NY Times Op-Ed: Bring Back the Autopsy

By Sandeep Jauhar 

I recently attended an autopsy at the hospital where I work, in a room in the basement adjoining the morgue. The corpse, a newborn baby, was lying peacefully, as if napping, on a steel table with rusted wheels. He had succumbed to hypoxia, low oxygen levels in the blood, a few minutes after a full-term delivery. Did he have a heart defect? Had the umbilical cord been compressed? Had he breathed meconium, the first stool, into his lungs? The purpose of the autopsy was to find out.

I’d come that afternoon to witness what has become a rarity: the non-forensic medical autopsy. “Autopsy” means to see for oneself, and that is exactly what doctors once did. Fifty years ago we performed autopsies on roughly half of patients who died in hospitals. Today, the number of autopsies has dwindled to less than 10 percent, with next to none in nonacademic hospitals.

The reasons for the decline are many. In 1971, the Joint Commission on Accreditation of Hospitals stopped requiring hospitals to perform autopsies. In 1986, Medicare, presumably skeptical of the usefulness of autopsies, stopped directly paying for them. Since then, private insurers have followed Medicare’s lead.

Perhaps the biggest reason for the autopsy’s demise is that doctors no longer seem to view the procedure as essential. We used to use autopsies as an educational tool, to learn what we’d gotten right and what we’d missed. Today, many doctors believe that medical tests will reveal all they need to know about how and why a patient died.

Nothing could be further from the truth. Despite technological improvements in medicine, diagnostic errors remain rampant. According to a recent Institute of Medicine report, misdiagnoses contribute to some 10 percent of patient deaths and account for 6 percent to 17 percent of patient harm in this country. Most Americans will be affected by at least one diagnostic error in their lifetime.

Strangely, diagnostic accuracy is rarely raised as a safety issue. Compared with treatment errors, such as medication overdoses or wrong-limb amputations, diagnostic errors receive little attention. Of course, they are harder to identify. Many doctors never find out about them because they never learn the outcomes of their cases.

This is where autopsies can be very useful. Even though modern testing can give us a lot of information, 10 percent to 30 percent of autopsies still reveal undiagnosed medical problems. Studies have found that patients at hospitals that perform more autopsies suffer fewer major diagnostic mistakes.

Still, it isn’t easy to ask a family for an autopsy. When I was a resident I took care of an elderly man in whom a mysterious fever had taken up residence like a malevolent squatter. His temperature was 102 degrees or higher for weeks. He lost a third of his body weight. All tests — bacterial, viral, oncologic, rheumatologic — were normal. They remained that way until he died. An autopsy would probably have been useful. It could have shown us what we had missed. But none of us wanted to ask his grieving wife for permission.

Feedback and reflection are critical in any human enterprise, but especially so in medicine, where the stakes are so high. We don’t always see the correct diagnosis when our patients are alive. However, it can be pretty obvious in the splayed-out tissues of an autopsy.

The autopsy I attended took two hours. The pathologist cut into the baby’s plum-size heart. It was normal. So, too, were the intestines, the liver and the kidneys.

After the initial examination, the pathologist placed the baby’s chest and abdominal organs on a neighboring table for closer inspection. It was there that we noticed the tiny lungs had yellowish spots on their surface. The pathologist didn’t know what to make of this, but when he sliced the lungs open, yellowish-green meconium was present in all the major breathing tubes out to the furthest branches. He said it was the worst case of meconium aspiration he had ever seen.

It occurred to me that the baby’s parents could perhaps now have some closure — as well as some potentially useful information. The diagnosis was not, say, a hereditary abnormality that could affect other children they might choose to have.

My hospital, I learned, is a regional autopsy center. It receives bodies from a dozen affiliated health care sites. (The baby in fact had been transferred from another hospital.) I was told that the hospital had increased the number of stored bodies, almost doubling the number of storage lockers. Yet the autopsy rate has not budged.

Reviving the autopsy would be a good thing, giving doctors a sorely needed tool to improve diagnosis. At the very least, Medicare and private insurers should start paying for autopsies again, so that financial considerations do not limit their use. Electronic hospital records should give doctors reminders to ask for them. Despite the emphasis on metrics and data in medicine today, we ignore perhaps the most important information of all: what we can see for ourselves.

Sandeep Jauhar, a cardiologist and a contributing opinion writer, is the author, most recently, of “Doctored: The Disillusionment of an American Physician.”

Tags: