[Skip to Content]
[Skip to Content Landing]
April 4, 1980

Hidden Bias in Research Design

JAMA. 1980;243(13):1365. doi:10.1001/jama.1980.03300390049024

I believe that most physicians have had the experience of seeing patients with what are generally considered serious diseases— diseases associated with substantial mortality or morbidity according to the textbooks—and were surprised at how many of these patients do remarkably well, not appearing to follow the morbid textbook course at all. Seen, perhaps, for some minor, unrelated problem, their serious disease may remain totally quiescent, indefinitely. Such experiences make one doubtful of the accuracy of original diagnoses or, in moments of skepticism, make one wonder whether patients referred to academic centers do worse than unreferred (and unreported) cases for reasons other than the pathophysiology of their disease.

An article published in this issue of The Journal by Ellenberg and Nelson (p 1337) may shed some light on this paradox. When studying children with febrile seizures, these investigators find that clinic-based studies may have reported on a biased spectrum of patients.