Kooperativer Bibliotheksverbund

UID:

Format: 1 online resource (x, 371 pages) : , digital, PDF file(s).

Edition: Second edition.

ISBN: 1-108-85085-5 , 1-108-85120-7 , 1-108-50011-0

Content: Medicine is becoming increasingly reliant on diagnostic, prognostic and screening tests for the successful treatment of patients. With new tests being developed all the time, a more informed understanding of the benefits and drawbacks of these tests is crucial. Providing readers with the tools needed to evaluate and interpret these tests, numerous real-world examples demonstrate the practical application and relevance of the material. The mathematics involved are rigorously explained using simple and informative language. Topics covered include the diagnostic process, reliability and accuracy of tests, and quantifying treatment benefits using randomized trials, amongst others. Engaging illustrations act as visual representations of the concepts discussed in the book, complementing the textual explanation. Based on decades of experience teaching in a clinical research training program, this fully updated second edition is an essential guide for anyone looking to select, develop or market medical tests.

Note: Title from publisher's bibliographic system (viewed on 08 May 2020). , Cover -- Half-title -- Title page -- Copyright information -- Dedication -- Contents -- Preface -- Acknowledgments -- Reference -- Chapter 1 Introduction: Understanding Diagnosis and Evidence-Based Diagnosis -- Diagnosis -- Evidence-Based Diagnosis -- Estimating Disease Probabilities -- Quantifying Treatment Effects -- Dichotomous Disease State (D+/D-): A Convenient Oversimplification -- Generic Decision Problem: Examples -- Preview of Coming Attractions -- Summary of Key Points -- References -- Problems -- References -- Chapter 2 Dichotomous Tests -- Introduction -- Definitions -- Sensitivity, Specificity, Positive, and Negative Predictive Value -- Prevalence, Pretest Probability, Posttest Probability, and Accuracy -- Importance of the Sampling Scheme -- Combining Information from the Test with Information about the Patient -- 2 x 2 Table Method for Updating Prior Probability -- Likelihood Ratios for Dichotomous Tests -- Necessary Digression: A Crash Course in Odds and Probability -- Deriving Likelihood Ratios (''Lite'' Version) -- Using the LR Slide Rule -- Treatment and Testing Thresholds -- Quantifying Costs and Benefits -- The Treatment Threshold Probability (PTT) -- Testing Thresholds for an Imperfect but Costless Test -- Visualizing Testing Thresholds -- Testing Thresholds for a Perfect but Risky or Expensive Test -- Testing Thresholds for an Imperfect and Costly Test -- Summary of Key Points -- Appendix 2.1 General Summary of Definitions and Formulas for Dichotomous Tests -- Appendix 2.2 Rigorous Derivation of Likelihood Ratios -- Appendix 2.3 Answers to Odds/Probability Conversions in Box 2.5 -- Appendix 2.4 Formulas for Testing Thresholds for Dichotomous Tests -- 2.4a For an imperfect but costless test -- Example Imperfect but costless test for influenza -- 2.4b For a perfect but costly test. , Example Perfect but costly test for influenza -- 2.4c For an imperfect and costly test -- Example Imperfect and costly test for influenza -- Appendix 2.5 Derivation of No Treat-Test and Test-Treat Probability Thresholds -- No Treat-Test Threshold -- 2.5a Geometry -- No Treat-Test Threshold -- 2.5b Algebra -- Test-Treat Threshold -- 2.5c Geometry -- Test-Treat Threshold -- 2.5d Algebra -- References -- Problems -- References -- Chapter 3 Multilevel and Continuous Tests -- Introduction -- Making a Continuous Test Dichotomous -- ROC Curves -- Area Under the ROC Curve (AUROC) -- ROC Curves for Continuous Tests -- The Walking Man Approach to Understanding ROC Curves -- Getting the Most Out of ROC Curves -- LRs for Multilevel Tests -- How ROC Curves Relate to LR -- Posterior Probability for Multilevel Tests -- Optimal Cutoff between Positive and Negative for a Multilevel Test -- ROC Curves and Optimal Cutoffs -- Regret Graphs and Multilevel Tests -- Summary of Key Points -- Appendix 3.1 Logarithms and the Likelihood Ratio Slide Rule -- How Does the LR Slide Rule Work? -- Mathematical Digression: Logarithms -- Natural Logarithms -- log(Odds) -- References -- Problems -- References -- Chapter 4 Critical Appraisal of Studies of Diagnostic Test Accuracy -- Introduction -- General Approach -- Important Biases for Studies of Diagnostic Test Accuracy -- Incorporation Bias -- Partial Verification Bias -- Effects on Sensitivity and Specificity -- Effects on Positive and Negative Predictive Value -- Differential Verification (aka Double Gold Standard) Bias -- Imperfect Gold Standard Bias -- Errors Are Conditionally Independent (Uncorrelated) -- Errors on the Gold Standard and Index Test Are Correlated -- What If There's No Gold Standard? -- Spectrum Bias -- Definition and Explanation -- Spectrum Bias vs. Disease Definition -- Underlying Continuous Disease Variable. , Underlying Categorical Disease Variable -- Potential Association between Prevalence and Spectrum of Disease -- Spectrum of Test Results - Exclusion of Intermediate Test Results -- Systematic Reviews of Diagnostic Tests -- Individual Patient Data Meta-Analysis -- Beyond Checklists -- Summary of Key Points -- References -- Problems -- References -- Chapter 5 Reliability and Measurement Error -- Introduction -- Types of Variables -- Measuring Interobserver Agreement for Categorical Variables -- Agreement -- Kappa for Dichotomous Variables -- Calculating Expected Agreement -- Understanding Expected Agreement -- Understanding the Kappa Formula -- Impact of the Marginals -- Balanced versus Unbalanced Disagreement -- Kappa versus Sensitivity and Specificity -- Kappa for Three or More Categories -- Unweighted Kappa -- Weighted Kappa -- Linear Weights -- Quadratic Weights -- Custom Weights -- Reliability of Continuous Measurements -- Test-Retest Reliability -- Within-Subject Standard Deviation and Repeatability -- Why Not Use Average Standard Deviation? -- Why Not Use the Correlation Coefficient? -- Measurement Error Proportional to Magnitude -- Method Comparison -- Calibration -- Using Studies of Reliability from the Literature -- Summary of Key Points -- Appendix 5.A Multi-Rater Kappa -- References -- Problems -- References -- Chapter 6 Risk Predictions -- Introduction -- Risk Predictions versus Diagnostic Tests -- Quantifying the Accuracy of Predictions -- Calibration -- Mean Bias, Mean Absolute Error, and Brier Score -- Discrimination -- ROC Curves and Calibration Plots -- Recalibration -- Risk Ratios, Rate Ratios, and Hazard Ratios -- Assessing the Value of Predictions -- Net Benefit Calculation -- Decision Curves -- Decision Curves vs. Regret Graphs -- Diagnostic Probability or Predicted Risk -- Critical Appraisal of Studies of Prediction. , Effects of Treatment -- Loss to Follow-Up -- Overfitting -- Publication Bias -- Quantifying New Information -- Genetic Tests -- Predicting Continuous Outcomes -- Summary of Key Points -- Appendix 6.1 Net Reclassification Index and Integrated Discrimination Improvement -- References -- Problems -- References -- Chapter 7 Multiple Tests and Multivariable Risk Models -- Introduction -- Test Independence -- Test Nonindependence and Spectrum Bias -- Combining the Results of Two Dichotomous Tests: An Example -- Combining the Results of Multiple Dichotomous Tests -- Classification Trees -- Logistic Regression -- Odds Ratios -- Logistic Regression Modeling -- Logistic Regression Using the Results of a Single Continuous Test -- Logistic Regression Using the Results of Two Continuous Tests -- Clinical Risk Models Developed Using Logistic Regression -- Selecting Tests to Include in a Risk Model -- Overfitting and the Importance of Validation -- K-Fold Cross-Validation -- Machine Learning -- The Clinician versus the Decision Rule -- Summary of Key Points -- References -- Problems -- References -- Chapter 8 Quantifying Treatment Effects Using Randomized Trials -- Introduction -- Why Do a Randomized Trial? -- Critical Appraisal of Randomized Trials -- Design and Conduct -- Authors and Funding Source -- Study Subjects -- Intervention and Comparison Group -- Blinding -- Levels of Blinding -- Assessment of Blinding -- Drawbacks of Blinding -- Outcomes -- Surrogate Outcomes -- Composite Endpoints -- Loss to Follow-Up -- Analysis -- Intention-to-Treat, As-Treated, and Per-Protocol Analyses -- Subgroup Analyses -- Multiple Outcomes -- Between-Groups versus within-Groups Comparisons -- Direction of Biases in Randomized Blinded Trials -- Quantifying Treatment Effects -- Continuous, Ordinal, and Count Outcome Variables -- Dichotomous Outcome Variables. , Relative versus Absolute Measures of Treatment Effect -- Inflating the Apparent Effect Size by Using the Odds Ratio -- Number Needed to Treat (NNT) -- Treatment Cost and Benefit per Bad Outcome Prevented (CBOP & -- BBOP) -- NNT and Treatment Threshold Probability (PTT) -- Treatment Cost per Good Outcome Caused -- Number Needed to Harm -- Summary of Key Points -- References -- Problems -- References -- Chapter 9 Alternatives to Randomized Trials for Estimating Treatment Effects -- Introduction -- Confounding by Indication -- Instrumental Variables -- Falsification Tests for Confounding or Bias -- Measuring Another Outcome -- Measuring Another Predictor -- Studying Another Patient Population -- Propensity Scores -- The Importance of Timing -- Summary -- References -- Problems -- References -- Chapter 10 Screening Tests -- Introduction -- Definition and Types of Screening -- Importance of a Critical Approach to Screening Tests -- Possible Harms from Screening -- Reasons for Excessive Screening -- Reasons for Underscreening -- Critical Appraisal of Studies of Screening Tests -- The Big Picture -- Observational Studies of Screening Tests -- Volunteer Effect (Confounding) -- Lead-Time Bias -- Length-Time Bias -- Stage Migration Bias -- Overdiagnosis (Pseudodisease) -- Randomized Trials of Screening Tests -- Total Mortality versus Cause-Specific Mortality -- Biases That Make Screening Tests Look Worse -- Back to the Big Picture -- Summary of Key Points -- References -- Problems -- References -- Chapter 11 Understanding P-Values and Confidence Intervals -- Introduction and Justification -- Background -- Two Kinds of Probability -- Review of Classical ''Frequentist'' Statistics -- Wrong and Right Definitions of P-Values -- Using Your Understanding of Diagnostic Tests to Understand P-Values -- Introduction to Bayesian Thinking: False-Positive Confusion. , Extending the Analogy.

Additional Edition: ISBN 1-108-43671-4

Language: English

URL: https://doi.org/10.1017/9781108500111