to structuring a decision, col
lecting relevant information about the probability and relative value
of outcomes, and making quantitative recommendations. Decision analysi
s includes the following basic components: (1) the decision model, (2)
the probabilities of clinical outcomes in the model, (3) the utilitie
s of clinical outcomes, and (4) the analysis and interpretation. To cr
itically interpret the results of a decision analysis, readers must co
nsider the validity of each component. The model should include all re
levant clinical strategies and all important clinical outcomes. Probab
ility estimates, whether derived from published studies or based on ''
clinical consensus,'' should be in general agreement with the reader's
clinical experience. Utilities (values assigned to outcomes) are ofte
n expressed in terms of quality-adjusted life expectancy. Methods used
to estimate life expectancy and to adjust for quality of life must be
scrutinized carefully. Within the analysis, readers should consider t
he effect of varying uncertain variables in the model (sensitivity ana
lysis) and, thus, the stability of the results. Finally, readers must
assess whether the magnitude of expected benefit from the favored clin
ical strategy is clinically important. CONCLUSIONS: As decision analys
is becomes more frequently used to influence clinical policy in surger
y, surgeons must learn to examine the technique more critically.