ebook img

Innovative Statistical Methods for Public Health Data PDF

354 Pages·2015·5.116 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Innovative Statistical Methods for Public Health Data

ICSA Book Series in Statistics Series Editors: Jiahua Chen · Ding-Geng (Din) Chen Ding-Geng (Din) Chen Jeff rey Wilson Eds. Innovative Statistical Methods for Public Health Data ICSA Book Series in Statistics SeriesEditors JiahuaChen DepartmentofStatistics UniversityofBritishColumbia Vancouver Canada Ding-Geng(Din)Chen UniversityofNorthCarolina ChapelHill,NC,USA Moreinformationaboutthisseriesathttp://www.springer.com/series/13402 Ding-Geng (Din) Chen • Jeffrey Wilson Editors Innovative Statistical Methods for Public Health Data 123 Editors Ding-Geng(Din)Chen JeffreyWilson SchoolofSocialWork ArizonaStateUniversity UniversityofNorthCarolina Tempe,AZ,USA ChapelHill,NC,USA ISSN2199-0980 ISSN2199-0999 (electronic) ICSABookSeriesinStatistics ISBN978-3-319-18535-4 ISBN978-3-319-18536-1 (eBook) DOI10.1007/978-3-319-18536-1 LibraryofCongressControlNumber:2015946563 SpringerChamHeidelbergNewYorkDordrechtLondon ©SpringerInternationalPublishingSwitzerland2015 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartof thematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation, broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformation storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodology nowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade. Printedonacid-freepaper SpringerInternational PublishingAGSwitzerlandispartofSpringerScience+Business Media(www. springer.com) To myparentsandparents-in-lawwho value higher-educationandhard-working,and to mywife, Ke, myson,JohnD. Chen, andmy daughter,JennyK. Chen, for theirloveand support.This bookisalsodedicated tomy passiontotheAppliedPublicHealth StatisticsSection inAmericanPublicHealth Association! Ding-Geng(Din)Chen, Ph.D. To mygrandson,WillemWilson-Ellis,andmy threedaughters,Rochelle,Roxanne,and Rhondafortheircontinuedloveandsupport. To thegraduatestatisticsstudentsatArizona StateUniversitypresentand pastfortheir directand indirectsupport. JeffreyWilson,Ph.D. Preface This book was originated when we were become part of the leadership of the AppliedPublicHealthStatisticsSection(https://www.apha.org/apha-communities/ member-sections/applied-public-health-statistics) in the American Public Health Association. Professor Chen was the Chair-Elect (2012), Chair (2013), and Past- Chair(2014)whileProfessorWilsonwastheChair-Elect(2011),Chair(2012)and Past-Chair(2013).Inaddition,ProfessorWilsonhasbeentheChairoftheEditorial BoardoftheAmericanJournalofPublicHealthforthepast3yearsandamember for5years.HehasbeenareviewerforthejournalandacontributortoStatistically Speaking. DuringourleadershipoftheStatisticsSection,wealsoservedasAPHAProgram PlannersfortheSectionintheannualmeetingsbyorganizingabstractsandscientific sessionsaswellassupportingthestudentpapercompetition.Duringourtenure,we gotaclose-upviewoftheexpertiseandtheknowledgeofstatisticalprinciplesand methods that need to be disseminated to aid the development and growth in the area of Public Health. We were convinced that this can be best met through the compilationofabookonthepublichealthstatistics. This book is a compilation of present and new developments in statistical methods and their applications to public health research. The data and computer programsusedinthisbookarepubliclyavailablesothereadershavetheopportunity toreplicatethemodeldevelopmentanddataanalysesaspresentedineachchapter. This is certain to facilitate learning and supportease of computationso that these newmethodscanbereadilyapplied. The book strives to bring together experts engaged in public health statistical methodology to present and discuss recent issues in statistical methodological developmentand their applications. The book is timely and has high potential to impactmodeldevelopmentanddataanalysisofpublichealthresearchacrossawide spectrumofthediscipline.Weexpectthebooktofostertheuseofthesenovelideas inhealthcareresearchinPublicHealth. Thebookconsistsof15chapterswhichwepresentinthreeparts.PartIconsists ofmethodstomodelclustereddata;PartIIconsistsofmethodstomodelincomplete ormissingdata;whilePartIIIconsistsofotherhealthcareresearchmodels. vii viii Preface Part I, Modelling Clustered Data, consists of five chapters. Chapter 1 is on MethodsforAnalyzingSecondaryOutcomesinPublicHealthCase–ControlStud- ies. This chapter unlike what is common does not deal with the analysis of the associationbetweentheprimaryoutcomeandexposurevariablesbutdealswiththe association between secondary outcomes and exposure variables. The analysis of secondary outcomes may suffer from selection bias but this chapter presents and compares a design-based and model-based approach to account for the bias, and demonstratesthemethodsusingapublichealthdataset. Chapter 2: Controlling for Population Density Using Clustering and Data WeightingTechniquesWhenExaminingSocialHealthandWelfareProblems.This chapter provides an algebraic weight formula (Oh and Scheuren 1983), in path analysistoelucidatetherelationshipbetweenunderlyingpsychosocialmechanisms and health risk behaviors among adolescents in the 1998 NLSY Young Adult cohort.Theoversamplingofunderrepresentedracial/ethnicgroupsiscontrolledby mathematically adjusting the design weights in the calculation of the covariance matricesforeachclustergroupwhilecomparingnon-normalizedversusnormalized path analysis results. The impact of ignoring weights leading to serious bias in parameterestimateswiththeunderestimationofstandardserrorsispresented. Chapter 3: On the Inference of Partially Correlated Data with Applications to PublicHealthIssues.ThischapterprovidesseveralmethodstocomparetwoGaus- sian distributed means in the two-sample location problem under the assumption ofpartially dependentobservations.For categoricaldata, tests of homogeneityfor partially matched-pairdata are investigated.Differentmethodsof combiningtests ofhomogeneitybasedonPearsonChi-squaretestandMcNemarchi-squaredtestare investigated.Inaddition,severalnonparametrictestingprocedureswhichcombine allcasesinthestudyareintroduced. Chapter4:ModelingTime-DependentCovariatesinLongitudinalDataAnalyses. This chapter discusses the effect of the time-dependent covariates on response variables for longitudinal data. The consequences of ignoring the time-dependent nature of variables in models are discussed by considering various common analysistechniques,suchasthemixed-modelingapproachortheGEE(generalized estimatingequations)approach. Chapter 5: Solving Probabilistic Discrete Event Systems with Moore-Penrose Generalized Inverse Matrix Method to Extract Longitudinal Characteristics from Cross-Sectional Survey Data. This chapter presents the Moore-Penrose (M-P) generalized inverse matrix theory as a powerful approach to solve an admissible linear-equation system when the inverse of the coefficient matrix does not exist. Thischapterreportstheauthors’worktosystemizetheProbabilisticDiscreteEvent Systemsmodelingincharacterizinghealthriskbehaviorswithmultipleprogression stages.Theestimatedresultswiththisapproacharescientificallystrongerthanthe originalmethod. PartII,ModellingIncompleteorMissingData,consistsoffourchapters.Chap- ter6:OntheEffectofStructuralZerosinRegressionModels.Thischapterpresents anextensionofmethodsinhandlingsamplingzerosasopposedtostructuralzeros when these zeros are part of the predictors. They present updated approaches Preface ix andillustrate the importanceof disentanglingthe structuraland samplingzerosin alcoholresearchusingsimulatedaswellasrealstudydata. Chapter7:ModelingBased onProgressivelyType-IIntervalCensoredSample. Inthischapter,severalparametricmodelingprocedures(includingmodelselection) are presented with the use of maximum likelihood estimate, moment method estimate, probability plot estimate, and Bayesian estimation. In addition, the model presentation of general data structure and simulation procedurefor getting progressivelytype-Iintervalcensoredsamplearepresented. Chapter8:TechniquesforAnalyzingIncompleteDatainPublicHealthResearch. This chapter deals with the causes and problems created by incomplete data and recommendstechniquesforhowtohandleitthroughmultipleimputation. Chapter9:AContinuousLatentFactorModelforNon-ignorableMissingData. This chapter presents a continuous latent factor model as a novel approach to overcome limitations which exist in pattern mixture models through the speci- fication of a continuous latent factor. The advantages of this model, including smallsamplefeasibility,aredemonstratedbycomparingwithRoy’spatternmixture model using an application to a clinical study of AIDS patients with advanced immunesuppression. In Part III, we present a series of Healthcare Research Models which consists of six chapters. Chapter 10: Health Surveillance. This chapter deals with the applicationofstatisticalmethodsforhealthsurveillance,includingthoseforhealth care quality monitoring and those for disease surveillance. The methods rely on techniquesborrowedfrom the industrialquality controland monitoringliterature. However,thedistinctionsaremadewhennecessaryandtakenintoaccountinthese methods. Chapter 11:StandardizationandDecompositionAnalysis:A UsefulAnalytical Method for Outcome Difference, Inequality and Disparity Studies. This chapter dealswithatraditionaldemographicanalyticalmethodthatiswidelyusedforcom- paring rates between populations with difference in composition. The results can be readily appliedto cross-sectionaloutcome comparisonsas well as longitudinal studies.WhileSDAdoesnotrelyontraditionalassumptions,itisvoidofstatistical significancetesting.Thischapterpresentstechniquesforsignificancetesting. Chapter 12:Cusp CatastropheModelingin Medicaland Health Research.This chapter presents the cusp catastrophe modeling method, including the general principleandtwoanalyticalapproachestostatisticallysolvingthemodelforactual dataanalysis:(1)thepolynomialregressionmethodforlongitudinaldataand(2)the likelihood estimation method for cross-sectional data. A special R-based package “cusp”isgivenforthelikelihoodmethodfordataanalysis. Chapter 13: On Ranked Set Sampling Variation and Its Applications to Public HealthResearch. Thischapterpresentsthe rankedset samplingasa cost-effective alternativeapproachtotraditionalsamplingschemes.Thismethodreliesonasmall fractionoftheavailableunits.Itimprovestheprecisionofestimation.InRSS, the desiredinformationisobtainedfromasmallfractionoftheavailableunits. Chapter 14: Weighted Multiple Testing Correction for Correlated Endpointsin Survival Data. In this chapter, a weighted multiple testing correction method for

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.