Abstract
Confidence intervals are commonly used to assess the precision of parameter estimations. Particularly in small clinical trials, such assessment may be used in place of a power calculation. We discuss `future' confidence interval prediction with binomial outcomes for small clinical trials and sample size calculation, where the term `future' confidence interval emphasizes the confidence interval as a function of a random sample that is not observed at the planning stage of a study. We propose and discuss three probabilistic approaches to future confidence interval prediction when the sample size is small. We demonstrate substantial differences among these approaches in terms of the interval width prediction and sample size calculation. We show that the approach based on the expectation of the boundaries has the most desirable properties and is easy to implement. In this chapter, we primarily discuss prediction of the Clopper-Pearson exact confidence interval, and then extend our discussion to other confidence interval methods. In particular, we discuss the arcsine transformation as a viable alternative to the exact confidence interval.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Agresti, A., & Coull, B. A. (1998). Approximate is better than “exact” for interval estimation of binomial proportions. The American Statistician, 52(2), 119–126.
Anscombe, F. J. (1948). The transformation of Poisson, binomial and negative-binomial data. Biometrika, 35(3/4), 246–254.
Blaker, H. (2000). Confidence curves and improved exact confidence intervals for discrete distributions. Canadian Journal of Statistics, 28(4), 783–798.
Blyth, C. R., & Still, H. A. (1983). Binomial confidence intervals. Journal of the American Statistical Association, 78(381), 108–116.
Bougnoux, P., Hajjaji, N., Ferrasson, M. N., Giraudeau, B., Couet, C., & Le Floch, O. (2009). Improving outcome of chemotherapy of metastatic breast cancer by docosahexaenoic acid: A phase II trial. British Journal of Cancer, 101(12), 1978–1985.
Breiman, L. (1992). Probability. Philadelphia, PA: SIAM.
Brown, L. D., Cai, T. T., & DasGupta, A. (2001). Interval estimation for a binomial proportion. Statistical Science, 16, 101–117.
Chang, M. N., & O’Brien, P. C. (1986). Confidence intervals following group sequential tests. Controlled Clinical Trials, 7(1), 18–26.
Clopper, C. J., & Pearson, E. S. (1934). The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika, 26(4), 404–413.
Freeman, M. F., & Tukey, J. W. (1950). Transformations related to the angular and the square root. The Annals of Mathematical Statistics, 21, 607–611.
Levy, E. I., Ecker, R. D., Horowitz, M. B., Gupta, R., Hanel, R. A., Sauvageau, E., et al. (2006). Stent-assisted intracranial recanalization for acute stroke: Early results. Neurosurgery, 58(3), 458–463.
Martin-Schild, S., Hallevi, H., Shaltoni, H., Barreto, A. D., Gonzales, N. R., Aronowski, J., et al. (2009). Combined neuroprotective modalities coupled with thrombolysis in acute ischemic stroke: A pilot study of caffeinol and mild hypothermia. Journal of Stroke and Cerebrovascular Diseases, 18(2), 86–96.
Newcombe, R. G., & Vollset, S. E. (1994). Confidence intervals for a binomial proportion. Statistics in Medicine, 13(12), 1283–1285.
Pires, A. M., & Amado, C. (2008). Interval estimators for a binomial proportion: Comparison of twenty methods. REVSTAT–Statistical Journal, 6(2), 165–197.
Rao, C. R. (1973). Linear statistical inference and its applications. New York: Wiley.
Rino, Y., Yukawa, N., Murakami, H., Wada, N., Yamada, R., Hayashi, T., et al. (2010). A phase II study of S-1 monotherapy as a first-line combination therapy of S-1 plus cisplatin as a second-line therapy, and weekly paclitaxel monotherapy as a third-line therapy in patients with advanced gastric carcinoma: A second report. Clinical Medicine Insights: Oncology, 4, CMO-S3920.
Sahai, H., & Ageel, M. I. (2000). Analysis of variance: Fixed, random and mixed models. Ann Arbor, MI: Sheridan Books.
Schiffer, F., Johnston, A. L., Ravichandran, C., Polcari, A., Teicher, M. H., Webb, R. H., et al. (2009). Psychological benefits 2 and 4 weeks after a single treatment with near infrared light to the forehead: A pilot study of 10 patients with major depression and anxiety. Behavioral and Brain Functions, 5(1), 46.
Smith, W. S., Sung, G., Starkman, S., Saver, J. L., Kidwell, C. S., Gobin, Y. P., et al. (2005). Safety and efficacy of mechanical embolectomy in acute ischemic stroke: Results of the MERCI trial. Stroke, 36(7), 1432–1438.
Sokalr, R. R., & Rohlf, F. J. B. (1981). The principles and practice of statistics in biological research. New York: WH Freeman.
Stallard, N. (1998). Sample size determination for phase II clinical trials based on Bayesian decision theory. Biometrics, 54, 279–294.
TIMI Study Group*. (1985). The thrombolysis in myocardial infarction (TIMI) trial: Phase I findings. New England Journal of Medicine, 312(14), 932–936.
Vollset, S. E. (1993). Confidence intervals for a binomial proportion. Statistics in Medicine, 12(9), 809–824.
Wang, W. (2014). An iterative construction of confidence intervals for a proportion. Statistica Sinica, 24, 1389–1410.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
The following R codes are to calculate the predicted width of confidence intervals proposed in Sects. 11.2 and 11.3. The parameters needed are n, p, and alpha for the sample size, true success rate, and 100(1-alpha)% confidence interval. The results of each function consist of predicted lower, upper bounds, and width.
R codes for Sect. 11.2
#####Approach 1###### approach1<-function(n,p,alpha){ x<-n*p lc<-1/(1+(n-x+1)/(x*qf(0.025,2*x,2*(n-x+1)))) rc<-1/(1+(n-x)/((x+1)*qf(0.975,2*(x+1),2*(n-x)))) cat(round(lc,4),round(rc,4),round((rc-lc),4), fill=T) } #Usage approach1(20,0.07,0.05) #####Approach 2-1#### approach2.1<-function(n,p,alpha){ x<-seq(0,n,1) probs<-pbinom(x,n,p) L<-max(x[probs<=(alpha/2)],-0.5)+0.5 U<-min(min(x[probs>=(1-alpha/2)])+1-0.5,n) cat(round(L/n,4),round(U/n,4), round((U/n-L/n),4), fill=T) } #Usage approach2.1(20,0.07,0.05) #####Approach 2-2#### #Functions to solve Eq. (11.6) incbetaL<-function(x,p,n,alpha){ duhaeyo<-0 ele1<-x+1 ele2<-n for(i in ele1:ele2){ duhaeyo<-duhaeyo+factorial(ele2)/(factorial(i)*factorial (ele2-i))* p^i*(1-p)^(ele2-i) } crit<-abs(duhaeyo-(1-alpha/2)) return(crit) } incbetaU<-function(x,p,n,alpha){ duhaeyo<-0 ele1<-x ele2<-n for(i in ele1:ele2){ duhaeyo<-duhaeyo+factorial(ele2)/(factorial(i)*factorial (ele2-i))* p^i*(1-p)^(ele2-i) } crit<-abs(duhaeyo-(alpha/2)) return(crit) } approach2.2<-function(nn,pp,alphaa){ nval<-nn pval<-pp alphaval<-alphaa x<-nval*pval clow<-1/(1+(nval-x+1)/(x*qf((alphaval/2),2*x,2*(nval-x+1)))) cupp<-1/(1+(nval-x)/((x+1)*qf((1-alphaval/2),2*(x+1), 2*(nval-x)))) lowval<-optimize(incbetaL, c(((clow-0.1)*nval), ((clow+0.1)*nval)),p=pval, n=nval, tol = 0.0001, alpha=alphaval) uppval<-optimize(incbetaU, c(((cupp-0.1)*nval), ((cupp+0.1)*nval)),p=pval, n=nval, tol = 0.0001, alpha=alphaval) lc<-max(lowval[[1]]/nval,0) rc<-min(uppval[[1]]/nval,1) cat(c(lc,rc,(rc-lc))) } #Usage approach2.2(20,0.07,0.05) ####Approach 3####### approach3<-function(n,p,alpha){ truep<-p x<-seq(0,n,1) probs<-dbinom(x,n,truep) counts<-0 clc<-c() crc<-c() for(i in 1:(n+1)){ lc<-binom.test(x[i], n, conf.level = (1-alpha))$conf.int[1] rc<-binom.test(x[i], n, conf.level = (1-alpha))$conf.int[2] clc<-c(clc,lc*probs[i]) crc<-c(crc,rc*probs[i]) } cat(round(sum(clc),4),round(sum(crc),4),round((sum(crc)-sum(clc)), 4), fill=T) } #Usage approach3(20,0.07,0.05)
R codes for Sect. 11.3
#Approach 1 approach1<-function(n,p,alpha){ c<-qnorm((1-alpha/2)) x<-n*p lc<-max(sin(asin(sqrt((3/8+x)/(n+3/4)))-c/(2*sqrt(n)))^2,0) rc<-min(sin(asin(sqrt((3/8+x)/(n+3/4)))+c/(2*sqrt(n)))^2,1) cat(round(lc,4),round(rc,4),round((rc-lc),4)) } #Approach 2 approach2<-function(n,p,alpha){ c<-qnorm((1-alpha/2)) lc<-max(((sin(asin(sqrt(p))-c/(2*sqrt(n)))^2)*(1+3/(4*n)) -3/(8*n)),0) rc<-min(((sin(asin(sqrt(p))+c/(2*sqrt(n)))^2)*(1+3/(4*n)) -3/(8*n)),1) cat(round(lc,4),round(rc,4),round((rc-lc),4)) } #Approach 3 approach3<-function(n,p,alpha){ c<-qnorm((1-alpha/2)) truep<-p x<-seq(0,n,1) probs<-dbinom(x,n,truep) counts<-0 clc<-c() crc<-c() for(i in 1:length(probs)){ lc<-max(sin(asin(sqrt((3/8+x[i])/(n+3/4)))-c/(2*sqrt(n)))^2,0) rc<-min(sin(asin(sqrt((3/8+x[i])/(n+3/4)))+c/(2*sqrt(n)))^2,1) clc<-c(clc,lc*probs[i]) crc<-c(crc,rc*probs[i]) } crc<-c(crc,1*probs[(n+1)]) cat(round(sum(clc),4),round(sum(crc),4),round((sum(crc) -sum(clc)),4)) }
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Yu, J., Vexler, A. (2018). Predicting Confidence Interval for the Proportion at the Time of Study Planning in Small Clinical Trials. In: Zhao, Y., Chen, DG. (eds) New Frontiers of Biostatistics and Bioinformatics. ICSA Book Series in Statistics. Springer, Cham. https://doi.org/10.1007/978-3-319-99389-8_11
Download citation
DOI: https://doi.org/10.1007/978-3-319-99389-8_11
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-99388-1
Online ISBN: 978-3-319-99389-8
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)