There were several interesting studies in Academic Medicine
this month….
The first study (1) was led by one of my favorite
educational researchers, Dr Geoffrey Norman. Dr Norman is one of the foremost
researchers in the area of cognitive reasoning. In this current study, his team
looked at resident physicians in Canada. Participants were second year residents from
three Canadian medical schools (McMaster, Ottawa, and McGill). They were
recruited right after they had taken the Medical Council of Canada (MCC)
Qualifying Examination Part II. They
were recruited in 2010 and 2011.
The researchers asked the residents to do one of two things
as they completed twenty computer-based internal medicine clinical cases. They
were instructed either to go through the case as quickly as possible without
making mistakes (Go Quickly Group; n=96) or to be careful, thorough, and
reflective (Careful Group; n=108). The results were interesting. There was no
difference in the overall accuracy (44.5% v. 45%; p=0.8, effect size (ES) =
0.04). The Go Quickly group, did that. They finished each case about 20 seconds
on average faster than the careful group (p<0.001). Interestingly, there was
an inverse relationship between the time on the case and diagnostic
accuracy—cases that were incorrect took longer for the participants to
complete.
Another interesting study about diagnostic errors came out
of the Netherlands (2). Dr Henk Schmidt asked an important question: does
exposure to information about a certain disease make doctors more likely to
make mistakes on subsequent diagnoses? In this study, internal medicine
residents were given an article from Wikipedia to review and critique. The
article was about one of two diseases (Legionnaire’s disease or Q fever). Half
of the residents received the Legionnaire’s article, the other half the article
on Q fever. Six hours later, they were tested on eight clinical cases in which
they were forced to make a diagnosis. Two of the cases (pneumococcal pneumonia
and community-acquired pneumonia) were superficially similar to Legionnaire’s disease.
Two were similar to the other disease from Wiki (acute bacterial endocarditis
and viral infection). The other four cases were “filler” cases that were not
related to either case from Wikipedia. (aortic dissection, acute alcoholic
pancreatitis, acute viral pericarditis, and appendicitis).
The results are a little scary. The mean diagnostic accuracy
scores were significantly lower on the cases that were similar to the ones that
they had read about in Wiki (0.56 v. 0.70, p=0.16). In other words, they were
more likely to make an error in diagnosis when they had read about something
that was similar but was not the correct diagnosis. The authors believed that
this demonstrates an availability bias because they were more likely to
misdiagnose the cases that were similar to ones that they had recently read
about. Availability bias can also be seen with students, think about the
student who comes from the Cardiology service. Every patient that they see in
clinic with chest pain is having a myocardial infarction.
The last article that caught my eye was another study out of
Canada. The authors, from the University of Calgary, wanted to determine if
students that were doing their clinical clerkships in a non-traditional longitudinal
fashion were learning as much as students in the traditional track. So they
looked at all of the students who completed their clinical training in a
Longitudinal Integrated Clerkship (n=34) and matched them to four students in
rotation-based clerkships. Students were matched based on grade point average
(GPA) and their performance on the medical skills examination in the second
year.
The outcomes that they studied were the Medical Council of
Canada Part 1 exam scores, in-training evaluation scores, and performance on
their clerkship objective structured clinical examinations (OSCE). They found
no significant differences between the two groups on the Part 1 exam score (p =
.8), in-training evaluation (p = .8), or the mean OSCE rating (p = .5). So,
apparently, students in a rural longitudinal rotation did just as well as those
who stayed at the University hospital for rotation-based clerkships.
References
(1) Norman G, Sherbino J, Dore K, Wood, T, et al. The
Etiology of Diagnostic Errors: A Controlled Trial of System 1 Versus System 2
Reasoning. Acad Med 2014; 89(2):
277-284.
(2) Schmidt H, Mamede S, van den Berge K, et al. Exposure to
Media Information About a Disease Can Cause Doctors to Misdiagnose
Similar-Looking Clinical Cases. Acad Med 2014; 89(2): 285-291.
(3) Myhre D, Woloschuk W, Jackson W, et al. Academic
Performance of Longitudinal Integrated Clerkship Versus Rotation-Based
Clerkship Students: A Matched-Cohort Study.
Acad Med 2014; 89(2), 292–295.
