• Hypothetical Study of Alcohol Consumption and Breast Cancer: You conduct a study to determine whether breast cancer in African-American and Hispanic women living in the Chicago area is associated with drinking alcohol.

Study Design

With help from the Illinois State Cancer Registry, you identify 250 cases of breast cancer in African-American women and 250 cases in Hispanic women diagnosed in January of this year.

Next, you identify 500 African-American and Hispanic women without a history of breast cancer from a list of persons responding to a Chicago area health survey administered in January of this year.

Drinking habits during the previous 5 years are compared, and the results are shown below:

Alcohol consumed/wk during the previous 5 years (average)	Breast Cancer
Yes	No
> 20 ounces	371	235	606
20 ounces or less	129	265	394
Totals	500	500

Calculate and interpret the appropriate measure of association and its 95% confidence interval.

Suppose you are concerned about smoking being a potential confounder of the above association. The odds ratio of the association between smoking and alcohol consumption is 2.8 with a 95% confidence interval of 1.9 3.7. The association between breast cancer and smoking is 0.7 and the standard error of the natural log of that odds ratio is 0.25. Using all of this information, determine whether smoking fits the definition of a confounder. Provide an explanation for your determination.