OBJECTIVE The objectives were to identify the components of a program to deliver early defibrillation that optimizes the effectiveness of automated external defibrillators (AEDs) in out-of-hospital and hospital settings, to determine whether AEDs are cost-effective, and if cost-effectiveness was determined, to advise on how they should be distributed in Ontario. CLINICAL NEED Survival in people who have had a cardiac arrest is low, especially in out-of-hospital settings. With each minute delay in defibrillation from the onset of cardiac arrest, the probability of survival decreases by 10%. (1) Early defibrillation (within 8 minutes of a cardiac arrest) has been shown to improve survival outcomes in these patients. However, in out-of-hospital settings and in certain areas within a hospital, trained personnel and their equipment may not be available within 8 minutes. This implies that "first responders" should take up the responsibility of delivering shock. The first responders in out-of-hospital settings are usually bystanders, firefighters, police, and community volunteers. In hospital settings, they are usually nurses. These first responders are not trained in reading electrocardiograms and identifying abnormal heart rhythms restorable by defibrillation. THE TECHNOLOGY An AED is a device that can analyze a heart rhythm and deliver a shock if needed. Thus, AEDs can be used by first responders to deliver early defibrillation in out-of-hospital and hospital settings. However, simply providing an AED would not likely improve survival outcomes. Rather, AEDs have a role in strengthening the "chain of survival," which includes prompt activation of the 911 telephone system, early cardiopulmonary resuscitation (CPR), rapid defibrillation, and timely advanced life support. In the chain of survival, the first step for a witness of a cardiac arrest in an out-of-hospital setting is to call 911. Second, the witness initiates CPR (if she or he is trained in CPR). If the witness cannot initiate CPR, or the first responders of the 911 system (e.g., firefighters/police) have arrived, the first responders initiate CPR. Third, the witness or first responders apply an AED to the patient. The device reads the patient's heart rhythm and prompts for shock when indicated. Fourth, the patient is handed over to the advanced life-support team with subsequent admission to an intensive care unit in a hospital. The use of AEDs requires developing and implementing a program at sites where the cardiac arrest rate is high, where a number of potential first responders are trained and retained, and where patients are transferred to an advanced care facility after initiating resuscitation. Obviously, placing an AED at a site where no cardiac arrests are likely to occur would be futile, as would placing an AED at a site where no one knows how to use it. Moreover, abandoning patients after initial resuscitation by not transferring them to an advanced care facility would negate all earlier efforts. Thus, it is important to identify the essential components of an AED program that might also optimize the effectiveness of AED use. METHODS There is a large body of literature on the use of AEDs in various settings ranging from closed environments such as hospitals, airlines, and casinos to open places such as sports fields and highways. There is little doubt regarding the effectiveness and safety of AEDs to treat people in cardiac arrest. It is intuitive that these devices should be provided in hospitals in areas that are not readily accessible to the traditional responders, the "code blue team." Similarly, it is intuitive to provide AEDs in out-of-hospital settings where the risk of cardiac arrest is high and a response plan involving trained first responders in the use of AEDs is in place. Thus, the Medical Advisory Secretariat reviewed the literature and focused on the components of an AED program in out-of-hospital settings that maximize the effectiveness and cost-effectiveness of the program in the management of cardiac arrest. Search engines included MEDLINE, EMBASE, EconLit and Web sites of other agencies that assess health technologies. Any study that reported results of an AED program in an out-of-hospital setting was included. Studies that did not use AEDs, had a physician-assisted emergency response plan, did not have a program for the use of AEDs, or did not include cardiac arrest as an outcome were excluded. SUMMARY OF FINDINGS A total of 133 articles were identified; 62 were excluded after reviewing titles and abstracts. Of the 71 articles reviewed, 8 reported findings of 2 large studies, the Ontario Prehospital Advanced Life Support (OPALS) study and the Public Access Defibrillation (PAD) trial. These studies examined the effect of a community program to respond to cardiac arrest with and without the use of AEDs. Their authors had reported a significant reduction in overall mortality from cardiac arrest with the use of AEDs. Factors That Improve the Effectiveness of an AED Program The PAD trial investigators reported a significant improvement in survival (P = .03) after providing AEDs in public access areas and training volunteers in CPR compared with training volunteers in CPR only. The OPALS study investigators reported odds ratios (ORs) and 95% confidence intervals (CIs) for significant predictors of survival, which were age (OR [age per 10 year], 0.8; CI, 0.8-0.9), arrest witnessed by bystander (OR, 3.9; CI, 2.7-5.5), CPR initiated by bystander (OR, 3.7; CI, 2.6-5.1), CPR initiated by first responder (OR, 1.6; CI, 1.1-2.3), and emergency medical service response within 8 minutes (OR, 3.0; CI, 1.8-5.1). The last 3 variables are modifiable and thus may improve the effectiveness of an AED program. For example, the rate of bystander-initiated CPR was only 14% in the OPALS study, but it was 100% in the PAD trial. This was because PAD trial investigators trained community volunteers whereas the OPALS study investigators did not. Cost-Effectiveness A systematic review of the literature suggests that cost-effectiveness varies from setting to setting. Most of the studies have estimated cost-effectiveness in American settings from a societal perspective; therefore, the results are not applicable to this report. However, results from this review suggest that the incidence of cardiac arrest in out-of-hospital setting in Ontario is 59 per 100,000 people. The mean age of cardiac arrest patients is 69 years. Eighty-five percent of these cardiac arrests occur in homes. Of all the cardiac arrests, 37% have heart rhythm abnormalities (ventricular tachycardia or ventricular fibrillation) that are correctable by delivering shock through an AED. Thus, in an out-of-hospital setting, general use of AEDs by laypersons would not be cost-effective. Special programs are needed in the out-of-hospital setting for cost-effective use of AEDs. One model for the use of AEDs in out-of-hospital settings was examined in the OPALS study. Firefighters and police were trained and provided with AEDs. The total initial cost (in US dollars) of this program was estimated to be $980,000. The survival rate was 3.9% before implementing the AED program and 5.2% after its implementation (OR, 1.33; 95% CI, 1.03-1.7; P = .03). Applying these estimates to cardiac arrest rates in Ontario in 2002, one would expect 54 patients of the total 1,395 cardiac arrests to survive without AEDs compared with 73 patients with AEDs; thus, 19 additional lives might be saved each year with an AED program. It would initially cost $51,579 to save each additional life. In subsequent years, however, total cost would be lower (about $50,000 per year), when it would cost $2,632 to save each additional life per year. One limitation of the OPALS study was that the authors combined emergency medical service response time and application of an AED into a single variable. Thus, it was not possible to tease out the independent effects of reduction in response time and application of an AED on the small improvement in survival. Nevertheless, the PAD study found that when response time was fixed, the application of AED improved survival. There are other delivery models for AEDs in casinos, sports arenas, and airports. The proportion of cardiac arrest at these sites out of the total cardiac arrests in Ontario is between 0.05% and 0.4%. Thus, an AED placed at these sites would likely not be used at all. Of the 85% cardiac arrests that occur in homes, 56% occur in single residential dwellings (houses), 23% occur in multi-residential dwellings (apartments/condominiums), and 6% occur in nursing homes. There is no program in place except the 911 system to reach these patients. Accordingly, the Medical Advisory Secretariat examined the cost-effectiveness of providing AEDs in hospitals, office buildings, apartments/condominiums, and houses. The results suggested that deployment of AEDs in hospitals would be cost-effective in terms of cost per quality adjusted life year gained. Conversely, deployment of AEDs in office buildings, apartments, and houses was not cost-effective. An exception, however, was noted for people at high risk of sudden cardiac arrest; these were patients with a left ventricular ejection fraction less than or equal to 0.35. CONCLUSIONS The OPALS study model appears cost-effective, and effectiveness can be further enhanced by training community volunteers to improve the bystander-initiated CPR rates. Deployment of AEDs in all public access areas and in houses and apartments is not cost-effective. Further research is needed to examine the benefit of in-home use of AEDs in patients at high risk of cardiac arrest.