Case Report Evidence

What it is

The ranking of study designs by how much confidence they can carry comes from evidence-based medicine, and the hierarchies are usually traced to the Canadian Task Force on the Periodic Health Examination in 1979. The book that brought the distinction into this vault does not name that origin, and the identification here is ours.

The structure of the argument is simple. A randomised trial can tell you whether an intervention caused an outcome because it has a comparison group and an allocation you did not choose. A case report has neither. What it has is detail, and detail is what lets someone notice that a thing has happened at all.

That is a real and honourable function. Case reports have a distinguished record of first signals, including the early reports that alerted doctors to thalidomide. Their limits are standard teaching rather than a criticism anyone needs to invent.

In effect

The failure mode this concept is built to catch is generalisation. One patient with damage to a specific area shows a striking deficit, and a chapter generalises from it to how everyone's brain works. The consumer version is the testimonial: one real customer, one real result, and no way to know whether it would happen to you.

The reference case in this library is Marc Dingman's Bizarre. About eighty-six per cent of its endnote slots are peer-reviewed, and about twenty-nine per cent are single-patient case reports or tiny series. Either figure quoted alone misleads. A reader told only the first would take the book to be well evidenced; a reader told only the second would take it to be poorly sourced. This publication reports both, which is the point of the tool.

The counting was done by hand: 229 endnote slots, of which about 67 are case reports or small series, and every chapter opens on one patient.

What it does not say

It does not say case reports are bad evidence. They are the right design for first signals and the wrong design for general claims.

It does not say a peer-reviewed source is unreliable. It says the peer review tells you about the venue and nothing about what the study could establish, which is the error that source-class inflation names.

It does not rest on research. This is a methodological category, not a finding, and nothing on this page is a testable claim.

---