CARES: Cardiac Arrest Registry to Enhance Survival

EMERGENCY MEDICAL SERVICES/CONCEPTS CARES: Cardiac Arrest Registry to Enhance Survival Bryan McNally, MD, MPH Allen Stokes, BS, EMT-P Allison Crouch, MPH Arthur L. Kellermann, MD, MPH For the CARES Surveillance Group* From the Department of Emergency Medicine, Emory University School of Medicine, Atlanta, GA. Despite 3 decades of scientific progress, rates of survival from out-of-hospital cardiac arrest remain low. The Cardiac Arrest Registry to Enhance Survival (CARES) was created to provide communities with a means to identify cases of out-of-hospital cardiac arrest, measure how well emergency medical services (EMS) perform key elements of emergency cardiac care, and determine outcomes through hospital discharge. CARES collects data from 3 sources 911 dispatch, EMS, and receiving hospitals and links them to form a single record. Once data entry is completed, individual identifiers are stripped from the record. The anonymity of CARES records allows participating agencies and institutions to compile cases without informed consent. CARES generates standard reports that can be used to characterize the local epidemiology of cardiac arrest and help managers determine how well EMS is delivering out-of-hospital cardiac arrest care. After pilot implementation in Atlanta, GA, and subsequent expansion to 7 surrounding counties, CARES was implemented in 22 US cities with a combined population of 14 million people. Additional cities are interested in joining the registry. CARES currently contains more than 13,000 cases and is growing rapidly. [Ann Emerg Med. 2009;54:674-683.] Provide feedback on this article at the journal s Web site, www.annemergmed.com. 0196-0644/$-see front matter Copyright 2009 by the American College of Emergency Physicians. doi:10.1016/j.annemergmed.2009.03.018 RATIONALE Out-of-hospital cardiac arrest is a leading cause of death in the United States. 1 Successful resuscitation depends on rapid performance of 4 critical actions: notification of emergency medical services (EMS), rapid provision of cardiopulmonary resuscitation (CPR), immediate defibrillation of victims found in ventricular fibrillation or pulseless ventricular tachycardia, and prompt access to definitive care. 2 The collective effect of these actions is so important that the American Heart Association (AHA) coined the term the chain of survival more than 15 years ago. 3 Despite universal acceptance of this concept and more than 3 decades of scientific progress in understanding the pathophysiology of cardiac resuscitation, the overall rate of survival from out-of-hospital cardiac arrest in the United States remains low. For reasons that are not always clear, communityspecific rates of survival vary dramatically. 4 Reported rates of survival from witnessed ventricular fibrillation range from 2% in Chicago to 46% in Seattle. 5,6 However, many EMS systems do not collect data, so their managers have no idea how well or how poorly they are doing. 7 In 2004, the AHA called on researchers to develop integrated methods of data collection that will allow for [U]niform data collection and tracking of data to facilitate *All members are listed in the Appendix. better continuous quality improvement in hospitals, emergency medical service (EMS) systems, and communities. The AHA noted that this will...enable comparison across systems for clinical benchmarking to identify opportunities for improvement. 8 In 2006, the Institute of Medicine (IOM) Committee on the Future of Emergency Care in the US Health System observed that many EMS agencies cannot document their effect on the communities they serve. To strengthen accountability, the IOM challenged EMS to collect, analyze and use performance improvement data. 7 In 2008, the AHA called for out-of-hospital cardiac arrest to be made a reportable event. It recommended that any out-of-hospital cardiac arrest reporting system include data on hospital outcomes. 9 The Cardiac Arrest Registry to Enhance Survival (CARES) was developed to serve as a central repository of data about cardiac arrests from EMS systems throughout the United States. Through its use, it is hoped that EMS systems of any size will be able to review key performance indicators about their own responses to out-of-hospital cardiac arrest, including response intervals, important aspects of care, and patients outcomes. Thoughtful evaluation of the information derived from CARES, including comparing local data with similar systems elsewhere, may identify opportunities to improve out-of-hospital care and achieve better outcomes. 674 Annals of Emergency Medicine Volume 54, NO. 5 : November 2009

McNally et al DESIGN Core Data Elements CARES represents a collaboration of Emory University, the Centers for Disease Control and Prevention (CDC), the Southeastern Affiliate of the AHA, and the Sansio Corporation (Duluth, MN). It collects out-of-hospital cardiac arrest related data from 3 sources that help define the continuum of care: 911 dispatch centers, EMS providers, and receiving hospitals. The number of mandatory data elements is limited to the minimum required to characterize an out-of-hospital cardiac arrest event and document its outcome. Candidate variables were drawn from 3 existing data sets that focus on out-of-hospital cardiac arrest (as opposed to inhospital arrests) the Utstein template, developed by an international consensus panel in 1991 and updated in 2004, 8,10 the National EMS Information System, created by a panel of US experts in 2001, 11 and the Resuscitation Outcomes Consortium, created by a network of National Institutes of Health funded research institutions in 2006. 12 Each variable was vetted by our team, with input from an ad hoc panel of national EMS experts (Appendix E1, available online at http://www.annemergmed.com). Four criteria were used: Is the variable necessary to characterize an out-of-hospital cardiac arrest event? Can it be clearly defined? Can it be objectively measured and reliably reported? Is it relevant to influencing or documenting the outcome? For the variable to be included in CARES, the answer to all 4 questions had to be yes. This iterative process produced a core data set of 37 elements, which is fewer than the elements required by Utstein, Resuscitation Outcomes Consortium, or National EMS Information System (Table 1). For the sake of consistency, National EMS Information System data definitions were used wherever possible. The CARES data dictionary is available at https://mycares.net. A sample form is depicted in Figure 1. Data Collection Once we defined our core data set, we devised a streamlined process to collect it. Because out-of-hospital cardiac arrest treatment spans a continuum of care, 3 sources are required to fully characterize an out-of-hospital cardiac arrest event: (1) 911 call center data (to document incident address and dispatch and unit arrival times), (2) EMS data (to capture presenting cardiac rhythm and the performance of key interventions) and (3) receiving hospital data (to document outcome at hospital discharge). To compile and collate data, CARES uses a secure, Health Insurance Portability and Accountability Act (HIPAA)- compliant, Internet-based data collection system developed by Sansio, Inc. and managed by CARES staff. This system provides participating 911 call centers, EMS agencies, and hospitals with ready access to their own data and the aggregate output of the registry, but users cannot scrutinize another agency s or hospital s data. Unauthorized parties and outside entities cannot gain access to individual records or the performance profiles of participating organizations. Because 911 call centers, EMS agencies, and hospitals capture and compile data differently, special procedures were needed to collect information from each source. Cardiac Arrest Registry to Enhance Survival 911 Data elements include incident address, the time each 911 call was received, the time the first responder and the transporting EMS unit were dispatched, and time each unit reached the scene. 911 Center managers can either upload data in batches or submit individual records for data entry. To ensure accurate reporting from communities with multiple public safety answering points, clocks are periodically synchronized. Once the relevant information is entered, CARES automatically calculates call processing and response-time intervals. EMS providers initiate the reporting process because they are best positioned to determine that an out-of-hospital cardiac arrest has occurred. To qualify as a case, an out-of-hospital cardiac arrest must be worked, meaning CPR was performed or defibrillation was attempted. If the victim is obviously dead or EMS personnel honor a do not attempt resuscitation request, the case is excluded. To be entered in the registry, an event must be presumptively due to cardiac disease. Cases of asphyxia, drowning, electrocution, drug overdose, trauma, or a primary respiratory event are excluded. EMS providers may submit data in any of 3 ways: A 1-page paper form can be fed into a special scanner that autopopulates CARES data fields, EMS services equipped with tablet or laptop computers can directly upload data to the CARES Web site, or EMS providers can directly file a CARES report online, using a Web-based application with embedded error checks to enhance the accuracy of data entry. To ensure that no cases are missed, each participating service compares the official records of worked resuscitations to cases entered in the CARES database. If a case was not reported or key data are missing, the providers who treated the patient are contacted and asked to resolve the problem. Hospital outcome is widely regarded to be the criterion standard for assessing the effectiveness of out-of-hospital cardiac care. 8 In most communities, the primary obstacle to securing hospital data appears to be unwarranted fears of violating federal patient confidentiality laws. 7 Some administrators resist providing outcome data because they fear that the process will be burdensome and costly. To allay these fears, we created a simple, HIPAA-compliant process that allows hospitals to submit data through the Web. Only 5 data elements are requested: emergency department (ED) outcome (admitted, died, or transferred); receiving hospital outcome information (same); patient disposition (discharged home, transferred to a second hospital or rehabilitation unit, released to the morgue); Cerebral Performance Score at discharge, a simple measure of functional status; and hypothermia treatment (if provided). To the degree possible, reporting is automated to expedite data collection. For example, whenever paramedics notify CARES of a new case, the system automatically e-mails the designated liaison at the receiving hospital that a CARES patient was transported to his or her facility. When the liaison accesses the hospital s site through the CARES portal, the case is there. It takes only a few clicks to report the patient s outcome. If no outcome data are reported by 15 days after an out-of-hospital cardiac arrest, the system automatically sends 2 additional e-mail reminders to the hospital during the next Volume 54, NO. 5 : November 2009 Annals of Emergency Medicine 675

Cardiac Arrest Registry to Enhance Survival McNally et al Table 1. Data elements comparison: CARES, National EMS Information System, Utstein, and Resuscitation Outcomes Consortium data sets related to cardiac arrest information. 10 Data Element National EMS Information System, National EMS Data Set Utstein Cardiac Arrest Nomenclature CARES Surveillance Registry Resuscitation Outcomes Consortium, Clinical Trial Patient information Patient s first (given) name No * No Patient s middle initial/name No No No Patient s last (family) name No * No Patient s age Age units No Patient s date of birth * No Patient s sex Patient s race/ethnicity No Scene information Incident address No No Incident city No No Incident county No No No Incident state No No Incident ZIP code No No Scene zone number No No No Scene GPS location No No No Geospatial location of event No No No Incident location type 911 responder information Date of cardiac arrest EMS agency ID No EMS vehicle ID No No Number of EMS personnel No No Service level No No EMS call number No No Fire/first responder service ID No No Other EMS agencies at scene No No No Other services at scene No No No Situation information Complaint reported by dispatch No No No Cardiac arrest (yes/no) No No Cardiac arrest cause Contributing factors No No No Evidence of implantable defibrillator No No No Arrest witnessed arrest witnessed by Arrest after arrival of 911 responder (by above question) Resuscitation attempted by 911 responder CPR information Who initiated CPR (by other CPR questions) No (by other CPR questions) Bystander CPR (by previous aid preformed by ) (by who initiated CPR ) CPR performed (by who initiated CPR ) Chest compressions No No No Chest compressions (by EMS) No No No Assisted ventilation (by below No No question) Type(s) of airway intervention(s) used No No Reason CPR discontinued No No No Monitor/defibrillator information First monitored rhythm of the patient No No No First arrest rhythm of the patient No No 676 Annals of Emergency Medicine Volume 54, NO. 5 : November 2009

McNally et al Cardiac Arrest Registry to Enhance Survival Table 1, continued. Data elements comparison: CARES, National EMS Information System, Utstein, and Resuscitation Outcomes Consortium data sets related to cardiac arrest information. 10 (continued) Data Element National EMS Information System, National EMS Data Set Utstein Cardiac Arrest Nomenclature CARES Surveillance Registry Resuscitation Outcomes Consortium, Clinical Trial First arrest rhythm of the patient (monitored by EMS) No No No Who first applied AED or monitor/defibrillator (by other No (by other monitor/ monitor/defibrillator defibrillator questions) questions) Bystander AED/defibrillator applied (by previous aid preformed (by who first applied AED or by ) monitor/defibrillator ) Defibrillation attempted (by who first applied AED or monitor/defibrillator ) Was an AED used during resuscitation No No No Advanced monitor used No No # Of AED shocks No No # Of manual shocks No No Other out-of-hospital treatment information Previous aid preformed by (before EMS) No No No Outcome of previous aid (before EMS) No No No Drug therapies No IV/IO line used No No Hypothermia No ROSC Sustained ROSC No Out-of-hospital disposition information Out-of-hospital disposition No No (by 4 questions below) Died at scene or en route No (by out-ofhospital disposition ) Reason not treated or why treatment halted No No Alive and not transported by EMS to hospital/ed No No No Transported to hospital/mode of transport No (by out-ofhospital disposition ) End of event/patient status at ED No Hospital information Date of ED/hospital arrival No Destination hospital ID No Cardiac rhythm on arrival at destination No No No ED outcome No No Interhospital transfer to another acute hospital No No (by ED outcome ) Date of interhospital transfer to another acute No No No hospital Hospital outcome Date of discharge or death No No Discharge destination No No No Neurologic status at discharge No CAD time information Agency CAD ID No No No CAD call number No No Ems call received time EMS dispatched time No EMS unit notified by dispatch time No No No EMS en route time No No EMS on scene time EMS on scene time (for first ALS crew) No No No EMS arrived at patient time # No EMS left scene time # No Volume 54, NO. 5 : November 2009 Annals of Emergency Medicine 677

Cardiac Arrest Registry to Enhance Survival McNally et al Table 1, continued. Data elements comparison: CARES, National EMS Information System, Utstein, and Resuscitation Outcomes Consortium data sets related to cardiac arrest information. 10 (continued) Data Element National EMS Information System, National EMS Data Set Utstein Cardiac Arrest Nomenclature CARES Surveillance Registry Transfer to aeromedical time No No EMS arrived at hospital/ed time # No EMS back in service time No No No EMS cancelled time No No No First responder call received time No No No First responder dispatched time No No No First responder on scene time date/time No No initial responder arrived on scene Treatment time information Incident or onset date/time No No No Estimated time of arrest before EMS arrival No No No Time of arrest (if EMS witnessed) No No Time arrest confirmed No # No No Time of first rhythm analysis/assessment of need No No No for CPR Time of first CPR No No No Time of first CPR (by EMS) No No Time when vascular access achieved No No No Time tracheal intubation achieved No # No No Time of EMS shock assessment No No Time of first defibrillation No No No Time of first defibrillation (by EMS) No No Time when medication given # No No Time of ROSC No No Time of end of ROSC No # No No Time of awakening No # No No Time CPR stopped/terminated time No resuscitation discontinued Time of death No No No Resuscitation Outcomes Consortium, Clinical Trial IV, Intravenous; IO, intraosseous, ROSC, return of spontaneous circulation. This table does not represent a full, comprehensive summary of all the fields and subfields in the National EMS Information System or Resuscitation Outcomes Consortium. It is intended only for general comparison between the data sets. *CARES only temporarily collects patient identifiers (name and date of birth) to ensure accuracy in matching the EMS records with the hospital outcomes. After the record is deemed complete by the CARES staff, the record is deidentified of name and date of birth. The Resuscitation Outcomes Consortium uses a combination of census tract, latitude/longitude, and universal transverse Mercator (UTM) to capture the location of the cardiac arrest incident. The Resuscitation Outcomes Consortium collects data for bag-valve-mask, esophageal-tracheal twin-lumen airway device (Combitube)/laryngeal mask airway/esophageal obturator airway, oral endotracheal, cricothyrotomy, ventilator, nasal endotracheal, continuous positive airway pressure, and rapid sequence intubation. CARES will start collecting hypothermia data starting summer 2008. Call received time (for EMS and first responder) may vary by organization, depending on local CAD structure. This may be a primary or secondary public safety answering point call time, depending on what times are locally recorded or available. Utstein: Indicates recommended as supplemental data element. # Utstein: Indicates in original data set but removed in newer version. few days. If outcome data are still missing 30 days after the event, the system prompts a CARES administrator to call the hospital contact. Data Linkage and Storage If a community s computer aided dispatch system autogenerates a common case identifier, CARES uses this number to link 911, EMS, and hospital reports to form a complete record. If a community s computer aided dispatch system lacks this capability, reports are linked by probabilistic matching. In most instances, reports are matched by patient name and the incident s time, date, and location. Once linkage is complete, the record is reviewed for accuracy. At this point, the case s individual identifiers are permanently stripped from the record. Once a record is permanently entered in the CARES database, no further edits can be made. The anonymity of CARES data is a major advantage. Because the registry contains only deidentified reports and uses 678 Annals of Emergency Medicine Volume 54, NO. 5 : November 2009

McNally et al Cardiac Arrest Registry to Enhance Survival Part A : Demographic Information 1 - Street Address (Where Arrest Occurred) Cardiac Arrest Registry 1 - City 1 - State 1 - Zip Code 2 - First Name 3 - Last Name 4 - Age Days Months Years Part B : Run Information 8 - Date of Arrest / / 5 - Date of Birth 6 - Gender Male Female 9 - Call # First Responding Agency Hospital Destination 11 - Fire/First Responder 12 - Destination Hospital 28 - Race/Ethnicity American-Indian/Alaska Hispanic/Latino Unknown Asian Black/African-American Native Hawaiian/Pacific Islander White Arrest Information 14 - Location Type 15 - Arrest Witnessed 16 - Arrest After Arrival of EMS 17 - Presumed Cardiac Arrest Etiology Home/ Residence Recreation/Sport Witnessed Arrest Presumed Cardiac Etiology Public Building Industrial Place Trauma Street/Hwy Unwitnessed Arrest No Farm Respiratory Nursing Home Mine / Quarry Drowning Residence/Institution Physician Office/Clinic Jail Electrocution Educational Inst. Airport Other Hospital Other Resuscitation Information 18 - Resuscitation Attempted by EMS 21 - Was an AED Used During Resuscitation 22 - Who First Applied Monitor/Defibrillator, AED Not Applicable No No Lay Person AED Present but not Used Lay Person Family Member 20 - Who Initiated CPR AED Malfunctioned Lay Person Medical Provider Not Applicable First Responder Lay Person # 0f AED Shocks Responding EMS Personnel Lay Person Family Member Lay Person Medical Provider First Responder # 0f Manual Shocks Responding EMS Personnel First Cardiac Arrest Rhythm of Patient and ROSC Information 23 - First Arrest Rhythm of Patient 24 - ROSC 26 - Out of Hospital Disposition 27 - End of the Event Ventricular Fibrillation Resuscitation not initiated at scene due to Dead in Field Ventricular Tachycardia No obvious signs of death, DNR, resuscitation Pronounced Dead in ED considered futile, or resuscitation is not Asystole required Ongoing Resuscitation in ED Idioventricular/PEA Unknown Shockable Rhythm Unknown Unshockable Rhythm 25 - Sustained ROSC No Resuscitation terminated at scene due to medical control order, protocol/policy requirements completed Transported to Hospital with or without ROSC SH3001 (1 of 1), Rev 3, 04/06 Copyright 2008 Sansio (Page 1) Figure 1. CARES EMS data form. HIPAA-compliant procedures, Emory University s institutional review board ruled that participating agencies and institutions may collect data without informed consent. Analysis and Production of Reports CARES allows its users to generate standard reports or conduct custom queries. Because CARES data are anonymous, aggregated reports can be shared without fear of compromising a patient s or service s confidentiality. Each organization has 24-hour access to its own profile but not individual access to others. However, participating organizations are able to compare their performance against that of their peers by benchmarking against a regularly generated summary national report. Outside parties cannot gain access to patient- or servicespecific data. They can only examine reports at the aggregate level. Standard CARES output includes histograms that graphically depict 911 call handling and response intervals. Call to Dispatch histogram provides insight into the efficiency of a community s 911 call center. The Dispatch to Unit Arrival, histogram depicts the intervals required for the first public safety provider to reach the scene (whether it is fire, police, or EMS). The Call to Unit Arrival interval depicts the system s overall response time (Figure 2). CARES also captures incident location data so that EMS system managers can assess where and in what type of locations out-of-hospital cardiac arrest occurs. This can be used to guide community-level interventions, such as neighborhood-specific CPR training programs, public access automated external defibrillators, and first-responding fire companies. In the near future, geographic information system software may be added to enable participating cities to map events and display them in Volume 54, NO. 5 : November 2009 Annals of Emergency Medicine 679

Cardiac Arrest Registry to Enhance Survival McNally et al Figure 2. A sample EMS response time report. The respective intervals are provided that allow for the dispatch process time to be viewed separately from the ambulance response time. A report for first responder times using these same intervals can also be generated. relation to neighborhood landmarks and various physical and social characteristics. PILOT IMPLEMENTATION Once the registry was created, we first implemented it in Atlanta, GA, a city served by a single advanced life support EMS service (Grady EMS), 1 first-responding fire department (the Atlanta Fire Department), a single 911 computer aided dispatch center (Fulton County 911), and a relatively small number of hospitals. After a 12-month period of pilot testing and refinement, during which CARES collected and linked data on more than 600 cases of out-of-hospital cardiac arrest, the registry was expanded to 7 surrounding counties, an area encompassing 2,000 square miles and a population of roughly 3 million. This phase provided a stringent test of the registry s capacity to identify and link cases in a complex, multijurisdictional setting. Historically, hospitals and EMS services in metro Atlanta were reluctant to share data. The sheer number and diversity of 911 centers, EMS agencies, fire departments, and hospitals posed a significant challenge. To satisfy the needs of participants, CARES had to be flexible enough to accommodate multiple methods of data submission and still link records from different jurisdictions. Nearly every EMS service that participated in this phase of the project works with multiple first-responding services. Many receive calls from more than one 911 center. Strategies were devised to overcome all of these difficulties (Appendix E2, available online at http://www.annemergmed.com) EARLY EXPANSION For many EMS systems, this is their first effort to systematically collect and evaluate data related to the continuum of care for out-of-hospital cardiac arrest. CARES was expanded to more than 21 communities outside Georgia. Additional communities have expressed an interest in joining. Currently, CARES is processing data from 32 computer aided dispatch centers, 108 first-responder agencies, 31 EMS systems, and more than 200 receiving hospitals in 13 states (Figure 3). Approximately 14 million people reside in communities that are participating in CARES. CASE REPORTING To date, CARES has compiled data on more than 13,000 cases of out-of-hospital cardiac arrest. The typical time from event to EMS data submission is 1 week (daily in a few sites with electronic patient care records). Delinquent reports and cases with missing data are requested at the end of each calendar month. According to initial audits and feedback from CARES site directors in 15 participating cities, we estimate that the registry is capturing at least 95% percent of eligible cases. With the exception of race, a social construct largely determined by self-report, 13 every element of the CARES data set is being consistently reported. The proportion of missing data ranges from a high of 25% for victim race to a low of 1% for patient name (permanently stripped from each record once data linkage is completed). Hospitals are reporting outcome data in more than 99% of cases. In 40 months of operation, there have been 680 Annals of Emergency Medicine Volume 54, NO. 5 : November 2009

McNally et al Cardiac Arrest Registry to Enhance Survival CARES SITES 2009 Designates sites active prior to December 1, 2008 Designates sites added since December 1, 2008 Figure 3. CARES national map of current sites with active data collection. no breaches of confidentiality and no patient or family member complaints. No call center, EMS service, hospital, or community has withdrawn from the registry. LIMITATIONS CARES greatest strength simplicity is also its greatest weakness. It is designed to collect the minimum number of variables required to characterize the treatment and outcome of an out-of-hospital cardiac arrest event. Participating systems may choose to collect additional data elements for their own use, but CARES is not intended to be all things to all people. Determining that a cardiac arrest is due to heart disease is subjective. The AHA has conceded that no uniformly applied definition of cardiac arrest exists. 9 In general, out-of-hospital cardiac arrest is ascribed to heart disease unless there is an obvious alternate cause, such as major trauma, drowning, electrocution, drug overdose, asphyxia, or exsanguination. Because few victims are autopsied, it is often impossible to assign a definitive cause of death. The anonymous nature of CARES records precludes the registry from being used for patient-specific inquiries or case reviews. CARES cannot be used to contact survivors for longterm follow-up. Had we desired this capability, we would have had to secure written, informed consent from every patient or the legally authorized representative, a prohibitive task. DISCUSSION The CDC defines public health surveillance as the ongoing, systematic collection, analysis, interpretation, and dissemination of data about a health-related event for use in public health action to reduce morbidity and mortality and to improve health. 14 According to the CDC, surveillance serves a number of public health functions, including supporting case detection and public health interventions, estimating the effect of a disease or injury, portraying the natural history of a health condition, determining the geographic distribution and spread of illness, generating hypotheses and stimulating research, evaluating prevention and control measures, and facilitating planning. 15 CARES satisfies many aspects of this definition. We designed it to facilitate the timely collection, analysis, and interpretation of out-of-hospital cardiac arrest data so it can be quickly shared with those who need to know. 16 The concept of making a noncommunicable disease such as out-of-hospital cardiac arrest a reportable event may be perplexing to some. Out-of-hospital cardiac arrest is not immediately preventable and, for individual victims, the event is final. But out-of-hospital cardiac arrest reporting can produce several benefits. First, consistent collection of performance and outcome data can pinpoint opportunities to improve treatment of the next cardiac arrest victim. Second, the system s benchmarking features, which allow communities to compare how well they perform relative to other communities, may foster greater support for EMS. And third, ongoing monitoring of the incidence and outcomes of this particularly lethal form of heart disease may help local public health officials focus the public s attention on the value of primary prevention. Over time, CARES may allow participating communities to document the Volume 54, NO. 5 : November 2009 Annals of Emergency Medicine 681

Cardiac Arrest Registry to Enhance Survival beneficial effects of population interventions, such as enactment of an indoor air ordinance or improved access to primary care. 17 As a public health surveillance system, CARES differs from typical research databases (Appendix E3, available online at http://www.annemergmed.com). Research databases are designed to answer hypotheses, rather than monitor population health. 18 They are generally more complex, costly, and labor intensive than a typical surveillance system. The Resuscitation Outcomes Consortium is a prime example of a research database. 12 Created by a national network of research institutions, with funding provided by the National Institutes of Health, the Resuscitation Outcomes Consortium is designed to test promising treatments for out-of-hospital cardiac arrest and life-threatening trauma. CARES was not created for this purpose. It was designed to help communities monitor their success at achieving each link in the AHA s chain of survival 3 to identify opportunities to improve rates of out-of-hospital cardiac arrest survival. If and when the Resuscitation Outcomes Consortium identifies a new technique for treating out-ofhospital cardiac arrest, it will be important to quickly disseminate it throughout the nation and assess its effect. CARES is ideally suited to this task. Its core data set can be supplemented at any time. For example, several cities currently participating in CARES have recently added a single data element to document out-of-hospital initiation of hypothermia. There is a clear need for tools to help communities monitor and improve delivery of emergency cardiac care. AHA has observed that [t]here is extensive variation in reported outcome after the onset of cardiac arrest...this disparity in survival rates reemphasizes that an effective EMS system can decrease disability and death from acute cardiovascular events in the out of hospital setting. 9 Nichol et al 4 observed that if out-of-hospital cardiac arrest survival could be increased throughout the United States to the level reported by the highest-performing community in their study of out-of-hospital cardiac arrest outcomes, 15,000 premature deaths could be prevented each year. 4 The National EMS Information System is not a research database, but it also differs markedly from CARES. Unlike CARES, which focuses on a single condition, the National EMS Information System is designed to capture the full spectrum of EMS encounters. It is therefore larger and much more complex. And the National EMS Information System has an important limitation: unlike CARES and the Resuscitation Outcomes Consortium, it does not collect data on hospital outcomes. Communities wary of tackling the National EMS Information System may make their initial foray into electronic data collection through CARES. A favorable experience may encourage them to adopt the National EMS Information System at a later date. In the meantime, many EMS managers may opt to use out-of-hospital cardiac arrest as a sentinel condition to assess their system s capabilities to manage timecritical events. In Emergency Medical Services at the Crossroads, 7 the IOM Committee on the Future of Emergency Care endorsed this approach: While a full-blown data collection and performance measurement and reporting system is the desired ultimate outcome, the committee believes a handful of key indicators of regional system performance should be collected and promulgated as soon as possible. These could include, for example, indicators of 911 call processing times, EMS response times for critical calls, and ambulance diversions. In addition, consensus measurement of EMS outcomes could be applied to two to three sentinel conditions. For example, emergency and trauma care systems across the country might be tasked with providing data on conditions such as cardiac arrest, pediatric respiratory arrest, and major blunt trauma with shock. Data from different system components would allow researchers to measure how well the system performs at each level of care (911, first response, and ED). 7 CONCLUSION An adage from the business world states, You can t manage what you can t measure. 19 This concept is equally applicable to EMS. The IOM Committee on the Future of Emergency Care envisions a day when our nation will be served by coordinated, regionalized and accountable emergency care systems. 7 This cannot happen without uniform procedures to collect and analyze performance improvement data. We created CARES to help communities of every size assess their treatment of out-of-hospital cardiac arrest, with the goal of improving rates of survival. For some of our participating systems, CARES represents their first effort to systematically collect, link, and evaluate data related to care of outof-hospital cardiac arrest. We hope that others will follow. Technical assistance from the CDC staff was provided by Michael D. Matters, MD, MPH, and Jing Fang, MD, MPH. Supervising editor: Theodore R. Delbridge, MD, MPH McNally et al Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article that might create any potential conflict of interest. See the Manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. Funding for the Cardiac Arrest Registry to Enhance Survival (CARES) is provided by cooperative agreement from the Centers for Disease Control and Prevention grant number MM-0917-05/ 05. The American Association of Medical Colleges is the grant administrator for CARES. Publication dates: Received for publication June 29, 2008. Revisions received December 30, 2008, and March 7, 2009. Accepted for publication March 11, 2009. Available online April 25, 2009. Address for reprints: Bryan McNally, MD, MPH, Department of Emergency Medicine, Emory University School of Medicine, 531 Asbury Circle Annex, Suite N340, Atlanta, GA 30322; E-mail bmcnall@emory.edu. 682 Annals of Emergency Medicine Volume 54, NO. 5 : November 2009

McNally et al Cardiac Arrest Registry to Enhance Survival REFERENCES 1. Gillum RF. Sudden coronary death in the United States. Circulation. 1989;79:756-765. 2. 2005 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circ. 2005;112: iv-20. 3. Cummins RO, Ornato JP, Thies WH, et al. Improving survival from sudden cardiac arrest: the chain of survival concept. A statement for health professionals from the Advanced Cardiac Life Support Subcommittee and the Emergency Cardiac Care Committee, American Heart Association. Circulation. 1991;83: 1832-1847. 4. Nichol G, Thomas E, Callaway C, et al. Regional variation in outof-hospital cardiac arrest incidence and outcome. JAMA. 2008; 300:1423-1431. 5. Rea TD, Eisenberg MS, Sinibaldi G, et al. Incidence of EMStreated out-of-hospital cardiac arrest in the United States. Resuscitation. 2004;63:17-24. 6. Eisenberg MS, Horwood BT, Cummins RO, et al. Cardiac arrest and resuscitation: a tale of 29 cities. Ann Emerg Med. 1990;19: 179-186. 7. Institute of Medicine of the National Academies Committee on the Future of Emergency Care in the United States Health System. Emergency Medical Services at the Crossroads. Washington, DC: National Academies Press; 2007:207-230. 8. Jacobs I, Nadkarni V, Bahr J, et al; International Liaison Committee on Resuscitation; American Heart Association; European Resuscitation Council; Australian Resuscitation Council; New Zealand Resuscitation Council; Heart and Stroke Foundation of Canada; InterAmerican Heart Foundation; Resuscitation Councils of Southern Africa; ILCOR Task Force on Cardiac Arrest and Cardiopulmonary Resuscitation Outcomes. Cardiac arrest and cardiopulmonary resuscitation outcome reports: update and simplification of the Utstein templates for resuscitation registries: a statement for healthcare professionals from a task force of the International Liaison Committee on Resuscitation (American Heart Association, European Resuscitation Council, Australian Resuscitation Council, New Zealand Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Councils of Southern Africa). Circulation. 2004;110:3385-3397. 9. Nichol G, Rumsfeld J, Eigel B, et al. Essentials features of designating out-of-hospital cardiac arrest as a reportable event. Circulation. 2008;117:2299-2308. 10. Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended guidelines for uniform reporting of data from outof-hospital cardiac arrest: the Utstein Style. A statement for health professionals from a task force of the American Heart Association, the European Resuscitation Council, the Heart and Stroke Foundation of Canada, and the Australian Resuscitation Council. Circulation. 1991;84:960-975. 11. National Emergency Medical Services Information System. NEMSIS Data Dictionary version 2.2.1. National Emergency Medical Services Information System Web site. Available at: http://www.nemsis.org/softwaredevelopers/downloads/ datasetdictionaries.html. Accessed March 6, 2007. 12. National Institutes of Health News. New federally funded research program aims to improve survival from cardiac arrest and severe trauma. US Department of Health and Human Services. Released March 24, 2006. 13. Mays VM, Ponce NA, Washington DL, et al. Classification of race and ethnicity: implications for public health. Annu Rev Public Health. 2003;24:83. 14. Centers for Disease Control and Prevention. Updated guidelines for evaluating public health surveillance systems: recommendations from the guidelines working group. MMWR Morb Mortal Wkly Rep. 2001;50:RR-1. 15. Teutsch SM, Churchill RE. Principles and Practice of Public Health Surveillance. 2nd ed. Oxford, England: Oxford University Press; 2000. 16. Teutsch S, Thacker S. Planning a public health surveillance system. Epidemiol Bull. 1995;16:21. 17. Cesaroni G, Forastiere F, Agabiti N, et al. Effect of the Italian smoking ban on population rates of acute coronary events. Circulation. 2008;117:1183-1188. 18. Thacker SB, Berkelman RL. Public health surveillance in the United States. Epidemiol Rev. 1988;10:164-190. 19. Alter M. You can t manage it if you can t measure it. National Federation of Small Business. Available at: http://www.nfib.com/ object/io_23800.html. Accessed April 3, 2009. APPENDIX. CARES Surveillance Group. Mike Levy, MD (Anchorage, Alaska (Anchorage Fire Department/EMS, Anchorage, AK); Ian Greenwald, MD, Earl Grubbs, MD, Eric Ossmann, MD (Clayton, Cobb, Douglas, Fulton, Gwinnett, Newton, and Rockdale Counties, Metropolitan Atlanta, Georgia ); Louis Gonzales, BS, EMT-P (Austin Travis County EMS, Austin, TX); David Hall, MD (Baytown EMS, Baytown, TX); Peter Moyer, MD, Sophia Dyer, MD, (Boston EMS, Boston, MA); Donald Locasto, MD (Cincinnati Fire Department EMS, Cincinnati, OH); David Ross, MD (Multiagency EMS, Colorado Springs, CO); David Keseg, MD (Columbus Fire Department EMS, Columbus, OH); Joe Barger, MD (Contra Costa County EMS, Contra Costa County, CA); Chris Colwell, MD (Denver Health Paramedic Division, Denver, CO); James Leaming, MD (Penn State Life Lion EMS, Hershey, PA); David Persse, MD (Houston Fire Department EMS, Houston, TX); Joseph Salomone, MD (Metropolitan Ambulance Services Trust, Kansas City, MO); Dave Slattery, MD (Las Vegas Fire-Rescue, Las Vegas, NV); Corey Slovis, MD (Nashville Fire Department EMS, Nashville, TN); Bob Swor, MD (William Beaumont Hospital, Oakland County, MI); Brent Myers, MD (Wake County EMS, Raleigh-Durham, NC); Joe Ornato, MD (Richmond Ambulance Authority, Richmond, VA); Karl Sporer, MD (San Francisco Fire Department EMS, San Francisco, CA); Jeff Luther, MD (Sioux Falls REMSA, Sioux Falls, SD); Ben Osborne, MD (Multi-agency EMS, Springfield, MA); Angelo Salvucci, MD(Multi-Agency EMS, Venture County, CA). Did you know? You can track the impact of your article with citation alerts that let you know when your article (or any article you d like to track) has been cited by another Elsevier-published journal. Visit www.annemergmed.com today to see what else is new online! Volume 54, NO. 5 : November 2009 Annals of Emergency Medicine 683

APPENDIX E1. CARES data element ad hoc panel (blinded). Mickey Eisenberg, MD, PhD, University of Washington School of Medicine Ray Fowler, MD, University of Texas Southwestern Ian Greenwald, MD, Emory University School of Medicine Richard Hunt, MD, Center for Disease Control and Prevention Greg Mears, MD, University of North Carolina School of Medicine Peter Moyer, MD, PhD, Boston University School of Medicine Eric Ossmann, MD, Emory University School of Medicine Arthur Yancey, MD, MPH, Emory University School of Medicine Appendix E2. Diagram: importance of data elements and linkage CARES. 683.e1 Annals of Emergency Medicine Volume 54, NO. 5 : November 2009

Appendix E3. Distinctions between public health surveillance and epidemiologic research. Reason for initiating data collection Frequency of data collection Method of data collection Amount of data collected per case Completeness of data collected Analysis of data Dissemination of data Use of data Public Health Surveillance Problem detection Problem description Identify cases for epidemiologic studies May be legally required Monitor geographic and temporal trends in disease occurrence Ongoing Established systems or procedures Many persons involved Traditionally depends on voluntary participation Usually minimal Often incomplete Traditionally simple Primarily to detect change in incidence Usually historical comparison groups Timely Regular Review in public health agency Targeted to public health and clinical audience Identifies a problem Triggers intervention Suggest hypotheses Commonly used to evaluate programs Estimates magnitude of a problem Hypothesis testing Problem description Usually time-limited Epidemiologic Research Special procedures tailored to hypotheses or questions of interest Fewer persons involved Depends on paid, supervised employees Can be considerable and usually detailed Usually complete Can be complex Hypothesis testing often requires statistical methods Concurrent controls Not timely Sporadic External review Targeted to academic as well as public health and clinical audience Describes a problem in detail Provides etiologic information Tests hypotheses, suggests additional hypotheses Less often used to evaluate programs Adapted from Table 1 in Thacker SB, Berkelman RL. Public health surveillance in the United States. Epidemiol Rev. 1988;10:164. Volume 54, NO. 5 : November 2009 Annals of Emergency Medicine 683.e2