Not all hospitals are created equal, and the differences in quality can be a matter of life or death.
In the first comprehensive study comparing how well individual hospitals treated a variety of medical conditions, researchers found that patients at the worst American hospitals were three times more likely to die and 13 times more likely to have medical complications than if they visited one of the best hospitals.
The study, published Wednesday in the academic journal PLOS One, shows “there is considerable variation in outcomes that really matter to patients, from hospital to hospital, as well as region to region,” said Dr. Thomas H. Lee, a longtime health care executive who was not involved in the research.
The study’s authors looked at 22 million hospital admissions, including information from both the federal Medicare program and private insurance companies, and analyzed them using two dozen measures of medical outcomes. Adjusting the results for how sick the patients were and other factors, like age and income, the researchers discovered widespread differences among hospitals. Even a hospital that had excellent outcomes for heart care might have poor outcomes in treating diabetes.
The study did not disclose which hospitals had which results. Under the terms of the agreement to receive the data, the researchers agreed to keep the identities of the hospitals confidential.
“Fundamentally, there is sort of an implicit assumption that every hospital is the same,” said Dr. Barry Rosenberg, the study’s lead author and a partner at the Boston Consulting Group in Chicago. But if someone has a heart attack, the closest hospital could have a death rate of 16 percent, compared with one a little farther away, where the rate was 4 percent, he said.
Earlier research examined the geographic variation in health care spending and how often patients received a medical procedure in a given market or hospital and found that there was wide variation. This study looked specifically at medical outcomes.
While factors like the health and income of a hospital’s patients contribute to its performance, Dr. Rosenberg emphasized that a large part is played by factors like the skill of the physicians and nurses, the culture at the hospital and how they chose to treat a given illness. “There is this other half of the story,” he said.
Hospitals that treated a high volume of cases were generally more successful than those that treated a low volume, but there were exceptions, Dr. Rosenberg said.
While the study underscored important differences among hospitals, the researchers also acknowledged that patients have little information about those differences. While consumers can use tools like Medicare’s Hospital Compare, which offers general quality information about individual hospitals, the data are very limited, Dr. Rosenberg said.
Many quality measures rely on reporting about whether the hospital gives patients an antibiotic, not whether they develop an infection, and they do not distinguish among different diseases. A hospital that is a top performer in heart surgery, for example, may be a poor place to choose to get a knee replacement.
The authors say patients need such information. “This paper raises the question of why don’t we have broader outcomes measurement and transparency around performance,” said Dr. Justin B. Dimick, one of the authors and a surgeon and researcher at the University of Michigan.
The researchers argue that this kind of information is necessary to judge the quality of the narrow network of hospitals increasingly offered by insurance companies. “The key thing about narrow networks is that they are created based on costs, negotiated prices and things like that,” Dr. Dimick said. “You need to pay attention to both cost and quality.”
Hospitals that excel in heart surgery or knee replacements should be rewarded by having more patients come and potentially being paid more for their care, Dr. Rosenberg said. “It’s an opportunity to improve health that has been underleveraged and underappreciated,” he said.
But researchers say obtaining information about outcomes is becoming increasingly difficult. Federal and state databases release less information than they used to.
If the incoming Trump administration wants to make health care function more like a market where people are encouraged to shop for medical care on their own, the researchers say, people will need that information.
“We’re going to need more access for people to find out these results about their hospitals and their care,” said Dr. Atul Gawande, one of the study’s authors and the executive director at Ariadne Labs in Boston, who has written frequently about the wide variation in medical practice in The New Yorker.
He points to two areas where information is publicly available: heart surgery by the Society of Thoracic Surgeons and cystic fibrosis. In those two areas, patients have the ability to make better-informed decisions and hospitals can use the data to improve their care.
“They are actually naming the outcomes by hospital,” he said. “The world did not end.”
Continue reading the main story