On the Epistemology of Pathology

Ed Uthman, MD

Diplomate, American Board of Pathology

15 Apr 1995

I have not posted much this last week, since I have been doing what I typically do after a newspaper exposee on "Pap mills," that is, digging out from behind the huge pile of "atypical" Paps generated by the psychological effects of bad press on the cadre of cytotechnologists. Of course, I myself am not immune to being spooked, so I find myself justifying in my mind why a given smear I have spent ten minutes reviewing needs to be called atypical, even though all the cells look normal to me.

In times like these I begin musing about the epistemological nature of diagnostic pathology in general. I am fond of saying that pathology is about as "scientific" as the field of art criticism. Whereas just about all competent art critics will agree that Van Gogh was a master, there are other areas of opinion where critics disagree violently with each other. The same is true of pathology. Whereas most of us agree that a given napkin-ring lesion of the colon is an adenocarcinoma, there are other areas, such as pigmented skin lesions, where establishing uniformity of opinion is very difficult, if not impossible. To cope with this problem, we have over the years developed a system of canon law, whereby cases which have generated disagreements at the diocese level are referred up through the ecclesiastical hierarchy to an archbishop or pope, whose final determination ex cathedra settles the issue. Final diagnoses rendered by recognized experts, in my opinion, lack even the logical underpinnings of a papal encyclical, in that the Pope typically cites specific reasons for his opinions, while a consultant's report rarely includes a detailed discussion of the differential diagnosis and the reasons why all the items but one were eventually ruled out.

That each diagnostic issue is settled does not alleviate the problem that behind pathologic determinations there is supposed to lie a logical backbone of science, which should be reproducible from one practitioner to another. I think that irreproducibility is the dirty little secret of pathology. Although failure of experts to agree on individual cases should be a fundamental and pivotal issue in our field, scant attention has been paid this in the literature. From years of reviewing Pap smears, I am convinced that psychological factors strongly determine how cytotechnologists evaluate smears, especially in the "squishy" area of "normal" versus "atypical" or "ASCUS." I have no reason to believe that similar factors do not influence the opinions of pathologists. I think this is an unexplored area of psychology, which would make an excellent field of research for behavioral scientists.

Commentary on this rant is also posted on this site.