Expert panel vs decision-analysis recommendations for postdischarge coronary angiography after myocardial infarction

Context Expert panels and decision-analytic techniques are increasingly use d to determine the appropriateness of medical interventions, but these 2 ap proaches use different methods to process evidence. Objective To compare expert panel appropriateness ratings of coronary angio graphy after myocardial infarction (from the time of hospital discharge to 12 weeks after infarction) with the health gains and cost-effectiveness pre dicted by a decision-analytic model. Design Comparison of the degree of importance of the clinical variables con sidered in expert panel appropriateness ratings vs a previously published d ecision-analytic model. Identification of 36 clinical scenarios from the ex pert panel that could be simulated by the decision-analytic model. Main Outcome Measures Appropriateness score and appropriateness classificat ion (expert panel) vs gain in quality-adjusted life-years (QALYs) and incre mental cost-effectiveness ratio (decision-analytic model). Results The most important clinical variables were similar in the 2 approac hes, with the exercise tolerance test result exerting the greatest leverage on strength of recommendation for angiography. Among the expert panel clin ical scenarios considered to be appropriate for coronary angiography that c ould be simulated in the decision-analysis model, the median (interquartile range) health gain and incremental cost-effectiveness ratio were 0.59 (0.4 1-0.76) QALYs and $27 000 ($23 000-$35 000) per QALY gained, respectively. Among the clinical scenarios that expert panels considered inappropriate, t he corresponding medians (interquartile ranges) were 0.24 (0.19-0.34) QALYs and $54 000 ($36 000-$58 000) per QALY gained. The Spearman rank correlati on between appropriateness score and QALY gain was 0.58 (P<.001) and betwee n appropriateness score and estimated incremental cost-effectiveness ratios was -0.66 (P<.001). Conclusions For the 36 expert panel scenarios that could be simulated by th e decision; analytic model, there was moderate to good agreement between th e appropriateness score and both the health gain and the incremental cost-e ffectiveness ratio of coronary angiography compared with no angiography in the convalescent phase of acute myocardial infarction, but several scenario s judged as inappropriate by the expert panel approach had cost-effectivene ss ratios comparable with many generally recommended medical interventions. Formal synthesis of expert judgment and decision modeling is warranted in future efforts at guideline development.