A Concise Guide to Observational Studies in Healthcare E-Book

Allan Hackshaw

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 490

Veröffentlichungsjahr: 2014

Beschreibung

A Concise Guide to Observational Studies in Healthcare provides busy healthcare professionals with an easy-to-read introduction and overview to conducting, analysing and assessing observational studies. It is a suitable introduction for anyone without prior knowledge of study design, analysis or conduct as the important concepts are presented throughout the text. It provides an overview to the features of design, analyses and conduct of observational studies, without using mathematical formulae, or complex statistics or terminology and is a useful guide for researchers conducting their own studies, those who participate in studies co-ordinated by others, or who read or review a published report of an observational study. Examples are based on clinical features of people, biomarkers, lifestyle habits and environmental exposures, and evaluating quality of care.

Leseprobe

CONTENTS

Cover

Title page

Copyright page

Foreword

Preface

Chapter 1: Fundamental concepts

1.1 Observational studies: purpose

1.2 Specifying a clear research question: exposures and outcomes

1.3 Types of observational studies

1.4 Strengths and limitations of the different types of study designs

1.5 Key design features

1.6 Interpreting and reporting the results and implication for public health or clinical practice

1.7 Translational research

1.8 Key points

References

Chapter 2: Outcome measures, risk factors, and causality

2.1 Types of measurements (endpoints)

2.2 ‘Counting people’ (risk)

2.3 ‘Taking measurements on people’

2.4 Time-to-event data

2.5 What could the true effect be, given that the study was conducted on a sample of people?

2.6 Understanding risk and risk factors

2.7 Risk factors and investigating association and causality

2.8 Key points

References

Chapter 3: Effect sizes

3.1 Effect sizes

3.2 What could the true effect be, given that the study was conducted on a sample of people?

3.3 Could the observed result be a chance finding in this particular study?

3.4 Simple statistical analyses

3.5 Key Points

References

Chapter 4: Regression analyses

4.1 Linear regression

4.2 Identifying and dealing with outliers

4.3 Different types of regressions

4.4 General comments on the different regression methods

4.5 Categorising exposure factors (variables)

4.6 Interpreting p-values for factors that have ≥ 3 levels

4.7 Examining several factors at the same time

4.8 Interactions between two exposures (effect modifiers)

4.9 Measuring the outcome measure more than once during the study

4.10 Checking the regression model

4.11 Missing data

4.12 Key points

References

Chapter 5: Cross-sectional studies

5.1 Purpose

5.2 Design

5.3 Measuring variables, exposures, and outcome measures

5.4 Collecting the data

5.5 Sample size

5.6 Analysing data and interpreting results

5.7 Outcome measures based on ‘counting people’ endpoints: Vocational dental practitioners (VDPs) and lifestyle habits (Box 5.2)

5.8 Outcome measures based on ‘counting people’ endpoints: COPD and passive smoking (Box 5.3)

5.9 Outcome measures based on ‘taking measurements on people’ endpoints: Quality of life in young cancer patients (Box 5.4)

5.10 Outcome measures based on ‘taking measurements on people’ endpoints: Salt intake and blood pressure among adolescents (Box 5.5)

5.11 Key points

References

Chapter 6: Case–control studies

6.1 Purpose

6.2 Design

6.3 Measuring variables, exposures, and outcomes

6.4 Collecting the data

6.5 Sample size

6.6 Analysing data and interpreting results

6.7 Outcome measures based on ‘counting people’ endpoints: Sudden infant death syndrome and sleeping factors (Box 6.2)

6.8 Measures based on ‘taking measurements on people’ endpoints: Alzheimer’s disease and head circumference (Box 6.3)

6.9 Outcome measures based on time-to-event endpoints: Venous thromboembolism (VTE) and survival (Box 6.4)

6.10 Key points

References

Chapter 7: Cohort studies

7.1 Purpose

7.2 Design

7.3 Measuring variables, exposures, and outcome measures

7.4 Collecting the data

7.5 Sample size

7.6 Analysing data and interpreting results

7.7 Outcome measures based on ‘counting people’ endpoints: Folic acid and ASD (Box 7.2)

7.8 Outcome measures based on ‘taking measurements on people’ endpoints: Lifestyle habits and body weight (Box 7.3)

7.9 Outcome measures based on time-to-event endpoints: Abdominal aorta size and hospital admission and mortality (Box 7.4)

7.10 Key points

References

Chapter 8: Quality of care studies

8.1 Purpose

8.2 Design

8.3 Measuring variables

8.4 Collecting the data

8.5 Sample size

8.6 Analysing data

8.7 Example 1: patient satisfaction with a service

8.8 Example 2: patient satisfaction with treatment

8.9 Example 3: assessing a service and improving service delivery

8.10 Key points

References

Chapter 9: Prognostic markers for predicting outcomes

9.1 Prognostic markers and models

9.2 Study design

9.3 Sample size

9.4 Analysing data and interpreting results

9.5 Examining several factors to develop a prognostic model

9.6 Genetic studies

9.7 Key points

References

Chapter 10: Systematic reviews and meta-analyses

10.1 Dealing with inconsistent study findings

10.2 Systematic combination of studies

10.3 What is a systematic review?

10.4 Interpreting systematic reviews

10.5 Considerations when reading a systematic review

10.6 Reporting systematic reviews

10.7 Key points

References

Chapter 11: Conducting and reporting observational studies

11.1 Before conducting the study

11.2 Study set-up

11.3 Conducting the study

11.4 End of study

11.5 Regulations

11.6 Reporting and publishing observational studies

11.7 Key points

References

Index

End User License Agreement

List of Tables

Chapter 02

Table 2.1 Life table for the survival data of nine participants on page 30.

Chapter 03

Table 3.1 Relative risk and absolute risk difference for disorders that have varying underlying death rates, using an observational study of smoking [2].

Table 3.2 Calculation of relative risk and odds ratio using the survey of VDPs (newly qualified dental graduates) (from Box 3.2), and how they can be similar (disorder is uncommon) or very different (disorder is common, >20%).

Table 3.3 Illustration of relative versus absolute effects.

Chapter 04

Table 4.1 Hypothetical results from an ordinal logistic regression, where the outcome measure has three levels (mild, moderate or severe asthma).

Table 4.2 The effect of three (exposure) factors on the outcome measure

‘risk of dying’

.

Table 4.3 Illustration of interaction for two exposure factors (age and sex) and an outcome measure (percentage of smokers who quit smoking).

Chapter 05

Table 5.1 Selected key results on smoking, alcohol consumption, and drug habits among male and female VDPs [3].

Table 5.2 Selected QoL results according to timing of death.

Table 5.3 Linear regression analyses between age (exposure) and ‘physical’ QoL outcome measure.

Table 5.4 Mean salt and sodium intake and systolic BP in each tertile of salt intake (in each age group).

Table 5.5 Linear regression analyses between systolic BP (outcome) and three exposure.

Chapter 06

Table 6.1 How selecting controls that have similar characteristics, and hence exposure (here smoking) status, to cases can underestimate the association, than if using general population controls.

Table 6.2 Summary table showing all three exposure groups and the odds ratios (ORs) taking into account the matching factors between the cases and controls.

Table 6.3 Factors associated with the sleeping environment.

Table 6.4 Estimated p-values for selected adjusted odds ratio in the ‘multivariable’ column of Table 6.3(using Box 6.8).

Table 6.5 Head circumference and two other potential confounding factors according to disease status. The table shows the mean values (standard deviations in brackets).

Table 6.6 Results of the multivariable linear regression analysis.

Table 6.7 Selected characteristics of cases and controls.

Chapter 07

Table 7.1 OR between two types of autistic spectrum disorder and folic acid use.

Table 7.2 Change in body weight (in pounds) over a 4-year period (averaged across several 4-year periods) in relation to selected dietary items.

Table 7.3 Selected baseline characteristics in each exposure group (size of abdominal aorta) from the study in Box 7.4.

Table 7.4 Hazard ratios for cause of death and first admission to hospital according to abdominal aorta size group (number of events shown in brackets).

Chapter 08

Table 8.1 Selected results from the study in Box 8.2.

Table 8.2 Selected results from the study in Box 8.3.

Table 8.3 Selected measures of life satisfaction and quality of care from the example in Box 8.4.

Chapter 09

Table 9.1 95% CIs for varying DR (sensitivity) and FPR (1-specificity), according to the number of individuals available for each measure.

Table 9.2 Findings from a study examining the performance of a prognostic marker (skull X-ray) in predicting the presence or absence of a cranial haematoma in patients with a head injury.

Table 9.3 Prognostic performance for fixed FPR cut-offs using Figure 9.2.

Table 9.4 Prognostic performance for SMA and its ability to predict death from oral cancer at either 1 or 3 years.

Table 9.5 Hazard ratios for selected markers from Example 9.3 (oral cancer patients).

Table 9.6 Results of a backward elimination analysis (Cox regression) of 15 potential prognostic markers for mortality among oral cancer patients.

Table 9.7 Association between a single SNP variant and a disorder (rheumatoid arthritis): Controls were healthy individuals without rheumatoid arthritis [29].

Table 9.8 Selected SNPs (and genes), found to have an association with the risk of multiple sclerosis [31].

Chapter 10

Table 10.1 Key design features and the main result from three observational studies of the association between passive smoking and the risk of lung cancer among female never-smokers.

Table 10.2 Studies examining the effect of maternal smoking on the risk of limb reduction defects (missing or deformed limbs, including missing, fused, or extra fingers or toes) [13].

List of Illustrations

Chapter 01

Figure 1.1 The effect of an exposure on an outcome, with a third factor, the confounder.

Figure 1.2 Hypothetical example of how a confounder can distort the results when examining the effect of an exposure on an outcome and how it can be allowed for.

Figure 1.3 Illustration of bias, using an example in which the aim is to estimate the proportion of people who smoke, based on self-reported measures.

Chapter 02

Figure 2.1 Histogram of the cholesterol values in 40 men, with a superimposed Normal (Gaussian) distribution curve.

Figure 2.2 Kaplan–Meier plot of the survival data for the nine participants on page 30, which can also be used to estimate survival rates and median survival: 4-year survival rate: A vertical line is drawn on the x-axis at 4, and the rate is the corresponding y-axis value where the line hits the curve, that is, 78%.Median survival: The time at which half the participants have died. A horizontal line is drawn on the y-axis at 50%, and the corresponding x-axis value (median) is where the line hits the curve, that is, 7.2 years.

Figure 2.3 Twenty studies, each estimating what the prevalence of binge drinking could be among

all

dental VDPs, with 95% CIs; 19 are expected to contain the true prevalence, but 1 (5%) is not, by chance alone.

Figure 2.4 Illustration of risk. There is no fixed value for a person; the value will change according to more information considered. The aforementioned factors relate to the risk of developing stroke, and knowing each of them will increase or decrease the risk value (like a sliding scale).

Figure 2.5 Features of a population that can influence incidence and prevalence and therefore how reliably they can be forecasted or predicted (prevalence will also depend on incidence).

Figure 2.6 Association between storks and the human birth rate (human and stork data for 17 European countries).

Figure 2.7 How time sequence is important in determining causality.

Chapter 03

Figure 3.1 Approximate guide to judging whether relative effect sizes are associated with small, moderate, or large effects.

Figure 3.2 Kaplan–Meier curves for comparing cancer patients with and without venous thromboembolism (cases and controls, respectively).

Figure 3.3 An example of survival curves that cross each other, so that the hazard ratio is unlikely to be an appropriate effect size. The hazard ratio is 0.84. The RMST up to 40 months is 13.6 months among the unexposed group and 11.9 months in the exposed group: a mean difference of +1.7 months, indicating that the unexposed group had a mean survival time 1.7 months longer than the exposed group.

Figure 3.4 How study size affects conclusions.

Figure 3.5 Two factors each influence the size of a p-value separately.

Figure 3.6 Definition of p-values.

Figure 3.7 Relationship between CIs and p-values where the distance from the effect size to the no effect value is measured in terms of standard errors. If the 95% CI excludes the appropriate no effect value then p-value should be <0.05.

Chapter 04

Figure 4.1 Scatter plot of 30 measurements of age and blood pressure.

Figure 4.2 Illustration of how a single outlier can create an association, when the other observations do not show any association. The 10 observations in (a) are the same as in (b).

Figure 4.3 Illustration of how a straight line cannot be fitted through risk (a), but it can through the logarithm of the odds (b).

Chapter 05

Figure 5.1 Bar chart showing the reported current use of recreational (including illegal) drugs among male and female VDPs separately. The bars are shown based on frequency of descending order. Current user is defined as someone who has taken cannabis more than once/twice, or uses it at least once per week, or uses any other drug at least once per month.

Figure 5.2 Hypothetical bar chart where there is a natural ordering to the categories (bars).

Figure 5.3 Diagram showing how the mean systolic BP (and 95% CI) changes with tertile of dietary salt intake, according to age group (the tertile categories are shown in the header of Table 5.5). An alternative figure is to display all the individual BP values, with the mean and standard deviations (or median and interquartile range) indicated; this would show the actual data, while the figure shows a summary of the data.

Chapter 06

Figure 6.1 General design of a case–control study. ‘Outcome’ could be any pre-defined event, such as a disorder, death, or other characteristic (e.g. did or did not stop smoking).

Figure 6.2 Information required for a sample size estimation for case–control studies in which the effect size is an OR.

Figure 6.3 How effect sizes for a factor could change after adjustment for confounders. The arrow shows the direction, moving from unadjusted (crude) estimate to adjusted (shown as bold). The 95% CIs also need to be interpreted, in terms of whether they include the no effect value or not.

Figure 6.4 The Kaplan–Meier curves. In (a), the unadjusted hazard ratio is 1.79 (95% CI 1.40–2.28), and the adjusted estimate is 1.55 (95% CI 1.21–2.00).

Chapter 07

Figure 7.1 General design of a cohort study.

Figure 7.2 How data were collected in the three cohort studies used as examples in this chapter. ‘National registry data’ do not directly involve the study participant.

Figure 7.3 Information required for a sample size estimation for cohort studies in which the effect size is a relative risk.

Figure 7.4 How body weight changes in relation to changes in diet and physical activity. Each food/drink item had a score , and a total score for ‘diet’ was calculated for each participant. A low quintile for diet indicates the 20% of participants who tend to have the ‘worst’ diets (i.e. consume foods or drinks that increase weight), and a low quintile for physical activity indicates the 20% of people who perform the least amount of physical activity.

Figure 7.5 Kaplan–Meier curves for time to hospital admission for circulatory disease, according to size of the abdominal aorta.

Chapter 08

Figure 8.1 Bar chart showing perceived reasons for discrimination or being treated badly among 706 patients out of the 963 in the study in Box 8.1 who reported this experience. The figure shows the absolute number of patients on the y-axis, but it is better to show percentages, for example, ‘Didn’t treat me equally’ would be 17.8% (126/706), and ‘Problem discussing about my expense’ would be 28.7% (203/706).

Figure 8.2 Prevalence of non-adherence to treatment among patients who switched from twice-daily drug (tacrolimus) to once daily (shift group) or those on cyclosporine or refused to switch (control group). T0 is baseline, and T3 and T6 are 3 and 6 months later, respectively. *p-value < 0.05, compared with the baseline value.

Chapter 09

Figure 9.1 Illustration of measures of prognostic performance using a study of people with insulin-dependent diabetes, in which the prognostic marker (or screening test) is a retinal photograph of the eye (abnormal or normal). The sensitivity is also 78%, and the specificity would be 99.6% (1 − FPR).

Figure 9.2 DR (sensitivity) and FPR (1 − specificity) in relation to burn size (% of TBSA). The lower diagram is a plot of DR against FPR, called an ROC curve, from which the AUC (A) is 0.81. Both diagrams can be used to develop Table 9.3.

Figure 9.3 Kaplan–Meier curves for the association between a tumour prognostic marker SMA and death from oral cancer, among oral cancer patients.

Figure 9.4 ROC curve (for 3 year survival) based on combining four markers in Example 9.3.

Figure 9.5 Validating a prognostic model: comparison of the mean predicted risk based on a model of five factors, when applied to a completely new dataset of 19,597 pregnancies, which included 47 with Down’s syndrome. The diagonal line represents perfect agreement between observed and expected prevalence.

Figure 9.6 Outline of a GWA study of multiple sclerosis, which aimed to find SNPs with strong associations with the risk of multiple sclerosis. Data from Ref. [30].

Chapter 10

Figure 10.1 Illustration of how changes in the exposure status could lead to an underestimated relative risk (using the study in Ref. [9]).

Figure 10.2 Example of a forest plot from a meta-analysis: observational studies investigating the association between dietary vitamin E intake and Parkinson’s disease (Box 10.3). The figure was obtained using RevMan. SE, standard error; cases, with Parkinson’s disease; unaffected, without Parkinson’s disease.

Figure 10.3 ‘Weight’ of a study in a meta-analysis.

Figure 10.4 Illustration of heterogeneity among four hypothetical studies. The results from Studies 1 to 3 are similar, but the result from Study 4 is clearly different from the other three.

Figure 10.5 Forest plot of mobile phone use of ≥ 10 years and the risk of cancer, based on a meta-analysis of 13 case–control studies [16], according to whether the interviewer (researcher) who obtained information from the participants (including exposure status) was blinded or not to their case–control status. .

Figure 10.6 Forest plot of observational studies examining the association between maternal smoking and the risk of limb reduction defects in babies.

Figure 10.7 Flow diagram showing how the case–control studies for the meta-analysis of mobile phone use and cancer were identified (see Figure 10.5).

Figure 10.8 Funnel plot of the seven studies of Parkinson’s disease shown in Figure 10.2. The x-axis is the odds OR, and the y-axis is the standard error based. The vertical line is the combined effect size (0.81). Small studies, with large standard errors, could be more likely to have large effects. Therefore, publication bias may be present if the funnel plot shows a very asymmetric pattern.

Chapter 11

Figure 11.1 Key aspects of a typical grant application used by a funding committee for evaluation purposes.

Guide

Cover

Table of Contents

Begin Reading

Pages

1ii

iv

vii

viii

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

30

31

32

33

34

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

69

70

71

72

73

74

75

76

77

79

80

81

135

136

137

140

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

A Concise Guide to Observational Studies in Healthcare

Allan Hackshaw

University College LondonLondonUK

This edition first published 2015, © 2015 by John Wiley & Sons, Ltd.

BMJ Books is an imprint of BMJ Publishing Group Limited, used under licence by John Wiley & Sons.

Registered OfficeJohn Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester,West Sussex, PO19 8SQ, UK

Editorial Offices9600 Garsington Road, Oxford, OX4 2DQ, UKThe Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK111 River Street, Hoboken, NJ 07030-5774, USA

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell

The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting a specific method, diagnosis, or treatment by health science practitioners for any particular patient. The publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. Readers should consult with a specialist where appropriate. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom.

Library of Congress Cataloging-in-Publication Data

Hackshaw, Allan K., author. A concise guide to observational studies in healthcare / Allan Hackshaw.  p. ; cm. Includes bibliographical references and index.

 ISBN 978-0-470-65867-3 (pbk.) I. Title. [DNLM: 1. Observational Study as Topic–methods–Handbooks. 2. Outcome and Process Assessment (Health Care)–Handbooks. 3. Epidemiologic Research Design–Handbooks. W 49] R852 610.72–dc23

    2014018407

A catalogue record for this book is available from the British Library.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Cover image: Stock Market Data: ©iStock.com /© The-Tor; Crowd: ©iStock.com/© AndreasKermann.

Foreword

Epidemiology is at the heart of medicine. Without knowledge of the epidemiology of disease and its methods of study, it can be impossible to interpret the results of observational studies. Epidemiology has an important role to play in determining causes of disease and in the interpretation of clinical tests since this depends on knowledge of the prevalence of the diseases for which such tests are done. Observational studies are the mainstay of epidemiology. Correctly interpreted, observational studies transform the unstructured natural variation of diseases and the exposures that cause them into intelligible insights that can be used to improve health and well-being. A Concise Guide to Observational Studies in Healthcare demonstrates how this is done and includes many practical examples.

It is easy to complicate epidemiology with mathematical formulae and specialist jargons that are difficult to understand. What are the differences between relative risk, odds ratio and hazard ratio? What is the difference between bias and confounding? How should a meta-analysis be presented and interpreted? Why are the terms detection rate and false positive rate better than sensitivity and specificity? What is the difference between a standard deviation and a standard error? Hackshaw carefully explains all these and more with elegance. The book succeeds in pulling together the essence of how observational studies can be used and interpreted in medical practice.

Hackshaw simplifies the principles and methods of the subject, covering a wide range of topics in a book short enough to be read over a weekend and one that will undoubtedly inspire readers to delve further into the subject.

A Concise Guide to Observational Studies in Healthcare is a useful companion to Hackshaw’s 2009 book on clinical trials. As with his previous book, this one is aimed at the general, medical and scientific reader, providing an introduction to the subject without requiring detailed specialist knowledge, an objective the author has accomplished with skill and rigour.

Professor Sir Nicholas Wald, FRS, FRCP

Wolfson Institute of Preventive Medicine

Barts and The London School of Medicine and Dentistry

Preface

Research studies are required for developing effective public health policies and clinical practice. Observational studies are perhaps the most common type of research, and they are essential for describing the characteristics of a group of people or finding ways to understand, detect, prevent or treat disease, or avert early death.

The purpose of the book is to provide researchers and health professionals with a focussed and simplified account of the main features of observational studies. It is important to first understand the key concepts. Specifics about the calculations involved in analyses should come after and are covered in other textbooks. The book is aimed at those who conduct their own studies or participate in studies coordinated by others, or to help review a published report. No prior knowledge of design, analysis or conduct is required. Examples are based on clinical features of people, biomarkers, lifestyle habits and environmental exposures, and evaluating quality of care.

This book is a companion to the book A Concise Guide to Clinical Trials (Hackshaw A, BMJ Books/Wiley-Blackwell). An overview of the key design and analytical features are provided in Chapters 1–4; then each study type is discussed using published studies (Chapters 5–8), showing how they were conducted and interpreted. Chapter 9 introduces prognostic markers, a topic which is often misunderstood, while Chapter 10 covers systematic reviews and how to deal with inconsistent results. Chapter 11 summarises how to conduct and publish an observational study.

One of the important goals of the book is to show that study features such as the design of questionnaires and interpreting results are common to most study types, so these topics are repeated throughout the chapters. By having many examples, the reader can see how a variety of study designs and outcomes can be interpreted in a similar way, which will help to reinforce key aspects.

The content is based on over 23 years of experience teaching evidence-based medicine to undergraduates, postgraduates, and health professionals; writing over 130 published articles in books and medical journals; and designing, setting up and analysing research studies for a variety of disorders. This background has provided the experience to determine what researchers need to know and how to present the relevant ideas.

I am most grateful to Jan Mackie, whose thorough editing of the book was invaluable. Final thanks to Harald Bauer.

Professor Allan Hackshaw

Deputy Director Cancer Research UK & UCL Cancer Trials CentreUniversity College London