# hw2key - BIOST/EPI 513 Spring Quarter 2011 Dr. McKnight...

This preview shows pages 1–3. Sign up to view the full content.

1 BIOST/EPI 513 Spring Quarter 2011 Dr. McKnight HOMEWORK 2 KEY See Appendix I for STATA commands and Appendix II for output. 1. Data from the cancer and steroid hormone (CASH) study, a case-control study examining the association between oral contraceptive (OC) use and the risk of breast cancer, are given in the table below. Table entries are total counts of case and control women in each of five age groups, and counts of women in each group with prior “exposure” to oral contraceptives. Using methods we have covered in class, analyze these data to determine how strongly the data in this table support the hypothesis that oral contraceptives are associated with breast cancer incidence in young women. Then turn in answers to the following questions: (a) What is the appropriate measure of association between OC use and breast cancer status for these data? In one sentence, why? The appropriate measure is OR, because it is the only estimable measure in a case-control study. (b) Provide an estimate of this association measure, with 95% confidence interval. 1.08 (0.92, 1.27) (c) Test the hypothesis that there is no association between OC use and breast cancer risk. i. State the name of the hypothesis test you use. Mantel-Haenszel test (test of whether the common odds ratio across age strata is 1) ii. Give the value of the test statistic: 0.89 iii. Name the probability distribution to which you compare the value of the statistic to obtain a P-value. Chi-squared distribution with 1 degree of freedom iv. Give the P-value: 0.345 (d) After adjusting for age group, women who had recently used oral contraceptives were estimated to be 1.1 times as likely as women without recent oral contraceptive use to develop breast cancer (95% CI: 0.92-1.3; p = 0.345).

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
2 2. In a clinical study of surgical mortality after coronary artery bypass graft (CABG), patients were categorized based on their left ventricular ejection fraction prior to the surgery. Numbers of subjects in each of five ejection-fraction categories who died and who survived their surgery are given in the table below: Total 15 92 297 707 3913 (a) Present a table of appropriate association measures and confidence intervals to describe the association between ejection-fraction category and operative mortality. Explain why you used the association measure(s) you used. Table 1. Relative risks (RR) and associated 95% confidence intervals (CI). Ejection fraction (%) RR 95 % CI <19 1.0 ref 20-29 0.65 (0.078, 5.5) 30-39 0.25 (0.031, 2.0) 40-49 0.47 (0.067, 3.2) ≥50 0.28 (0.042, 1.9) I chose to present death risks relative to that of the lowest ejection fraction group because risks are estimable from a prospective study, and I had expected mortality to change with increased ejection fraction. Note that RRs might be more precise if a reference category with more subjects than the <19% group is used. (b)
This is the end of the preview. Sign up to access the rest of the document.

## hw2key - BIOST/EPI 513 Spring Quarter 2011 Dr. McKnight...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online