# hw4key - 1. (50) A clinical trial to compare a...

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

1. (50) A clinical trial to compare a cholesterol-lowering drug with placebo among individuals with coronary disease or other occlusive disease, but with below average levels of LDL cholesterol (the "bad" cholesterol) is planned. The investigators hypothesize that further lowering of cholesterol in this high risk group is beneficial. The primary endpoint is total mortality and the planned follow-up is 5 years. It is expected that 10% of participants given placebo will die in 5 years. The cholesterol lowering treatment is estimated to lower the LDL cholesterol level by 40 mg/dl more than placebo. The full effect of this cholesterol lowering treatment is assumed to result in 20% fewer deaths than the placebo group. It is assumed that the full effect of the treatment on mortality will not be reached for 2 1/2 years. It is also assumed that the non-adherence rate to the capsules prescribed will be 30% in the group that receive the cholesterol-lowering drug by the end of the 5-year trial. a. (5) Using Halperin's model and Table 1 in his paper, what is the adjusted value for p e ? Answer: Halperin’s model: specifiy p c, K, dropout rate (d), lag (f) For the study above p c = 0.1, K =0.2, d=0.3 and f=0.5 T From the table I obtain adjusted p e = 0.087 Without adjustment for lag and non-adherence, p e = 0.08 b. (5) What is the combined sample size required for this study based on your estimate of p e in part a. See 1 st formula on slide #7 of Lecture 17). Assume the type I error is 0.05 (2-sided) and power = 0.90. Answer: P = = 0.0935 n = 2 = 2 = 10528, 2n = 21056 c. (5) If there was no lag, what would the adjusted value of p e be (assume the non-adherence rate to the capsules for the cholesterol-lowering drug is 30% as in part a.)? Answer: d =0.3, K = 0.2, F= 0, Pc = 0.1, So Adjusted Pe = 0.083 d. (5) If compliance was perfect (100%) in the cholesterol-lowering group, what would the adjusted value of p e be if the lag was 2 1/2 years? Answer: d =0, K = 0.2, F= 0.5T, Pc = 0.1, So Adjusted Pe = 0.085

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

View Full Document
e. (5) Would sample size be larger or smaller than the estimate in part b. if compliance to the cholesterol lowering treatment was 100% (zero dropouts in Halperin’s terminology)? Answer : Smaller f. (5) What would be the impact on sample size (larger or smaller than your answer in part b.) for this study if the estimated death rate was 7.5% for the placebo group instead of 10% (assume all other assumptions in the study description above are the same)? Answer : Larger g. (5) The investigators predict that 15% of patients assigned placebo will begin taking a cholesterol lowering treatment during follow-up. If this is taken into account in the determination of sample size, what impact would it have (just state larger, smaller, or no change in sample size estimated for part b.)? Answer
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 11/21/2011 for the course PUBH 7420 taught by Professor Ph7420 during the Spring '07 term at Minnesota.

### Page1 / 7

hw4key - 1. (50) A clinical trial to compare a...

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

View Full Document
Ask a homework question - tutors are online