Laura Landro’s column in the Wall Street Journal today is a must-read: Adding Up Diagnosis Errors. Why? Let me count the ways:
- She is one of the most thorough, informed health care reporters around.
- She is covering an important topic that should be of interest to everyone.
- The study is behind a paywall so media coverage is the only way the general public can learn of the findings.
Her opening line:
Mistakes diagnosing patients are the most common, costly and dangerous errors made by doctors in the U.S. and result in permanent injury or death for as many as 160,000 patients annually, a new study found.
My eyebrows shot up and I immediately searched for the full-text article, which is unfortunately behind a paywall:
25-Year summary of US malpractice claims for diagnostic errors 1986–2010: an analysis from the National Practitioner Data Bank (BMJ Quality and Safety; April 22, 2013)
I also spent a few minutes digging through my own tweets since I remembered a similar study in JAMA recently. I should have just listened first to the podcast accompanying Landro’s column, since she mentions it:
Types and Origins of Diagnostic Errors in Primary Care Settings (JAMA Internal Medicine; March 25, 2013)
And an accompanying commentary:
Measuring Diagnostic Errors in Primary Care: The First Step on a Path Forward
The quote that jolted me when I read it in March:
“Most diagnostic errors are linked to clear defects in bedside history taking, physical examination…”
Which led me to tweet:
- Is it any wonder that 35% of U.S. adults go online to get a jump on a diagnosis?
- Is it any wonder that so many U.S. adults track some aspect of health, esp those with serious conditions?
But now, looking back, I’m intrigued by this line (visible in the free summary of the commentary):
“Missed, wrong, and delayed diagnoses have been underappreciated by internal peer review, autopsy reports, and examination of malpractice claims.”
Today’s news, of course, focuses on malpractice claims, so I have lowered one of my eyebrows (I’m an optimistic skeptic, so this is a pretty frequent facial exercise).
I also feel like a paywall ninja for clicking through to read a correction to the JAMA commentary, which contains this key paragraph:
“Nevertheless, with more than half a billion primary care visits annually in the United States, if these data from Singh et al are generalizable, at least 500 000 missed diagnostic opportunities occur each year at US primary care visits, most resulting in considerable harm. If error rates at the half-billion non–primary care visits are similar, the total could be more than 1 million. If even 10% of these (100 000) are easily prevented, and combining this figure with autopsy-based estimates of preventable US hospital deaths from diagnostic errors (40 000/y to 80 000/y), more than 150 000 patients per year in the United States might have undergone preventable misdiagnosis-related harm.”
That is pretty close to the 160,000 patients per year estimate cited in today’s study.
Finally, I have to point out another favorite quote from Landro’s column since it captures why I love research:
“There isn’t going to be one magic bullet to solve the problem of diagnostic error, but we can start by consistently monitoring and measuring it,” says David Newman-Toker, lead author of the study and an associate professor of neurology at Hopkins.
Indeed. Swap out “problem of diagnostic error” and swap in “question of participatory medicine” and you have my raison d’être. Understanding how people gather, share, and create information is an essential ingredient to understanding the future of medicine.
I can’t help but reflect – and project – on prospects for IBM Watson to assist in this area … and I suspect I’m not the only one thinking along these lines 🙂
Watson: A Love Story
Ha! And yes. High hopes for us to figure this out — and quickly.
I should mention that anyone interested in pursuing this topic further should check out the Society to Improve Diagnosis. Their website isn’t fancy but the board and their annual conference are stellar.
Via Twitter, Urmimala Sarkar, MD, shared another study:
“Our survey shows that MDs feel there is insufficient thinking time for accurate diagnosis…” with a link to Challenges of making a diagnosis in the outpatient setting: a multi-site survey of primary care physicians (BMJ Quality & Safety, May 2012)
Here are the results, from the abstract:
“Of 1817 physicians surveyed, 1054 (58%) responded; 848 (80%) respondents primarily practiced in outpatient settings and had an assigned patient panel (inclusion sample). Inadequate knowledge (19.9%) was the most commonly reported cognitive factor. Half reported >5% of their patients were difficult to diagnose; more experienced physicians reported less diagnostic difficulty. In adjusted analyses, problems with information processing (information availability and time to review it) and the referral process were associated with greater diagnostic difficulty. Strategies for improvement most commonly involved workload issues (panel size, non-visit tasks).”
I understand that these studies are meant primarily for a professional audience, but what can non-professionals do to help? Did any primary care physicians praise (or complain about) patients who brought in their own research or self-tracking records? What can people do to stack the deck in their favor?
Susannah,
Hi. Nice post. While I haven’t read Urmimala’s article yet, one of the oft overlooked contributing factors to misdiagnosis and medical error (I guess they are the same thing) is poor communication on the physician’s part.
Here’s what I mean.
It’s been said for years that a physician’s communication skills are as important as their clinical expertise. That’s because the accuracy of the diagnosis and treatment is dependent upon the physician’s ability to 1) elicit the patient’s story (which is where the patient talks about “what’s wrong” and 2) the physician’s ability to listen to the patient’s story fully without interruption. As Sir William Osler, the Father of Modern Medicine once said: “If you listen to the patient they will tell you what’s wrong.”
Today physicians interrupt patients within the “first 18 seconds” of telling their story…without usually coming back to their story…or capturing vital contextual data possibly related to a diagnosis. Listen to any physician and they bemoan the fact that they “don’t have time to talk to patients” and “don’t get paid to talk to patients.” Given that “talking” is essential to diagnostic accuracy…and given this “anti-talk” attitude on the part of some clinicians…it wouldn’t surprise me if the misdiagnosis rate was much higher than the estimates suggest.
Steve Wilkins
“Which led me to tweet:
Is it any wonder that 35% of U.S. adults go online to get a jump on a diagnosis?”
Although I’m 100% in favor of any actions patients take to better inform and empower themselves, “jump on a diagnosis” is a bit of a two-edged sword.
For my part, I try my best to know very little about a patient before I go in the room (whaaaat??)…in the EHR days, this is almost impossible. But it’s true…because what you know is not always knowledge…sometimes it’s just bias. Something as simple as seeing that someone has COPD in their past medical history has already biased you towards assuming their shortness of breath is related to COPD; a consultation note re: the complaint from a colleague cannot help but bias the reader into coming to the same conclusion.
Every day everyone has a litany of symptoms: a little dizziness on standing, a pain here and there, a cough, a sneeze, chills, a headache, etc. Coming into an office with a preconceived notion (whether consciously or not) of what someone has causes you to unconsciously trump up the symptoms that confirm what you believe and minimize those that go against it.
Patient who go online are in the same boat. It’s very easy to put a bunch of symptoms into Google and find something…if the symptoms of that something are quite subjective (i.e. they’re not “you’re pinky finger over the course of 20-30 minutes turns indigo”), then it’s very easy for patients to end up in the same confirmation bias situation…leading to a greater emphasis on those symptoms that confirm the diagnosis and less emphasis on those that go against it. I’ve even seen people give a nearly word by word textbook description in the office of a symptom in a certain condition…this is not “lying”…rather it’s the phenomenon of someone reading something and going “That’s IT!!! That’s Me!!!”
The challenge is that this changes, inevitably and subconsciously, how they then relay their symptoms to their care provider.
This is something all diagnosticians, including the patients themselves who are engaging in self-diagnosis, need to actively be aware of and work against all the time.
I work against it by trying to get two elements of the story out before any “context”: a) in your own words, what is the issue? b) describe the issue in as much detail as you can re: onset, duration, severity, things that make it worse, things that make it better, things that seem to always happen at the same time, etc.
I then take the other info: past medical history, medications, allergies, old consult notes and look at them through the lens of why the person is here (rather than the other way around). I inevitably get the odd “why do I have to go through this again?? don’t you have my old notes??”…to which I always say “I want to hear what you have to say first…then I’ll look at the old stuff.”.
Is this less “efficient”? Perhaps, if you gauge efficiency purely by the time it takes to get to “an” answer. It may be more efficient, however, if the extra time then leads to the “best” answer sooner.
Thanks so much for making the jump from Twitter!
I’d love to unpack this a bit and hear from more people about their experiences, as patients and as clinicians. I suspect there is a big difference between people who are living with rare conditions, for example. I also wonder what the relative advantages we have if we share vocabulary in common with our clinicians. In fieldwork I’ve learned how people go online to find out what words they can use to describe their (or their child’s) symptoms so that a health professional can understand.
When it’s posted, Sally Okun’s TEDMED talk will be a must-watch because she describes how PatientsLikeMe tackles this issue, mapping how people describe their symptoms (in their own words) so they can be categorized and tracked (in clinical terms).
Like magic: Sally Okun tweeted me a link to her excellent presentation slides from Medicine 2.0 Boston. Highly recommended:
http://www.slideshare.net/sallyokun/medicine-20-okun
Awesome discussion and insights. As a member of the “difficult to diagnose >5%” it’s pretty easy to observe that doctors are incentivized to “diagnose, treat, bill”. The difficult cases gum up the works and don’t let them get past step one, leading to coping mechanisms like “You look healthy! Bye!” For normal problems, doctors are great!
But if the major issues are “clear defects in bedside history taking, physical examination, and test ordering” then I would restate that first item as “scarcity of human intelligence”. Medicine is messy and human intelligence is especially adapted to cope with messy. Even though doctor intelligence is scarce, human intelligence is not (among patient peers). And by human intelligence, I mean someone else really taking the time to get to know you, understand your personal case, and then offer an informed opinion.
Patient communities should be great for this, but the main scarcity there is not intelligence, but trust. If you join a community with avatars and psuedonyms, it takes other users weeks to figure out if they should invest the energy to care about you (i.e. are you legit or just another sign on?).
How do you unleash patient peer human intelligence?
1) Make it easy to get to know people (30-second condition overview page)
2) Quickly establish trust (like we have for millennia with names and faces)
Sorry, you asked for a patient perspective. 😀
Thank you! Sensible, smart, and (this is key) actionable.
You cut to the very core of it! Actionable. (Speaking about scarcity in health care, “actionable change” would be at the top of the list…). Crohnology is a really good example of this. I think Sean Ahrens understands the importance of establishing trust in a patient network where you are the “real you.” Hopefully we’ll see more of this to come.
More fuel for the fire: a new article on misdiagnosis in the Washington Post today:
Misdiagnosis is more common than drug errors or wrong-site surgery by Sandra G. Boodman and Kaiser Health News
Very interesting. Thanks for sharing.
160,000 injuries or deaths per year is very scary. It reinforces what Best Doctors talks about. As I tweeted you earlier, I wonder why more companies don’t mandate a secondary review of complex cases for their members. I would think this was in the best interest of the member, the employer (payer), and the physician.
For example, we work with some clients in oncology to require a second opinion after someone is diagnosed. They can’t begin therapy without having their physician talk to one of our oncologists. They provide a review of the diagnosis and staging of the disease. They also compare their care plan to NCCN guidelines (evidence-based care). The oncology nurses are then able to support the provider and the patient with case management support and ongoing advocacy. This should minimize much of this when fully integrated with medical and pharmacy claims along with a copy of all the genetic markers and other labs that are done.
Eventually, I think this type of personalized medicine and safety net process is important in “partnering” with the provider since without technology and second opinions it’s often hard for evidence-based care to be adopted without the traditional 16 year lag. Hopefully, EMRs and their use with smart algorithms will change this over time looking for care gaps and other risks, but until then, there is a great need for a virtual care team and their collaboration with the front-line provider and the patient.
> I wonder why more companies don’t mandate a secondary review
> of complex cases for their members
Do you mean employers, or insurance companies?
I’ve been profoundly amazed by the utter lack of interest I’ve often seen among insurance companies, regarding any idea that could lower costs. It’s a mystery, with only a few proposed explanations, none of them encouraging to patient interests.
Of course not all insurance companies are purely evil in this way, but if you think about every insurance pricing scandal in recent years, they’re all accompanied by “Well gosh, costs keep going up – what do you expect?” I hate to think that executives and boards at those companies might actually welcome errors and avoidable costs but more than a few people I’ve met at conferences have proposed that – because, they say, health insurance is pretty much a commission business: when costs go up, total margin does too. GOOD for business, perversely.
Either. I think it’s an easier sell to employers, but I think some insurers are starting to see the value here.
The challenge is that with High Deductible Plans that there is a push of responsibility to the consumer without the appropriate safety nets (IMHO). I think the Stop Loss Carriers get it and may eventually put some pressure on the self-funded groups to help the consumer be more responsible with their money by overseeing how it’s spent, but today that’s generally not happening.
The problem is also that a lot of people don’t understand how to blend cost and care in an integrated solution. Ultimately, it’s about the Triple Aim – cost, care, experience. That’s not easy. I have hope in a few things that I’m working on, but there’s a long lift to make that the norm.
Riddle me this everyone…
Another apparent yet woefully unreported indicator of scope of the “medical error” and misdiagnosis problem came to my attention recently when I learned the following:
Specialty centers like MD Anderson Cancer Center (where my wife is currently being managed for a recurrence of her Stage IV lung cancer) and the University of Michigan do not accept radiology results (meaning digital renditions of MRIs and CT and PET/CT exams) done at other institutions. Why? Not because they want to charge for repeating the procedure…but because the quality of the digital scans are often so bad that they can not be reliably interpreted.
Now think about this. If these scans being done in local community hospitals are of such questionable quality that “experts” can’t read them..what do you imagine the implications are for the quality of the interpretation being done by local radiologists and interventionists?? I am appalled that no one has looked into this area …at least in the published literature. On a related note, I was once told by a researcher at MD Anderson that 50% of the pathology samples (for locally diagnosed leukemia cases) reviewed at MD Anderson on a second-opinion basis were ms-diagnosed at the local community path labs/doctors.
The problem is huge…and because doctor and hospitals don’t “call one another out” on such errors (professional courtesy) the problem continues.
Steve Wilkins
Steve, THANK YOU for that. I’ve wondered because when I got my annual CT scan last year, in my own OsiriX viewer it looked to be noticeably lower resolution than the ones I’d gotten at Beth Israel Deaconess. But nobody at BID said anything. I guess I should ask them!
I would LOVE to see a good & bad comparison for a particular image so the public can see the difference. We’re all savvy enough to know when a photo is low res, and I’m assuming(?) there would be visible differences in an image of a chest lesion.
Do you think someone at MD Anderson would be willing to help do that? Like, read one of the ones they reject, and point out (with screen captures) what’s sloppy about the rejected one and clear on the in-house one?
Fascinating. This is great! I just met with a French company that’s trying to come to the US to solve this problem (which I didn’t realize existed until Monday). They stated that it’s very difficult to read scans from different machines. They’ve built a universal reader that works with all the big players across multiple types of scans to allow for a longitudinal online record and analysis of MRI, CT, PET, etc.
This would allow scans to be used and tracked by the providers across sites and technology platforms. They’re at 700 sites in 25 countries and being tried at a few places like MD Anderson so maybe there is hope!
Awesome comment Dave! I’m concerned also that with changes coming from the Affordable Care Act, insurers will by law have to spend a fixed amount of the premiums they receive on reimbursements. The only way to grow their profit will be to — as you said — increase their costs or premiums. I don’t have a low opinion of insurance companies, they just want to maximize their control over value like most of the rest of us, IMHO. It is a shame though, that so much value has been taken away from patients and is instead controlled by employers, insurance companies, doctors, and the government. Consumer control over value is more protected in the auto insurance market, and as a consequence those companies are far more beholden to the little guys.