1 / 29

Design and Analysis of Clinical Study 4. Sample Size Determination

Design and Analysis of Clinical Study 4. Sample Size Determination. Dr. Tuan V. Nguyen Garvan Institute of Medical Research Sydney, Australia. Practical Difference vs Statistical Significance. Outcome Group A Group B Improved 9 18 No improved 21 12 Total 30 30

calix
Download Presentation

Design and Analysis of Clinical Study 4. Sample Size Determination

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Design and Analysis of Clinical Study 4. Sample Size Determination Dr. Tuan V. Nguyen Garvan Institute of Medical Research Sydney, Australia

  2. Practical Difference vs Statistical Significance Outcome Group A Group B Improved 9 18 No improved 21 12 Total 30 30 % improved 30% 60% Outcome Group A Group B Improved 6 12 No improved 14 8 Total 20 20 % improved 30% 60% Chi-square: 5.4; P < 0.05 “Statistically significant” Chi-square: 3.3; P > 0.05 “Statistically insignificant”

  3. The Classical Hypothesis Testing • Define a null hypothesis (H0) and a null hypothesis (H1) • Collect data (D) • Estimate p-value = P(D | H0) • If p-value > a, accept H0; if p-value < a, reject H0

  4. P-value là gì ? “Alendronate treatment was associated with a 5% increase in BMD compared to placebo (p<0.05)” • It has been proved that alendronate is better than placebo? • If the treatment has no effect, there is less than a 5% chance of obtaining such result • The observed effect is so large that there is less than 5% chance that the treatment is no better than placebo • I don’t know

  5. P value is NOT • the likelihood that findings are due to chance • the probability that the null hypothesis is true given the data • P-value is 0.05, so there is 95% chance that a real difference exists • With low p-value (p < 0.001) the finding must be true • The lower p-value, the stronger the evidence for an effect

  6. P-value • Grew out of quality control during WWII • Question: the true frequency of bad bullets is 1%, what is the chance of finding 4 or more bad bullets if we test 100 bullets? • Answer: With some maths (binomial theorem), p=2% So, p-value is the probability of getting a result as extreme (or more extreme) than the observed value given an hypothesis

  7. Process of Reasoning The current process of hypothesis testing is a “proof by contradiction” If the null hypothesis is true, then the observations are unlikely. The observations occurred ______________________________________ Therefore, the null hypothesis is unlikely If Tuan has hypertension, then he is unlikely to have pheochromocytoma. Tuan has pheochromocytoma ______________________________________ Therefore, Tuan is unlikely to have hypertension

  8. What do we want to know? • Clinical P(+ve | Diseased): probability of a +ve test given that the patient has the disease P(Diseased | +ve): probability of that the patient has the disease given that he has a +ve test • Research P(Significant test | No association): probability that the test is significant given that there is no association P(Association | Significant test): probability that there is an association given that the test statistic is significant

  9. Diagnostic and statistical reasoning

  10. Case study • Một phụ nữ 45 tuổi • Không có tiền sử ung thư vú trong gia đình • Đi xét nghiệm truy tầm ung thư bằng mammography • Kết quả dương tính • Hỏi: Xác suất người phụ nữ này bị ung thư là bao nhiêu? - Tần suất ung thư trong quần thể Các yếu tố liên quan cần biết để có câu trả lời: - Độ nhạy của test chẩn đoán - Độ đặc hiệu của test chẩn đoán

  11. Giải đáp 10000 Phụ nữ Kết quả +ve= +ve mammography -ve= -ve mammography Tần suất ung thư trong quần thể: 1% Ung thư 100 Không 9900 Độ nhạy của xét nghiệm: 95% Độ đặc hiệu xét nghiệm: 90% +ve 95 -ve 5 +ve 990 -ve 8910 Tổng số bệnh nhân +ve: 95 + 990 =1085 Xác suất bị K nếu có kết quà dương tính: 95/1085= 0.087

  12. Suy luận trong nghiên cứu khoa học 10000 nghiên cứu Vit C +ve: p <0.05 -ve: p>0.05 Vit C = giả dược, 50% Hiệu quả 5000 Không 5000 Power: 80% Alpha: 5% +ve 4000 -ve 1000 +ve 250 -ve 4750 β 1-β α 1-α Sai lầm loại II Sai lầm loại I Tổng số kết quả nghiên cứu +ve: 4000+250 =4250 Xác suất Vit C có hiệu quả vói điều kiện +ve kết quả: 4000/4250= 0.94

  13. What Are Required for Sample Size Estimation? • Parameter (or outcome) of major interest • Blood pressure • Magnitude of difference in the parameter • 10 mmHg is an important difference / effect • Variability of the parameter • Standard deviation of blood pressure • Bound of errors (type I and type II error rates) • Type I error = 5% • Type II error = 20% (or power = 80%)

  14. The Normal Distribution 0.95 Prob. Z1 Z2 0.80 0.84 1.28 0.90 1.28 1.64 0.95 1.64 1.96 0.99 2.33 2.81 0.95 Z2 Z1 1.64 -1.96 1.96 0 0 0.05 0.025 0.025

  15. The Normal Deviates Alpha Za/2 c 0.20 1.28 0.10 1.64 0.05 1.96 0.01 2.81 Power Z1-b 0.80 0.84 0.90 1.28 0.95 1.64 0.99 2.33

  16. Study Design and Outcome • Single population • Two populations • Continuous measurement • Categorical outcome • Correlation

  17. Sample Size for Estimating a Population Proportion • How close to the true proportion • Confidence around the sample proportion. • Type I error. • N = (Za/2)2p(1-p) / d2 • p: proportion to be estimated. • d: the accuracy of estimate (how close to the true proportion). • Za/2: A Normal deviate reflects the type I error. • Example: The prevalence of obesity is thought to be around 20%. We want to estimate the preference p in a community within 1% with type I error of 5%. • Solution N = (1.96)2 (0.2)(0.8) / 0.012 = 6146 individuals.

  18. Effect of Accuracy • Example: The prevalence of disease in the general population is around 30%. We want to estimate the prevalence p in a community within 2% with 95% confidence interval. • N = (1.96)2 (0.3)(0.7) / 0.022 = 2017 subjects.

  19. Sample Size for Difference between Two Means • Hypotheses: Ho: m1 = m2 vs. Ha: m1 = m2 + d • Let n1and n2be the sample sizes for group 1 and 2, respectively; N = n1+ n2; r = n1/ n2;s: standard deviation of the variable of interest. • Then, the total sample size is given by: • If we let Z = d/s be the “effect size”, then: • If n1 = n2 ,power = 0.80, alpha = 0.05, then (Za + Z1-b)2 = (1.96 + 1.28)2 = 10.5, then the equation is reduced to: Where Za and Z1-b are Normal deviates

  20. Sample Size for Two Means vs.“Effect Size” For a power of 80%, significance level of 5%

  21. Sample Size for Difference between 2 Proportions • Hypotheses: Ho: p1 = p2 vs. Ha: p1 = p2 + d . • Let p1and p2 be the sample proportions (e.g. estimates of p1 and p2) for group 1 and group 2. Then, the sample size to test the hypothesis is: Where: n = sample size for each group ; p = (p1 + p2) / 2 ; Za and Z1-b are Normal deviates A better (more conservative) suggestion for sample size is:

  22. Sample Size for Difference Between 2 Prevalence • For most diseases, the prevalence in the general population is small (e.g. 1 per 1000 subjects). Therefore, a difference formulation is required. • Let p1and p2 be the prevalence for population 1 and population 2. Then, the sample size to test the hypothesis is: Where: n = sample size for each group; Za and Z1-b are Normal deviates.

  23. Sample Size for Two Proportions: Example • Example: The preference for product A is expected to be 70%, and for product B 60%. A study is planned to show the difference at the significance level of 1% and power of 90%. • The sample size can be calculated as follows: • p1= 0.6; p2= 0.7; p = (0.6 + 0.7)/2 = 0.65; Z0.01 = 2.81; Z1-0.9 = 1.28. • The sample size required for each group should be: • Adjusted / conservative sample size is:

  24. Sample Size for Two Proportions vs. Effect Size Difference from p1 by: 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 P1 0.1 424 131 67 41 28 19 14 10 0.2 625 173 82 47 30 20 14 9 0.3 759 198 89 50 30 19 13 8 0.4 825 206 89 47 28 17 . . 0.5 825 198 82 41 22 . . . 0.6 759 173 67 31 . . . . 0.7 625 131 45 . . . . . 0.8 424 73 . . . . . . Note: these values are “unadjusted” sample sizes

  25. Sample size for Estimating an Odds Ratio • In case-control study the data are usually summarized by an odds ratio (OR), rather then difference between two proportions. • If p1 and p2 are the proportions of cases and controls, respectively, exposed to a risk factor, then: • If we know the proportion of exposure in the general population (p), the total sample size N for estimating an OR is: • Where r = n1 / n2 is the ratio of sample sizes for group 1 and group2; p is the prevalence of exposure in the controls; and OR is the hypothetical odds ratio. If n1 = n2 (so that r = 1) then the fomula is reduced to:

  26. Sample Size for an Odds Ratio: Example • Example: The prevalence of vertebral fracture in a population is 25%. It is interested to estimate the effect of smoking on the fracture, with an odds ratio of 2, at the significance level of 5% (one-sided test) and power of 80%. • The total sample size for the study can be estimated by:

  27. Some Comments • The formulae presented are theoretical. • They are all based on the assumption of Normal distribution. • The estimator [of sample size] has its own variability. • The calculated sample size is only an approximation. • Non-response must be allowed for in the calculation.

  28. Computer Programs • Software program for sample size and power evaluation • PS (Power and Sample size), from Vanderbilt Medical Center. This can be obtained from me by sending email to (t.nguyen@garvan.org.au). Free. • On-line calculator: • http://ebook.stat.ucla.edu/calculators/powercalc/ • References: • Florey CD. Sample size for beginners. BMJ 1993 May 1;306(6886):1181-4 • Day SJ, Graham DF. Sample size and power for comparing two or more treatment groups in clinical trials. BMJ 1989 Sep 9;299(6700):663-5. • Miller DK, Homan SM. Graphical aid for determining power of clinical trials involving two groups. BMJ 1988 Sep 10;297(6649):672-6 • Campbell MJ, Julious SA, Altman DG. Estimating sample sizes for binary, ordered categorical, and continuous outcomes in two group comparisons. BMJ 1995 Oct 28;311(7013):1145-8. • Sahai H, Khurshid A. Formulae and tables for the determination of sample sizes and power in clinical trials for testing differences in proportions for the two-sample design: a review. Stat Med 1996 Jan 15;15(1):1-21. • Kieser M, Hauschke D. Approximate sample sizes for testing hypotheses about the ratio and difference of two means. J Biopharm Stat 1999 Nov;9(4):641-50.

More Related