Slow to Diagnose Endocarditis After Repeat Visits

A 43-year-old woman saw her new primary care physician one time. Shortly after that visit, she saw a covering physician for an urgent care appointment with complaints of intermittent fever and fatigue for one week. Blood work, including Lyme serology, urinalysis, and urine culture, were ordered. She was instructed to call or return if her symptoms worsened.

Two days later, she emailed her PCP with a complaint of continued fever, fatigue, and a new rash. She made an appointment for the following day. In her medical record was a history of congenital bicuspid aortic valve, but her PCP did not see it. She repeated the lab work, and treated the patient for a presumed urinary tract infection.

Initially, the patient improved, but eight days after her visit, she called the office—and emailed her PCP—with complaints of increased fever (to 102.5), joint pain, and a rash that had spread to her chest. The patient requested additional blood work and a treatment plan but was not seen in person. Her lab work was repeated and she was also referred to infectious disease for evaluation of her persistent fever and fatigue.

One week after that, upon exam the infectious disease physician detected a heart murmur. A subsequent echocardiogram led to a diagnosis of aortic valve endocarditis. The patient was admitted for IV antibiotics and required a valve replacement. She did not experience any long-term sequelae.

The patient sued her PCP, alleging that the defendant’s failure to recognize her medical history led to the delayed diagnosis and treatment of endocarditis. The case was settled in the medium range.

To discuss the patient safety and risk management aspects of this case Dr. Jonathan Einbinder joins us now. Dr. Einbinder is Vice President, Advanced Data Analytics and Coding, at CRICO. Jonathan is also an attending internal medicine physician at Brigham and Women’s Hospital.

Q.) Jonathan, thank you for joining us.

A.) Thank you, Tom. Glad to be here.

Q.) You’ve heard the case. Where do you start to see things go sideways a little bit?

A.) Well, Tom, the first thing I would say is that this is the kind of case that scares every doctor, because it’s a pretty routine case, a patient having a fever, and 99 times out of 100 it turns out to be nothing. And it’s that one time where it isn’t is the one that we all worry about missing, and that’s what happened here. I think because fever is such a common complaint, I don’t think that it really went wrong at the initial evaluation, where the patient first presented and was evaluated, had some testing done. I think it really started to go wrong when the patient presented with the same complaint that was persisting, and the evaluation didn’t go down the path of possible of other diagnoses, like endocarditis.

Q.) A lot of times these medical liability cases feature continued patient complaints for the same issue. How does that signal play a part in this case? What kind of principle would you think of in these situations?

A.) So when a patient returns with a repeated complaint, in this case with the repeated complaint of fever and even some new symptoms, such as the rash, it’s important to always think about what could be going on and whether your working model for the diagnosis is the correct one. One thing which seemed to be missing in this particular evaluation was certainly a formal differential diagnosis that may have been present in the physician’s brain, but wasn’t documented, and there’s no evidence that it was really thought about. For example, in this patient, a differential might include things like a simple viral illness, it could also have included Lyme disease, and testing was done for Lyme disease, or a urinary tract infection, but it also could have included things like endocarditis, pneumonia, other occult infections, or even other inflammatory conditions. And then when the patient continued to come back, or came back again, and the symptoms persisted over time, additional testing and additional evaluation should have been done for those other possibilities.

Q.) It  also didn’t seem like the patient history was as thorough as it needed to be for this case.

A.) Yeah, and this is a really important point. This is a read between the lines a little bit, because we’re missing a few of the details, but the patient is a 43-year-old woman who had a new primary care physician. And it doesn’t say whether she was new to the practice or just new to this physician, but presumably the physician at the initial encounter takes a medical history, may or may not review prior records, if those are available to her, is going to do some documentation, including writing a note, putting medications on a medication list, allergies on an allergy list, and problems on a problem list. And in particular for this patient, there’s a history of a bicuspid aortic valve, which is a condition that increases the risk for a condition, like endocarditis, and is a condition that should be reflected, should be known to the physician, and should be documented on the problem list, and it is completely understandable to me, as a practicing physician, how at an initial visit I might not be aware of it and I might not get it on the problem list.

What that means is, in this case the patient’s second visit, when she had the fever, was with a covering provider, it was with an urgent care physician. That physician would have been relying on the problem list to understand what that patient’s conditions were, and in fact if the bicuspid valve were present, might have increased the suspicion for endocarditis. So the documentation of that problem is absolutely crucial; it’s completely understandable why that might have fallen through the cracks; and it’s one of those situations that’s a set up for what happened to this patient.

Q.) How would you see the ideal kind of communication? What would that look like with the patient throughout the case to improve the chance for a better outcome?

A.) Well, the patient calls back after eight days, saying she’s still having a high fever, 102.5 degrees Fahrenheit, and now having joint pain, still having a rash, which was getting worse, still feeling fatigued, and so now a fever for eight days has a different differential diagnosis than a fever for one or two days. Unlikely to be the flu or a cold or a cough or even pneumonia; something else is going on. Again, the patient’s doing her job, she’s notifying her care team, and this is where the care team needs to recognize these symptoms as being concerning and requiring really additional, acute evaluation.

She should have been seen in person, so somebody could listen to her chest and see if she has a heart murmur. They could have thought of additional testing, like a chest x-ray or an echocardiogram, so there’s definitely things that I think could or should have been done at that point. And, to the point you made earlier, Tom, if they realized that she had an bicuspid aortic valve, I am confident they would have thought of this diagnosis and done additional testing, so that was really a critical piece of missing information.

I think I guess the two things that I would just emphasize is that fever is a common complaint, and most of the time it is not something serious. So the initial evaluation doesn’t need to go overboard most of the time with lots of expensive and invasive testing, referrals and so on. However, if the patient has risk factors, such as a bicuspid valve in this case, or if the patient is not getting better, then there needs to be a differential diagnosis and there needs to be a clear plan of treatment and evaluation. I think that would be sort of one point to make.

The other point, and this just relates to differential diagnosis, is just to saying that I always think of when I’m seeing patients—I’ll give credit to Jerome Groopman from Beth Israel Deaconess Medical Center for this principle—but when I’m seeing a patient, even if I think I know what is going on, it is always helpful to ask the question, ‘What else could this be?’ And so if I’m seeing a patient who’s having a high fever and I think it’s a urinary tract infection, I should at least be asking the question, ‘What else could it be?’ I don’t necessarily need to be doing diagnostic testing at that point, but I should have a differential diagnosis, and if the patient is not promptly improving, be ready to look further down that list of diagnoses.

Lastly, from a risk management perspective, the documentation of those issues is critical in demonstrating the sort of thought process of the clinician and helping to prevent or defend in the event of a malpractice claim being filed.

Q.) Well, thank you. Dr. Jonathan Einbinder is Vice President, Advanced Data Analytics and Coding at CRICO. For MedMal Insider, I’m Tom Augello.