Diagnostic & Treatment Errors: The Banana in the Guacamole

We just finished celebrating the 50th anniversary of the U.S. landing on the moon, only to have the story marred by the news that astronaut Neil Armstrong, the first man to walk on the moon, likely died from what at least looks to be major medical errors in his treatment.

What a tragic end to an American hero. We hear about these stories from time to time, mostly when famous people are involved. But the truth is that medical errors are still so rampant in our medical industry that it is incredibly upsetting. Every time I have to set foot in the health system, I worry. I have witnessed, in my own circle of friends and family, numerous medical errors in diagnosis, treatment and the management of bad side effects that it sometimes feels like a helpless mission to try to fix it. Hopefully that’s not the case, but it is definitely a question much on my mind.

Is it possible for mere mortals to turn this around? Why, so many years after the 1999 Institute of Medicine report “To Err is Human” pointed this problem out for all to see, is it still so rampant?  Here we are 20 years later, and the headlines just don’t change. It is downright depressing (but please don’t try to diagnose my depression – I don’t want you to get it wrong!).

In addition to the Neil Armstrong story, there were a few others I’ve seen in the last few weeks that don’t make me feel so good. Specifically, I just saw this article from 2016-17 that describes how showing the same MRI imaging study to 10 different radiologists resulted in massive dissonance and large numbers of diagnostic errors. The study’s conclusion was this:

This study found marked variability in the reported interpretive findings and a high prevalence of interpretive errors in radiologists’ reports of an MRI examination of the lumbar spine performed on the same patient at 10 different MRI centers over a short time period. As a result, the authors conclude that where a patient obtains his or her MRI examination and which radiologist interprets the examination may have a direct impact on radiological diagnosis, subsequent choice of treatment, and clinical outcome.

So, basically, we are talking some combination of machine variance (some MRI machines are more advanced than others) and pilot error. I don’t know what happened to the patient, but I sure hope she picked the right interventionalist and not just the one with the nicest waiting room.

Reading this gave me a vision of walking into a grocery store and asking 10 different people to point you to the avocados, only to find that there are 10 different things people think are avocados…and some of those things are actually beef jerky and bananas. Jeez.

The other article, sent to me by my husband who clearly thinks I am getting too much sleep at night, was THIS ONE, which focuses on a Johns Hopkins study finding that diagnostic errors account for 1/3 of malpractice claims that result in serious harm or death:

I mean, it’s logical. Malpractice claims are likely to be the result of errors. But given that many cases go wrong and don’t result in malpractice claims, I can only imagine the real number. The article says that diagnostic errors are “among of the most common mistakes in medicine.” Swell. If you diagnose the problem wrong, one would think that your odds of treating it correctly probably begin to approach zero. As previously noted, if I start my guacamole recipe with a banana, I am going to have some very disappointed amigos.

But hey, as they say, it gets worse. There is an overwhelming array of evidence that, faced with a correct diagnosis, physicians also struggle to pick the right treatment anyway. This article points out that physicians presented with multiple treatment alternatives did not much better than laypeople in selecting high value vs. low value treatments. Now granted, this ARTICLE is more about whether high deductible plans and value-based pricing drive better treatment behavior for consumers, but it was rather surprising to me that when physicians as consumers were the test population, they often failed to select the most cost-effective, high value options. That does not bode well for what they will recommend to their patients.

There are many young companies working hard to try to stem the tide of diagnostic errors, treatment errors and related events (two immediately on my mind are Health Reveal and Covera, though there are many others). I have invested in several of them and served as an advisor to others (disclosure: I’m on the board of Health Reveal). I continue to believe, despite what continues to be my realized experience, that this quest to reduce medical errors will matter to people who matter. And it does matter to some of those people. Employers seem to really get it, wanting to invest in products, technologies and services that will improve the diagnosis, treatment and monitoring of their most valued asset: employees. Other payers sometimes get it, but don’t often like to get too involved in the path of clinical care even when they know there are issues. Some provider organizations get it, but all too often provider organizations are more focused on adopting products and services that increase revenue rather than decrease error – it’s a lot easier to spend money for growth, I guess, than to spend money to make existing services more profitable, less expensive, better overall.

I have walked away from so many board meetings, pitch meetings and other similar interactions with emerging companies shaking my head because what I have learned is this: entrepreneurs struggle to get buyers excited enough to spend money to materially improve diagnosis and treatment, despite those buyers knowing full well there is ample room for material improvement. There are a myriad of companies seeking to use artificial intelligence, machine learning, and other technology techniques to identify and reduce error, either prospectively or retrospectively, and it is often the same old story – customers love the idea but would also like to see that thing behind the curtain which increases revenue instead – maybe if that works out they will hear you out on that error reduction thingy later on.

Sometimes increasing the number of procedures is absolutely right. We should give heart valve replacements to those who need new valves but haven’t been properly diagnosed as needing them. We should perform breast biopsies on those who have had mammograms where there is a questionable lesion, particularly when it was missed the last time. Hell, I’ll even go out on the politically incorrect limb of saying that we should give more expensive pharmaceutical interventions to people when they are more effective and have an overall better impact on both outcome and cost over the longer haul.

But this isn’t the problem I’m worried about. I’m worried about the lack of accuracy and the inability to train physicians in a way that changes this dynamic – what are medical schools doing to change this sad set of headlines? I’m worried about the difficulty of injecting technologies into clinical organizations so they can deliver better diagnoses and treatments because the IT department is too busy futzing with their EMR or whatever. I cannot tell you how often I hear that story. It often amounts to this: I can’t get around to making the guacamole right because, well, you know, the chips aren’t here yet. Um, What?!? I am worried about those physicians who resist tools to make them more accurate because, hey, they are trained professionals who know stuff; I don’t hear a lot about accountants refusing to adopt new financial software because, well, they can count just fine on their fingers. In virtually every other professional field, tools that make you better, faster, and more accurate are welcomed. In medicine, they are often considered inconveniences because you might have to change the workflow. When my grandmother died due to a medical error, I was not super interested in hearing about the physician’s workflow.

Seriously, how do we let this continue? CMS spends an extreme amount of time trying to stop fraud (which they should be doing – I’m not arguing this is a bad idea), but at least the same amount of money is going out the door due to medical error and the energy about that is not the same. There is really no point in measuring cost-reduction if it doesn’t come with attendant quality improvement. It’s easy to reduce costs – don’t pay for things. We have a whole industry built on that model. What we need is a will to not pay for things that are bad – if we got home and realized our avocados were bananas, we would go back to the store and get our money back.

I am cranky about this and just want my damn banana-free guacamole. And I don’t want to hear from all the entrepreneurs who are working on this problem – I actually know you are out there, and I applaud your mission, your passion and your blind willingness to keep trying in the face of much resistance. I want to hear from the big provider organizations who are 100% committed and actually acting to address physician and nurse error, even if it costs some money. I want to hear from payers who have made a serious commitment to install technologies that identify when patients are getting the wrong care. I want to hear from those creating new curricula at medical and nursing schools about how they are going to turn out better diagnosticians or at least great clinicians who agree that validated tools that actually enable them to perform better are welcomed. Let’s get some positive headlines flowing by people who are actually committed to fixing this problem (actually, not theoretically) to show the rest of the industry how it’s done.

The solutions are, at least in part, out there. But they only work when we admit we have a problem. And the problem isn’t fixed by artificial intelligence until the inputs to the AI machine are accurate, folks.  Buyer beware.

Afterward: After I wrote this piece above (written at 30,000 feet somewhere over the midwest), two things happened on the same (next) day:

1) I read a very compelling article in CNN about Doug Lindsay, a renown scientist who, before he was a scientist, had a rare, undiagnosable disease that rendered him bedridden for 11 years; the frustration and inability to find relief led him to commit his life to figuring out the medical mystery AND solving it by inventing the surgical cure; and

2) I went to see the new Tarantino movie Once Upon a Time in Hollywood (two thumbs up). But what was far more interesting was a pre-movie trailer for an upcoming TNT show, hosted by Ann Curry, called Chasing the Cure. The show is literally a medical diagnosis crowdsourcing show where patients can submit their incorrectly diagnosed or undiagnosed medical problems and have the crowd, both medical and laypeople, seek to identify the problem in hopes of finding the treatment that will change their lives. Wow! Talk about medicine hitting the mainstream! There have been attempts at this online and in other forums, but nothing as broad-based as a prime time TV show with a recognizable host and serious publicity attached.

The show is set up almost like one of those cold case thriller shows where patient can submit their “case file” and seek the wisdom and input of the public and a chosen panel of medical experts. Obviously, it wouldn’t be good TV if they didn’t find the correct diagnosis and cure for some of these people, so I imagine we will see some of those on every show. It would be a pretty short-lived series, I suspect, if every episode ended in, “Well no idea, sorry to say, sucks to be you.”

I was thinking about these two occurrences (Lindsay article and Chasing the Cure promo) in the context of the earlier content of this blog. And it made me wonder (cue Led Zeppelin…) Why all of a sudden is pop culture taking on this issue of our medical diagnosis challenge? Is the need for this TV show further demonstration of the problem of the poor diagnostic capabilities of our medical system or is it more demonstrable of the inconsistent access to good care? Are these instances evidence that peer-to-peer healthcare (a la Susannah Fox’s assertion) is essential to best outcomes or is it evidence that science is still a long way from nirvana? I’m not sure, but what I know is that that hundreds of people have already uploaded their case files to the TV show’s website, so it definitely means something.