Death risk is lowest at busiest emergency centers

Graphic: Look inside an ambulance

The study suggests opportunity for new ways to organize and rate care; © Roberto Marinello/

When a medical emergency strikes, our gut tells us to get to the nearest hospital quickly. But a new study suggests that busier emergency centers may actually give the best chance of surviving – especially for people suffering life-threatening medical crises. In fact, the analysis finds that patients admitted to a hospital after an emergency had a 10 percent lower chance of dying in the hospital if they initially went to one of the nation's busiest emergency departments.

The risk of dying differed even more for patients with potentially fatal, time-sensitive conditions. People with sepsis had a 26 percent lower death rate at the busiest emergency centers compared with the least busy, even after the researchers adjusted for a range of patient and hospital characteristics. For lung failure patients, the difference was 22 percent. Even heart attack death rates differed.

The new findings, based on national data on 17.5 million emergency patients treated at nearly 3,000 hospitals, appear in an Annals of Emergency Medicine paper by a University of Michigan Medical School team. Using U-M Department of Emergency Medicine funding, they analyzed data from the Nationwide Inpatient Sample database compiled by the Agency for Healthcare Research and Quality.

The authors calculate that if all emergency patients received the kind of care that the busiest emergency centers give, 24,000 fewer people would die each year.

"It is too early to say that based on these results, patients and first responders should change their decision about which hospital to choose in an emergency," says Keith Kocher, M.D., MPH, the lead author of the new study and a U-M Health System emergency physician. "But the bottom line is that emergency departments and hospitals perform differently, there really are differences in care and they matter."

This is the first time a relationship has been shown on a national, broad-based scale between the volume of emergency patients seen at a hospital and the chance those patients will survive their hospital stay.

With half of all hospital inpatients now entering via the emergency department, data and lessons from the best-performing hospitals could improve patients' chances of leaving any hospital alive. It could also help guide the development of regional systems for emergency care, and specific measures of emergency care quality that could be used to rate hospitals and spur them to perform better.

In addition to survival for all patients admitted to the hospital from the emergency department, Kocher and his colleagues focused on eight high-risk, time-critical conditions. They were: Pneumonia, congestive heart failure, sepsis, the type of heart attack known as an acute myocardial infarction, stroke, respiratory failure, gastrointestinal bleeding and acute respiratory failure.

All require emergency providers to use a certain level of diagnostic skill and technology, and successful treatment depends on the ability of emergency and inpatient teams to deliver specialized treatment. All carry a death risk of at least 3 percent, and rank among the top 25 reasons emergency patients get admitted to a hospital.

The findings mirror surgery studies by U-M teams and others: The higher the volume, or number, of patients a center treats, the better the outcomes – even after adjusting for complicating factors. So, Kocher says, the survival effect in emergency care may be due to many things – the experience of the diagnosing emergency physicians, the availability of specialists, the skill and staffing levels of emergency and inpatient teams, the technologies available at the hospital, the patients' own health and socioeconomic background, and the location and nature of the hospital. He and his colleagues adjusted for these factors as much as possible before calculating survival rates.

Their results do not give insights into why the differences in survival occur – but for the first time, they show that they occur, so that further research can probe deeper.; Source: University of Michigan Medical School