14: Cognitive Autopsy of a groin lump

One, two, three, testing, testing, cocaine flame in my bloodstream. Oh, I love that song. Welcome to I Am Reasoning, a podcast about diagnosis and clinical reasoning, and other important topics that affect physicians and trainees everywhere. Welcome back to I Am Reasoning. This is episode 14, wow. And we're going to try something a little bit different today, maybe just as an experiment if it works, maybe we'll do it again. And we're going to do something called what we're calling a cognitive autopsie. And so we're going to take a case and we'll explain where we got the case in a second. And we're going to go through it in some degree of detail and analyze it, I guess, what went wrong? It's a case of a diagnostic error with a, unfortunately, a poor outcome. There's a few things we probably have to say before we begin. And the first thing probably is that this is a case that happened some time ago at a New Zealand hospital. And whenever we're looking at something retrospectively, there's a certain bias that you bring to the analysis of that case. And we also have a tendency, I think, at times to perhaps look at a bad outcome and say, oh, gosh, I wouldn't have done that. I'm a better doctor than that. And I would just ask listeners to sort of suspend that thinking and that feeling for the time being. And put yourself in the position of some of the clinicians involved in this case, because I think, although, hopefully, not many of us have had this kind of outcome, we certainly have been at risk for this kind of outcome, I think, because of systems errors and cognitive errors. Lucky. Much of the time. Well, absolutely. Stumble upon diagnoses, and in the meantime, nothing bad happens. Yeah. And I think this is a case, as we unfold it, where that Swiss cheese model of a bad outcome is really highlighted where all those holes line up, and there's no safety net that catches the system before it collapses. Right. In retrospect, we get to see the whole block of cheese, and it's much easier to pick the points where things went wrong. So the idea of this cognitive autopsy is not to cast aspersions, not to pose as infallible, but to actually learn from actual cases that have gone before. Okay. So first of all, I'll tell you a little bit about where this case came from, where we got it from. So in New Zealand, we've got this thing called the Health and Disability Commission, and we're called the HDC. And at its head is the commissioner, and this is like the highest arbiter of claims around the quality or appropriateness of care provided by anyone in the healthcare sector. So you'll all be familiar with your hospital's own patient relations department here. It's called Consumer Liaison. And basically, they take all the complaints and compliments from patients and their families, and they respond to those complaints, and they collect information from all the clinicians and nurses and everyone else who was involved to formulate these responses. And it's never meant to be a defensive interchange. It's meant to be a sincere and an honest effort by the hospital to be accountable to their patients and the care that they provide to their patients. But you can imagine that patients are not happy at the end of this every single time, sometimes they're left frustrated. And certainly in our hospital, we then, when it's beyond what we're able to do for them in terms of satisfaction, we recommend that they go to the HDC and put a claim through there, and then it can be sorted out further. Now patients can go to the HDC on their own, they don't have to go through the hospital process, and there's no restrictions on that at all. Either way, when the claim gets launched at the HDC, the HDC investigates the complaint in much the same way that the hospital would first by gathering all the information and responses from all the people that were involved in the care, it also gets expert advisors. And I've had the opportunity to be an expert advisor on several of these cases that the HDC is investigating from around the country. So as an advisor, my task is to give an opinion about the care provided by my counterparts. In other words, the physicians or the doctors involved in that case at that other hospital. And I have to say whether I felt it was a reasonable standard that was maintained. And in the same way, the HDC would solicit the advice of other experts such as nursing experts or radiology experts or whatever the case calls for. And so at the end of this very serious process, the HDC publishes a report, and this report is published on their website. It's available to the public, obviously, and it gives its recommendations on this report. And the ultimate goal, of course, is to inform and educate the public and particularly the hospitals and the healthcare providers so that we can all avoid making some of the same mistakes or errors in the future. We're really quite fortunate in New Zealand that the first person to whom someone runs is not a lawyer that compensation, financial or otherwise, is handled separately by a no-fault accident compensation commission. So we'll start with the executive summary of this case and then we'll sort of delve a little bit more into each day or each significant day during the course of the hospitalization talk a little bit about what we think are some of the cognitive errors in particular that have led to a poor outcome. And systems errors probably emphasize a little bit less only in so far as they may have affected the cognitive issues at stake. Okay, so as you suggested, Art, first thing to do is to go through this executive summary so that everyone's on the same page. So the bottom line for this case is that this patient will call her Mrs. A as they did in the report. She's 78 years old. She came into that she was referred into the hospital for a possible hernia and she ended up dying and the final diagnosis wasn't incarcerated hernia but this was never actually discovered through most of her admission. So just to give you the brief, she's 78 years old. She's usually, she was usually active, independent and generally healthy. And as I said, she was referred into the hospital by her GP with the following concerns, vomiting, dehydration, actually the new onset of an irregular pulse and a lump in her groin. And the GP had been worried about an incarcerated hernia and wrote that in some of her consultation notes prior to the referral. So she's seen initially by the medical registrar resident and also by the medical consultant on the same day. She's provisionally diagnosed with a "abdominal malignancy" unquote or lymphoma. And she's admitted for fluids with an eventual plan to discharge her once the nausea has settled. Now, she does not improve in her symptoms, she continues to vomit. She starts having significant respiratory symptoms. She, at that point, also gets misdiagnosed again as a communicard pneumonia when, in fact, she was probably aspirating. And finally, she has this catastrophic vomit and aspiration that results in hypoxic respiratory failure, cardiac arrest. And she dies from basically not waking up from a noxic brain injury. And in the interim, of course, she is, she does have a CT scan done. She's taken to the operating theater where they diagnosed this incarcerated hernia and deal with it. But by that point, the damage had been done. OK, so I think what we'll do is we'll break up the case into phases and we'll just analyze each phase in turn rather than go through the entire case in detail from the beginning. So I'll just start summarizing the first phase, which we'll call the GP referral. And there's not much to say here in terms of detail, except that, you know, as we mentioned the GP referred this woman into the hospital. And the referral letter said, quote, thank you for seeing Mrs. A. regarding the following problems. In brackets, she wrote difficult historian, close brackets. And then she had three points dehydration with ongoing vomiting/coff. A second point was new atrial fibrillation. And the third point was groin lump. So two points I'd like to raise just even at this very early stage in the case, because I think the seeds of what ultimately goes wrong with this patient are sewn here. And the first is the fact that in the referral letter to the hospital, there was really actually no differential diagnosis raised. It was merely a list of the problems for which the patient was being referred. Now, interestingly, during the course of the preceding couple of days when Mrs. A was being evaluated by the GP, the GP did in fact raise the question of whether or not this new groin lump was a hernia and had organized an outpatient ultrasound of that area because the GP was concerned about a hernia, but didn't actually relay that information in the form of sort of a differential diagnosis for this groin lump. Although she did include her consultation notes from the previous visits and they clearly documented that she had had the thought of hernia. And to be fair, later on in the case, it becomes an issue of who has time to read what. And I think in a busy hospital practice where you're seeing lots of patients in ED, I think most of what's important needs to be encapsulated in that referral letter because myself included most people that I know of don't go through days of preceding notes for a GP visit. No, not routinely, yeah, and it's exactly the issue of framing that we talk about. On the one hand, framing can be a minus. But on the other hand, if you're concerned about specific issues or specific diagnoses, presumably you would include those concerns in the one piece of communication you're sending to your colleagues. The second thing I'd like to point out is this bracketed comment on the patient that's included in the GP referral note and that is calling the patient a difficult historian. And I have trouble with these kinds of terms that we use, either a vague historian or difficult or poor historian. And I'm not entirely sure why that was included in this referral letter, but what it tends to do, it's almost like a fundamental attribution bias in some ways where you're almost blaming the patient for not being able to get a good history. And I think it does two things, A, I think it labels the patient, as I just mentioned. But the second thing it does is that it gives everyone who comes after you and out for not getting a good history because, oh, this patient's just a difficult historian and I'm not going to waste too much time trying to get a history from her because she can't describe her symptoms. And I think it's a bit of a cop out. And it's interesting that it's even termed in that way. I'm sure other people have heard of this concept as well, but if you think about it, the patient is not the historian. A historian is the person who does the research and puts together the history. That's how it works for history books. And so, you know, the patient is just a subject and your job because you're the professional is to elicit the information from that subject and you put the history together. So in fact, you're the historian. Maybe what she meant is that she's a difficult historian. And I was frequently told as a junior house officer back in my training that this was by a particular mentor of mine who said, "There are no difficult historians. There's only bad history takers." And we all know that that's not completely true. I mean, there are people who have difficulty describing their symptoms. They don't quite remember. When you ask them about A, they tell you about B. But I think when we label a patient as a difficult historian, we are just sowing the seeds for everybody who comes afterward to just not pursue the history enough. And I think that, as we'll see, is this case unfolds. I think that's certainly part of the problem. I always remember the quintessential vague answer to a question, I remember one of my professors once brought this up. It's like when you ask a patient, so when did you start having chest pain or shortness of breath or whatever it was? And their answer is something like ever since my mother died. So it's not that you're not going to get that information eventually, but it's just one extra question and your eyes start rolling already when you've got to ask that one extra question. Right. Right. This phase of this is when the patient actually rolls up to the emergency department department with this letter. And she is evaluated by a junior registrar. Now in the initial assessment in the emergency department, the triage person, records on the form that the patient's symptoms include a "new lump" in the left inguinal region. And for whatever reason Mrs. A, the patient turned up to ED by herself, so there was no one else to give corroborative history at that point in time. But the patient was assessed by a junior registrar who completed a standard sort of admission documentation. And under the heading titled "Reason for Presenting to Hospital," he wrote 78-year-old woman who is a referral from the community GP for vomiting and new onset of atrial fibrillation. So this whole issue of a groin lump kind of falls off the radar at this point in time. It will ultimately come back on, but at this point in time in these very formative moments of trying to come up with an assessment, the groin lump is sort of nowhere to be found. The registrar notes that she'd had some prior history of cough with a bit of sputum, some decreased appetite, and this one to two week history of vomiting with at that point in time normal bowel habits according to the history that was obtained then. Under the heading that says "Presenting Problems, Symptoms, Duration," et cetera, it says that Mrs. A, the patient had mentioned some lumps on her neck and left groin, as well as some recent but not significant weight loss. In his examination, the registrar did note some lumps in the right posterior neck and in the left groin, which was described at that point in time as four centimeters in diameter, which were, that was the groin lump, which were mobile and non-tender. The abdominal exam at that point in time was unremarkable. The registrar also noted that the patient has already got an appointment for a lymph node biopsy in the community, which in fact was not correct. She had an ultrasound that was ultimately scheduled, but not a lymph node biopsy. A chest X-ray was done as part of the assessment, and it was noted, I believe, by the radiologist in the official report that there were some short air fluid levels noted in the upper abdomen, but nowhere as far as I can tell from the documentation was that ever. You know, the chest X-ray was not documented in that initial admission note by the registrar. Right. So I'm not sure whether it wasn't looked at or whether perhaps they were focusing on the lungs because of the cough, and so didn't actually note what was happening in the upper abdomen, but clearly there were some signs there that could have been used to ultimately make the right diagnosis. And so in the final problem list, the registrar notes ongoing vomiting, mild dehydration, new AF, the lumps reappear, but now that left groin lump is lumped together, no pun intended, with the lumps behind the neck, and there's no differential included there. It's just query cancer. So query is a very common notation used. Yeah, it's just question mark cancer. And so the patient because she had been vomiting during that assessment was given some IV metaclopromide, and that was sort of where we want to bring this initial assessment to a close and comment on it. Yeah, well, there's quite a bit to comment on there. I guess the first point that I want to make you mentioned that the husband was not able to come with her for that initial assessment. It wasn't in the room when the registrar assessor. And so there wasn't this possibility for corroborating history. That happens all the time. I don't think that in itself is a huge problem unless, of course, the patient is unable to give a history because of delirium or dementia or some other communication issue. Which I don't think was the case here, but having said that, I don't know if you experienced this regularly, but I find that my accountability level goes up when there's family in the room. And it's not obviously not an intentional thing. I want to get the right history and I want, but I feel like there's a lot more fudging that can happen, especially if the history is difficult to take than if when there's a family in the room where you have a monitor there who is going to make sure that the history is actually right. So that's point number one and I don't know how much that affected. I guess the second point is the framing. You know, again, framing comes back. You make your own frame before you even start taking history and I think this may have happened here. You mentioned that the registrar initially recorded or wrote the problem that he was dealing with as 78-year-old female GP referral for vomiting and nuance at atrial fibrillation. Now that's not exactly a frame that elicits any intuitive singular diagnosis for me, but it also doesn't capture everything that was happening at that point. And so possibly the framing there, if that's what he was concentrating on, is sending him down a different rabbit hole. I can't help but say that it was a problem with history taking. I'm not sure what bias that is, but there was a lot of history taking, guessing, or you mentioned this unpacking principle. You mentioned this. When you hear pain, it comes with a package and you unpack the pain and you get all the classic features of that pain, same thing with vomiting, for example. This didn't seem to happen here, there wasn't much preoccupation with all the features of the case and they didn't unpack all the symptoms. No, and again, we're looking at this in retrospect and it's difficult to know exactly what questions were asked and when, but there doesn't seem to be a lot that had been recorded around the history of the vomiting. And if you don't get a sufficient history about a particular symptom, it's very difficult to come up with a differential diagnosis. You don't know whether it's acute, whether it's chronic, when it happens, is it post-prandial, is it not post-prandial. I think here, really, certainly there was an unpacking, a lack of unpacking here in terms of history which I think perhaps went astray. And again, I think it may get back to the fact that she's a poor historian, so what am I going to bother asking these questions for? She's probably not going to be able to give me an accurate answer. And that's fine potentially as a practical strategy, if there truly is a difficulty getting the history, as long as you're going to address the things that you see. But there's more. There's more here and more that they didn't address. You mentioned the assertion that there's a biopsy arranged for later in the week, which was obviously not the case. It wasn't a biopsy. It was an ultrasound. That was an error in history that was obtained. But the phenomenon here, which I think you were getting at, is the diffusion of responsibility. Absolutely. And we've been in the situation before. If we've got five different issues that we're dealing with, we're going to concentrate on the urgent one. And if there's something that's not urgent that maybe isn't quite an inpatient worthy issue, and someone else is already looking into it and has made arrangements for investigations, et cetera, there is a sense of relief almost like, okay, fine, that's one less thing I have to worry about. Someone else is already looking at that. And I think there was an element of that happening here. Yeah, I completely agree, completely agree. And the next for me anyway, the next big cognitive error I think was this automatic assumption, this patient, I'm not sure if we'd previously mentioned this, but this woman had had some epidermal cysts in her neck. They had previously biopsied, although I'm not sure the registrar knew at that point in time that they'd previously been looked at and were benign cysts. But she did have a history of necklumps. And so the cognitive error was sort of putting all this together in relation to this groin lump. So that necklumps, if there's lumps in the neck, they must be the same as the lumps in the groin. But in fact, the lump in the groin was significantly larger than all of the other lumps. And new. And short, relatively new. So it seems a cognitive error to just sort of come to that premature closure that this groin lump is the same as the necklumps. And you could also say that that's perhaps availability bias. But she had lumps in the neck, so why not just put them all together. It's Occam's razor trying to make one diagnosis out of things that may not necessarily be one diagnosis. And so there's lots of things here where I think there was a cognitive bias at play. And you can take your pick, which one was the most important one. But the ultimate result is that now this groin lump is linked with lumps elsewhere. Every cancer. Yeah. And, yeah, the hiccums dictum would have been a more appropriate bias to employ in this particular situation. It's hard to tell, you know, at the beginning what you're supposed to do better not to have any biases at all. The hiccums dictum being the idea that a patient can have as many diseases as they damn well pleased. I think that's how it was stated in natural law that says they all have to be simplified into one. Now, you mentioned, I don't know if you mentioned posterior probability error just a second ago, but that also came up for me because let's face it, on the medical service, an incarcerated hernia is not usually one of those diagnoses that comes onto our turf. So it's not a diagnosis that we think much about. So if we're going to think about lumps here and there, wherever, we're more likely to be thinking of something like lymphoma, which is something that comes up on our radar a lot more commonly than an incarcerated hernia. So there may have been an element of being biased just by what service they've been referred to and what you normally see. And it's a kind of availability bias as well, isn't it? You've seen lymphomas before. You may not have seen incarcerated hernias before and so immediately you're thinking based upon what you've previously seen, that it's that, when it may not be. Yep. Okay. So let's move on to the next phase of the admission, which was the consultant reviewing Mrs. A. And just to be clear, he did this on the afternoon of the same day of admission. I think Mrs. A was reviewed by the register around 2 p.m. And then the consultant came around and reviewed Mrs. A early on that afternoon, sometime between 4 and 6. And the reason, because this is a little bit uncommon, usually what the consultant does is, you know, they're on for a period of time in terms of the patients that they take. And then the following morning is when they do their post take round with the register. It wasn't done in this case that way because the consultant wasn't going to be around the following day and so he was sort of using this time in the afternoon as a substitute to the following day's round. It's important to note that he had to see, I think that afternoon he saw 8 patients within those 2 hours that he was that he was rounding. This review by the consultant was pretty standard. Usually what happens is the, as all over the world, I would imagine the register who's admitted the patient presents the history to the consultant, they probably have a discussion about it and then they both together go to visit the patient and any other historical features or a physical examination findings are re-illicited and discussed further. It's documented in the summary that the consultant spent a considerable amount of time with Mrs. A clarifying the nature of the lump and he stated that he was reassured at the time because Mrs. A had actually just finished eating her dinner and she wasn't really experiencing much in the way of problems either abdominal pain or vomiting after that. So the consultant documented his examination findings, his significant lymph node on the left groin, the abdomen was soft, nontender and the pulse was irregular and apparently he had an impression at that time of malignancy, likely lymphoma or intradominal malignancy. I think that was his best guess but there wasn't much in the way of a rationale documented for how he came up with this diagnosis. Incidentally we didn't mention before that on admission the serum sodium was measured to be 116, it wasn't even mentioned in the admission note. I'm assuming that they had all sorts of explanations based on the vomiting as to why the sodium ended up that low but it didn't seem to be much of a concern clinically, it didn't even make it into the problem list. So the final plan documented after that review by the consultant was first of all that Mrs. A would probably be able to go home the following day when she was rehydrated and her nausea had settled and that the rest of the investigations could be done as an outpatient and he specifically asked the registrar to speak to Mrs. A's GP and get a further history and just to confirm what that procedure was that was booked, the one that was written as lymph node biopsy, they were going to confirm what that was actually all about. And the written plan was to give fluids with IV saline a range for that biopsy of the left inguinal lump, although she presumably already had an appointment for this, they were going to repeat some blood tests, the electrolytes, the thyroid functions, do some malignancy screening in the way of serum protein electrophoresis and depending on those quote biopsy results, they were going to potentially plan for a CT scan of the abdomen and maybe even a bone marrow biopsy, I guess in line with what they were thinking the diagnosis was. So I think there's a couple of issues again, not surprisingly, that arise here. One of them I think is largely a system issue, a system problem, then that the consultant was actually seeing the patient, not that long after the registrar had seen the patient. So didn't have the benefit of perhaps seeing that the following day she had still been nauseated, still vomiting, etc. because he was actually rostered to be at a clinic elsewhere. And so certainly I think he lost the benefit of seeing the patient over some degree of time and he was relying on the registrar's assessment the following day to alert him to anything that was amiss. It always reminds me a little bit that we're, sometimes we think uncharidably about some of the admissions that we have to see the following morning, but we have to remember that we always have the benefit of those 24 hours of observation and in retrospect we say, "Oh, this patient didn't really need to get admitted maybe, it's very hard to make that determination in the first few hours," and so often patients do get kept. Yeah, I mean time is one of the greatest diagnosticians known to any of us. I rely on it heavily. Yeah, absolutely. The second thing I wanted to just comment on is what I think was very significant premature closure in the sense that his, the examination findings were labeled as a significant lymph node in the left groin, because I think if you think that that's a lymph node, oh, she's had, she's had lumps in her neck as well, oh, they must all be lymph nodes and so this is got to be a lymphoma. We talk a lot on this podcast and to our juniors about the absolute need for differential diagnosis. People diagnosis being the hallmark, the cornerstone of type 2 analytic reasoning, and it prevents us from jumping to premature conclusions, and I think in this case there were no differentials really offered at any point in time, certainly there was no differential for the vomiting. So when I think about vomiting, I think could this be an intracerebral cause of vomiting, something triggering the chemoreceptor, trigger zone in the brain, could this be something toxic metabolic, systemic, could this be something gastrointestinal, like an obstruction? And I sort of go through that almost intuitively at this point in time when I see somebody with persistent, intractable vomiting, which is really what this lady had, there was no diagnosis at any point offered for that. Even if you thought it was secondary to an intradominable malignancy, what is the mechanism for that? It would almost certainly need to be an obstructive picture if you're considering the possibility of this lump being an intradominable malignancy, either metastatic or due to a lymphoma. Obviously, the other place where there was no differential diagnosis offered is in the realm of what this groin lump is. Again, this is not a difficult differential to come up with, I think anatomically. So certainly in the femoral region, a lump that had been there for a week or a month, it doesn't really matter, still could be a hernia. Whether or not it's incarcerated is a different story, but certainly could be a hernia. Could be an aneurysm. Could be a lymph node. An abscess. It could be an abscess. It could be hematoma. It could be a sarcoma. I mean, it could be anything, really, and you need to do a differential diagnosis. And I wonder if perhaps a hernia had been included on that differential diagnosis, whether or not that might have clued something in someone around the cause of the vomiting. Now there was mention that at some point in time, before he had a direct conversation with the GP, the registrar was aware of the GP querying a hernia. Yet despite that, that was never included in a differential diagnosis. And we sometimes see this as well when we get referred patients from the community, you know, to the grand hotel, the grand hybrid or ivory tower, that we know best, really, the GP is kind of out there doing the not very important stuff, but any of the really diagnostic important stuff we're doing. And so we tend to discount information that other people might have available for us. And perhaps the registrar paid a bit more attention to the GP's concerns. Maybe there would have been a differential diagnosis generated instead of just jumping to the conclusion that this was a lymph node. When the registrar was asked about this later on when the HTC was gathering information, I mean, I think they rightly said, look, we make our own judgments about patients and histories, et cetera. And I accept that. But I think when you're making a few assumptions and jumps to get to a single diagnosis, I'm sure that registrar would also have said that this diagnosis, this putative diagnosis of an intraptoma malignancy or lymphoma, was not 100% certain. And whenever it's not 100% certain, surely you would want to propose other possibilities. You know, unless, well, only because there's no benefit in keeping those private, right? It's very possible that they had a differential going on in their mind. They just didn't bother to document it because they knew that, okay, if it doesn't turn out to be what we think it is after we biopsy the hernia, then we'll move on to the next thing on my private list, you know, that I keep in my head. But there's no benefit in doing that. You might as well document it and give not only yourself a reminder for what else you might have been thinking when you reviewed the case in detail, but your colleagues would appreciate very much to have an idea of all the cognitive work that you've already done. Right. And by not listing a differential diagnosis, as you suggest, you make it seem like this is a certain diagnosis, and so it just adds to that diagnostic momentum and makes it very, very difficult for anyone to veer off of that predetermined pathway. Well, I wanted to mention that there was an interesting comment from the consultant post-hawk after the diagnosis had already been out, et cetera. And he said, quote, "A hernia presenting under those circumstances is very uncommon, and it is likely at that stage that many other doctors would have made the same misdiagnosis." And I don't disagree that with a benign abdomen and tolerating food and in that moment when he assessed her not vomiting. These are all reassuring features, and I agree that maybe at that point an incarcerated hernia would have been considered less likely. But the phenomenon which we've talked about before is this, and it's related to not documenting a differential diagnosis, but it's this unwillingness to put forward and document unlikely diagnoses and rather just keep all that back, so you either, the attitude is you either know what the diagnosis is and put that forward, or you don't know what the diagnosis is, in which case you say nothing and only plan for CT scans, et cetera, et cetera. How often do we see no impression just a plan with some investigations planned? And part of a differential diagnosis is an ordering of that differential list in terms of likely to less likely. So you don't stop just by listing the possibilities, you have to make some assessment about which is most likely, which is less likely but still associated with a high stakes diagnosis, which is very unlikely and not associated with significant stakes. So again, you also do have to somehow prioritize that differential once you've created it. So we'll move on to the next phase, which is probably the post-acute day or the day after the patient was admitted to hospital. The nursing notes say that the patient is doing somewhat better, although intermittently still vomiting, intermittently requiring intravenous metaclopromide. To the registrar's credit, he did actually ring the GP and they had a conversation in which case, rather during which the GP recalled mentioning again or querying again, the possibility that this lump was a hernia and asking about the cause of the vomiting for which the team really hadn't provided any legitimate explanation at that point in time. The registrar also did have an opportunity to speak with family members. And apparently, during that conversation with family members, there were several things that came out of that. I think the registrar presented the possible diagnosis of a lymphoma, of a cancer. And related them to the lumps on the patient's neck, at which point in time, there is documentation here in the final report that the family informed the registrar that those had previously been biopsyed and were found to be benign epidermal cysts. There were some other things that the family states they conveyed to the registrar, but the registrar didn't seem to have much recollection of. And that is that the patient actually hadn't had a bowel motion for a couple of weeks. Now that's probably a little bit unlikely, given the fact that her abdomen was soft and non-understandard. But the family, at least, was aware of some alteration in the patient's bowel habits, which I'm not sure went anywhere other than expressing that. And the family was concerned that she was still vomiting. And they felt as though it was a pretty dramatic change from her baseline, and that nothing really had been explained. The registrar and the consultant had a phone call as well, at which point in time, the consultant states that he was reviewing the patients who had been admitted the day before. And that the registrar apparently told him that the patient had elected to stay in hospital. But it didn't mention anything at that point in time about the phone call to the GP or the conversation that he'd had with the family. So the consultant at that point in time was led to believe that the patient was stable and there was no further discussion about the patient at that point in time. Okay, so I'm sure our listeners have queued into what some of the major problems are with this phase. But what keeps up coming up for us is a bad case of confirmation bias and diagnostic momentum. I mean, the registrar is receiving all sorts of new information from the GP and from the family, but somehow is not reengineering their thought process. They continue to ignore all these important pieces of information because it's not fitting with the diagnosis that they've already put forward. So the GP continues to challenge the registrar to explain the vomiting, which still hasn't been done, and even questions whether this all could represent a hernia. She's actually giving him the diagnosis as something to think about, but this is not taken into account in any serious way to then carry it forward. And then the family, the conversation with the family, they have a few concerns here and there. But most importantly, they've given the registrar the diagnosis of those previous necklumps. If you didn't know this before from the records, they've just confirmed for him that in fact those necklumps have been diagnosed, they've been biopsied and they were epidermal cysts. That's not lymphoma, that's not metastatic intradominal malignancy, and that's unrelated to a new lump that is different that shows up in the groin. And yet, we're burying ahead with this diagnosis of query lymphoma, query interdominal malignancy, and though those necklumps are out of the equation, should be out of the equation altogether, they're not out of the equation in the minds of the registrar. The last thing I think is worth mentioning is that this line of thinking is actually manifested in what the registrar says to the consultant in that the patient has "elected to stay in the hospital." I don't know how most other hospitals work, but we get a lot of pressure here to discharge patients when they're ready to be discharged and it has to be extenuating circumstances to change that attitude. Patients don't usually elect it to stay in hospital if they're feeling well, and the note says that she was feeling better, but she was still nauseated, and I do wonder in what way was she feeling better, exactly, and what really caused her to stay in hospital. So the next phase after this is not a good one. Essentially, the patient began to vomit more copious amounts and more frequently than she previously had been vomiting. She was being treated symptomatically for that, but it didn't seem to be making much in the way of a difference. At some point in time, her blood pressure began to fall, and that was noted. No early warning score alarms were triggered at that point in time. The next phase of her illness happened as you can probably predict in that she began to experience some respiratory compromise. She began to have worsening cough. She began to have some hypoxia, and a chest x-ray was ordered by the on-call house officer, who when they saw some infiltration in the right middle lobe called this "community acquired pneumonia," and put her on antibiotics for that, but in fact her hypoxia and her symptoms began to worsen. And certainly, there was a systems issue here in terms of how quickly people responded to what was a pattern of worsening vital signs. But in fact, the patient Mrs. A. Underwent suffered a cardiopulmonary arrest from which she was revived. At that point in time, a decision was made to do an abdominal CT scan, which did confirm the diagnosis of an incarcerated hernia with a bowel obstruction. She went to surgery, but never really woke up, and ultimately died a few days later after suffering massive anoxic brain injury, so a very, very tragic case. So in the final analysis, just to conclude, you know, I think this was a case where the clinicians over employed their type 1 intuitive thinking came up with a diagnosis that seemed to be rational at the time, but didn't come up with any sort of differential diagnosis either for the vomiting or the cause of the lump, and then fell prey to premature closure, diagnostic momentum, confirmation bias, and all of the other biases that we've highlighted so far in this episode. And, you know, you and I have discussed at times, you know, does every presentation, every problem require a differential? Are there some problems that are so barned or and so obvious that they don't need a differential diagnosis? My retort to you has always been, and I'll share it with the listeners, is that I'm not smart enough to know when this is a barn door, a slam dunk diagnosis or not. And so I think I should be coming up with a differential diagnosis for every problem, for every presentation, even if it seems obvious, because I at least should be asking myself, is there anything else this could be? Is there anything else that I need to make sure it's not within the next 12 to 24 hours? Even if it's something as simple as cellulitis, you know, and I'm sure of the diagnosis, I still should be asking myself, you know, could this be necrotizing fasciitis? Is there any evidence here of a compartment syndrome? Something else to trigger my analytic brain into action? Yeah, and we've mentioned this before. I mean, I'm on the same page, and I, you know, certainly for my practice, I never feel so confident in my intuition that I'm going to just forego the question, you know, what else could I be dealing with? But even if you could convince me that there are cases that some clinicians can just jump to the diagnosis and really it's so clear that they, you know, they don't need to go any further, even if that was the case. I don't think that this was one of those cases, and I think that, well, certainly in retrospect, the clinicians involved in this case would agree, you know, we've mentioned this before, but one of the things that I often ask my team when they're presenting cases to me is, okay, so this is the diagnosis that you think it is. How certain are you that this is the diagnosis that that is what you're dealing with? And you'd have to be over 99% certain if you're going to not give me a list of other possibilities. I mean, imagine if you're 90% certain, I mean, that is pretty darn good. I'm rarely 90% certain about the diagnosis. And even there, one in 10 of those patients, you're going to get it wrong. So there's very little excuse to not do, and this clearly was a case that there was no certainty. I mean, there was all sorts of assumptions made and conclusions jumped to. I think they probably recognized that they were taking a little bit of a guess at saying things like lymphoma or intradominal malignancy, and yet there seemed to be a reluctance to document their thoughts early on, and they were just going to go in series and hopefully stumble upon the diagnosis as they went. Yeah, and I'm not sure that we certainly wouldn't want to be, have been in their shoes. I think we're not raising these issues in any way just to reiterate in a punitive way or to cast aspersions on anyone's clinical skills, et cetera. But I do think for me, this case highlights a lot of what we've talked about in this podcast. And if people find these cognitive autopsies helpful, please let us know on the website. We certainly have access to other cases that we could unfortunately, countless other cases from the HDC. Right. So hopefully this was useful, and hopefully we can do something. I like that. Are we done? Yeah.