1. Suppose that 50,000 women who have never had a breast cancer diagnosis are available and willing to participate in a new screening program of computerized digital mammography. At the first screen, 1132 women were positive on the screening test. All of these positive women had a breast biopsy (the gold standard for breast cancer detection) and 136 were diagnosed with breast cancer. Of the 48,868 who screened negative on the test, 24 were in fact found to have breast cancer (determined by follow-up examinations).

a.Fill in the two-by-two table showing these screening test results.

b.Calculate and interpret the sensitivity and specificity of this new screening test.

c.Calculate and interpret the predictive value of a positive test result for this new screening test.

d.A colleague in another city reports that the same screening test (with the same sensitivity and specificity) is achieving a predictive value positive of only 6%. What is the explanation for this low positive predictive value?

2. Suppose that a study of oral contraceptive (OC) use and development of bacteriuria was conducted among 2,390 women, all of whom were initially free from bacteriuria. At the start of the study 482 of the subjects were using OCs and 1908 were not. By the end of the follow-up period, a total of 27 of the OC users and 77 of the non-users had developed bacteriuria.

a.Is this a cohort study or an experimental study?

b.Set up the two-by-two table for these data.

c.Calculate the risk ratio describing the strength of the relationship between OC use and bacteriuria.

3. A case-control study of dust lead levels and lead poisoning in children was conducted. Of the 200 cases of lead poisoning, 50 had high dust lead levels in their homes. Of the 200 controls, 30 had high dust lead levels in their homes.

a.What measure of comparison should be used to estimate the strength of the relationship between high dust lead level and lead poisoning?

b.Set up the two-by-two table and calculate this measure of comparison.

c.The 95% confidence interval for the measure of comparison was 0.8 to 3.0. Give your interpretation of this confidence interval.

4. In a cohort study of smoking and lung cancer, the incidence of lung cancer among smokers was found to be 7/1,000 and the incidence among nonsmokers was 2/1,000. From another source we know that 35% of the total population were smokers.

a.What is the etiologic fraction (i.e., the proportion of lung cancer among smokers that are attributable to smoking)?

EXTRA CREDIT

Suppose that we undertake a cohort study to examine the possible association of high salt intake and hypertension. Our study population consists of male and female Hispanics and Caucasians. No matching or restriction is performed. Lets assume that the following facts are true: Men consume more salt than women. Irrespective of their salt consumption, men have a greater risk of developing hypertension than women. Smokers consume the same amount of salt as non-smokers. Smoking is an independent risk factor for hypertension. Hispanics consume more salt than Caucasians. Irrespective of their salt consumption, Hispanics have the same risk of hypertension as Caucasians. Based on the above information, which of the following are potential confounders in this study?

Select all that apply.

A. gender

B. smoking

C. race/ethnicity

D. None of the above