A group of Rhode Island Hospital surgeons investigated medical conditions associated with failure to rescue — death after a treatable complication — using statistics from the hospital, said Andrew Stephen, assistant professor of surgery and coauthor of the study. The paper was recently published online in the Journal of the American College of Surgeons.
The study was led by Larissa Chiulli, a fifth-year surgical resident, with contributions from Stephen, Assistant Professor of Surgery Daithi Heffernan and Associate Professor of Surgery Thomas Miner. Chiulli started the project approximately two years ago, leading to the presentation of the research at the 95th Annual Meeting of the New England Surgical Society in September 2014, Miner said.
The researchers tried to identify comorbidities — the presence of two or more conditions in a patient — that contributed to high and low rates of failure to rescue, Stephen said. Failure to rescue is used to grade and compare the success of hospitals or surgeons in fixing problems and caring for their patients, Miner said.
But the rate of failure to rescue does not tell the whole story.
“What that missed was the degree of difficulty of taking care of these patients,” Stephen said. Though a small hospital may have a lower failure to rescue rate, this can be attributed to the fact that patients who require more complex and high-risk surgeries are often sent to larger hospitals in the city, he said.
The researchers conducted the study using the National Surgical Quality Improvement Program database for Rhode Island Hospital, which listed 7,763 patients, according to the article. The database includes preoperative factors such as age, gender and preexisting conditions as well as operative factors such as length of the operation and amount of blood loss. The data also includes information on postoperative complications that occur within 30 days of the surgery, Miner said.
The researchers first examined links between preexisting conditions and postoperative complications, followed by links between postoperative complications and failure to rescue, according to the article.
The study found that comorbidities related to organ damage, such as liver failure or sepsis — the body’s possibly lethal response to a bacterial infection — were more closely associated with high failure to rescue rates, while other comorbidities like diabetes and hypertension were not, according to the article.
A common mistake in these types of studies is that people assume causation from correlation, Miner said. The presence of a certain comorbidity is not a definite prediction of a postoperative complication, he added.
Though these studies will not likely change health policy, they can be used to improve overall care to lower post-surgery deaths, Stephen said. Results from these studies “can lead to increased screening or vigilance about certain patients that present to the hospital,” he said. “You may have to be more careful about these patients or mobilize more resources for them.”
Such tactics include ordering certain laboratory tests earlier, monitoring the treatment of the patient’s comorbidities and using more intensive care unit resources, he added.
Despite the current focus on lowering failure to rescue rates, these rates may not always be indicative of a struggling hospital or doctor who made a mistake, Miner said.
“There (are) times when failure to rescue isn’t a failure. It can be sound clinical judgment and a hallmark of a compassionate practitioner,” Miner said. As a surgical oncologist, Miner said cancer patients may suffer less if they undergo fewer procedures and interventions at the end of their lives.
Joseph Hyder, assistant professor of anesthesiology at Mayo Medical School in Rochester, Minn., and Elliot Wakeam, a surgical resident at the University of Toronto, are authors of another article examining failure to rescue that was published in June in the same journal. They studied data from the Nationwide Inpatient Sample from hospitals in the lowest mortality quintile and the highest mortality quintile from 2007 to 2011, according to the article.
Hyder and Wakeam reached similar conclusions to those in the RIH study, as both articles suggested the importance of identifying preoperative factors and being more attentive to patients with certain comorbidities, Hyder said.
“What my study tried to do is to ask whether there are new ways to take care of sick people,” Hyder said.
In their article, Hyder and Wakeam also emphasize the difficulty of communication when failure to rescue patients are identified postoperatively and care is provided too late.
“Identifying the patient in need of rescue is a challenge, but the second challenge is communicating about that patient,” Hyder said. “There are many people who care for that patient who need to know whether that patient is really sick or not.”
Though identifying high-risk patients preoperatively takes away some of the guesswork in solving postoperative complications, it is not always enough, Wakeam said.
“To prevent these deteriorations from happening is as important as managing them after they occur,” Wakeam said. “We’ll always have to do both. Our predictions can be imperfect.”
