Study objective: Our objective is to evaluate the incremental cost-effectiveness of use of cardiopulmonary resuscitation (CPR) and automated external defibrillators (AEDs) by lay responders (CPR+AED) versus CPR only for cardiac arrest during a multicenter randomized trial. Methods: This was a prospective trial from July 2000 to September 2003 that randomly assigned 993 community units (eg, office buildings, public areas) in 24 sites to an emergency response system, using lay volunteers trained in CPR only or CPR+AED. Cost and quality of life data were collected with effectiveness data. The primary analysis evaluated the incremental cost-effectiveness of defibrillator use in public locations by using Markov modeling. Results: CPR only had 14 survivors to discharge and CPR+AED had 29. CPR only had a mean of 0.58 (95% confidence interval [CI] 0.28 to 0.88) quality-adjusted life-years and a mean $42,400 (95% CI $22,100 to $62,600) costs. CPR+AED had mean 1.14 (95% CI 0.44 to 1.83) quality-adjusted life-years, mean $68,400 (95% CI $28,300 to $108,400) costs, and a long-term cost of mean $46,700 (95% CI $23,100 to $68,600) per quality-adjusted life-year. Results were sensitive to the effectiveness of the intervention, time horizon, location of arrest, and other factors. Conclusion: Training and equipping lay volunteers to defibrillate in public places may have an incremental cost-effectiveness that is similar to that of other common health interventions.