Evidence-based Medicine: What Does it Really Mean?

Traditionally, medical practice has been based on observations, knowledge, and experience. It is clear that each of these factors remains, and always will remain, extremely important. But, today, there is a growing concern that these three factors alone are not enough if medical practice is to serve the needs of individuals in a more focused and effective manner.

Observations, for instance, have normally been personal observations drawn from individual clinical practice. It is inevitable that for any particular condition they are in most cases limited in number, and usually restricted to a particular population group.

Knowledge of the basic mechanisms of disease and pathophysiology is essential to guide practice. Health care providers need information but many of their needs are not met. In many parts of the world, and particularly in developing countries, the textbooks used during training are often the only information health care providers have. Moreover, the textbooks are frequently out of date and there are limited, if any, opportunities for health workers to participate in continuing education. One landmark meta-analysis showed, for instance, that over 13 years elapsed between presentation of clear proof that thrombolytic drugs save lives after heart attack and the recommending of this treatment in the majority of cardiology texts and review articles. The cardiology authors simply could not keep abreast with randomized controlled trials appearing in nearly 100 different journals. Clinical guidelines can be helpful or dangerous, depending on how they are developed.

Health workers who wish to increase their knowledge and understanding often have little access to opportunities to update what they learned after their basic training. Medical journals are few and far between in developing countries, and those mailed from developed countries are expensive and arrive erratically. In any case, information on a particular topic may be scattered in different papers in numerous journals, making it very difficult for health practitioners to get a good overview of all the data available on a given subject. Access to medical information over the Internet is only now beginning to increase. While a Medline search provides information that is easy to access for those with the relevant technology, the output is hard to synthesize. In addition, the Medline content is limited to abstracts, which may be incomplete. Consequently, the knowledge and performance of health care workers in developing countries are given few opportunities to improve.

The health worker also seeks the opinions of experts who have more experience in a particular area. When a recent survey asked physicians what they would do when faced with a challenging clinical problem, the most common response was to consult a respected authority. Unfortunately, authorities are not always the best source of advice and may not be completely objective in their views.

The shift to evidence-based medicine reflects not only the constant desire to improve practice but also the developments that have taken place in clinical research, methodology and the increase in medical literature. 

Randomized controlled trials

Clinical research, for instance, now makes extensive use of randomized controlled trials because these are the most reliable method of assessing the effectiveness of preventive or therapeutic health care interventions. Randomized controlled trials are experiments in which investigators randomly allocate eligible people or health care units into groups to receive, or not to receive, the intervention(s) that are being compared. When the sample size is adequate, randomization ensures baseline comparability of prognostic variables, whether known or unknown.

However, the most appropriate research design depends on the health problem or question that is faced. For example, if the objective is to evaluate the accuracy of a diagnostic test, cross-sectional studies of patients suspected of having the disorder are required. Similarly, questions about prognosis can be answered by follow-up studies of patients having the disorder and corresponding controls.

Meta-analysis

In terms of methodology, meta-analysis is used to integrate results from more than one study to produce a summary estimate of the treatment effect across a number of studies. Meta-analysis is an application of a statistical technique used in observational studies (case-control studies and cohort studies) during stratified analysis. In a meta-analysis in a systematic review of randomized controlled trials, however, each stratum is an individual randomized controlled trial.

Meta-analysis is only an analytical tool in a systematic review and not all systematic reviews necessarily include a meta-analysis. In the presence of disparities among trials, meta-analysis can help by stratifying different characteristics so as to identify the sources of the disparities. Meta-analysis is conducted in a systematic review when the review includes more than one trial, although it does not necessarily follow that a summary estimate of the treatment effect is obtained. When there are clinical or biological disparities (heterogeneity) between trials, then using meta-analysis to produce a single summary estimate may be misleading and should be avoided.

Increase in medical literature 

The world has many hundreds of medical journals, and thousands of journals with content that relates to medical issues and concerns. The amount of material being made available is rapidly increasing, particularly as the Internet makes it possible to publish more and to publish more quickly. This explosion in the amount of medical literature has been accompanied by the development of electronic databases so that material can be better classified and more easily retrieved. At the same time, methods of searching papers have improved so that it is now easier to find a large range of papers on a specific topic. However, when conducting a review of papers it must be remembered that the results will only be robust and conclusive if the trials included are of sufficient quality.

Evidence-based medicine

Evidence-based medicine is the systematic, scientific and explicit use of current best evidence in making decisions about the care of individual patients. It is based on the following assumptions:

• clinical experience is crucial but systematic observations are necessary in order to summarize evidence;

• knowledge of the basic mechanisms of disease is also necessary but is an insufficient guide for selecting treatments for clinical practice; and

• understanding certain rules of evidence is necessary in order to interpret the literature correctly.

The shift to evidence-based medicine reflects not only the constant desire to improve practice but also the developments that have taken place in clinical research, methodology and the increase in medical literature.

Evidence-based health care means that the policies and practices employed in the prevention and treatment of health problems are based on principles that have been proven through appropriate scientific methods. It must always be remembered, of course, that proving the clinical effectiveness of a procedure is not sufficient. It needs to be complemented by evidence that both health care provider and user are satisfied, and that the procedure is both cost-effective and feasible in different settings.

M.Sc. Programmes at the London School of Hygiene and Tropical Medicine REPRODUCTIVE AND SEXUAL HEALTH RESEARCH This course includes a broad introduction to the concepts and methods of reproductive health research and a solid foundation in the key aspects of the field including: family planning; obstetric health; AIDS/STDs; and sexual health. The course is structured to provide an attractive combination of formal teaching and practical experience. MEDICAL DEMOGRAPHY This course offers broad training in the theories and methods of demography and the population sciences and their application to health, social welfare and economic development. It is designed for individuals with interests in developing and developed countries. Enquiries: Registry, London School of Hygiene & Tropical Medicine, 50 Bedford Square, London WC1B 3DP, England (Tel: +44 171 299 4648; Fax: +44 171 323 0638; e-mail: registry@lshtm.ac.uk; web site: www.lshtm.ac.uk)

Go to Progress