literature review january–march 2009

3
PHARMACEUTICAL STATISTICS Pharmaceut. Statist. 2009; 8: 170–172 Published online 11 May 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/pst.379 Literature Review January– March 2009 S. Krishna Padmanabhan 1 and Andrew Stone 2, ,y 1 Wyeth Research and Development, Collegeville, PA, USA 2 AstraZeneca Pharmaceuticals, Clinical Development, Alderley Park, Macclesfield, UK INTRODUCTION This review covers the following journals received during the period from January to March 2009: Applied Statistics, volume 58, part 1. Biometrical Journal, volume 51, issue 1. Biometrics, volume 65, issue 1. Biometrika, volume 96, issue 1. Biostatistics, volume 10, part 1. Clinical Trials, volume 6, part 1. Contemporary Clinical Trials, volume 30, parts 1 and 2. Drug Information Journal, volume 43, parts 1 and 2. Journal of Biopharmaceutical Statistics, volume 19, parts 1 and 2. Journal of the Royal Statistical Society, Series A, volume 172, part 1. Journal of the Royal Statistical Society, Series B, volume 71, part 1. Statistics in Medicine, volume 28, parts 1–8. Statistical Methods in Medical Research, volume 18, part 1. SELECTED HIGHLIGHTS FROM THE LITERATURE Multiplicity The January issue of the Journal of Biopharmaceu- tical Statistics contains a series of papers related to multiplicity. The main paper discusses a number of multiple testing issues in regulatory applications, which are also discussed in separate articles by a number of prominent authors in the field. Hung HMJ, Wang SJ. Some controversial multiple testing problems in regulatory applications. Journal of Biopharmaceutical Statistics 2009; 19:1–11. Potential applications in clinical trials are pre- sented when there are co-primary endpoints present. Qian H. Li. Evaluating co-primary endpoints collectively in clinical trials. Biometrical Journal 2009; 51:137–145. Meta-analyses This paper provides comprehensive review of approaches to random-effects meta-analysis. The authors provide a clear set of recommendations, describe the possible objectives of a random-effects meta-analysis and potential pitfalls in interpretation Higgins JPT et al. A re-evaluation of random effects meta-analysis. Journal of the Royal Statistical Society A 2009; 172:137–159. y E-mail: [email protected] *Correspondence to: Andrew Stone, AstraZeneca Pharma- ceuticals, Clinical Development, Alderley Park, Macclesfield, SK10 4TG, UK. Copyright r 2009 John Wiley & Sons, Ltd.

Upload: s-krishna-padmanabhan

Post on 06-Jul-2016

217 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Literature Review January–March 2009

PHARMACEUTICAL STATISTICS

Pharmaceut. Statist. 2009; 8: 170–172

Published online 11 May 2009 in Wiley InterScience

(www.interscience.wiley.com) DOI: 10.1002/pst.379

Literature Review January–March 2009

S. Krishna Padmanabhan1 and Andrew Stone2,�,y1Wyeth Research and Development, Collegeville, PA, USA2AstraZeneca Pharmaceuticals, Clinical Development, Alderley Park,

Macclesfield, UK

INTRODUCTION

This review covers the following journalsreceived during the period from January to March2009:

� Applied Statistics, volume 58, part 1.� Biometrical Journal, volume 51, issue 1.� Biometrics, volume 65, issue 1.� Biometrika, volume 96, issue 1.� Biostatistics, volume 10, part 1.� Clinical Trials, volume 6, part 1.� Contemporary Clinical Trials, volume 30, parts

1 and 2.� Drug Information Journal, volume 43, parts 1

and 2.� Journal of Biopharmaceutical Statistics,

volume 19, parts 1 and 2.� Journal of the Royal Statistical Society,

Series A, volume 172, part 1.� Journal of the Royal Statistical Society,

Series B, volume 71, part 1.� Statistics in Medicine, volume 28, parts 1–8.� Statistical Methods in Medical Research,

volume 18, part 1.

SELECTED HIGHLIGHTS FROMTHE LITERATURE

Multiplicity

The January issue of the Journal of Biopharmaceu-tical Statistics contains a series of papers related tomultiplicity. The main paper discusses a number ofmultiple testing issues in regulatory applications,which are also discussed in separate articles by anumber of prominent authors in the field.

� Hung HMJ,Wang SJ. Some controversial multipletesting problems in regulatory applications. Journalof Biopharmaceutical Statistics 2009; 19:1–11.

Potential applications in clinical trials are pre-sented when there are co-primary endpoints present.

� Qian H. Li. Evaluating co-primary endpointscollectively in clinical trials. BiometricalJournal 2009; 51:137–145.

Meta-analyses

This paper provides comprehensive review ofapproaches to random-effects meta-analysis. Theauthors provide a clear set of recommendations,describe the possible objectives of a random-effectsmeta-analysis and potential pitfalls in interpretation

� Higgins JPT et al. A re-evaluation of randomeffects meta-analysis. Journal of the RoyalStatistical Society A 2009; 172:137–159.yE-mail: [email protected]

*Correspondence to: Andrew Stone, AstraZeneca Pharma-ceuticals, Clinical Development, Alderley Park, Macclesfield,SK10 4TG, UK.

Copyright r 2009 John Wiley & Sons, Ltd.

Page 2: Literature Review January–March 2009

The following article describes meta-analyses oflongitudinal data and present the advantagesobtained when individual patient data are incor-porated due to the ability to correctly model thecorrelation between repeated observations.

� Jones AP et al. Meta-analysis of individualpatient data versus aggregate data from long-itudinal clinical trials.Clinical Trials 2009; 6:16–27

Missing data

The handling of missing data is a crucial con-sideration in any analysis. The following papercompares two approaches to the handling ofmissing data at a pre-specified timepoint followinga series of repeat observations per patient. LastObservation Carried Forward (LOCF) and aMixed-Effects Regression Model (MMRM) arecompared. The MMRM analysis is a repeatedmeasures approach that allows estimation ofseparate treatment effects for each timepoint bymeans of the inclusion of a treatment-by-timeinteraction. The paper presents a series of simula-tions and re-analyses of 48 trials submitted inNDAs for neurological psychiatric drug productsdatasets.

� Siddiqui O et al. MMRM vs. LOCF: acomprehensive comparison based on simula-tion study and 25 NDA datasets. Journal ofBiopharmaceutical Statistics 2009; 19:227–246.

Adaptive designs

There is an ever expanding body of literature onadaptive designs. An example adaptive doseresponse design is presented for a phase I, dose-escalating study, where, subject to some con-straints, the number of patients per dose level ischosen to optimize estimation of specific para-meters from an Emax model. In doing so theauthors also provide a useful summary of variousversions of the Emax model.

� Leonov S, Miller S. An adaptive optimaldesign for the Emax model and its applicationin clinical trials. Journal of BiopharmaceuticalStatistics 2009; 19:360–385.

In many trials the design goal is to find the doseassociated with a certain target toxicity rate(Phase I) or the dose with a certain mean valueof a continuous response (dose-finding). In thisarticle, a unified dose-finding design for any dose-finding study with a target dose is described.

� Ivanova A, Kim SH. Dose finding for con-tinuous and ordinal outcomes with a mono-tone objective function: a unified approach.Biometrics 2009; 65:307–315.

This paper is presented as a ‘Tutorial inBiostatistics’ with a focus on adaptive designs forconfirmatory clinical trials. The authors reviewadaptive design methodologies for a single nullhypothesis and how to perform adaptive designswith multiple hypotheses using closed test proce-dures. They report the results of an extensivesimulation study evaluating the operating char-acteristics of the various methods.

� Bretz F et al. Adaptive designs for confirma-tory clinical trials. Statistics in Medicine2009;28:1181–1217.

Pharmacoepidemiology

The February edition of Statistical Methods inMedical Research contains a series of articles onPharmacoepidemiology that is becoming increas-ingly relevant to the pharmaceutical statistician.The issue contains a series of articles that serve asan excellent introduction to the area especially thefollowing.

� Hanley JA, Dendukuri N. Efficient samplingapproaches to address confounding in data-base studies. Statistical Methods in MedicalResearch 2009;18:81–105.

Miscellaneous (communication)

The authors recommend a format for commu-nicating an estimate with its standard error orconfidence interval. The format reinforces that the

Copyright r 2009 John Wiley & Sons, Ltd. Pharmaceut. Statist. 2009; 8: 170–172DOI: 10.1002/pst

Literature Review 171

Page 3: Literature Review January–March 2009

associated variability is an inseparable componentof the estimate and it substantially improvesclarity in tabular displays.

� Louis TA, Zeger SL. Effective communicationof standard errors and confidence intervals.Biostatistics 2009; 10:1–2.

Data analysis

Analytical data are often subject to left-censoringwhen the actual values to be quantified fall belowthe limit of detection. The primary interest of thispaper is statistical inference for the two-sampleproblem. The authors develop a nonparametricpoint and interval estimation procedure for thelocation shift model. A large set of simulationscompares 14 methods including naive, parametric,

and nonparametric methods. Additionally, a realdata example is given followed by discussion.

� Zhang D et al. Nonparametric methods formeasurements below detection limit. Statisticsin Medicine 2009; 28:700–715.

Regulatory

The author considers and compares an indirectmethod and a direct method to evaluate thetreatment effect in a bridging study by borrowingthe strength of effect observed in the foreign trial.Issues considered include a prospectively plannedevaluation of global treatment effect along with apre-specified region-specific treatment effect.

� Wang SJ. Bridging study versus prespecifiedregions nested in global trials. Drug Informa-tion Journal 2009;43:27–34.

Copyright r 2009 John Wiley & Sons, Ltd. Pharmaceut. Statist. 2009; 8: 170–172DOI: 10.1002/pst

172 Literature Review