OBJECTIVES: Prior results from the TACTICS-TIMI 18 trial demonstrated that an early invasive approach to the management of patients with unstable angina (UA) or non-ST-segment elevation myocardial infarction (NSTEMI) was cost-effective, with an estimated cost per life-year gained of $8371; for high-risk patients with ST-segment changes at baseline, $3224 per life-year gained. We examined the joint influence of diabetes and ST-segment changes on the cost-effectiveness of the early invasive strategy. METHODS: Inpatient hospital costs for the 1722 US, non-VA patients for the 6-month trial period were obtained from the UB92 and Medicare cost/charge ratios. Other costs included: physician, outpatient, medication, and productivity costs. Life-expectancy estimates for patients with acute MI/coronary heart disease from the Framingham Heart Study were used for patients who survived the trial with/without experiencing a non-fatal MI. Cost-effectiveness was evaluated in terms of cost per life-year gained. Regression analyses of the net monetary benefit across a range of ceiling ratios were used to obtain cost-effectiveness acceptability curves for subgroups defined by diabetes and ST-change status. RESULTS: Net monetary benefit regression analyses revealed a three-way interaction between diabetes, ST-changes and treatment group which approached significance (p<0.10) for models based on a ceiling ratio of $14,000-$49,000. Associated cost-effectiveness acceptability curves suggest a probability that the invasive strategy is cost-effective for non-diabetic patients with ST-segment changes of > 95% for ceiling ratios $4,000 per life-year gained. At a ceiling ratio of $50,000, the probability of cost-effectiveness for diabetic patients without ST-changes is 56%, for diabetic patients with ST-changes, 73%, and for non-diabetic patients without ST-changes, 74%. CONCLUSIONS: In addition to ST-segment changes at presentation, the cost-effectiveness of an invasive strategy for patients with UA or NSTEMI varies by DM status. This analysis demonstrates the usefulness and efficiency of net-benefit regression for the evaluation of cost-effectiveness for different patient subgroups.