Principles of Epidemiology Questions 1-4: In planning a case-control study of the relationship between smoking and cancer of the cervix, you need to determine if age is a confounder and/or an effect modifier in this relationship. Of 217 cases, 130 are smokers and 198 of the 243 controls are nonsmokers. The age distribution for cases and controls who are smokers is as follows: Smokers Age Cases Controls Age 20-29 41 6 Age 30-39 66 25 Age 40+ 23 14 The age distribution for cases and controls who are nonsmokers is as follows: Nonsmokers Age Cases Controls Age 20-29 13 53 Age 30-39 37 83 Age 40+ 37 62 1. Complete the 2 x 2 table and calculate a crude odds ratio for the case-control study relating smoking to cervical cancer. Label the columns& rows and interpret the result. (1 pts) Yes No Yes No Total 2. Complete the series of 2 x 2 tables and calculate an odds ratio for each age strata. Label the columns& rows. (6 pts) Yes No Yes No Yes No Total Yes Homework #3 Page 1 of 4 No Total Yes No Yes No Total 3. 4. Is age an effect modifier? Explain your answer. (1 pt) An age-adjusted odds ratio, the Mantel-Haenszel odds ratio (MHOR), was calculated using the data from the age stratified 2 X 2 tables. The MH OR was 6.3. Compare the crude odds ratio that you calculated in #1 above with this MH OR. Is age a confounder? Explain your answer. (1 pt.) Questions 5-8: Consider the following events in one man's life. Note that some events are hypothetical and some are real. Event Age (years)______ Birth Prostate cancer begins Prostate cancer is detectable by screening Man is screened, cancer is detected, and treatment begins If no screening, symptoms would have developed and cancer would have been detected If no screening, death would have occurred Death occurs 0 45 50 55 60 70 85 _____________________________________________________________________ _ 5. 6. Assume that this man did not get screened. Compute the total preclinical phase of this man's prostate cancer. (1 pt) Assume that this man did not get screened. Compute the total detectable preclinical phase of this man's prostate cancer. (1 pt) Homework #3 Page 2 of 4 7. Compute the lead time for this patient. (1 pt) 8. Did screening increase the life span of this patient? Justify your answer. (1 pt) Questions 9-12: Suppose a new screening test has been developed. The test was evaluated in a probability sample in three communities with the following results: Mooreville Test Result Positive Negative Total Health Problem Present Absent 887 888 99 7989 986 8877 Total 1775 8088 9863 Positive Negative Total Health Problem Present Absent 3634 269 404 2423 4038 2692 Total 3903 2827 6730 Positive Negative Total Health Problem Present Absent 2866 56 318 506 3184 562 Total 2922 824 3746 Short City Test Result Epiville Test Result 9. Calculate the prevalence of the health problem in each of these communities. (3 pts) 10. Calculate and interpret the sensitivity of the test for each community. (3 pts) 11. Calculate and interpret the specificity of the test for each community. (3 pts) Homework #3 Page 3 of 4 12. Calculate and interpret the predictive value positive for each community. (3 pts) Homework #3 Page 4 of 4