[Skip to Navigation]
November 1991

Some Cautions in Labeling Effects

Author Affiliations

Stanford, Calif

Arch Intern Med. 1991;151(11):2313. doi:10.1001/archinte.1991.00400110141028

To the Editor.—  Havas et al1 recently reported a neutral labeling effect in the Massachusetts Model Systems for the blood cholesterol screening project. Their report, although provocative and timely, leaves several unresolved questions of concern.First, patient selection factors are largely undefined. The authors provide no details about how patients were recruited from the four sociodemographically diverse Massachusetts communities, nor whether those participating in the last phase (3489/ 10500 screenees, 33%; September 1988 through January 1989) differed in any measurable way from those earlier in the screening sequence (October 1987 through August 1988). Most importantly, a critical reader would need to know the proportion of self-selected vs "reluctant participants."22 The latter are reportedly more likely to display adverse labeling effects.2-4Second, the investigators' intervention may not have been very powerful, either positively or negatively. They describe spending only 5 minutes per participant and still managed to address

Add or change institution