US Regional Variability Seen in Incidence, Outcomes of In-Hospital Cardiac Arrest

The odds of patients experiencing in-hospital cardiac arrest and, if they do, surviving to discharge vary depending on where they live, suggests a US database study published online March 19, 2015, ahead of print in Circulation.

Another Take“A national surveillance program… could help identify additional patient- and hospital-level factors responsible for the observed geographic differences, in order to develop targeted interventions to enhance the overall quality of resuscitation and postresuscitation care and improve [in-hospital cardiac arrest] outcomes,” write Gregg C. Fonarow, MD, of the University of California at Los Angeles (Los Angeles, CA), and coauthors. 

The investigators looked at 838,465 patients (mean age 67.2 years; 45.4% female) who underwent cardiopulmonary resuscitation (CPR) for in-hospital cardiac arrest. Data were derived from the 2003-2011 databases of the Nationwide Inpatient Sample (NIS).

Overall incidence of in-hospital cardiac arrest was 2.85 per 1,000 admissions but varied by region (P < .001):

  • Northeast (19.4% of cases): 2.75                                                                                                                                        
  • Midwest (19.0%): 2.33                                                                                                                                    
  • South (37.7%): 2.81                                                                                                                                                          
  • West (23.9%): 3.73                                                                                                                                                      

There was also significant variation among states, which had incidences ranging from 0.86 to 6.31 cases per 1,000 admissions.

Compared with patients in the Northeast, those in the Midwest, South, and West were younger. The Midwest had a higher proportion of Caucasians, the South a higher percentage of African-Americans, and the West a higher proportion of Hispanics and Asian/Pacific Islanders than did the Northeast. In addition, patients in the West were more likely to have Medicaid vs Medicare as the primary payer.

Prevalence of comorbidities was similar across regions, except for more smoking and obesity outside the Northeast and higher prevalence of deficiency anemia and fluid/electrolyte disorders in the South and West.

Ventricular tachycardia/ventricular fibrillation was the cardiac arrest rhythm in 21% of patients overall and was less common in the Northeast than in other regions.

One-quarter (24.7%) of all arrest patients survived to hospital discharge, although the odds of doing so were lower in the Northeast than in the other regions (P < .001). After adjustment, hospital survival after arrest remained higher in the Midwest (OR 1.33; 95% CI 1.31-1.36), South (OR 1.21; 95% CI 1.19-1.23), and West (OR 1.25; 95% CI 1.23-1.27) compared with the Northeast. Individual states also showed risk-adjusted differences in survival, with New York having the lowest rate (20.4%) and Wyoming the highest (40.2%) 

Odds of Survival Improved Over Time

Over the 8-year study period, there was a trend toward improved survival both overall (adjusted OR per year 1.05; 95% CI 1.04-1.05) and in each region (P for trend < .001 for all).

Factors predicting worse odds of survival to hospital discharge were:

  • African-American, Hispanic, or Asian/Pacific Islander race/ethnicity                                                                                
  • Medicare, Medicaid, or no insurance                                                                                                                         
  • Weekend admission                                                                                                                                                
  • Treatment at an urban hospital                                                                                                                                              

In contrast, having a median household income above the 26th percentile and being treated at a small- or medium-size hospital increased survival chances.

There was no association between the incidence of in-hospital cardiac arrest and survival at the regional level, but states with a lower incidence had higher post-arrest survival rates (P = .001).

Most survivors of in-hospital arrest were discharged to a skilled nursing facility (40.4%). Compared with survivors in the Northeast, those in the other regions were more likely to be discharged home and less likely to require home health care or transfer to a skilled nursing facility (P < .001).

According to the study authors, the incidence of in-hospital cardiac arrest is a function of both the severity of the patient’s illness and the institutional response and process of care for treating patients and preventing arrest. The fact that comorbidities were similarly distributed across the regions suggests that the burden of illness may contribute little to the regional differences in arrest rates, they say.

Editorial Calls for Mandatory Reporting

“These findings are disturbing and clearly signal that where you live and where you arrest matter,” writes Raina M. Merchant, MD, MSHP, of the University of Pennsylvania (Philadelphia, PA), in an accompanying editorial.

“Ultimately, improving arrest care is dependent on the ability to quantify and change it,” she says. “A mandatory standardized system for reporting and surveillance could be an important first step.”

Dr. Merchant adds that “[i]f the state level information presented [here] also existed at the hospital level in a readily accessible, easily interpretable report card it may be more likely to be used by patients for making choices about where to receive care and by other stakeholders (eg, providers, researchers, policy makers) seeking to track and improve care.”

Data May Be Skewed

“Even though the authors took great care to analyze the data as well as they could, there are so many confounders and ways in which data are manipulated that it is not easy to understand what [the findings] mean,” Sorin J. Brener, MD, of New York Methodist Hospital (New York, NY), commented in a telephone interview with TCTMD.

As a result, he said, the data may be skewed. For example, some hospitals may dilute their admission pool by admitting less sick patients, resulting in a lower incidence of cardiac arrest. As for outcomes data, Dr. Brener noted, “there is a large number of patients on whom we do CPR but we just go through the motions—we don’t expect or want it to succeed.” These are individuals who are terminally ill but for whatever reason lack a Do Not Resuscitate (DNR) order, he explained. On the other hand, some hospitals may have a disproportionately high number of DNR patients, he added.

Unfortunately, Dr. Brener said, due to unavailability of data from the NIS, the researchers were unable to look at the outcome that patients and doctors care about most:  survival without neurological deficit.

“In the hospital we would expect to get to patients and restore circulation quickly, and hence they wouldn’t have much neurological damage,” he explained. However, the outcome “survival to discharge” provides no clue to a patient’s condition. “If the patient is transferred to a skilled nursing facility because of significant neurological deficit, I consider that a failure,” he added.

Dr. Brener endorsed participation in quality-improvement efforts such as the American Heart Association’s Get With The Guidelines-Resuscitation registry, but he was skeptical about the idea of reducing data from hospital records of cardiac arrest to a “report card” that might help guide patients’ choices. “It is very difficult to control for all the relevant factors, and making things simpler makes them less likely to be accurate,” he commented.

Ultimately, close patient monitoring and the ability to respond immediately and perform resuscitation correctly are key to improving survival from cardiac arrest, Dr. Brener said. That includes training for nurses, he noted, because they are often first on the scene.

Meanwhile, he observed, “the most sobering finding is that even though in-hospital survival is twice as good as out-of-hospital survival, it’s still only about 20%.”



1. Kolte D, Khera S, Aronow WS, et al. Regional variation in incidence and outcomes of in-hospital cardiac arrest in the United States. Circulation. 2015;Epub ahead of print.

2. Merchant RM. Public report cards for in-hospital cardiac arrest: empowering the public with location-specific data [editorial]. Circulation. 2015;Epub ahead of print.

  • Dr. Merchant reports receiving grant support from the National Institutes of Health and pilot funding from Cardiac Science, Philips Medical, Physio-Control, and Zoll Medical.

Related Stories:


US Regional Variability Seen in Incidence, Outcomes of In-Hospital Cardiac Arrest

The odds of patients experiencing in-hospital cardiac arrest and, if they do, surviving to discharge vary depending on where they live, suggests a US database study published online March 19, 2015, ahead of print in Circulation.
  • Drs. Fonarow and Brener report no relevant conflicts of interest.