Medical errors in America kill more people than AIDS or drug overdoses. Here’s why.

by Sarah Kliff

1) The American health care system kills tens of thousands of patients each year

Medical errors kill more people than car crashes or new disease outbreaks. They kill more people annually than breast cancer, AIDS, plane crashes, or drug overdoses. Depending which estimate you use, medical errors are either the 3rd or 9th leading cause of death in the United States. Those left dead as a result of their medical care could fill an average-sized Major League Baseball stadium — sometimes twice over.

Medical errors tend to fall into two buckets. There are the mistakes that happen when doctors set a wrong plan: when they prescribe the wrong medication, for example, thinking it was the right treatment. Then there are the errors that occur when doctors set the right plan but don’t follow it — when messy handwriting means a patient gets the wrong drug dosage, for example, or when a surgeon operates on the wrong body part (yes, this actually happens).

“Something like 2 to 3 percent of people who go into the hospital are going to have some pretty severe harm as a result,” says Don Berwick, the Obama administration’s former Medicare administrator. “Australian studies show that the rate might be as high as 12 percent. The harder you look, and the more you study the issue, the more errors you find.”

There are two main studies that try to estimate how many people our health care system kills annually. One is To Err is Human, a seminal research project published in 1999. It used two studies — one conducted in New York in 1984, and another from Colorado and Utah in 1992— that tried to estimate how frequently preventable errors happen in medicine.


In the Utah and Colorado study, researchers found that 2.9 percent of patients experienced a medical error during a hospital trip. In New York, it was 3.7 percent. The New York study also found that, of those who experienced a medical error, 13.7 percent died because of it. In other words: of 1,000 patients who enter a hospital, about 30 or 40 would experience a mistake in their care. About four would end up dead as a result.

To Err is Human expanded those figures to include the entire population. It estimated that “at least 44,000 and perhaps as many as 98,000 Americans die in hospitals each year as a result of medical errors.”

This number is huge — far higher than many people realize. One survey found that both doctors and the general public, when asked to estimate the number of deaths related to medical errors, ballpark the figure around 5,000.

The report was a bombshell in the medical community. It told doctors, nurses, and everyone who provides health care that the work they did was, at a minimum (using the lower-bound figure of 44,000) the 8th most frequent killer in the United States. Medical prescription errors alone — just the instances where patients accidentally get the wrong drug or dosage — were killing 7,000.

For decades, the To Err is Human was accepted as the best guess of how the health care system inflicts harm — even though those who worked on the data said it likely understated the size of the problem.

“My own feeling was the problem was twice as bad,” says Harvard’s Lucian Leape, one of the nation’s top experts on medical errors who authored the study on error rates in New York. “They were based on what was written on medical errors, and we all know that one of the things doctors often leave off a medical record is a mistake.”

A newer analysis, published in 2013, made a similar claim: that To Err is Human had counted too few medical errors. The research, published in the Journal of Patient Safety, estimated that medical errors contribute to the deaths of between 210,000 and 440,000 patients. At the lower bound, that’s the equivalent of nearly 10 jumbo jets crashing every week — or theentire population of Birmingham, Alabama dying every year.

“There is no single medical intervention that will ever save as many lives as patient safety improvement,” Berwick says. “There is so much harm going on.”

2) Bed sores are a huge source of harm in the health care system

Some errors in medicine are stunningly bad. One study, published in the journal Surgery, found that surgeons operated on the wrong part of the body 2,413 times between 1990 and 2010. They left foreign objects behind in the body (typically sponges) 4,857 times. In 27 cases, they operated on the wrong patient altogether.

These errors are terrible, and easy to recognize. But they aren’t what cause the most harm in American health care. It’s the less stunning, more quotidian mistakes that are the biggest killers. Take, for example, bed sores.

Bed sores are one of the more mundane complications of modern medicine. They’re called “pressure ulcers” in medical jargon, and are the open wounds that patients develop when they have not moved for long periods of time. The skin literally cracks under the pressure of the body weighing down on it.

We know a lot about why pressure ulcers happen: because immobile patients stay immobile too long. Bedsores are an especially large problem in long-term care and nursing facilities, where residents can spend days or weeks lying in the same positions.

click to view

A foot with bed sores (BSIP/UIG Via Getty Images)

A 2006 government survey found that more than half a million Americans are hospitalized annually for bed sores that are the result of other care they have received. 58,000 of those patients die in the hospital during that admission.

Does this mean that pressure ulcers killed all those patients? No — these are typically frail, elderly patients battling other conditions ranging from pneumonia to dementia. But did bed sores mean some of these patients died who otherwise wouldn’t have? Experts say that’s almost certainly the case.

“The number of people at risk for pressure ulcers is growing with the elderly population living longer,” says Madhuri Reddy, a gerontologist at Beth Israel Deaconess Medical Center in Boston, who has done extensive research on pressure ulcer prevention.

Some experts say bedsores should never happen. We know the cause (immobility) and effect (bedsores). “The goal for health care facilities to reduce pressure ulcers is admirable but not adequate,” writes Kathy Duncan, a faculty member and nurse at the Institute for Healthcare Improvement. “The goal for pressure ulcer incidence should be zero.”

Why do pressure ulcers persist as a major problem? There is a bit of a split in the research community about that. Leape argues its a matter of diligence: this is a problem that we could easily solve if we cared more about it.

“You have to put adequate padding on the feet and legs, and you have to turn the patient frequently, but that requires vigilance,” Leape says. “This requires people to adhere to a really strict schedule of checking the patient, often in nursing homes that are poorly staffed. It’s really all about providing rigorous attention, and that doesn’t always happen.”

Reddy argues that the issue a bit more complicated than Leape presents it. There are risks, for example, to moving already-fragile patients. “Pressure is the biggest risk factor for an ulcer, but not the only one,” Reddy says. “Another issue is friction,” when the skin rubs against the bed as the patient is being turned.

One problem, Reddy says, is that there just isn’t great research available on how to balance those risks. Right now, most health providers adhere to a two-hour standard for how often to move immobile patients. But they don’t really know whether there would be more or fewer ulcers if they moved patients more often. Nobody has done that study.

“Our guidelines come from one several-decades-old study from the 1980s,” says Reddy. “We really need better evidence in prevention and management literature.”

3) Medical errors haunt and embarrass doctors — and that can make them less likely to tell patients

Compounding the problem of medical errors is the fact that often, patients don’t even know when a mistake has occurred. Only one-third of patients are told about a medical error when it happens, multiple surveys find.

Harvard University’s Robert Blendon conducted the first survey on the topic in 2002. He found that 30 percent of the patients who said they experienced medical harm were told about the problem. Doctors reported a similar rate when they received health care (31 percent).

A separate review study, published in 2004, found that about one in five doctors said they had disclosed their most serious medical error of the last year. This could be one reason that, despite medical harm being a leading cause of death, many don’t see it as a pressing issue for the health care system to address.

The lack of disclosure often gets chalked up to America’s medical malpractice laws: doctors don’t want to tell patients when something went wrong because it increases the likelihood that they could get sued.

“When we first started doing the research, we thought that it was fear of being sued that was the major problem,” says Tom Gallagher, a bioethicist at the University of Washington who studies doctor disclosure of medical errors.

“It’s really tough in a medical environment to talk about things going wrong”

But Gallagher started looking at other countries, like New Zealand, where the legal system doesn’t allow patients to sue doctors for medical harm, or Canada, where doctors get sued about 80 percent less often than in the United States. And he saw that doctors in those countries were just as unlikely to tell patients when something had gone wrong in their care.

This suggests that malpractice reform — changing our laws to look more like New Zealand’s — wouldn’t necessarily make providers more likely to tell patients when they screwed up. Gallagher has, instead, focused much of his work looking at the more psychological and personal factors that make it difficult for doctors to talk about care gone wrong.

“Think about the times that you’ve harmed somebody personally,” Gallagher says. “Doctors go through all the same rationalization and defense mechanisms as the rest of us. It’s painful and embarrassing to say something significantly went wrong with a patient’s care, and that you may have been partially responsible for that.”

Doctors typically went into their profession to help save lives and mend bodies. When they end up doing the opposite, it can be shameful and embarrassing. Doctors talk about replaying the moment of the error in their mind, over and over again, asking: how could I have screwed that up? How is it possible I got that wrong?

Danielle Ofri, a practicing physician in New York City, wrote an essay for the journal Health Affairs about the shame she felt in the wake of a near-fatal medical error, where she gave a patient too little insulin — a mistake that nearly killed the patient:

When I think back to that moment in the ER when the senior resident berated me for my error, it was shame that overpowered me. Of course I felt guilty—that was the easy part. But it was the shame that was paralyzing. It was the shame of realizing that I wasn’t who I thought I was, that I wasn’t who I’d been telling my patient and my intern I was. Up until that moment, I’d thought I was a competent, even excellent, doctor. In one crashing moment of realization, that persona shattered to bits.

“It’s really tough in a medical environment to talk about things going wrong,” says Gallagher.

Much of Gallagher’s research looks at how to increase disclosure rates. One thing that he’s found to work are disclosure coaches: hospital staff members who are on call 24/7, and whose job it is (as the title implies) to help doctors work through the process of telling patients and their family that something went wrong.

4) Sometimes doctors won’t even know when an error has occurred. That’s part of what makes it a hard problem to fix.

When a plane, car, or any other form of transportation crashes, we know that something went wrong.

But spotting medical errors — particularly those that kill people — is much more difficult. Patients die in hospitals every day, typically at no fault of their doctors. So unlike plane crashes, where a dead body is a sure sign that something went wrong, medical errors can often blend into the lives and deaths that happen routinely in the medical system.

“If you see a dead person on the road, you know something is wrong,” says Ian Leistikow, a Dutch expert on patient safety who works for the government’s improvement efforts. “You don’t have to study anything to know this. But in hospitals, you see dead people all the time.”

There are some errors that are easy to see: when doctors amputate the wrong leg, for example, or prescribe 100 milligrams of a drug when he meant to write down 10. But there are likely dozens more subtle, quick decisions that doctors make each day and that are difficult to trace through a patient’s care.

Errors often happen unbeknownst to the doctors making them. When he was a practicing psychiatrist, Leistikow got a phone call one day from his boss, asking him to prescribe 10 mg of Haldol, an anti-psychotic medication. Leistikow thought that was a large dosage, as did the pharmacist he sent it to. The pharmacist asked whether he should halve the dosage. Leistikow said yes — and didn’t think much of it, until the police found his patient wielding a giant sword in his neighbor’s garden. Turns out, he really did need the larger Haldol dosage.

Lots of other mistakes, however, don’t lead to sword-wielding patients. “It’s only when the harm was very evident, and very clear, that I would find out about a mistake,” he says. “There are probably many mistakes I made that were never noticed.”

Medical care is complex, often involving many different doctors and nurses. In the United States particularly, these people are spread across different hospitals, with different owners, that don’t share electronic medical records. If a patient turns up at a different emergency room or seeks care elsewhere, it’s possible their original doctor may never know.

5) The financial structure of the health care system makes it difficult to reduce errors

When the Institute of Medicine published To Err is Human, David Blumenthal thought he saw a business opportunity to help hospitals make less mistakes. It seemed like a no-brainer: with this big, landmark report calling out a major issue, the C-suite should be desperate to make their hospital look better.

“I thought it was this great consulting opportunity,” says Blumenthal, who was then at Harvard and is now the president of the Commonwealth Fund.* “But there was just no interest.”

The financial incentives of the health care system make it hard to do the right thing. Right now, most health care providers get paid on a fee-for-service basis: each time they do a surgery or check-up, the insurance company gives them a set amount of cash.

Doctors’ mistakes can create perverse incentives. A harmed patient typically needs more care — and the health care provider will profit.

This is not to say health care providers harm patients to profit. That type of behavior is both rare and criminal. But the way doctors get paid suggests one reason why it’s difficult for hospitals to commit to reform: they’re being asked to work against their own financial incentives.

A harmed patient typically needs more care — and the health care provider will profit

“When the chief executive has a cabinet meeting in the C-suite, the number one item is going to be how to keep the hospital occupied and making revenue,” says Berwick. “If that’s what they’re talking about — how do we get more referrals, how do we get to be the biggest MRI provider — that’s where the attention will go.”

The federal government has done some work to try to change this. Part of the Affordable Care Act penalizes hospitals if their patients come back within 30 days of their initial visit for something that went wrong the first time. Hospitals can now lose as much as 3 percent of their Medicare revenue if they have high readmission rates. There is another Obamacare program that penalizes specific types of errors, like when patients catch infections in the hospital.

Experts are divided on whether these financial incentives are strong enough to combat the other incentives to provide lots and lots of care. But the aim of these programs, at least, is clear: to make reducing errors both the right and more lucrative thing to do.

6) There are signs of hope that, with serious effort, the health care system can reduce errors

Less than two decades ago, health care providers accepted the fact some patients who got a catheter inserted into their heart would end up with a bloodstream infection. In the mid-2000s, this happened to an estimated 80,000 people annually, and caused as many as 28,000 patient deaths. At the time, this appeared to be an unavoidable risk of a complicated heart surgery.

In 2006, a researcher at Johns Hopkins University proved this all wrong. Peter Pronovost published a landmark paper in the New England Journal of Medicine with evidence that the health care system didn’t have to accept these infections as the cost of doing business. For three months, 103 hospital emergency rooms in Michigan implemented five relatively simple changes in how they inserted catheters. Surgeons had to wash their hands with an antibacterial called chlorhexidine, for example.

After three months of implementing these changes, an amazing thing happened: the bloodstream infection rate at these 103 Michigan ICUs fell to zero.

The change in protocol spread quickly: in early 2015, the federal government reported a 46 percent decrease in central line infections between 2008 and 2015.

The story of catheter-related bloodstream infections is an example of real success in making the health care system a safer place. It showed that some of the things we accept as necessary risk in health care aren’t actually that necessary at all — and, while they take work to fix, don’t require a Herculean effort.

“Some people will say that some amount of error is unpreventable, but I disagree,” says Harvard’s Leape. “I say that the Pronovost example shows that’s not true: it’s just an error we haven’t learned how to prevent.”

7) But it’s hard to know on a broad scale whether things are getting better

Patient safety experts say it’s incredibly hard to measure whether the health care system is getting safer on a larger scale — if, as each year passes, it kills more people or fewer. This is mostly due to the fact that widespread, system-wide data on medical errors is hard to come by.

Researchers have tried to quantify it, sometimes to disappointing results. A decade after To Err is Human was published, a team of researchers led by Harvard University’s Chris Landrigan looked at North Carolina hospitals to see whether health care had become safer. The researchers chose North Carolina because the hospitals there had been especially interested in improving care and signed up, for example, for a national patient safety campaign.

The findings, published in 2010 in the New England Journal of Medicine, “did not show any reduction in patient harm despite substantial national attention and allocation of resources to improve the safety of care,” the researchers wrote.

The Obama administration published more optimistic findings during the end of 2014: it estimated that reductions in specific medical errors had saved 50,000 lives between 2010 and 2013. Some researchers, including Pronovost, say they’re skeptical of these claims, in part because they didn’t standardize the data that hospitals had to submit — making it possible that hospitals could choose records that might make their error rates look lower.

In general, researchers say they’re skeptical of any sweeping claims about the amount of harm in health care going up or down — mostly because we know so little about how much it happens in the first place.

This article is the first in Vox’s yearlong series on fatal medical errors. Reporting for this series is sponsored by by the Association of Health Care Journalists’ Reporting Fellowship on Health Care Performance and supported by The Commonwealth Fund.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s