Understanding Clinical Data Analysis: Learning Statistical Principles from Published Clinical Resear

Randomness. Basis of All Scientific MethodsIntroductionWhy Mankind Is Not Fond of Thinking in Terms of RandomnessWhy Randomness Is Good for YouThe Term Randomness Has a Different Meaning in Different Clinical Research Communications Including Research Pape rs, Study Protocols , ConsensusesRandom resultRandomized trialRandom sampleRandom variableRandom errorRandom accessRandom selectionRandom assignmentRandomnessQuasi-randomnessRandomization testsDrawing randomlyRe-randomization testsUnrandomnessStratified randomizationPresence of Randomness in Everyday LifeWhat Does the Scientific Method Look Like, Who Were the Inventors of It and Why Is It NeededRandomness Is Very Much the Basis of the Scientific MethodNull Hypothesis Testing as Compared to the Devil’s AdvocacyConclusionsReferencesRandomized and Observational Research. Writing Protocols, Making Study Data FilesIntroductionScientific RigorTrial ProtocolTypes of Protocols, GeneralCase-Control StudiesCohort StudiesDifference Between Odds Ratio and Risk RatioOther Forms of Observational ResearchRandomized ResearchMaking a Data FileSAS (Statistical Analysis System)SSPS (Statistical Package for Social Sciences)The Variables in a Data FileConclusionsReferencesRandomized Clinical Trials, History, Designs. Questionable Use of Placebos and Questionable Lack of Placebos, and Stepped Wedge DesignsIntroductionRandomized Controlled Trials (RCTs) Are Highly RegulatedClinical Trial DefinitionHistoryMain Use of Clinical Trials: Causal InferenceCounterfactual Assertion ExperimentControl in Clinical Trials by RandomizationBlinding and PlacebosRandomization MethodsClinical Trial ClassificationsExperimental Study DesignsConclusionsReferencesRandomized Clinical Trials, Analysis Sets, Statistical Analysis, Reporting Issues. Principal Features Analyses, the Cochrane Risk-of-Bias-ToolIntroductionIntention to Treat and Per Protocol AnalysesFirst Example (BMC 2007; 7: 3-10)Second Example (Curr Med Res Opin 2008; 24: 2151-7)Third Example (Neth J Med 2015; 73: 23-9)Statistical PrinciplesHypothesesStratificationsMissing ValuesSafety and TolerabilityCONSORT (Consolidated Statement of Randomized Trials)Reporting Issues Including Reporting BiasConclusionsReferencesDiscrete Data Analysis, Failure Time Data Analysis. Better Assessments of Biological and Pharmaceutical AgentsIntroductionFour Step Data Analysis, Different Hypothesis TestsHypothesis Testing One Sample Z-TestHypothesis Testing Two Sample Z-TestHypothesis Testing Two Sample Chi-Square TestHypothesis Testing Two Sample Fisher’s Exact TestSample Size Considerations for a Two-Group Clinical TrialHypothesis Testing One Sample Two MeasurementsHypothesis Testing One Sample Multiple Repeated MeasurementsFailure-Time DataConclusionsReferencesQuantitative Data Analysis. Modeling for False Positive Findings, Using Median Absolute DeviationsIntroductionA Real Data Example, Losartan Reduces Aortic Dilatation in Marfan SyndromeStep One, Data SummariesStep Two, Determining the Reliability of the Above StatisticsStep Three, Hypothesis TestingConclusionsReferencesSubgroup Analysis. European Medicines Agency’s and American Food Drug Administration’s DirectivesIntroductionInternational GuidelinesRegression Models, Many Possibilities, General FormRegression Modeling, for the Purpose of Increasing PrecisionRegression Modeling, to Deal with StratificationRegression Modeling, to Correct for ConfoundingRegression Modeling, for Assessment of Interactions/ SynergismsGood ModelsConclusionsReferencesInterim Analysis. Alpha Spending Function ApproachIntroductionIncreased Risk of Type I ErrorMethods for Lowering the Type I Error (a), the Armitage/Pocock Group Sequential MethodMethods for Lowering the Type I Error (a), the Group Sequential Method with a-Spending Function ApproachMethods for Lowering the Type I Error (a), the Group Sequential Method with Adaptive DesignsContinuous Sequential TrialsConclusionsReferencesMultiplicity Analysis. Gate Keeping Strategies and Closure PrinciplesIntroductionA Brief Review of Some Basic Hypothesis Testing Methodology with a Single Outcome VariableNull Hypothesis Testing with Multiple Outcome VariablesThe Gate Keeping Procedures for Null Hypothesis Testing with Multiple Outcome VariablesMultiple ComparisonsConclusionsReferencesMedical Statistics: A Discipline at the Interface of Biology and Mathematics. Equating Subjective Feelings with Probabilities and Providing Quality Criteria for Diagnostic TestsIntroductionStatistics Is to Prove Prior HypothesesStatistics Is to Improve the Quality of Your ResearchStatistics Is a Discipline at the Interface of Biology and MathematicsStatistics Is to Better Understand the Limitations of Clinical ResearchStatistics Is for Testing (Lack of) RandomnessStatistics Is for Providing Quality Criteria for Diagnostic Tests, General RemarksStatistics Is for Providing Quality Criteria for Diagnostic Tests, Validity, Reproducibility, and Precision of Qualitative TestsValidity of Qualitative TestsReproducibility of Qualitative TestsPrecision of Qualitative TestsStatistics for Providing Quality Criteria for Diagnostic Tests, Validity, Reproducibility, and Precision of Quantitative TestsValidity of Quantitative TestsReproducibility of Quantitative TestsPrecision of Quantitative TestsConclusionsReferencesExercise
Next >