Education and debate

Using research findings in clinical practice

S E Straus, deputy director,  D L Sackett, director. 

NHS Research and Development Centre for Evidence Based Medicine, Nuffield Department of Clinical Medicine, Oxford Radcliffe Hospital NHS Trust, Oxford OX3 9DU

Correspondence to: Dr Straus sharon.straus{at}clinical-medicine.ox.ac.uk

Series editors: Andrew Haines and Anna Donald

In clinical practice caring for patients generates many questions about diagnosis, prognosis, and treatment that challenge health professionals to keep up to date with the medical literature. A study of general practitioners in North America found that two clinically important questions arose for every three patients seen.¹ The challenge in keeping abreast of the medical literature is the volume of literature. General physicians who want to keep up with relevant journals face the task of examining 19 articles a day 365 days a year.²

One approach to meeting these challenges and avoiding clinical entropy is to learn how to practise evidence based medicine. Evidence based medicine involves integrating clinical expertise with the best available clinical evidence derived from systematic research.³ Individual clinical expertise is the proficiency and judgment that each clinician acquires through clinical experience and practice. Best available clinical evidence is clinically relevant research which may be from the basic sciences of medicine, but especially that derived from clinical research that is patient centred, that evaluates the accuracy and precision of diagnostic tests and prognostic markers, and the efficacy and safety of therapeutic, rehabilitative, and preventive regimens. This paper focuses on what evidence based medicine is and how it can be practised by busy clinicians.

The practice of evidence based medicine is a process of lifelong self directed learning in which caring for patients creates a need for clinically important information about diagnoses, prognoses, treatment, and other healthcare issues. The box at the bottom of the next page illustrates the five steps necessary to the practice of evidence based medicine.

	Asking answerable clinical questions

Formulating clear, focused clinical questions is a prerequisite to answering them. Four components of the question must be specified: the patient or problem being addressed; the intervention being considered (a cause, prognostic factor, or treatment); another intervention for comparison, when relevant; and the clinical outcomes of interest.⁴ The intervention could be from a clinical trial (for example, a drug) or from nature (for example, sex or age).

To illustrate how many questions may arise in the treatment of one patient consider a 65 year old man with a history of cirrhosis and ascites secondary to alcohol abuse who presents to accident and emergency with haematemesis. The patient is taking a diuretic. On examination he is disoriented and looks unwell but is afebrile. His blood pressure is 90/60 supine and 70/50 while seated; his heart rate is 100 beats per minute while supine. In addition to spider naevi and gynaecomastia he has ascites. Bowel sounds are present.

Dozens of questions may arise in treating this patient; some are summarised in the box opposite. The questions cover a wide spectrum: clinical findings, aetiology, differential diagnosis, diagnostic tests, prognosis, treatment, prevention, and self improvement.⁴ Given their breadth and number, and knowing that clinicians are likely to have only about 30 minutes in a week to address any of them,⁵ it is necessary to pare the questions down to just one. This can be done by considering the question that would be most important to the patient's wellbeing and balancing it against a number of factors including which question appears most feasible to answer, which question is most interesting to the clinician , and which question is most likely to be raised in subsequent patients and could provide information for them. For this patient the question becomes: in a patient with cirrhosis and suspected variceal bleeding does treatment with somatostatin decrease the risk of death?

Questions to be asked in treating patient with cirrhosis and haematemesis Clinical findings Which is the most accurate way of diagnosing ascites on physical examination: fluid wave or shifting dullness? Aetiology Can gastrointestinal bleeding cause confusion in a patient with cirrhosis and ascites? Differential diagnosis In a patient with cirrhosis and ascites which is most likely to cause gastrointestinal bleeding, variceal haemorrhage or peptic ulcer disease? Diagnostic tests In a patient with suspected alcohol abuse is the use of the CAGE questionnaire specific for diagnosing alcohol abuse?6 Prognosis Does gastrointestinal bleeding increase the risk of death in a patient with cirrhosis? Treatment Does treatment with somatostatin decrease the risk of death in a patient with cirrhosis and variceal bleeding? Prevention Does treatment with a  blocker decrease the risk of morbidity and mortality in a patient with cirrhosis, ascites, and varices? Self improvement To improve my understanding of the pathophysiology of ascites would I gain more from spending an hour in the library reading a textbook or spending 15 minutes on the ward computer looking at the CD ROM version of the same textbook?

	Searching for the best evidence

A focused question sharpens the search for the best evidence. Strategies that increase the sensitivity and specificity of searches have been developed and are available both in paper⁴ and electronic forms (for example, at the site established by the NHS Research and Development Centre for Evidence-Based Medicine at URL: http://cebm.jr2.ox.ac.uk). Librarians also may be helpful in guiding and assisting searches.

The types and number of resources are rapidly expanding and some of them have already undergone critical appraisal during development. Most rigorous of these are the systematic reviews on the effects of health care that have been generated by the Cochrane Collaboration, readily available as The Cochrane Library on compact disc,⁷ and accompanied by abstracts for critically appraised overviews in the Database of Abstracts of Reviews of Effectiveness created by the NHS Centre for Reviews and Dissemination.⁷ A systematic review from The Cochrane Library is exhaustive and therefore takes years to generate; reviews from the Database of Abstracts of Reviews of Effectiveness can be generated in months. Still faster is the appearance of clinical articles about diagnosis, prognosis, treatment, quality of care, and economics that pass both specific methodological standards (so that their results are likely to be valid) and clinical scrutiny for relevance and that appear in evidence based journals such as the ACP Journal Club, Evidence-Based Medicine, and Evidence-Based Cardiovascular Medicine. This selection process reduces the amount of clinical literature by 98% to the 2% that is most methodologically rigorous and useful to clinician.⁸ In these journals, the evidence is summarised in structured abstracts and a clinical expert adds commentary to the article which allows the reader to place the findings in context.

An electronic publication, Best Evidence, combines the contents of the ACP Journal Club with the contents of Evidence-Based Medicine in an easily searched compact disc.⁹ In caring for the patient with cirrhosis and gastrointestinal bleeding a search of The Cochrane Library using the term "variceal bleed" identified the Cochrane review that evaluated the use of somatostatin versus placebo or no treatment in acute bleeding oesophageal varices.¹⁰

Some evidence based materials also appear on the internet, including those of the Cochrane Collaboration (URL: http://hiru.mcmaster.ca/COCHRANE) and some sites include clinically useful evidence about diagnosis, prognosis, and treatment. For example, the site established by the NHS Research and Development Centre for Evidence-Based Medicine (URL given above) permits browsers to apply the specificity of shifting dullness and the sensitivity of a history of ankle swelling to diagnose patients thought to have ascites; this information could be used to answer some of the questions posed in the diagnosis of the patient with cirrhosis. If the foregoing strategies for gaining rapid access to evidence based medicine fail clinicians can resort to the time honoured and increasingly user friendly systems for accessing the current literature via Medline and Embase, employing methodological quality filters to maximise the yield of high quality evidence.

	Critically appraising the evidence

Once clinicians find potentially useful evidence it has to be critically appraised and its validity and usefulness determined. Guidelines have been generated to help clinicians evaluate the validity of evidence about diagnostic tests (was there an independent, blind comparison with a gold standard of diagnosis?), treatment (was the assignment of patients to treatments really randomised?), prognostic markers (was an appropriate sample of patients assembled at a uniform point in their illness?), and clinical guidelines or other strategies for improving the quality of care. ^{4 11} Worksheets for applying guidelines to determine whether findings are valid are also available (see the address of the NHS Research and Development site given above). The trend towards publishing more informative abstracts also makes it easier for clinicians to determine whether research findings are applicable to their patients.

For the patient with cirrhosis and haematemesis, an assessment of the Cochrane review finds that it is valid, and the results showed that somatostatin did not have a statistically significant effect on survival. The confidence interval for the effect on mortality was wide, suggesting that larger studies need to be done to find definitive answers.⁹

After finding an article and determining if its results are valid and useful, it is often helpful to file a summary so that it can be referred to again or passed along to colleagues. One way to do this is to prepare a one page summary that includes information on the patient, the evidence, and the clinical bottom line organised as a critically appraised topic (CAT).¹² CATmakers (for constructing, storing, and printing information on critically appraised topics, and for calculating likelihood ratios and numbers needed to treat) are becoming more widely available, as are websites where they can be stored or retrieved (see the NHS site described earlier). Information on critically appraised topics are more useful to those who produce them (clinicians who produce them become more effective in searching and critically appraising evidence) than to potential users (since the summaries undergo little peer review and may be useful mainly for their citations).

	Applying the evidence

Applying the results of critical appraisals involves the essential second element of evidence based medicine: integrating the evidence with clinical expertise and knowledge of the unique features of patients and their situations, rights, and expectations. Only after these things have been considered can we then decide whether and how to incorporate the evidence into the care of a particular patient. In the case of the patient with cirrhosis and haematemesis, there was insufficient power in the review to determine if the risk of mortality would be reduced with the use of somatostatin. The study did report that one unit of blood was saved in treating each patient, but this is unlikely to be considered cost effective. Another factor to consider is whether endoscopic services are available for sclerotherapy or ligation of varices, and if somatostatin should be used in the interim if endoscopy is not readily available. Accordingly, the decision of whether to treat the patient with somatostatin would have to grow out of a therapeutic alliance with the patient who would have to be informed about the potential risks and benefits of this treatment.

	Evaluating your performance

To complete the cycle of practising evidence based medicine clinicians should evaluate their own performance. Clinicians can evaluate their progress at each stage by asking whether their questions were answerable, by asking if good evidence was found quickly, by asking if evidence was effectively appraised, and by asking whether the integration of the appraisal with clinical expertise and the patient's unique features left them with a rational, acceptable management strategy. This fifth step of self evaluation allows clinicians to focus on earlier steps that may need improvement in the future. For example, for the patient with cirrhosis and haematemesis we can assess the application of the evidence about the treatment of variceal bleeding and determine whether we discussed the risks and benefits of treatment with the patient and whether the patient's own values were incorporated into our discussion.

	Conclusions

Can medical practice be evidence based? Recent audits have been encouraging; a general medicine service at a district general hospital affiliated with a university found that 53% of patients admitted to the service received primary treatments that had been validated in randomised controlled trials or systematic reviews of randomised controlled trials; an additional 29% of patients received care based on convincing non-experimental evidence.¹³ Three quarters of all of the evidence had been immediately available in the form of critically appraised topic summaries, and the remaining quarter was identified and applied by asking answerable questions at the time of admission, rapidly finding good evidence, quickly determining its validity and usefulness, swiftly integrating it with clinical expertise and each patient's unique features, and offering it to the patients. Similar results have been found in a study performed at a psychiatric hospital,¹⁴ general practitioners' office,¹⁵ and a paediatric surgery department.¹⁶

Practising evidence based medicine is one way for clinicians to keep up to date with the exponential growth in medical literature, not just by more efficient browsing but by improving our skills in asking questions, finding the best evidence, critically appraising it, integrating it with our clinical expertise and our patients' unique features, and applying the results to clinical practice. When added to conscientiously practised clinical skills and constantly developing clinical expertise, sound external evidence can be applied efficiently and effectively to our patients' problems.

	Acknowledgments

Funding: SES is supported by the R Samuel McLaughlin Foundation.

Conflict of interest: None.

	References

1. Covell DG, Uman GC, Manning PR. Information needs in office practice: are they being met? Ann Intern Med 1985; 103: 596-599.
2. Davidoff F, Haynes RB, Sackett DL, Smith R. Evidence-based medicine: a new journal to help doctors identify the information they need. BMJ 1995; 310: 1085-1086[Free Full Text].
3. Sackett DL, Rosenberg WMC, Gray JAM, Richardson WS. Evidence based medicine: what it is and what it isn't. BMJ 1996; 312: 71-72[Free Full Text].
4. Sackett DL, Richardson WS, Rosenberg WMC, Haynes RB. Evidence-based medicine: how to practice and teach EBM. London: Churchill Livingstone , 1997.
5. Sackett DL. Using evidence-based medicine to help physicians keep up-to-date. Serials 1996; 9: 178-181.
6. Bush B, Shaw S, Cleary P, Delbanco TL, Aronson MD. Screening for alcohol abuse using the CAGE questionnaire. Am J Med 1987; 82: 231-235[Medline].
7. The Cochrane Library [database on disk and CD ROM]. Cochrane Collaboration. Oxford: Update Software; 1998. Updated quarterly.
8. Sackett DL, Haynes RB. 13 steps, 100 people, 1 000 000 thanks. Evidence Based Med 1997; 2: 101-102.
9. Best Evidence [database on CD ROM]. Philadelphia: American College of Physicians , 1996.
10. Gotzsche PC. Somatostatin or octreotide vs placebo or no treatment in acute bleeding oesophageal varices. In: Gluud C, Jorgenson T, Koretz RL, Morabito A, Pagliaro L, Poyunard T, et al, eds. Hepato-biliary module, Cochrane Database of Systematic Reviews [updated 23 April 1997]. The Cochrane Library. Cochrane Collaboration; Issue 2. Oxford: Update Software; 1998. Updated quarterly.
11. Oxman AD, Sackett DL, Guyatt GH, for the Evidence-based medicine working group. How to get started. JAMA 1993; 270: 2093-2095[Free Full Text].
12. Sauve JS, Lee HM, Farkouh ME, Sackett DL. The critically appraised topic: a practical approach to learning critical appraisal. Ann R Coll Physicians Surg Canada 1995; 28: 396-398.
13. Ellis J. Mulligan I, Rower J, Sackett DL. Inpatient general medicine is evidence based. Lancet 1995; 346: 407-410[Medline].
14. Geddes JR, Game D, Jenkins NE, Peterson LA, Pottinger GR, Sackett DL. What proportion of primary psychiatric interventions are based on randomised evidence? Qual Health Care 1996; 5: 215-217[Abstract].
15. Gill P, Dowell AC, Neal RP, Smith N, Heywood P, Wilson AK. Evidence based general practice: a retrospective study of interventions in our training practice. BMJ 1996; 312: 819-821[Abstract/Free Full Text].
16. Kenny SE, Shankar KR, Rentala R, Lamont GL, Lloyd DA. Evidence-based surgery: interventions in a regional paediatric surgical unit. Arch Dis Child 1997; 76: 50-53[Abstract/Free Full Text].