I'm following citations to our paper explicating Hickam's Dictum (HD) and Ockham's Razor (OR) as applied to diagnostic decision making, published two years ago. It was my hope that this paper would lead to a more thoughtful consideration of these allegedly conflicting clinical saws, but alas I was too sanguine about that prospect. It would seem that most people citing our paper are not even reading it. Rather, it appears that they come to the conclusion that they have identified a case of HD, write a case report about it, then after the fact seek supporting references and cite our paper naively, without applying its insights to their case.
I'm happy to have our paper cited for any perspicatious and curious readers of these case reports. And, the problem of careless citation practices is rampant. Often when I check citations to see if they support the referenced statements, I'm gobsmacked: not only do many citations fail to support the authors' statements, sometimes they directly undermine them. Commonly, an author makes claims as though starting from truth rather than uncertainty, then searches for supporting references, but does not evaluate his claim or bring evidence to bear upon it. Or, even read the papers he's referencing. Rather, he reads the abstract and determines if a superficial and incomplete "reading of the literature" (reading the abstract) can be used to bolster his claims to his likewise unwary lectors. This is not scholarship, this is confirmation bias and motivated reasoning in action. It is abetted and exacerbated by the tendency of authors to overextend themselves in their abstracts, making conclusions that are broader and firmer than warranted by the underlying data. As this phenomenon has a tendency to magnify and replicate flimsy claims, we wind up with wholly unsupported concepts that persist for decades, viz, immune dysregulation in sepsis. (Just think about this for a second and decide if it makes sense:"mamallian immunological evolution stalled, failed, or overshot bacterial host defenses, thus immunity is dysregulated in this disease.") Yet this concept accompanies almost every single article about sepsis for more than a generation; you gotta reference it in your sepsis paper. So, go to pubmed, type it in, find an abstract that talks about it, import into EndNote. Done. Specious concept propagated again.
In this case report alleging the simultaneous presentation of Guillain-Barre Syndrome (GBS) and a pancreaticoduodenal artery aneurysm PDAA with impending rupture, the author invokes Hickam's dictum becaues the patient appears to have two diseases that were discovered at the same time. The schema outlined in Figure 3 of our paper shows that there are four possibilities for the relations between two (or more) disease: incidentalomas, pre-existing diseases, diseases causally connected to the principal diagnosis, and lastly, two symptomatic, independent, temporally coincident diseases. This last possibility is decidecly the least common, but it's what the author settled upon.
The patient had a viral illness (predisposing condition for GBS) and a week later developed numbness and back pain, two common symptoms of GBS. She went to an ER and because of the back pain, had a CT of the abdomen and a PDAA was discovered. The case report describes it as having impending rupture but we are not told how this determination was made (can it be made?). It was embolized and her back pain (maybe? kinda?) resolved.
Not described in the report is the well-established fact that the vast majority of unruptured PDAA are totally asymptomatic and discovered incidentally on imaging as in this case. Furthermore, when they are symptomatic they cause epigastric and abdominal pain, not back pain. Risk of rupture is not clearly correlated with size.
Just that background knowledge is enough to establish, at least in retrospect, that this is an incidentaloma. In real time, knowing that these aneurysms take years to form and are mostly incidentally discovered and don't cause back pain should lead the wary clinician to question whether this is an incidentaloma and the real disease is still lurking in the background. Her symptoms should be more carefully explored, information about the patient, and the diagnostic possibilities further investigated and synthesized.
In other circumstances, if we find incidentalomas and treat them with invasive therapies, we will cause many iatrogenic problems. Though PDAAs warrant embolization because of a high risk of later rupture, we may not want to place it high on the list of management priorities depending on the acute disease (here, GBS) we're dealing with. Imagine that this lady had presented just a bit later when she had respiratory muscle weakness, and had been taken for embolization, which should have been elective, then she coded during conscious sedation for the procedure. Or, imagine if, as almost happened in this case, that the principal diagnosis (the one which explains the chief complaint; GBS) was missed and she had respiratory arrest because PDAA with impending rupture was misidentified as the principal problem and her weakness was misattributed to residual sedation effects.
Analysis of this case report makes clear that this is not a purely pedagogical or pedantic exercise. Knowledge structured according to our Figure 3 heightens awareness that we can always be chasing incidentalomas or anchoring on known diseases, missing the true principal diagnosis, or failing to make connections between causally related facets of the presenting illness, invoking multiple diagnoses rather than a unifying one.
Medicine is hard.
