Statistical Methods for Evaluating Safety in Medical Product Development

385 Pages·2015·5.61 MB·English

by  A. Lawrence Gould

STATISTICSINPRACTICE
SeriesAdvisors
HumanandBiologicalSciences
StephenSenn
CRP-Santé,Luxembourg
EarthandEnvironmentalSciences
MarianScott
UniversityofGlasgow,UK
Industry,CommerceandFinance
WolfgangJank
UniversityofMaryland,USA
FoundingEditor
VicBarnett
NottinghamTrentUniversity,UK
StatisticsinPracticeisanimportantinternationalseriesoftextswhichprovidedetailedcover-
ageofstatisticalconcepts,methodsandworkedcasestudiesinspecificfieldsofinvestigation
andstudy.
With sound motivation and many worked practical examples, the books show in down-to-
earthtermshowtoselectanduseanappropriaterangeofstatisticaltechniquesinaparticular
practicalfieldwithineachtitle’sspecialtopicarea.
The books provide statistical support for professionals and research workers across a range
ofemploymentfieldsandresearchenvironments.Subjectareascoveredincludemedicineand
pharmaceutics;industry,financeandcommerce;publicservices;theearthandenvironmental
sciences;andsoon.
Thebooksalsoprovidesupporttostudentsstudyingstatisticalcoursesappliedtotheabove
areas. The demand for graduates to be equipped for the work environment has led to such
coursesbecomingincreasinglyprevalentatuniversitiesandcolleges.
It is our aim to present judiciously chosen and well-written workbooks to meet everyday
practicalneeds.Feedbackofviewsfromreaderswillbemostvaluabletomonitorthesuccess
ofthisaim.
Acompletelistoftitlesinthisseriesappearsattheendofthevolume.
Statistical Methods for Evaluating Safety
in Medical Product Development
Editor
A. Lawrence Gould
MerckResearchLaboratories,USA
Thiseditionfirstpublished2015
©2015JohnWileyandSonsLtd
Registeredoffice
JohnWiley&SonsLtd,TheAtrium,SouthernGate,Chichester,WestSussex,PO198SQ,UnitedKingdom
Fordetailsofourglobaleditorialoffices,forcustomerservicesandforinformationabouthowtoapplyforpermissionto
reusethecopyrightmaterialinthisbookpleaseseeourwebsiteatwww.wiley.com.
TherightoftheauthortobeidentifiedastheauthorofthisworkhasbeenassertedinaccordancewiththeCopyright,Designs
andPatentsAct1988.
Allrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystem,ortransmitted,inanyform
orbyanymeans,electronic,mechanical,photocopying,recordingorotherwise,exceptaspermittedbytheUKCopyright,
DesignsandPatentsAct1988,withoutthepriorpermissionofthepublisher.
Wileyalsopublishesitsbooksinavarietyofelectronicformats.Somecontentthatappearsinprintmaynotbeavailablein
electronicbooks.
LimitofLiability/DisclaimerofWarranty:Whilethepublisherandauthorhaveusedtheirbesteffortsinpreparingthisbook,
theymakenorepresentationsorwarrantieswithrespecttotheaccuracyorcompletenessofthecontentsofthisbookand
specificallydisclaimanyimpliedwarrantiesofmerchantabilityorfitnessforaparticularpurpose.Itissoldontheunder-
standingthatthepublisherisnotengagedinrenderingprofessionalservicesandneitherthepublishernortheauthorshallbe
liablefordamagesarisingherefrom.Ifprofessionaladviceorotherexpertassistanceisrequired,theservicesofacompetent
professionalshouldbesought.
LibraryofCongressCataloging-in-PublicationData
Statisticalmethodsforevaluatingsafetyinmedicalproductdevelopment/[editedby]LawrenceGould.
p.;cm.
Includesbibliographicalreferencesandindex.
ISBN978-1-119-97966-1(cloth)
I.Gould,Lawrence(A.Lawrence),editor.
[DNLM: 1. DrugEvaluation–methods. 2. ClinicalTrialsasTopic. 3. Models,Statistical. 4. Pharmacovigilance.
5. SafetyManagement–methods. 6. Technology,Pharmaceutical–standards. QV771]
R853.C55
615.5072′4–dc23
2014025599
AcataloguerecordforthisbookisavailablefromtheBritishLibrary.
ISBN:9781119979661
Typesetin10/12ptTimesbyLaserwordsPrivateLimited,Chennai,India
1 2015
Contents
Preface xiii
ListofContributors xv
1 Introduction 1
A.LawrenceGould
1.1 Introduction 1
1.2 Backgroundandcontext 2
1.3 Afundamentalprincipleforunderstandingsafetyevaluation 3
1.4 Stagesofsafetyevaluationindrugdevelopment 4
1.5 Nationalmedicalproductsafetymonitoringstrategy 5
1.6 Adverseeventsvsadversedrugreactions,andanoverallviewof
safetyevaluation 5
1.7 Abriefhistoricalperspectiveonsafetyevaluation 7
1.8 Internationalconferenceonharmonization 8
1.9 ICHguidelines 9
References 11
2 Safetygraphics 22
A.LawrenceGould
2.1 Introduction 22
2.1.1 Exampleandgeneralobjectives 22
2.1.2 Whatisthegraphictryingtosay? 25
2.2 Principlesandguidanceforconstructingeffectivegraphics 26
2.2.1 Generalprinciples 26
2.3 Graphicaldisplaysforaddressingspecificissues 26
2.3.1 Frequencyofadverseeventreportsoroccurrences 26
2.3.2 Timingofadverseeventreportsoroccurrences 33
2.3.3 Temporalvariationofvitalsignand
laboratorymeasurements 36
vi CONTENTS
2.3.4 Temporalvariationofcombinationsofvitalsignandlaboratory
measurements 39
2.3.5 Functional/multidimensionaldata 44
2.3.6 Multivariateoutlierdetectionwithmultiplicityadjustmentbased
onrobustestimatesofmeanandcovariancematrix 48
2.3.7 Monitoringindividualpatienttrends 53
2.4 Discussion 53
References 60
3 QSARmodeling:predictionofbiologicalactivityfromchemicalstructure 66
AndyLiawandVladimirSvetnik
3.1 Introduction 66
3.2 Data 67
3.2.1 Chemicaldescriptors 67
3.2.2 Activitydata 68
3.3 Modelbuilding 69
3.3.1 Randomforests 69
3.3.2 Stochasticgradientboosting 70
3.4 Modelvalidationandinterpretation 71
3.5 Dataexample 74
3.6 Discussion 76
References 81
4 Ethicalandpracticalissuesinphase1trialsinhealthyvolunteers 84
StephenSenn
4.1 Introduction 84
4.2 Ethicalbasics 85
4.3 Inferentialmatters 86
4.3.1 Analysisofseriousside-effects 87
4.3.2 Timingofevents 87
4.4 Designforsubjectsafety 88
4.4.1 Dosinginterval 88
4.4.2 Contemporarydosing 88
4.5 Analysis 89
4.5.1 Objectivesoffirst-in-mantrials 89
4.5.2 (In)adequacyofstatisticalanalysisplans 89
4.5.3 ‘Formal’statisticalanalyses 90
4.6 Designforanalysis 90
4.6.1 Treatmentassignmentsandtheroleofplacebo 90
4.6.2 Dose-escalationtrialdesignissues 91
4.6.3 Precisionatinterimstages 93
4.7 Somefinalthoughts 94
4.7.1 Sharinginformation 94
4.8 Conclusions 96
4.9 Furtherreading 96
References 97
CONTENTS vii
5 Phase1trials 99
A.LawrenceGould
5.1 Introduction 99
5.2 Dosedeterminedbytoxicity 101
5.2.1 Algorithmic(rule-based)approaches 101
5.3 Model-basedapproaches 104
5.3.1 BasicCRMdesign 104
5.3.2 Adaptiverefinementofdosagelist 105
5.3.3 Hybriddesigns 106
5.3.4 Comparisonswithrule-baseddesigns 107
5.4 Model-baseddesignswithmorethanonetreatment(ornon-monotonic
toxicity) 108
5.5 Designsconsideringtoxicityandefficacy 110
5.5.1 Binaryefficacyandtoxicityconsideredjointly 110
5.5.2 Useofsurrogateefficacyoutcomes 112
5.5.3 Reductionofefficacyandtoxicityoutcomestoordered
categories 112
5.5.4 Binarytoxicityandcontinuousefficacy 113
5.5.5 Timetooccurrenceofbinarytoxicityandefficacyendpoints 114
5.5.6 Determiningdosageandtreatmentschedule 115
5.6 Combinationsofactiveagents 117
5.7 Software 117
5.8 Discussion 117
References 118
6 Summarizingadverseeventrisk 122
A.LawrenceGould
6.1 Introduction 122
6.2 Summarizationofkeyfeaturesofadverseeventoccurrence 123
6.3 Confidence/credibleintervalsforriskdifferencesandratios 126
6.3.1 Metrics 126
6.3.2 Coverageandinterpretation 126
6.3.3 Binomialmodel 127
6.3.4 Poissonmodel 140
6.3.5 Computationalresults 142
6.4 Screeningforadverseevents 142
6.4.1 Outlineofapproach 146
6.4.2 Distributionalmodel 146
6.4.3 Specificationofpriors 148
6.4.4 Example 149
6.5 Discussion 151
References 177
7 Statisticalanalysisofrecurrentadverseevents 180
LiqunDiao,RichardJ.CookandKer-AiLee
7.1 Introduction 180
viii CONTENTS
7.2 Recurrentadverseeventanalysis 181
7.2.1 Statisticalmethodsforasinglesample 181
7.2.2 Recurrenteventanalysisanddeath 183
7.2.3 Summarystatisticsforrecurrentadverseevents 184
7.3 Comparisonsofadverseeventrates 185
7.4 Remarksoncomputingandanapplication 186
7.4.1 Computingandsoftware 186
7.4.2 Illustration:Analysesofbleedinginatransfusiontrial 188
7.5 Discussion 190
References 191
8 Cardiovasculartoxicity,especiallyQT/QTcprolongation 193
ArneRingandRobertSchall
8.1 Introduction 193
8.1.1 TheQTintervalasabiomarkerofcardiovascularrisk 193
8.1.2 AssociationoftheQTintervalwiththeheartrate 194
8.2 Implementationinpreclinicalandclinicaldrugdevelopment 194
8.2.1 Evaluationsfromsponsorperspective 194
8.2.2 RegulatoryconsiderationsonTQTtrials 196
8.3 Designconsiderationsfor“ThoroughQTtrials” 198
8.3.1 Selectionoftherapeuticandsupra-therapeuticexposure 198
8.3.2 Single-versusmultiple-dosestudies;co-administrationof
interactingdrugs 199
8.3.3 Baselinemeasurements 199
8.3.4 Parallelversuscross-overdesign 200
8.3.5 TimingofECGmeasurements 200
8.3.6 Samplesize 200
8.3.7 Complexsituations 200
8.3.8 TQTtrialsinpatients 201
8.4 Statisticalanalysis:thoroughQT/QTcstudy 201
8.4.1 Data 201
8.4.2 Heartratecorrection 203
8.4.3 AgeneralframeworkfortheassessmentofQTprolongation 208
8.4.4 Statisticalinference:Proofof“LackofQTprolongation” 211
8.4.5 MixedmodelsfordatafromTQTstudies 212
8.5 ExamplesofECGtrialdesignsandanalysesfromtheliterature 215
8.5.1 Paralleltrial:Nalmefene 215
8.5.2 Cross-overtrial:Linagliptin 216
8.5.3 Cross-overwithminorQTceffect:Sitagliptin 217
8.5.4 TQTstudywithheartratechangesbutwithoutQTceffect:
Darifenacin 218
8.5.5 TrialwithbothchangesinHRandQT(c):Tolterodine 218
8.5.6 Boostingtheexposurewithpharmacokineticinteractions:
Domperidone 219
CONTENTS ix
8.5.7 DoubleplaceboTQTcross-overdesign 220
8.6 Otherissuesincardiovascularsafety 220
8.6.1 Rosiglitazone 221
8.6.2 RequirementsoftheFDAguidance 221
8.6.3 Impactonthedevelopmentofantidiabeticdrugs 223
8.6.4 Generalimpactonbiomarkervalidation 224
References 224
9 Hepatotoxicity 229
DonaldC.Trost
9.1 Introduction 229
9.2 Liverbiologyandchemistry 230
9.2.1 Liverfunction 230
9.2.2 Liverpathology 232
9.2.3 Clinicallaboratorytestsforliverstatus 235
9.2.4 Otherclinicalmanifestationsofliverabnormalities 240
9.3 Drug-inducedliverinjury 240
9.3.1 Literaturereview 240
9.3.2 Livertoxicology 241
9.3.3 Clinicaltrialdesign 243
9.4 Classicalstatisticalapproachestothedetectionofhepatictoxicity 245
9.4.1 Statisticaldistributionsofanalytes 245
9.4.2 Referencelimits 245
9.4.3 Hy’sruleandotherempiricalmethods 252
9.5 Stochasticprocessmodelsforliverhomeostasis 253
9.5.1 TheOrnstein–Uhlenbeckprocessmodel 253
9.5.2 OUdataanalysis 258
9.5.3 OUmodelappliedtoreferencelimits 263
9.6 Summary 265
References 266
10 Neurotoxicity 271
A.LawrenceGould
10.1 Introduction 271
10.2 Multivariatelongitudinalobservations 272
10.3 Electroencephalograms(EEGs) 275
10.3.1 Specialconsiderations 275
10.3.2 Mixedeffectmodels 279
10.3.3 Spatialsmoothingbyincorporatingspatialrelationshipsof
channels 281
10.3.4 Explicitadjustmentformuscle-induced(non-EEG)artifacts 282
10.3.5 Potentialextensions 285
10.4 Discussion 285
References 289
x CONTENTS
11 Safetymonitoring 293
JayHerson
11.1 Introduction 293
11.2 Planningforsafetymonitoring 294
11.3 Safetymonitoring-sponsorview(masked,treatmentgroupspooled) 297
11.3.1 Frequentistmethodsformaskedorpooledanalysis 297
11.3.2 Likelihoodmethodsformaskedorpooledanalysis 298
11.3.3 Bayesianmethodsformaskedorpooledanalysis 299
11.4 Safetymonitoring-DMCview(partiallyorcompletelyunmasked) 301
11.4.1 DMCdatareviewoperations 301
11.4.2 Typesofsafetydataroutinelyreviewed 301
11.4.3 Assaysensitivity 302
11.4.4 Comparingsafetybetweentreatments 304
11.5 Futurechallengesinsafetymonitoring 312
11.5.1 Adaptivedesigns 312
11.5.2 Changesinthesettingofclinicaltrials 313
11.5.3 Newtechnologies 313
11.6 Conclusions 313
References 314
12 Sequentialtestingforsafetyevaluation 319
JieChen
12.1 Introduction 319
12.2 Sequentialprobabilityratiotest(SPRT) 320
12.2.1 WaldSPRTbasics 320
12.2.2 SPRTforasingle-parameterexponentialfamily 321
12.2.3 Aclinicaltrialexample 322
12.2.4 Applicationtomonitoringoccurrenceofadverseevents 323
12.3 Sequentialgeneralizedlikelihoodratiotests 325
12.3.1 SequentialGLRtestsandstoppingboundaries 325
12.3.2 ExtensionofsequentialGLRteststomultiparameter
exponentialfamilies 327
12.3.3 ImplementationofsequentialGLRtests 327
12.3.4 ExamplefromSection12.2.3,continued 328
12.4 Concludingremarks 330
References 331
13 Evaluationofpost-marketingsafetyusingspontaneousreportingdatabases 332
IsmaïlAhmed,BernardBégaudandPascaleTubert-Bitter
13.1 Introduction 332
13.2 Datastructure 333
13.3 Disproportionalitymethods 334
13.3.1 Frequentistmethods 334
13.3.2 Bayesianmethods 335
13.4 Issuesandbiases 337
13.4.1 Notorietybias 337