gms | German Medical Science

GMDS 2012: 57. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie

16. - 20.09.2012, Braunschweig

Utilizing historical information from adult and pediatric clinical trials: Results, limitations, and extensions of the meta-analytic predictive approach

Meeting Abstract

Search Medline for

  • Raphael Koch - Institute of Biostatistics and Clinical Research (IBKF), University of Münster, Deutschland
  • Joachim Gerß - Institute of Biostatistics and Clinical Research (IBKF), University of Münster, Deutschland
  • Andreas Faldum - Institute of Biostatistics and Clinical Research (IBKF), University of Münster, Deutschland
  • Wolfgang Köpcke - Institute of Biostatistics and Clinical Research (IBKF), University of Münster, Deutschland

GMDS 2012. 57. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e.V. (GMDS). Braunschweig, 16.-20.09.2012. Düsseldorf: German Medical Science GMS Publishing House; 2012. Doc12gmds132

DOI: 10.3205/12gmds132, URN: urn:nbn:de:0183-12gmds1328

Published: September 13, 2012

© 2012 Koch et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by-nc-nd/3.0/deed.en). You are free: to Share – to copy, distribute and transmit the work, provided the original author and source are credited.


Outline

Text

In pediatric clinical trials the number of available patients is often restricted. The necessary sample size for the usual requirements of confirmatory clinical trials can’t be achieved. Thus, in a future pediatric study it is a valuable option to utilize historical information from previous adult and pediatric trials to increase the power or decrease the required sample size. Bayesian methods represent a flexible way to utilize external information in new clinical trials [1], [2]. Neuenschwander [3] and Gerss [4] use a meta-analytic predictive approach to summarize historical information and to determine the amount of evidence. The aim is to assess the number of patients which can be saved for future recruitment (“prior effective sample size”).

Hierarchical models and Markov chain Monte Carlo (MCMC) methods are applied to perform Bayesian meta-analyses and to estimate the study-specific effect in a future pediatric trial. The posterior predictive distribution of the study-specific effect will be used as informative prior distribution in the new trial. The information contained herein is translated into the “prior effectiv sample size”. The “prior effective sample size” is directly affected by the between-study variation of the historical trials. Non-informative prior distributions for hyperparameters are applied to emphasize the observed historical data [5], [6]. Hence, model fitting problems could occur, especially in the case of few historical trials or very heterogeneous study outcomes. The prior distributions affect the posterior predictive distributions more than the observed data. Consequently, the Markov chains of the estimated parameters contain several extreme values. The corresponding means, variances and therefore the “prior effective sample size” are biased.

Various results of the meta-analytic predictive approach will be presented based on clinical study data. Further, the influence of different prior distributions on the “prior effective sample size” will be illustrated. In the case of normal-distributed endpoints, a trimming method for MCMC simulations is suggested. The Markov chains of all parameters are α-trimmed (α=0.001) with respect to the predictive study effect in a future pediatric trial. By this approach extreme values in the simulated Markov chains are removed.


References

1.
Spiegelhalter DJ, Abrams KR, Myles JP. Bayesian Approaches to Clinical Trials and Health-Care Evaluation. New York: Wiley; 2004.
2.
Goodman SN, Sladky JT. A Bayesian approach to randomized controlled trials in children utilizing information from adults: the case of Guillain-Barre. Clin Trials. 2005;2(4):305-10.
3.
Neuenschwander B, Capkun-Niggli G, Branson M, Spiegelhalter DJ. Summarizing historical information on controls in clinical trials. Clin Trials. 2010;7(1):5-18. DOI: 10.1177/1740774509356002 External link
4.
Gerss J, Male C, Benda N and Koepcke W. Ein Bayes-Ansatz zur Planung einer pädiatrischen klinischen Studie, in deren Auswertung die vorliegenden Daten aus Studien an Erwachsenen eingebracht werden. In: Schmuecker P, Ellsässer KH, Hayna S, editors. Abstractband der 55. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (GMDS) e.V.; 2010 Sep 5-9; Mannheim, Germany. Dietzenbach: Antares-Computer-Verl.; 2010. p. 500-1.
5.
Lambert PC, Sutton AJ, Burton PR, Abrams KR, Jones DR. How vague is vague? A simulation study of the impact of the use of vague prior distributions in MCMC using WinBUGS. Stat Med. 2005;24(15):2401-28. DOI: 10.1002/sim.2112 External link
6.
Gelman A. Prior distributions for variance parameters in hierarchical models. Bayesian Analysis. 2006;1:1-19.
7.
Lunn D, Spiegelhalter D, Thomas A, Best N. The BUGS project: Evolution, critique and future directions. Stat Med. 2009;28(25):3049-67. DOI: 10.1002/sim.3680 External link