To the Editor.—
Barnard1 presented an interesting study of persons attending the Pritikin Longevity Center between 1977 and 1988. Barnard is to be commended for attempting this study on the relationship between diet and serum lipid levels. The Center's goals, likewise, are laudable.There are, however, difficulties with this study, including well-known biases (sampling, selection, measurement, and confounding), ecologic fallacy, and generalizability.2 It is an uncontrolled, unblinded "intervention study" in one sense, but actually it resembles a variation of a "historical cohort." There is selection bias, since these persons are highly motivated volunteers. Because it was an unblinded trial, measurement bias cannot be dismissed.2 There are the potential problems of an "unpredictable outcome" (since biologic variation cannot be evaluated without a control group), "Hawthorne effect," and regression to the mean.2 Since a retrospective search of records was done, measurement bias is likely and incomplete records are
Chapdelaine PA. Life-style Changes and Lipid Lowering. Arch Intern Med. 1992;152(8):1721–1722. doi:10.1001/archinte.1992.00400200145030
Customize your JAMA Network experience by selecting one or more topics from the list below.