Hospital-acquired sepsis, pneumonia kill 48,000 each year

A new study is the latest to document the high costs and deadly consequences of preventable nosocomial infections.

Researchers examined 69 million hospital discharges in 40 states from 1998 to 2006, looking at two conditions, sepsis and pneumonia, often caused by drug-resistant infections. Nosocomial sepsis and pneumonia kill 48,000 patients annually and cost $8.1 billion to treat, said the study, published in the Feb. 22 Archives of Internal Medicine.

"The first thing is to show causation -- that the hospital caused the infection and the infection caused the death," said study co-author Anup Malani, PhD, professor of law and Aaron Director Research Scholar at the University of Chicago Law School. "The reason why causation is important is because if there's no causation, you can't make the opposite claim, which is that these deaths can be prevented."

Researchers examined claims information from the Nationwide Inpatient Sample, a database run by the Agency for Healthcare Research and Quality. While lacking definitive clinical data, they used two techniques to isolate sepsis and pneumonia caused by infections acquired in the hospital and to show that the infections were deadly.

First, researchers looked at sepsis and pneumonia that developed in the hospital after elective surgery, assuming that surgery in nearly all cases would have been delayed if the illnesses were present before the procedures. Second, they matched patients who died after developing sepsis or pneumonia and compared them with similar patients who had the same procedures to differentiate the mortality rate from what otherwise would have been expected.

David J. Murphy, MD, said death and cost estimates are debatable, but that the toll exacted by hospital-acquired infections demands action.

"When you're dealing with 69 million discharges, you've got to make some systematic assumptions," said Dr. Murphy, who co-authored a commentary on the study in the same Archives issue. "These decisions can affect your specific effects. Whether it's $8.1 billion in costs or $7 billion or $10 billion, we're still talking about a whole lot of money. ... What the study tells us is that health care-associated infections continue to be a big problem, and it provides more evidence that we have to continue to work as a system to eliminate this harm."

Timely information needed

To make better progress in protecting patients, doctors and nurses need actionable clinical data at the local level, said Dr. Murphy, a member of the national team for On the CUSP: Stop BSI, a quality initiative aimed at preventing bloodstream infections that kill as many as 62,000 patients annually. Hospitals in 27 states are participating in the program, based on a project in Michigan that has cut the average catheter-related bloodstream infection rate 83% by implementing a checklist of infection-control procedures.

"This study provides yet another picture of the harm that's going on, but it's limited in that it's administrative data and it is difficult to translate that into actions," said Dr. Murphy, a pulmonology and critical care fellow at the Johns Hopkins University School of Medicine. "In our project, we have real-time data about things like the time since the last infection. The frontline staff can see the fruits of their labor or, conversely, the problems they still have to deal with."

While quality initiatives can yield benefits, Malani argued that hospitals lack the financial impetus to implement them, despite the Centers for Medicare & Medicaid Services' "no pay" policy for certain hospital-associated conditions, such as catheter-associated urinary tract infections.

"The Medicare solution doesn't go far enough," Malani said. "Because hospitals get paid for caring for the infections, they don't have a strong incentive to eliminate them. ... We want to break that connection."

The American Medical Association, the American Hospital Assn. and others have argued that it is unfair to deny payment for care associated with conditions that cannot always be prevented, even when evidence-based guidelines are followed.