The Many Modes of Meta
- PDF / 1,115,564 Bytes
- 15 Pages / 504 x 719.759 pts Page_size
- 2 Downloads / 196 Views
oO92-8615/2000 Copyright 0 2000 Drug Information Association Inc.
THE MANY MODES OF META STEPHEN SENN,BA, MSc, PHD Professor of Pharmaceutical and Health Statistics, Departments of Epidemiology and Public Health and Statistical Science, University College London, London, United Kingdom
Many of the reservations that might attach to the use of meta-analysis generaLly uor example, regarding publication bias) do not apply in the specific context of drug development. A meta-analysis is, in fact, a highly natural and appropriate way to summarize the results of a drug development program, as has been recognized in the International Conference on Harmonization (ICH) E9 guideline. Since a sponsor will have access to all original data, the data from a set of clinical trials in a drug development program have a very similar (hierarchical) structure to the data from a set of centers in a single multicenter trial. Curiously, however; the controversies over analyzing multicenter trials have often been different from those in the field of meta-analysis. In this paper; the options open to the meta-analyst in drug development are examined and comparisons to approaches used in analyzing multicenter trials are made in an attempt to provide some unifying insights, in particular as regards the handling of models with interactions. Key Words: Meta-analysis; Multicenter trials; Fixed effects; Random effects; n-of- 1 trials
INTRODUCTION
fever’ (paratyphoid or typhoid fever) in the
IN THE CONTEXT OF clinical research, meta-analysis is “the process of summarizing, through formal statistical means, the results from a number of clinical trials” (1). More generally, its goal “is to combine the results of previous studies to arrive at summary conclusions about a body of research” (2). Although the term ‘meta-analysis’ was first coined by Glass (3) less than a quarter of a century ago in the field of education, the general approach goes back much further and earlier medical examples are to be found. Karl Pearson (4) reported a quantitative summary of the results of inoculation for ‘enteric
Reprint address: Stephen Senn, Professor of Pharmaceutical and Health Statistics, Department of Epidemiology and Public Health, Department of Statistical Science, University College London. 1-19 Tomngton Place, London WClE 6BT, UK. E-mail: Stephens@ public-health.ucl.ac.uk.
British Medical Journal of 1904. Typhoid was also the subject of an early (1907) analysis by Goldberger. See Winkelstein for discussion (5). In agriculture, of course, RA Fisher introduced methods for analyzing complex experiments (6) early on (in 1921), which, as will be explained below, have similarities with groups of simple experiments. Healy (7) draws attention to an important paper by Yates and Cochran (8) of 1938 which explicitly addresses the analysis of such groups. Tippet in 1931 (9) and Fisher in 1932 (10) had earlier introduced techniques for combining P-values. See Sonneman for an excellent review of such methods (11). There is no question that meta-analysis, despite fre
Data Loading...
