Skip to main content

Diagnostic Errors Common, Costly, and Harmful

 |  By John Commins  
   April 23, 2013

Wrong site surgeries and medication mishaps get all the gory malpractice headlines, but diagnostic errors are a larger, deadlier, and costlier problem, according to a study released Monday by Johns Hopkins researchers.

"We have said repeatedly that the fundamental premise is if you don't have the diagnosis right you can't possibly get the treatment right unless occasionally you get lucky," says David E. Newman-Toker, MD, an associate professor of neurology at the Johns Hopkins University School of Medicine and leader of the study being published online in BMJ Quality and Safety.

"The correct premise is that correct therapy begins with correct diagnosis. Unfortunately the entire medical profession operates under the collective delusion that diagnoses are almost always right," Newman-Tucker says. "There are all kinds of things we track in the hospitals; quality measures of one kind or another. But nobody is tracking whether or not their diagnoses are right. It's either ironic or scary but it's not good."

The Johns Hopkins researchers reviewed 350,706 malpractice claim payouts between 1986 and 2010 and found that diagnostic errors accounted for 28.6% of the claims, the most severe patient harm, and punitive payouts that amounted to $38.8 billion.

The study examined only a subset of claims that prompted a malpractice payout, but the researchers estimated that between 80,000 to 160,000 patients suffer misdiagnoses-related, potentially preventable, significant permanent injury or death annually across the nation.

Newman-Toker spoke with HealthLeaders Media about his findings. The following is an edited transcript of the interview.

HLM: Were the results you got the results you expected?
N-T: We were anticipating most of the results. There were certain aspects of this that we expected, but there wasn't prior literature. For instance: The high frequency with which claims were associated with disability and not just death. That has not been previously looked at and that is one of the most important findings from the data we analyzed.
It really indicates that the magnitude of the problem as has been previously assessed through autopsy studies is a significant underestimate of the serious negative impact and harm associated with diagnostic error.

HLM: What seems to be the pitfall here?
N-T: There are no quality measures that systematically ask the question 'What is our accuracy of diagnoses?' 'How often do we make diagnostic errors?' 'What is the impact of those diagnostic errors?' The call to action message from this paper and many others that we have tried to publish over the last five to 10 years to shed light on this problem is that we must systematically measure diagnostic errors. It must be a requirement and it must be part of the quality reporting guidelines in order for us to start tackling this problem."

HLM: What is the source of this diagnostic disconnect?
N-T: I don't think it is arrogance, although some of it is overconfidence in the sense that we have a system that does not give us a great deal of feedback on the accuracy of our diagnosis. Not every patient gets the same follow-up. The patients who come back to see us come back because they are satisfied and they like us and they send us Christmas cards and knit us shawls and we think of ourselves as good diagnosticians as a result.

Unfortunately, the ones who don't come back are those who sought care elsewhere and got better treatment or unfortunately some of them ended up in a body bag. We never find out about those patients and that is worse in places like the emergency department, where they get virtually no follow-up on any of their patients either.

But it is also bad in primary care and other places where there isn't a mechanism whereby the system ensures that you will get follow up on all the patients who don't come back. We developed this delusional belief that because we are seeing patients in follow-up who are doing well that we must be right most or all of the time.

The other thing that is the fool's gold here is that most of the time our diagnostic mistakes don't harm patients. So we probably have a huge rate of diagnostic error. If you just look at autopsy data where the studies are reasonably clear and crisp around this issue 5% of autopsies done in hospitals suggest a Class 1 diagnostic error.

That means those are patients who died but had their diagnosis been correct they 'coulda, shoulda, woulda' left the hospital alive instead of in a body bag. That seems like a relatively small number. Certainly you can say 95% were right and that sounds good. But there weren't 95% all right. There is another 20% Class 2 error rate where we made major or significant misdiagnoses.

Then you start getting into other misdiagnoses that get further away from harm the rates get up into the 50% plus range. We are probably making diagnostic errors all the time, but we are lucky because the conditions don't kill people or it gets discovered later.

Because we don't see the harm very often, both because they are infrequent as a fraction of all the diagnostic errors we make and because the ones who die don't come back to see us when we systematically missed those events we have this kind of collective delusion that we are usually right.

HLM: What can hospitals or physicians do to improve on these misdiagnoses?
N-T: Although there are a number of solutions that have been proposed, none of them are magic bullets. They range from simple check lists all the way to highbrow, computer-based diagnostic decision support. Those solutions are probably going to have to be used in concert in very dedicated ways to solve particular types of diagnostic error problems.

None of those solutions, all interventions described out there to reduce diagnostic errors, have ever been properly tested at the highest level of did 'I actually save anybody?' They have been proposed and described and tested on paper cases but never actually studied in practice. My view is there are a lot of potential solutions out there. But in order for those solutions to be properly tested we need to measure this on an ongoing daily regular basis.

HLM: Why aren't more hospitals or physicians taking it upon themselves to address this issue?
N-T: There is right now no incentive to track diagnostic errors. In fact there is a disincentive. If no one is required to report it why would a hospital want to look under the rug and find all these dust bunnies; that their diagnostic error rate is 10% to 15 % of all their encounters?

That would look bad. There would be mass panic and confusion, particularly if they are the only ones saying it. That is the problem. Everyone is afraid to look at this issue and as long as no one forces them to look at it they are going to keep it under the rug as long as they can.

Pages

John Commins is a content specialist and online news editor for HealthLeaders, a Simplify Compliance brand.

Tagged Under:


Get the latest on healthcare leadership in your inbox.