Defining and Improving Survival Rates From Cardiac Arrest in US Communities

Since emergency medical services (EMS) programs began in the early 1970s, much has been learned about the management of out-of-hospital cardiac arrest. The key therapeutic interventions, metaphorically characterized as 4 links in a chain of survival,¹ are rapid access, rapid cardiopulmonary resuscitation (CPR), rapid defibrillation, and rapid advanced life support, including endotracheal intubation, intravenous access, and medication. Many researchers propose a fifth link—timely postresuscitative care, namely, hypothermia.² In addition, event factors such as witnessed collapse, cardiac rhythm, and comorbidity are associated with the likelihood of resuscitation. Each of these therapeutic and event factors helps explain why an individual may live or die following cardiac arrest. Yet the predictive power of each of these factors pales in comparison with one easily characterized feature. The most important, powerful, and underappreciated factor is the community in which the cardiac arrest occurs.

Since 1990, fewer than 24 US communities have reported their survival rates from ventricular fibrillation (VF), the most common rhythm associated with cardiac arrest. Reports in peer-reviewed journals from individual communities provide a wide range of survival rates from 0% to 46%—selected examples include Detroit, Michigan, 0%³ ; Chicago, Illinois, 3%⁴ ; New York, New York, 5%⁵ ; Dallas, Texas, 10%⁶ ; Milwaukee, Wisconsin, 26%⁶ ; Rochester, Minnesota, 46%⁷ ; and Seattle and King County, Washington, 46%.⁸ For the denominator, some of these rates used all cases of VF, some used witnessed VF, and some excluded cases of VF that had occurred after arrival of EMS personnel. Despite these methodological differences, the variation in rates is extraordinary. Similar ranges have been reported in other studies.^{9 - 10} A recent study⁶ used common case definitions and methods to report survival rates following out-of-hospital cardiac arrest in 10 US and Canadian communities. For VF, survival to discharge ranged from 3% to 39.5%.⁶ In short, methodological features of previous studies do not account for the striking geographic variation in survival.

Two modifiable factors responsible for this variation are the time from the moment of collapse to the delivery of CPR and the time to defibrillation. When CPR and defibrillation are delivered within minutes of collapse, the patient has an increased chance of survival and good neurological recovery.¹¹ When therapy cannot be delivered in a timely fashion, death is the inevitable outcome. Other factors are also likely to explain some of the wide variation in community survival rates. For example, factors such as quality of care, strong medical and administrative control, ongoing quality improvement, and programmatic culture also contribute to the survival rates. Nevertheless, if communities with poor survival rates could reduce the average times from collapse to CPR and defibrillation, their citizens would likely experience an increase in the rate of survival after VF. Therapeutic and programmatic factors are not mutually exclusive and may be synergistic—good programs likely deliver good therapy.

The striking paucity of studies reporting cardiac arrest survival serves as an impediment to identifying or reducing geographic variation. The first step is to define the magnitude of the problem, not only in the handful of communities that have reported their survival rates, but also throughout the nation. A 2008 scientific statement from the American Heart Association argued that “National annual reports on key indicators of progress in managing acute cardiovascular events in the out-of-hospital setting should be developed and made publicly available.”¹² Although the collection of additional data on survival rates in more communities would provide local information, a standardized national approach to the reporting of cardiac arrest is the optimal method for identifying geographic variation.

Previous work has resulted in an internationally accepted method for reporting cardiac arrest data, the Utstein criteria.¹³ The core summary statistic is defined as survival, defined as discharged alive from hospital, following witnessed cases of VF in which the arrest occurs before the arrival of EMS personnel. The Utstein criteria acknowledge the importance of neurological recovery as the ultimate goal of cardiac resuscitation and recommend that neurological status be determined at discharge using a simplified scale such as the cerebral performance category scale or other simplified scale.¹³ The widespread availability of automated external defibrillators that electronically interpret the patient's rhythm provides a reliable method to assess the presence of VF and thus accurately identify cases. The denominator for calculating a community's survival rate is the number of patients with VF (witnessed collapse and collapse before emergency personnel arrival) for which an automated external defibrillator delivers a defibrillatory shock. The numerator is the number who survive to discharge from a hospital. Many states allow release of hospital discharge information to EMS agencies without patient consent provided the information remains confidential and is part of quality improvement activities. This simple resource-light approach to documenting system performance can be achieved by most any EMS system.

The first and simplest action is committing to the identification and reporting of survival rates from VF cardiac arrest. The local leadership required for this effort should come from the medical directors of EMS. The medical director is the individual charged with issuing medical protocols and supervising the medical care by paramedics, and also is responsible for medical quality improvement.

Collecting and reporting data on survival is only the first step. Once survival is defined, efforts to improve it will require more detailed measures. The key times, measured to the second, and the key intervals include time of call to 911 center, interval to process the call and dispatch units, interval for responders to leave their quarters, interval from leaving quarters to arrival at the scene, interval from arrival at scene to arrival at patient, time of CPR, time of first defibrillatory shock, estimated time interval from collapse to CPR, estimated time interval from collapse to defibrillatory shock, and delivery of telephone CPR instructions.

With these additional data in hand, medical directors can scrutinize their system's performance in managing VF, implement interventions for improvement, and evaluate those interventions over time. Importantly, annual survival rates should be shared with the community. Cardiac arrest and VF survival rates are medical matters and should be treated using a medical approach. Whenever a patient who was expected to live dies, the appropriate question is why. This approach is as appropriate for VF as it is for any other unexpected adverse medical outcome. A large proportion of patients with VF should survive; and if they do not, the questions “Why did this patient die? What could have been done to save this life?” should be asked. Not every patient with VF will survive because comorbidity or illness severity may preclude successful resuscitation; but when CPR and defibrillation can be provided in a timely fashion, community survival rates can approach 50%.^{7 - 8}

Compared with urban and suburban communities, rural communities are sometimes considered to be at a disadvantage in their ability to improve outcomes from cardiac arrest because of long response times from contacting the EMS system to arrival at the patient in cardiac arrest. However, long intervals to care are also likely in urban communities given traffic congestion and delays to reaching patients in high-rise condominiums and office buildings. Scrutiny of performance—whether in a rural, suburban, or urban system—can pinpoint the reasons for poor VF survival and suggest strategies for improvement.

Emergency medical services have long been an orphan in the federal system, relegated to a rung far down the administrative ladder of the US Department of Transportation, within the National Highway Traffic and Safety Administration. A 2006 Institute of Medicine report¹⁴ offered 20 specific recommendations to improve prehospital emergency care. One recommendation was to establish a lead federal agency for prehospital emergency care within the US Department of Health and Human Services. Such an agency would be an ideal home for gathering data and reporting on the incidence of VF arrests and survival rates from VF across the country.

The US Census Bureau and other federal agencies collect and regularly report an immense amount of data about the state of the economy, but far less is known about the health of the public. Although the National Cancer Institute uses the Surveillance, Epidemiology, and End Results network to provide information on cancer incidence,¹⁵ the monitoring of the incidence of cardiovascular events in the United States, except through death certificates, is not systematic, typically occurring through occasional studies or projects. Ventricular fibrillation arrests and survival rates would be an excellent place to start, largely because highly effective interventions that are known to save lives are available. In contrast with most community-based interventions, for which the benefit is estimated by an abstract quantity such as changes in event rates over time, the persons resuscitated by EMS are individually identifiable as neighbors, friends, and family members.