Developing and testing health warnings about alcohol and risk for breast cancer: Results from a national experiment with young adult women in the United States.

BACKGROUND

This study sought to identify effective health warnings about alcohol consumption and breast cancer risk among young adult female participants.

METHODS

We tested a pool of health warnings in a national pilot study. We used the most effective designs from the pilot in the main experiment where young (ages 21-29) U.S. adult female participants (N = 1038) reporting past 30-day alcohol consumption were randomly assigned into 1 of 4 conditions where they viewed a health warning about (1) mortality, (2) mastectomy, (3) hair loss, or (4) control (non-health warning message). Participants were then randomly assigned to view 1 of 2 message types within each condition: text-only or pictorial. Warnings were shown apart from products. Outcomes were message reactions (attention to and cognitive elaboration of warnings, fear, hope, and perceived message effectiveness), attitudes and beliefs (perceived severity and susceptibility to alcohol harms, and perceived response and self-efficacy to prevent alcohol harms), and behavioral intentions to stop or to reduce alcohol consumption in the next month.

RESULTS

Multivariate analysis of covariance (MANCOVA) models testing between warning conditions showed estimated marginal means (EMM) for every health warning condition were significantly higher than the control for attention (control = 5.80 vs. mortality = 6.63, mastectomy = 6.81, hair loss = 6.83, all ps < 0.05), fear (control = 2.45 vs. mortality = 4.11, mastectomy = 4.16, hair loss = 4.02, ps < 0.05), perceived message effectiveness (control = 3.44 vs. mortality = 5.75, mastectomy = 5.82, hair loss = 6.09, ps < 0.05), and perceived severity of alcohol harms (control = 5.51 vs. mortality = 6.25, mastectomy = 6.09, hair loss = 6.35, ps < 0.05). There were no significant differences between the health warnings about cancer effects for perceived message effectiveness. EMMs for intentions to reduce alcohol consumption in the next month were significantly higher in the mortality (6.44) and hair loss (6.35) conditions versus control (5.61, ps < 0.05).

CONCLUSION

Exposure to health warnings about alcohol consumption and breast cancer risk (vs. control) resulted in greater attention, fear, perceived message effectiveness, perceived severity of alcohol harms, and intentions to reduce alcohol consumption.