This concise, easy-to-read pocket guide offers medical trainees, researchers, and clinicians at every level the perfect resource on Evidence Based Medicine (EBM). Based on the author’s many years of experience teaching EBM to medical students and medical residents at Columbia University, this handy title addresses not only all the basic concepts and issues in EBM, but also takes an example-based approach and is replete with numerous illustrations. This brief book provides readers with all the tools needed to tell the good from the bad in healthcare research. It discusses every type of study design, from the assessment of diagnostic tests to clinical trials and meta-analysis. The work also introduces readers to novel methods, such as the Bayesian analysis of clinical trials. In addition, to help readers better retain the information, the guide includes thought-provoking review questions and answers in an appendix. In all, Pocket Evidence-Based Medicine: A Survival Guide for Clinicians and Students is an ideal resource for anyone who encounters statistics in their studies or career, including clinicians, researchers, trainees in medicine and graduate students in a wide range of other disciplines
Author(s): Walter R. Palmas
Publisher: Springer
Year: 2023
Language: English
Pages: 220
City: Cham
Preface
Contents
1: The Most Basic Concepts in Biostatistics
1.1 Statistical Inference: From a Sample to the Population
1.2 Internal and External Validity
1.3 Null Hypothesis Significance Testing
1.4 The Almighty P-Value
1.5 Limitations of P-Values
1.6 More About the Meaning of the P-Value
1.7 Two-Sided P-Values Are the Norm in Medicine
1.8 Confidence Intervals
1.9 Type 1 and Type 2 Statistical Errors
1.10 We All Do Frequentist Statistics
1.11 Bayesian Statistics: Credible Intervals, Probability Estimates
1.12 Confounding
1.13 Interaction or Effect Modification
1.14 Collider Bias
2: Assessment of Diagnostic Tests
2.1 Sensitivity, Specificity, Predictive Values
2.2 Likelihood Ratios
2.3 Receiver Operating Characteristic (ROC) Curves
2.3.1 Effect of Threshold Changes on Predictive Values
2.4 F-Score
2.5 Common Biases in the Evaluation of Diagnostic Tests
2.5.1 Spectrum Bias
2.5.2 Post-Test Referral Bias
2.5.3 Biased Gold Standard
2.5.4 Highly Selected Populations
2.6 Avoiding Biases
2.7 Checklist for the Assessment of a Diagnostic Test
2.8 Screening
3: Use of a Diagnostic Test
3.1 The Two-by-Two Table in Different Scenarios
3.2 Pretest Probability Estimates
3.3 Likelihood Ratios and Fagan Nomogram
3.4 Use of Predictive Models
4: Observational Studies
4.1 Observational Studies. General Considerations
4.2 Cross-Sectional Studies
4.3 Odds Ratio
4.4 Case Control Studies
4.4.1 Importance of the Control Group
4.4.2 Ascertainment or Diagnostic Bias
4.4.3 Recall Bias
4.4.4 Interviewer Bias
4.4.5 Nested Case-Control Studies
4.5 Prospective Cohort Studies
4.5.1 Selection Bias
4.5.2 Confounding by Indication (Prescription Bias)
4.5.3 Immortal Time Bias
4.5.4 Attrition of those Susceptible
4.5.5 Protopathic Bias
4.5.6 Chronology (Secular) Bias
4.5.7 Non-Randomized Outcomes Studies
4.5.8 Checklist for Observational Studies
5: Commonly Used Statistics
5.1 Relative Risk
5.2 Relative Risk Reduction
5.3 Number Needed to Treat
5.4 Number Needed to Harm
5.5 Censoring
5.6 Kaplan-Meier Curves
5.6.1 Incorrect Kaplan-Meier Formatting
5.6.2 Censoring and Numbers at Risk
5.7 Hazard Ratio
5.7.1 Benefits of an Adjusted Hazard Ratio
5.7.2 Assessing Palliative Treatments
5.7.3 The Proportionality Assumption
5.8 The Log-Rank Statistic
5.8.1 Observed Vs. Expected Event Rates in Month 1
5.8.2 Observed Vs. Expected Event Rates in Month 2
5.8.3 Observed Vs. Expected Event Rates in Month 3
5.8.4 Obtain a Global Chi-Square Test
5.9 Odds Ratio
5.10 Vaccine Efficacy
5.11 Attributable Proportion
6: Randomized Clinical Trials
6.1 Why Do We Need Randomized Trials?
6.2 Types of Clinical Trials by Phase
6.3 Clinical Trial Registration and Compliance with Guidelines
6.4 Inclusion and Exclusion Criteria
6.5 Internal and External Validity
6.6 Type 1 and Type 2 Statistical Errors
6.7 Sample Size and Power Estimates
6.8 Randomization: The R in RCT
6.8.1 Preventing Bias
6.8.2 Reducing Confounding
6.8.3 Stratified Randomization
6.8.4 Randomization by Blocks
6.8.5 Did Randomization Prevent Confounding?
6.8.6 Clustered Randomization
6.8.7 Fixed Unequal Allocation
6.8.8 Dynamic Allocation: Adaptive Randomization and Minimization
6.9 Blinding
6.9.1 Involuntary Unmasking
6.10 Crossovers
6.11 Completeness of Follow-Up
6.12 Intention to Treat Analysis
6.13 Sequential Stopping Boundaries
6.14 Improvements in Trial Monitoring
6.15 Primary and Secondary Outcomes
6.16 Hierarchical Testing of Secondary Outcomes
6.16.1 Stepwise Hierarchical Testing
6.17 Subgroup Analysis
6.18 Adaptive Clinical Trials
6.19 Checklist for a Randomized Clinical Trial
6.20 Assessing a Negative Clinical Trial
6.21 Checklist for a Negative Clinical Trial
7: Non-inferiority Clinical Trials
7.1 Why Do We Need Non-Inferiority Trials?
7.2 A Quick Reminder about Superiority Trials
7.3 Hypotheses in Superiority Trials
7.4 Type 2 Error in Superiority Trials
7.5 A Whole New World: Hypotheses in Non-Inferiority Trials
7.6 The Relative Risk (RR) Non-Inferiority Margin
7.7 The Absolute Risk Difference (ARD) Non-Inferiority Margin
7.8 Both Relative and Absolute Risk Are Clinically Important
7.9 Type 1 Error in Non-Inferiority Trials
7.10 As Treated and Intention to Treat Analysis
7.11 Superiority Analysis in a Non-Inferiority Trial
7.12 ABSORB III: A Non-Inferiority Study that Used an Absolute Risk Difference Margin
7.13 Is a Non-Inferiority Trial Acceptable when the Outcome Is Death?
7.14 Checklist for Non-Inferiority Trials
8: Bayesian Analysis of Clinical Trials
8.1 Limitations of Conventional Statistics
8.2 Similarities with the Diagnostic Process
8.3 Prior Probability, Data, and Posterior Probability
8.4 Types of Prior Probability
8.5 Prior Probabilities and Clinical Common Sense
8.6 The Use of Excessively Optimistic Priors Should be Avoided
8.7 Advantages of the Uninformative Prior
8.8 Navigating Statistical Lingo in the Methods Section
8.9 Clinical Interpretation of Posterior Probabilities
8.10 Irrelevance of the Null Hypothesis
8.11 The Added Value of Bayesian Methods
9: Health Economics Studies
9.1 Basic Definitions
9.2 Commonly Used Outcome Measurements
9.3 Controversial Issues
9.4 Sponsorship Bias
9.5 Role of Health Economics Studies in Healthcare Policy Making
9.6 Checklist for Health Economics Studies
10: Meta-Analysis
10.1 Systematic Reviews and Meta-Analysis
10.2 Publication Search
10.3 Publication Bias
10.3.1 Funnel Plot
10.3.2 Significance Tests
10.4 Selective Reporting
10.5 Quality of Individual Studies—Risk of Bias
10.6 Testing for Heterogeneity
10.6.1 How Much Do Individual Study Results Differ from each Other?
10.7 Obtaining a Pooled Estimate
10.7.1 Fixed Effect Model
10.7.2 Random Effects Model
10.7.3 Fixed Versus Random Effects
10.8 Sensitivity Analysis
10.9 Meta-Regression
10.10 Network Meta-Analysis
10.10.1 Qualitative Assessment of Transitivity
10.10.2 Quantitative Assessment of Consistency (a.K.a. Coherence)
10.11 Meta-Analysis Checklist
11: Introduction to Artificial Intelligence Methods
11.1 Basic Definitions
11.2 Performance Analysis of Machine Learning Models
11.3 Biases in Artificial Intelligence
11.4 Quality of Medical Research Using Artificial Intelligence
11.5 Checklist for Artificial Intelligence Methods
12: Finding the Best Evidence
12.1 The Essential Components of Clinical Practice
12.2 Five Step EBM Model
12.3 The Hierarchy of Evidence
12.4 The Cochrane Collaboration
12.5 PubMed Queries
12.6 Other Reliable Sources
12.7 Google Scholar
12.8 The Future of EBM
13: Ethics of Clinical Research
13.1 Relevance
13.2 The Belmont Report
13.3 Beneficence
13.4 Justice
13.5 Respect for Autonomy
13.6 Institutional Review Boards
13.7 Conflicts of Interest
Appendix: Self-Assessment Test
Answers to Practice Questions
Open Access Evidence Based Clinical Knowledge
Online EBM Resources That Require Institutional Access
EBM Calculators, Decision Making Tools
