The autopsy, long out of favor in non-forensic cases, is enjoying new life in some hospitals as a tool for uncovering cause of death and improving quality of care.
This article appears in the April 2014 issue of HealthLeaders magazine.
For years, the hospital autopsy has been dying.
In fact, experts say most hospital policies regard the autopsy as already dead—at best a calcified relic—since the mid-1970s, when The Joint Commission lifted its requirement that hospitals perform autopsies on 20% of inpatient deaths to gain accreditation. Even teaching hospitals have reduced the number they do for training medical students.
Rare is the hospital where the autopsy rate is greater than 5% of nonforensic deaths. At most it's usually only 1%, several pathology experts say.
A big deterrent is the autopsy's cost, $1,000–$3,500 or more, which the hospital usually has to absorb as part of its quality assurance requirements. Medicare embeds this in the episode payment for its beneficiaries without a line item, per se, which means that hospitals can't bill separately for the procedure, although every beneficiary has the right to one if requested.
Revealing information on misdiagnosis
But against all this, a few hospital leaders are resisting the trend, trying to keep the autopsy alive, and even revive it as they campaign for their doctors and families to allow much wider usage. The autopsy can answer questions about contributing causes of death and thus improve quality of care, they say.
"We thought you had x but you ended up dying of y, and that means somewhere between 60,000–80,000 people a year die in the intensive care unit because of misdiagnosis," says Peter J. Pronovost, MD, senior vice president for patient safety and quality at 1,059-bed Johns Hopkins Hospital in Baltimore, Md. Pronovost also is director of the Armstrong Institute for Patient Safety and Quality, also in Baltimore.
A staunch autopsy advocate whose research discovered those statistics, Pronovost says missed infections and unrecognized vascular complications are the most common diagnostic errors resulting in death, and yet many of these complications are preventable.
Recognizing such risks can prompt ultrasound testing to find deep vein thrombosis, which puts patients at risk of potentially lethal pulmonary embolism, he says. The autopsy "is a valuable tool to help us learn how to prevent misdiagnoses in the future," Pronovost says.
Hospitals with good autopsy track records might consider showcasing that as evidence that they are committed to transparency and aren't afraid to learn from their mistakes.
The caveat, however, is that the autopsy must be part of a feedback loop that gets that diagnostic information back to the clinical team, he says. "Where we go wrong, and where people question the autopsy's value, is if the result just sits in pathology and doesn't get back to the team to show ways to improve."
Even when autopsies are performed on patients for whom the cause of death was not in question—even patients with multiple known diseases and advanced age—up to 35% of the examinations revealed something unknown about the patient's disease or condition that could have affected their course of care, says Elizabeth Burton, MD, The Johns Hopkins Hospital's autopsy division deputy director.
That's one reason that Johns Hopkins makes autopsies a priority, asking the medical representatives of every patient who dies to allow the procedure, even when the cause of death is thought to be well understood and there are no concerns about a preventable adverse event.
"At Hopkins, 20%–22% of the patients who die here get an autopsy," or about 350 cases a year, Burton says, noting that these days, that's an extraordinarily high percentage.
Over the years of her experience at other hospitals, Burton recalls autopsy findings that patients diagnosed with cancer never had it to begin with, or had a different type, or that their primary cancer type was different than the one being treated, or that they had multiple types of cancer when only one was under treatment.
Her first case years ago in Texas involved an alcoholic with cirrhosis, which was the presumed cause of death. The autopsy revealed he had a common and treatable type of liver cancer, but no one knew.
Fungal infections can run rampant quickly through a hospital, causing death that might not be recognized but for the autopsy, Burton adds. When you find the first patient who died from that, "you can feed that back to the hospital epidemiologist and treat other patients prophylactically."
Potential cost savings
Gregory J. Davis, MD, FCAP, has three jobs in which he advocates for more autopsies. He's the medical director of the Forensic Consultation Service of the Department of Pathology and Laboratory Medicine at 871-bed University of Kentucky HealthCare in Lexington, the assistant state medical examiner for the Commonwealth of Kentucky, and a spokesperson for the College of American Pathologists.
"I'm astounded that many protocols, like for cancer treatment, have nothing about providing funds for autopsies should the patient die. What better way do you have to determine whether the treatment was successful, or on its way toward being successful, but to look at the effects of that treatment on the disease?" he asks.
Davis says the University of Kentucky performs about 150 autopsies each year at no cost to the family, which is about 10% of inpatient deaths. Most hospitals, however, only do 1%, he says.
"There are many hospitals around the country where the autopsy is viewed as nothing more than a financial drain. Many hospitals have built new buildings without even including an autopsy suite and have no place to store bodies. So the funeral home comes directly to the floor to pick up," he says.
That's the case with many Kentucky hospitals, which contract with the university for autopsies on an ad hoc basis when an issue comes up, he says.
Though he knows hospital quality leaders who want to do more autopsies have a tough fight, they should make the argument about how the autopsy can actually save hospitals money. Consider how they can deflect costly litigation, he says.
"The vast majority of cases in which the family is ready to bring legal action for malpractice go away when an autopsy is performed," Davis says. "That's because in civil court, all the plaintiff needs is preponderance of evidence … and a lawyer with a gift of gab who convinces the jury on a coin toss.
"A good autopsy can produce the facts to make the erroneous case go away, but without that, plaintiffs can claim pretty much what they please."
Davis says another reason there is resistance against doing more autopsies is "frankly, hubris. People think they know more than what they do because they have a CT scanner and an MRI, and thus a good handle on things."
And pathologists who might want to do more don't push the point "because they're already superbusy, with more biopsies and consults than they can deal with. If the payment structure changed so that pathologists were paid by the autopsy, like any other human endeavor, I think you'd have more people wanting to do them."
Davis adds, "there's a mentality in this country that once the patient is dead, there's nothing more you can do, when in actuality there is more. You can get information on a micro level to help the patient and the family and the clinical team, and on a macro level to look at quality measures" within the hospital.
Hospital leaders may have an added incentive to learn from inpatient deaths in order to prevent them, as federal health reform regulations increasingly impose financial penalties on hospitals with higher rates of complication and 30-day mortality.
Validating technologies, processes
Michael Fishbein, MD, professor of pathology and medicine, as well as director of autopsy pathology, at Ronald Reagan UCLA Medical Center, says his hospital's autopsy rate is about 15% of all inpatient deaths, "although we'd like to have it much higher."
That's because in his experience, 30%–40% of the time, autopsies reveal something about why the patient died. They're also much more revealing than an MRI or CT, with which some hospitals have replaced the traditional invasive autopsy, he says.
The traditional autopsy can reveal adverse effects from new technologies and medications that don't yet have a long track record of safety in patient care.
"The autopsy surprises us by showing us a new type of complication," Fishbein says. Many new surgical techniques have not been thoroughly validated and there, too, the autopsy can uncover surprises about how effective the technologies were.
For example, he says, robotic surgery for prostate cancer is said to be "nerve sparing. Do they really spare the nerve? How can you find out? Well, you can look and examine the nerve."
The autopsy has revealed that assist devices on patients waiting for heart transplants may perform less optimally depending on how they are attached, he says. "We're learning through the autopsy that there are certain ways these are attached that make them more efficient, with less risk of infection and less risk of bleeding."
Looking at deaths among patients with artificial heart valves is another effectiveness check, Fishbein says. "How are you going to know if they died because something was wrong with that valve or because they have some other cardiac problem? The only way is going to be through the autopsy."
George Lundberg, MD, chief medical officer at cancer data analytics company CollabRx, editor-at-large of Medscape, and former editor-in-chief of the Journal of the American Medical Association, has harsh words about the decline of the autopsy. He says hospitals today don't perform them because "they don't really want to know" why a patient died. "There's a culture of silence and repression or suppression of truth. And many doctors will tell you they're afraid of being sued if someone finds something they missed or did wrong."
That's extremely misguided, he says. "The patient safety movement depends on learning if something goes wrong. You can't have a proper safety initiative without a significant autopsy rate."
Lundberg also says the cost of autopsies is a huge deterrent to doing them. "Hospitals don't want to spend money on dead people," he says.
So, why did The Joint Commission lift its accreditation requirement for hospitals to perform autopsies on one in five inpatient deaths?
Ana Pujols McKee, MD, the organization's executive vice president and chief medical officer, explains that The Joint Commission found that hospitals "may perform autopsies to meet the minimum requirement, or if families object to an autopsy it may be difficult for an organization" to meet the minimum rate. "In addition, there is no consensus on what is the best rate."
The Joint Commission dropped the requirement in the early 1970s. Hospitals using Joint Commission accreditation for deemed-status purposes are still required to develop criteria for autopsies, and of course should "attempt to secure autopsies in all cases of unusual death" and cases of "medical, legal, and educational interest," McKee says.
At 894-bed Northwestern Memorial Hospital in Chicago, "rumors of the demise of the autopsy are premature," says Jon Lomasney, MD, director of autopsy service. There, autopsies are performed for about 20% of patients who died in the hospital and for one in five of those, the autopsy always reveals something important involving the accuracy of the diagnosis or missed diagnoses, the efficacy of the therapy or treatment, and whether the treatment or care caused any harm.
In any event, "you learn something from every case," Lomasney says.
"We learn just as much from doing an autopsy on a young patient who drops dead while shopping on Michigan Avenue as we do from a patient who is 86 with a 20-year history of congestive heart failure."
Approaching the family
A challenge to performing autopsies is getting consent from the family without making anyone suspicious that the surgeon or intensive care unit did something wrong.
Families and physicians often think the autopsy is intrusive or macabre. There are misconceptions that an autopsy will significantly delay release of remains, preclude an open-casket service, or is forbidden by their religious faith, says Lomasney.
At Ronald Reagan UCLA Medical Center, Fishbein says, physicians are instructed that it is their "obligation to request an autopsy" from family members. "It should be emphasized to the family that the body will be treated with respect and that the autopsy will not interfere with viewing or open casket funeral services."
All families are given a UCLA pamphlet asking them to consider an autopsy for the deceased. It reads, in part, "Sometimes families feel 'guilty' that they or the doctors might not have done everything possible. The autopsy may provide findings that help alleviate this guilt."
The autopsy can identify infectious or inherited diseases, occupational illnesses, or environmental hazards that pose a threat to others, and may assist the family in collecting insurance benefits. The brochure says many diseases that affected large numbers of people were first defined at autopsy, "including AIDS, Legionnaire's disease, and toxic shock syndrome."
Lomasney says that often the autopsy results can yield information to set at ease family members who may fear that they didn't bring their loved one in for medical attention soon enough.
At Northwestern, the autopsy has saved lives in many ways, Lomasney says.
For example, a few years ago in an older building, "we were having unexplained deaths on the bone marrow transplant service. And we discovered by autopsying those patients that they were inhaling aspergillus spores through the duct work and getting pulmonary aspergillosis. Everyone coming into their rooms wore masks, but the patients didn't.
"The hospital shut down the transplant units, cleaned the ductwork, and put on HEPA filters."
Lomasney acknowledges that some clinicians "fear malpractice litigation" that might result from an autopsy finding. But the bottom line is that everyone "needs to know the facts. Sometimes families will be unhappy and seek to sue. But if there's a therapeutic misadventure, we need to know. It will be confirmed or discovered by the autopsy … and shared with the family immediately.
"But it also gives us an opportunity to prevent it from happening to another patient."
Reprint HLR0414-9
This article appears in the April 2014 issue of HealthLeaders magazine.