Penn: Here's how we prevented 11,000 patient deaths

Data, physician engagement proved key to quality improvement

Writing in Modern Healthcare, Ashok Selvam explains how the University of Pennsylvania Health System (UPHS) leveraged its own data and clinicians to identify root causes of inpatient mortality and develop strategies to improve clinical outcomes.

Seven years ago, the three-hospital system's mortality rate was 2.5%—significantly higher than the national average of 2.2%. To tackle the problem, CMO Patrick Brennan assembled a team of physicians, nurses, and administrators to analyze outcomes and develop strategies to improve the system's performance.

"We decided that we should get organized in a very deliberate way and study every death," he recalls.

Using data to identify 'hot spots'

According to the Institute for Healthcare Improvement's Don Goldmann, many factors can drive up mortality rates, and hospitals must account for these "hot spots" in order to pinpoint areas for improvement.

An analysis of the system's electronic health record (EHR) system revealed two glaring hot spots at UPHS:

  • Excess mortality among patients with sepsis; and
  • Excess mortality among patients with hospital-acquired delirium, which was triggering falls and breathing problems that led to unnecessary deaths.

The dangers of sepsis
NYT spotlights how a missed infection can kill
Sepsis hospitalization rates double across last decade
Nine Calif. hospitals see major drop in sepsis mortality

Physician engagement critical for success

UPHS began the improving process by using data from the University HealthSystem Consortium to set benchmarks. The system also enhanced its data collection process by inputting deceased patient information into its EHR system.

At the outset, UPHS's new mortality review committee understood that it would need to enlist physicians and develop a peer-based action strategy. Physicians submitted their opinions on the root causes of patient deaths and whether or not they were preventable; those surveys were used to identify major problems that led to poor outcomes and areas for improvement.

For example, the committee found that:

  • Sepsis patients required faster treatment, as their condition often deteriorated as they waited for a doctor;
  • Staff needed training to better recognize symptoms of sepsis; and
  • Certain processes, including unsafe transports or prolonged waits when transitioning from the ED to intensive care, may trigger episodes of delirium.

Additionally, the committee learned that antibiotics were being underutilized, so UPHS developed an algorithm for administering drugs to treat sepsis and added a pharmacist to the rapid-response team.

Surgeon: 'We had to lead by example'

As a result of the improvement efforts, hospital administrators estimate there have been 11,000 fewer patient deaths than expected since 2006. Now, UPHS' mortality rate is 1.6%, well below the national average.

"We had to lead by example," says Rachel Kelz, an assistant professor of surgery at UPHS, adding, "I don't think that people have a firm understanding of the difference an individual can make" (Selvam, Modern Healthcare, 10/26 [subscription required]).

Next in the Daily Briefing

Behind bars: The downfall of the nation's busiest cardiologist

Read now