How to Read a Paper: The Basics of Evidence-Based Medicine (3 page)

In the 18 years since the first edition of this book was published, evidence-based medicine has waxed and waned in popularity. Hundreds of textbooks and tens of thousands of journal articles now offer different angles on the ‘basics of EBM’ covered briefly in the chapters that follow. An increasing number of these sources point out genuine limitations of evidence-based medicine in certain contexts. Others look at evidence-based medicine as a social movement—a ‘bandwagon’ that took off at a particular time (the 1990s) and place (North America) and spread dramatically quickly with all sorts of knock-on effects for particular interest groups.

When preparing this fifth edition, I was once again minded not to change too much apart from updating the examples and the reference lists, as there is clearly still room on the bookshelves for a no-frills introductory text. In the last (fourth edition), I also added two new chapters (on quality improvement and complex interventions), and in this latest edition I have added two more—one on applying evidence-based medicine with patients (the science of shared decision making) and another on common criticisms of EBM and responses to those. As ever, I would welcome any feedback that will help make the text more accurate, readable and practical.

Trisha Greenhalgh
January 2014

Acknowledgements

I am not by any standards an expert on all of the subjects covered in this book (in particular, I am very bad at sums), and I am grateful to the people listed here for help along the way. I am, however, the final author of every chapter, and responsibility for any inaccuracies is mine alone.

1.
To Professor Sir Andy Haines and Professor Dave Sackett who introduced me to the subject of evidence-based medicine and encouraged me to write about it.

2.
To the late Dr Anna Donald, who broadened my outlook through valuable discussions on the implications and uncertainties of this evolving discipline.

3.
To Jeanette Buckingham of the University of Alberta, Canada, for invaluable input to Chapter 2.

4.
To various expert advisers and proofreaders who had direct input to this new edition or who advised me on previous editions.

5.
To the many readers, too numerous to mention individually, who took time to write in and point out both typographical and factual errors in previous editions. As a result of their contributions, I have learnt a great deal (especially about statistics) and the book has been improved in many ways. Some of the earliest critics of
How to Read a Paper
have subsequently worked with me on my teaching courses in evidence-based practice; several have co-authored other papers or book chapters with me, and one or two have become personal friends.

6.
To the authors and publishers of articles who gave permission for me to reproduce figures or tables. Details are given in the text.

7.
To my followers on Twitter who proposed numerous ideas, constructive criticisms and responses to my suggestions when I was preparing the fifth edition of this book. By the way, you should try Twitter as a source of evidence-based information. Follow me on
@trishgreenhalgh
– and while you're at it you could try the Cochrane Collaboration on
@cochrancollab
, Ben Goldacre on
@bengoldacre
, Carl Heneghan from the Oxford Centre for Evidence Based Medicine on
@cebmblog
and the UK National Institute for Health and Care Excellence on
@nicecomms
.

Thanks also to my husband, Dr Fraser Macfarlane, for his unfailing support for my academic work and writing. My sons Rob and Al had not long been born when the first edition of this book was being written. It is a source of great pride to me that they have now read the book, applied its messages in their own developing scientific careers (one in medicine) and made suggestions on how to improve it.

Chapter 1

Why read papers at all?

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

Evidence-based medicine (EBM) is much more than just reading papers. According to the most widely quoted definition, it is ‘the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients’ [1]. I find this definition very useful but it misses out what for me is a very important aspect of the subject—and that is the use of mathematics. Even if you know almost nothing about EBM, you probably know it talks a lot about numbers and ratios! Anna Donald and I decided to be upfront about this in our own teaching, and proposed this alternative definition:

Evidence-based medicine is the use of mathematical estimates of the risk of benefit and harm, derived from high-quality research on population samples, to inform clinical decision-making in the diagnosis, investigation or management of individual patients
.

The defining feature of EBM, then, is the use of figures derived from research on
populations
to inform decisions about
individuals
. This, of course, begs the question ‘What is research?’—for which a reasonably accurate answer might be ‘Focused, systematic enquiry aimed at generating new knowledge’. In later chapters, I will explain how this definition can help you distinguish genuine research (which should inform your practice) from the poor-quality endeavours of well-meaning amateurs (which you should politely ignore).

If you follow an evidence-based approach to clinical decision-making, therefore, all sorts of issues relating to your patients (or, if you work in public health medicine, issues relating to groups of people) will prompt you to ask questions about scientific evidence, seek answers to those questions in a systematic way and alter your practice accordingly.

You might ask questions, for example, about a patient's symptoms (‘In a 34-year-old man with left-sided chest pain, what is the probability that there is a serious heart problem, and, if there is, will it show up on a resting ECG?’), about physical or diagnostic signs (‘In an otherwise uncomplicated childbirth, does the presence of meconium [indicating fetal bowel movement] in the amniotic fluid indicate significant deterioration in the physiological state of the fetus?’), about the prognosis of an illness (‘If a previously well two-year-old has a short fit associated with a high temperature, what is the chance that she will subsequently develop epilepsy?’), about therapy (‘In patients with an acute coronary syndrome [heart attack], are the risks associated with thrombolytic drugs [clot busters] outweighed by the benefits, whatever the patient’s age, sex and ethnic origin?'), about cost-effectiveness (‘Is the cost of this new anti-cancer drug justified, compared with other ways of spending limited healthcare resources?’), about patients' preferences (‘In an 87-year-old woman with intermittent atrial fibrillation and a recent transient ischaemic attack, does the inconvenience of warfarin therapy outweigh the risks of not taking it?’), and about a host of other aspects of health and health services.

Professor Sackett, in the opening editorial of the very first issue of the journal
Evidence-Based Medicine
summarised the essential steps in the emerging science of EBM [2]:

1.
To convert our information needs into answerable questions (i.e. to formulate the problem);

2.
To track down, with maximum efficiency, the best evidence with which to answer these questions—which may come from the clinical examination, the diagnostic laboratory, the published literature or other sources;

3.
To appraise the evidence critically (i.e. weigh it up) to assess its validity (closeness to the truth) and usefulness (clinical applicability);

4.
To implement the results of this appraisal in our clinical practice;

5.
To evaluate our performance.

Hence, EBM requires you not only to read papers but to read the
right
papers at the right time, and then to alter your behaviour (and, what is often more difficult, influence the behaviour of other people) in the light of what you have found. I am concerned that how-to-do-it courses in EBM too often concentrate on the third of these five steps (critical appraisal) to the exclusion of all the others. Yet if you have asked the wrong question or sought answers from the wrong sources, you might as well not read any papers at all. Equally, all your training in search techniques and critical appraisal will go to waste if you do not put at least as much effort into implementing valid evidence and measuring progress towards your goals as you do into reading the paper. A few years ago, I added three more stages to Sackett's five-stage model to incorporate the patient's perspective: the resulting eight stages, which I have called a
context-sensitive checklist for evidence-based practice
, are shown in Appendix 1 [3].

Box 1.1 Web-based resources for Evidence-based medicine

Oxford Centre for Evidence-Based Medicine
: A well-kept website from Oxford, UK, containing a wealth of resources and links for EBM.
http://cebm.net
.

National Institute for Health and Care Excellence
: This UK-based website, which is also popular outside the UK, links to evidence-based guidelines and topic reviews.
http://www.nice.org.uk
/.

National Health Service (NHS) Centre for Reviews and Dissemination
: The site for downloading the high-quality evidence-based reviews is part of the UK National Institute for Health Research—a good starting point for looking for evidence on complex questions such as ‘what should we do about obesity?’
http://www.york.ac.uk/inst/crd/
.

Clinical Evidence
: An online handbook of best evidence for clinical decisions such as ‘what’s the best current treatment for atrial fibrillation?' Produced by BMJ Publishing Group.
http://clinicalevidence.bmj.com
.

If I were to be pedantic about the title of this book, these broader aspects of EBM should not even get a mention here. But I hope you would have demanded your money back if I had omitted the final section of this chapter (Before you start: formulate the problem), Chapter 2 (Searching the literature), Chapter 15 (Implementing evidence-based practice) and Chapter 16 (Applying evidence with patients). Chapters 3–14 describe step three of the EBM process: critical appraisal—that is, what you should do when you actually have the paper in front of you. Chapter 16 deals with common criticisms of EBM.

Incidentally, if you are computer literate and want to explore the subject of EBM on the Internet, you could try the websites listed in Box 1.1. If you're not, don't worry at this stage, but do put learning/use web-based resources to on your to-do list. Don't worry either when you discover that there are over 1000 websites dedicated to EBM—they all offer very similar material and you certainly don't need to visit them all.

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

Critics of EBM might define it as ‘the tendency of a group of young, confident and highly numerate medical academics to belittle the performance of experienced clinicians using a combination of epidemiological jargon and statistical sleight-of-hand’ or ‘the argument, usually presented with near-evangelistic zeal, that no health-related action should ever be taken by a doctor, a nurse, a purchaser of health services, or a policymaker, unless and until the results of several large and expensive research trials have appeared in print and approved by a committee of experts’.

The resentment amongst some health professionals towards the EBM movement is mostly a reaction to the implication that doctors (and nurses, midwives, physiotherapists and other health professionals) were functionally illiterate until they were shown the light, and that the few who weren't illiterate wilfully ignored published medical evidence. Anyone who works face-to-face with patients knows how often it is necessary to seek new information before making a clinical decision. Doctors have spent time in libraries since libraries were invented. In general, we don't put a patient on a new drug without evidence that it is likely to work. Apart from anything else, such off-licence use of medication is, strictly speaking, illegal. Surely we have all been practising EBM for years, except when we were deliberately bluffing (using the ‘placebo’ effect for good medical reasons), or when we were ill, overstressed or consciously being lazy?

Well, no, we haven't. There have been a number of surveys on the behaviour of doctors, nurses and related professionals. It was estimated in the 1970s in the USA that only around 10–20% of all health technologies then available (i.e. drugs, procedures, operations, etc.) were evidence-based; that figure improved to 21% in 1990, according to official US statistics [4]. Studies of the interventions offered to consecutive series of patients suggested that 60–90% of clinical decisions, depending on the specialty, were ‘evidence-based’ [5]. But as I have argued elsewhere, such studies had methodological limitations [3]. Apart from anything else, they were undertaken in specialised units and looked at the practice of world experts in EBM; hence, the figures arrived at can hardly be generalised beyond their immediate setting (see section ‘Whom is the study about?’). In all probability, we are still selling our patients short quite most of the time.