How to Read a Paper. The Basics of Evidence-Based Healthcare

Foreword to the first edition by Professor Sir David Weatherall xii

Preface to the seventh edition xiv

Preface to the first edition xvii

Acknowledgements xix

Chapter 1 Why read papers at all? 1

Does ‘evidence- based medicine’ simply mean ‘reading papers in medical journals’? 1

Why do people sometimes groan when you mention evidence- based healthcare? 4

Before you start: formulate the problem 11

Exercises based on this chapter 13

References 14

Chapter 2 Searching the literature 15

The information jungle 15

What are you looking for? 16

Levels upon levels of evidence 17

Synthesised sources: systems, summaries and syntheses 18

Pre-appraised sources: synopses of systematic reviews and primary studies 21

Specialised resources 22

Primary studies: tackling the jungle 23

One-stop shopping: federated search engines 25

Using artificial intelligence to search the literature 25

Asking for help and asking around 26

Online tutorials for effective searching 26

Exercises based on this chapter 27

References 28

Chapter 3 Getting your bearings: what is this paper about? 30

The science of ‘trashing’ papers 30

Three preliminary questions to get your bearings 32

What are randomised controlled trials and why do they matter? 34

What are cohort studies? 38

What are case–control studies? 40

What are cross-sectional surveys? 40

What are case reports? 41

The traditional hierarchy of evidence 42

Exercises based on this chapter 43

References 43

Chapter 4 Assessing methodological quality 45

Was the study original? 45

Who is the study about? 46

Was the design of the study sensible? 47

Was bias avoided or minimised? 49

Was assessment ‘blind’? 54

Were preliminary statistical questions addressed? 55

A note on ethical considerations 58

Summing up 59

Exercises based on this chapter 60

References 60

Chapter 5 Statistics for the non-statistician 63

How can non-statisticians evaluate statistical tests? 63

Have the authors set the scene correctly? 65

Paired data, tails and outliers 71

Correlation, regression and causation 72

Probability and confidence 74

The bottom line (quantifying the chance of benefit and harm) 77

Summary 79

Exercises based on this chapter 79

References 80

Chapter 6 Papers that report clinical trials of simple interventions 82

What is a clinical trial? 82

Drug trials: ‘evidence’ and marketing 83

Making decisions about therapy 86

Surrogate endpoints 87

What information to expect in a paper describing a randomised controlled trial: the CONSORT statement 91

Getting worthwhile evidence from pharmaceutical representatives 91

A note on vaccine trials 94

Exercises based on this chapter 95

References 95

Chapter 7 Papers that report trials of complex interventions 99

Complex interventions 99

Ten questions to ask about a paper describing a complex intervention 101

Exercises based on this chapter 106

References 107

Chapter 8 Papers that report diagnostic or screening tests 109

Ten suspects in the dock 109

Validating diagnostic tests against a gold standard 110

Ten questions to ask about a paper that claims to validate a diagnostic or screening test 115

Likelihood ratios 119

Clinical prediction models 122

Exercises based on this chapter 124

References 125

Chapter 9 Papers that summarise other papers (systematic reviews and meta-analyses) 128

When is a review systematic? 128

Evaluating systematic reviews: five questions to ask 131

Meta-analysis for the non-statistician 137

Explaining heterogeneity 142

New approaches to systematic review 145

Exercises based on this chapter 146

References 146

Chapter 10 Papers that advise you what to do (guidelines) 151

The great guidelines debate 151

Ten questions to ask about a clinical guideline 155

Exercises based on this chapter 162

References 162

Chapter 11 Papers that estimate what things cost (health economic evaluations) 164

What is an economic evaluation? 164

Health economics studies: two key approaches 166

Costs and benefits of health interventions 167

Measuring the value of health states 168

Quality-adjusted life-years 169

Low-value health: choosing wisely 171

Twelve questions to ask about a health economic evaluation 172

Conclusion 176

Exercises based on this chapter 176

References 177

Chapter 12 Papers that go beyond numbers (qualitative research) 179

What is qualitative research? 179

Summarising and synthesising qualitative research 183

Nine questions to ask about a qualitative research paper 184

Conclusion 191

Exercises based on this chapter 192

References 192

Chapter 13 Papers that report questionnaire research 195

The rise and rise of questionnaire research 195

Ten questions to ask about a paper describing a questionnaire study 196

Exercises based on this chapter 205

References 206

Chapter 14 Papers that report quality improvement case studies 208

What are quality improvement studies and how should we research them? 208

Ten questions to ask about a paper describing a quality improvement initiative 210

Conclusion 217

Exercises based on this chapter 217

References 218

Chapter 15 Papers that describe genetic association studies 220

The three eras of human genetic studies (so far) 220

What is a genome-wide association study? 222

Clinical applications of genome-wide association studies 225

Direct- to- consumer genetic testing 226

Mendelian randomisation studies 227

Epigenetics: a space to watch 228

Ten questions to ask about a genetic association study 230

Exercises based on this chapter 234

References 234

Chapter 16 Applying evidence with patients 237

The patient perspective 237

Patient- reported outcome measures 239

Shared decision- making 240

Option grids 243

n-of-1 trials and other individualised approaches 244

Exercises based on this chapter 246

References 247

Contents xi

Chapter 17 Papers on artificial intelligence in healthcare 249

Introduction 249

Artificial intelligence 251

Big data 253

Machine learning 254

Generative artificial intelligence: large language and multimodal models 254

Ethical principles for the use of artificial intelligence for health 255

Appraising artificial intelligence papers: a plethora of checklists 256

Ten questions to ask about a paper that reports AI studies in healthcare 260

Summary 264

Exercises based on this chapter 264

References 265

Chapter 18 EBM+: the importance of mechanistic evidence 268

What is mechanistic evidence? An example 268

The many types of mechanistic evidence and a preliminary hierarchy 269

EBM+ means ‘both and’, not ‘either or’ 270

Mechanistic evidence in the COVID-19 pandemic 272

Exercises based on this chapter 275

References 276

Chapter 19 Papers that report consensus exercises 278

Why are consensus method papers important? 279

How do experts choose and reach consensus on a specific topic? 279

Consensus methods 281

Ten questions to ask about a paper that reports a consensus statement 285

Exercises based on this chapter 290

References 291

Chapter 20 Criticisms of evidence-based healthcare 293

What’s wrong with evidence-based healthcare when it’s done badly? 293

What’s wrong with evidence-based healthcare when it’s done well? 296

Why is ‘evidence-based policymaking’ so hard to achieve? 299

Exercises based on this chapter 301

References 301

Appendix 1 Checklists for finding, appraising and implementing evidence 304

Appendix 2 Assessing the effects of an intervention 316

Index 317

Learn to assess published research in this best-selling introduction to evidence-based healthcare

Evidence-based practices have revolutionized medical care. Clinical and scientific papers have something to offer practitioners at every level of the profession, from students to established clinicians in medicine, nursing and allied professions. Novices are often intimidated by the idea of reading and appraising the research literature. How to Read a Paper demystifies this process with a thorough, engaging introduction to how clinical research papers are constructed and how to evaluate them. Now fully updated to incorporate new areas of research, readers of the seventh edition of How to Read a Paper will also find:

• A careful balance between the principles of evidence-based healthcare and clinical practice
• New chapters covering consensus methods, mechanistic evidence, big data and artificial intelligence
• Detailed coverage of subjects like assessing methodological quality, systemic reviews and meta-analyses, qualitative research, and more.

How to Read a Paper is ideal for all healthcare students and professionals seeking an accessible introduction to evidence-based healthcare – particularly those sitting undergraduate and postgraduate exams and preparing for interviews.

Trisha Greenhalgh is a general practitioner and Professor of Primary Care Health Sciences and Fellow of Green Templeton College at the University of Oxford.

Paul Dijkstra is a sport and exercise medicine physician and Director of Medical Education at Aspetar Orthopaedic and Sports Medicine Hospital in Doha, Qatar. He has an academic affiliation with the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences at the University of Oxford.