Predicting tobacco sales in community pharmacies using population demographics and pharmacy type.

OBJECTIVE

To determine whether the population demographics of the location of pharmacies were associated with tobacco sales in pharmacies, when controlling for pharmacy type.

DESIGN

Retrospective analysis.

SETTING

Iowa.

PARTICIPANTS

All retailers in Iowa that obtained tobacco licenses and all pharmacies registered with the Iowa Board of Pharmacy in 2003. MAIN OUTCOME MEASURE AND INTERVENTIONS: Percentage of pharmacies selling tobacco (examined by pharmacy type using chi-square analysis); median income and distribution of race/ethnicity in the county for pharmacies that did or did not sell tobacco (t tests); predictors of whether a pharmacy sold tobacco (logistic regression using the independent variables county-level demographic variables and pharmacy characteristics).

RESULTS

County gender composition, race/ethnicity make-up, and income levels were different for tobacco-selling and -nonselling pharmacies. Logistic regression showed that whether a pharmacy sold tobacco was strongly dependent on the type of pharmacy; compared with independent pharmacies (of which only 5% sold tobacco products), chain pharmacies were 34 times more likely to sell tobacco products, mass merchandiser outlets were 47 times more likely to stock these goods, and grocery stores were 378 times more likely to do so. Pharmacies selling tobacco were more likely to be located in counties with significantly higher numbers of multiracial groups.

CONCLUSION

The best predictor of whether an Iowa pharmacy sells tobacco products is type of pharmacy. In multivariable analyses, population demographics of the county in which pharmacies were located were generally not predictive of whether a pharmacy sold tobacco.