A case report describes one patient in detail: what happened, in what order, with what was measured.
A case report describes one patient in detail: what happened, in what order, with what was measured. It sits at the bottom of every evidence hierarchy and it is not worthless, because what it does establish it establishes firmly.
A well-documented case shows that something can happen. That is a real contribution for rare events and for first observations, and no amount of statistical work can substitute for the first careful description of something new.
One case says nothing about how often. A report of a dramatic response is compatible with that response occurring once in fifty thousand attempts, and reporting bias ensures the dramatic cases are the ones written up.
Without a comparison, the sequence of events is all there is. Recovery after an intervention is compatible with recovery regardless of it, and single cases cannot separate the two however detailed the description.
Several similar cases collected together are more informative and still lack a control group. A series of twenty improvements is twenty uncontrolled observations, not a small trial.
A clear temporal link between an exposure and an unusual harm, especially where rechallenge reproduced it, is genuinely persuasive. Much pharmacovigilance works this way, because harms are often too rare for trials to detect.
A single story is more readable than a trial, so case reports are over-represented in secondary coverage relative to their evidential weight. The vividness is exactly what makes them memorable and misleading.
As a reason to look for better evidence. If a case report is the strongest thing supporting a claim after several years, the absence of anything stronger is the more informative fact.
Where an exposure was stopped, the effect resolved, the exposure was resumed and the effect returned, a single case carries unusual weight. The sequence is difficult to explain by coincidence and it is the pattern that has driven several genuine safety findings from individual reports.
Journals accept cases that are novel or instructive, so the published set is not a sample of what clinicians see. A pattern across published cases can reflect what editors found publishable rather than what occurs, and this is the main reason case series cannot support frequency claims.
The claims on this page about one person, described in detail, and the narrow claim that supports can be checked against sources outside this site: