Internal Medicine Residency

Dr. Grodon Guyatt

Gordon Guyatt is a Professor in the Departments of Clinical
Epidemiology and Biostatistics and Medicine at McMaster University.

Evidence-Based Medicine

Evidence-based medicine is about solving clinical problems. [1] In 1992, a group of clinician-teachers from McMaster University described EBM as a shift in medical paradigms. [2] In contrast to the traditional paradigm of medical practice, EBM acknowledges that intuition, unsystematic clinical experience, and pathophysiologic rationale are insufficient grounds for clinical decision-making; and it stresses the examination of evidence from clinical research. In addition, EBM suggests that a formal set of rules must complement medical training and common sense for clinicians to interpret the results of clinical research effectively. Finally, EBM places a lower value on authority than the traditional medical paradigm does.

A complementary way of thinking about EBM focuses on two key principles that we will enunciate in the next section.

Two Fundamental Principles of EBM

As a distinctive approach to patient care, EBM involves two fundamental principles. First, evidence alone is never sufficient to make a clinical decision. Decision-makers must always trade the benefits and risks, inconvenience, and costs associated with alternative management strategies, and in doing so consider the patient’s values and preferences. [1] Second, EBM posits a hierarchy of evidence to guide clinical decision-making.

Clinical Decision-Making: Evidence Is Never Enough

Picture a patient with chronic pain resulting from terminal cancer. She has come to terms with her condition, has resolved her affairs and said her goodbyes, and she wishes to receive only palliative therapy. The patient develops pneumococcal pneumonia. Now, evidence that antibiotic therapy reduces morbidity and mortality from pneumococcal pneumonia is strong. Almost all clinicians would agree, however, that even evidence this compelling does not dictate that this particular patient should receive antibiotics. Despite the fact that antibiotics might reduce symptoms and prolong the patient’s life, her values are such that she would prefer a rapid and natural death.

Now envision a second patient-- an 85-year-old man with severe dementia, who is incontinent, contracted, and mute, without family or friends, and who spends his days in apparent discomfort. This man develops pneumococcal pneumonia. Although many clinicians would argue that those responsible for this patient’s care should not administer antibiotic therapy because of his circumstances, others would suggest that they should do so. Again, evidence of treatment effectiveness does not automatically imply that treatment should be administered. The management decision requires a judgment about the trade-off between risks and benefits; and because values or preferences differ, the best course of action will vary from patient to patient and among clinicians.

Finally, picture a third patient-- a healthy, 30-year-old mother of two children who develops pneumococcal pneumonia. No clinician would doubt the wisdom of administering antibiotic therapy to this patient. However, this does not mean that an underlying value judgment has been unnecessary. Rather, our values are sufficiently concordant, and the benefits so overwhelm the risks, that the underlying value judgment is unapparent.

In current health care practice, judgments often reflect clinicial or societal values concerning whether intervention benefits are worth the cost. [2] Consider the decision regarding administration of clopidigrel versus aspirin to patients with transient ischemic attacks. Here, evidence from large randomized trials suggests that the more expensive agent is more effective. Many authorities, however, recommend first-line treatment with the less expensive, less effective drug, presumably because they believe society’s resources would be better used in other ways. Implicitly, they are making a value or preference judgment about the tradeoff between deaths and strokes prevented, and resources spent.

By values and preferences, we mean the underlying processes we bring to bear in weighing what our patients and our society will gain--or lose-- when we make a management decision. The explicit enumeration and balancing of benefits and risks that is central to EBM brings the underlying value judgments involved in making management decisions into bold relief.

Acknowledging that values play a role in every important patient care decision highlights our limited understanding of eliciting and incorporating societal and individual values. Health economists have played a major role in developing a science of measuring patient preferences. [3] [4] Some decision aids incorporate patient values indirectly: If patients truly understand the potential risks and benefits, their decisions will likely reflect their preferences. [5] These developments constitute a promising start. Nevertheless, many unanswered questions remain concerning how to elicit preferences and how to incorporate them in clinical encounters already subject to crushing time pressures. Addressing these issues constitutes a challenging frontier for EBM.

A Hierarchy of Evidence

What is the nature of the “evidence” in EBM? We suggest a broad definition: Any empirical observation about the apparent relation between events constitutes potential evidence. Thus, the unsystematic observations of the individual clinician constitute one source of evidence; and physiologic experiments constitute another source. Unsystematic observations can lead to profound insight, and experienced clinicians develop a healthy respect for the insights of their senior colleagues in issues of clinical observation, diagnosis, and relations with patients and colleagues. Some of these insights can be taught, yet rarely appear in the medical literature.

At the same time, unsystematic clinical observations are limited by small sample size and, more importantly, by deficiencies in human processes of making inferences. [6] Predictions about intervention effects on clinically important outcomes based on physiologic experiments usually are right, but occasionally are disastrously wrong. The expectation that encainide and flecainide reduce lethal arrhythmias, when randomized trials show that the drugs increase sudden death, constitutes an example of this phenomenon.

Given the limitations of unsystematic clinical observations and physiologic rationale, EBM suggests a hierarchy of evidence. Table 1 presents a hierarchy of study designs for treatment issues; very different hierarchies are necessary for issues of diagnosis or prognosis. Clinical research goes beyond unsystematic clinical observation in providing strategies that avoid or attenuate spurious results. Because few-- if any-- interventions are effective in all patients, we would ideally test a treatment in a patient to whom we would like to apply it. Numerous factors can lead clinicians astray as they try to interpret the results of conventional open trials of therapy. These include natural history, placebo effects, patient and health worker expectations, and the patient’s desire to please.

Table1: A Hierarchy of Strength of Evidence for Treatment Decisions

  • N of 1 trial
  • Systematic reviews of randomized trials
  • Single randomized trial
  • Systematic review of observational studies addressing patient-important outcomes
  • Single observational study addressing patient-important outcomes
  • Physiologic studies ( Studies of blood pressure, cardiac output, exercise capacity, bone density, etc…)
  • Unsystematic clinical observations

The same strategies that minimize bias in conventional therapautic trials involving multiple patients can guard against misleading results in studies involving single patients. [7] In the N of 1 RCT, patients undertake pairs of treatment periods in which they receive a target treatment in one period of each pair, and a placebo or alternative in the other. Patients and clinicians are blind to allocation, the order of the target and control are randomized, and patients make quantitative ratings of their symptoms during each period. The N of 1 RCT continues until both the patient and clinician conclude that the patient is, or is not, obtaining benefit from the target intervention. When conditions are right, N of 1 randomized trials are often feasible, [8] [9] can provide definitive evidence of treatment effectiveness in individual patients, and may lead to long-term differences in treatment administration. [10]

When considering any other source of evidence about treatment, clinicians are generalizing from results in other people to their patients, inevitably weakening inferences about treatment impact and introducing complex issues of how trial results apply to individual patients. Inferences may nevertheless be very strong if results come from a systematic review of methodologically strong RCTs with consistent results. However, inferences generally will be somewhat weaker if only a single RCT is being considered, unless it is very large and has enrolled a diverse patient population (see Table 1). Because observational studies may under- or, more typically, over-estimate treatment effects in an unpredictable fashion, [11] [12] their results are far less trustworthy than those of randomized trials. Physiologic studies and unsystematic clinical observations provide the weakest inferences about treatment effects.

This hierarchy is not absolute. If treatment effects are sufficiently large and consistent, for instance, observational studies may provide more compelling evidence than most RCTs. For instance, observational studies have allowed extremely strong inferences about the efficacy of insulin in diabetic ketoacidosis or that of hip replacement in patients with debilitating hip osteoarthritis. At the same time, instances in which RCT results contradict consistent results from observational studies reinforce the need for caution. Defining the extent to which clinicians should temper the strength of their inferences when only observational studies are available remains one of the important challenges for EBM. The challenge is particularly important given that much of the evidence regarding the harmful effects of our therapies comes from observational studies.

The hierarchy implies a clear course of action for physicians addressing patient problems: They should look for the highest available evidence from the hierarchy. The hierarchy makes clear that any statement to the effect that there is no evidence addressing the effect of a particular treatment is a non sequitur. The evidence may be extremely weak – it may be the unsystematic observation of a single clinician, or a generalization from physiologic studies that are only indirectly related – but there is always evidence.

Next we will briefly comment on additional skills that clinicians must master for optimal patient care and the relation of those skills to EBM.

Clinical Skills, Humanism, Social Responsibility, and EBM

The evidence-based process of resolving a clinical question will be fruitful only if the problem is formulated appropriately. A secondary care internist developed a lesion on his lip shortly before an important presentation. He was quite concerned and, wondering if he should take acyclovir, proceeded to spend the next two hours searching for the highest quality evidence and reviewing the available RCTs. When he began to discuss his remaining uncertainty with his partner, an experienced dentist, she quickly cut short the discussion by exclaiming, “But, my dear, that isn’t herpes!”

This story illustrates the necessity of obtaining the correct diagnosis before seeking and applying research evidence in practice, the value of extensive clinical experience, and the fallibility of clinical judgment. The essential skills of obtaining a history and conducting a physical examination and the astute formulation of the clinical problem come only with thorough background training and extensive clinical experience. The clinician makes use of evidence-based reasoning -- applying the likelihood ratios associated with positive or negative physical findings, for instance-- to interpret the results of the history and physical examination. Clinical expertise is further required to define the relevant treatment options before examining the evidence regarding the expected benefits and risks of those options.

Finally, clinicians rely on their expertise to define features that impact on the generalizability of the results to the individual patient. We have noted that, except when clinicians have conducted N of 1 RCTs, they are attempting to generalize (or, one might say, particularize) results obtained in other patients to the individual patient before them. The clinician must judge the extent to which differences in treatment (local surgical expertise or the possibility of patient noncompliance, for instance), the availability of monitoring, or patient characteristics (such as age, comorbidity, or concomitant treatment) may impact estimates of benefit and risk that come from the published literature.

Thus, knowing the tools of evidence-based practice is necessary but not sufficient for delivering the highest quality patient care. In addition to clinical expertise, the clinician requires compassion, sensitive listening skills, and broad perspectives from the humanities and social sciences. These attributes allow understanding of patients' illnesses in the context of their experience, personalities, and cultures. The sensitive understanding of the patient bears on evidence-based practice in a number of ways. For some patients, incorporation of patient values for major decisions will mean a full enumeration of the possible benefits, risks, and inconvenience associated with alternative management strategies that are relevant to the particular patient. For some of these patients and problems, this discussion should involve the patients’ family. For other problems – the discussion of screening with prostate-specific antigen with older male patients, for instance – attempts to involve other family members might violate strong cultural norms.

Many patients are uncomfortable with an explicit discussion of benefits and risks, and object to having what they perceive as excessive responsibility for decision-making being placed on their shoulders. [13]. Because such patients would tell us they want the physician to make the decision on their behalf, the physician’s responsibility is to develop insight to ensure that choices will be consistent with patients’ values and preferences. Understanding and implementing the sort of decision-making process patients desire, and effectively communicating the information they need, requires skills in understanding the patient’s narrative and the person behind that narrative. [14] [15] A continuing challenge for EBM -- and for medicine in general -- will be to better integrate the new science of clinical medicine with the time-honored craft of caring for the sick.

Ideally, evidence-based physicians' technical skills and humane perspective will lead them to become effective advocates for their patients both in the direct context of the health system in which they work and in broader health policy issues. Most physicians see their role as focusing on health care interventions for their patients. Even when they consider preventive therapy, they focus on individual patient behavior. However, this focus may be too narrow.

Observational studies have documented the strong and consistent association between socio-economic status and health. Societal health may be more strongly associated with income gradients than with the total wealth of the society. In other words, the overall health of the populace may tend to be higher in poorer countries with a relatively equitable distribution of wealth than in richer countries with larger disparities between rich and poor. These considerations suggest that physicians concerned about the health of their patients as a group, or about the health of the community, should consider how they might contribute to reducing poverty.

Observational studies have shown a strong and consistent association between pollution levels and respiratory and cardiovascular health. Physicians seeing patients with chronic obstructive pulmonary disease will suggest that they stop smoking. But should physicians also be concerned with the polluted air that patients are breathing? Perhap they should.

Additional Challenges for EBM

The skills necessary for evidence-based practice include the ability to precisely define a patient problem and to ascertain what information is required to resolve the problem, conduct an efficient search of the literature, select the best of the relevant studies, apply rules of evidence to determine their validity, extract the clinical message, and apply it to the patient problem as the skills necessary for evidence-based practice. [1] to understand how the patient’s values impact the balance between advantages and disadvantages of the available management options, and the ability to appropriately involve the patient in the decision.

The biggest challenge to evidence-based practice: time limitation. Fortunately, new resources to assist clinicians are available and the pace of innovation is rapid. One can consider a classification of information sources that comes with a mnemonic device, 4S: the individual study, the systematic review of all the available studies on a given problem, a synopsis of both individual studies and summaries, and systems of information. By systems we mean summaries that link a number of synopses related to the care of a particular patient problem (acute upper gastrointestinal bleeding) or type of patient (the diabetic outpatient) (Table 2). Evidence-based selection and summarization is becoming increasingly available at each level.

Table 2. A Hierarchy of Preprocessed Evidence

Studies Preprocessing involves selecting only those studies that are both highly relevant and that are characterized by study designs that minimize bias and thus permit a high strength of inference.
Summaries Systematic reviews provide clinicians with an overview of all of the evidence addressing a focused clinical question.
Synopses Synopses of individual studies or of systematic reviews encapsulize the key methodologic details and results required to apply the evidence to individual patient care.
Systems Practice guidelines, clinical pathways, or evidence-based textbook summaries of a clinical area provide the clinician with much of the information needed to guide the care of individual patients.

This book deals primarily with. Evidence-based approaches can inform decision-making at the level of the individual patient, health policy-making, [16] day-to-day decisions in public health, and systems-level decisions such as those facing hospital managers. In each of these arenas, EBM can support the appropriate goal of gaining the greatest health benefit from limited resources. On the other hand, evidence --- as an ideology, rather than a focus for reasoned debate --- has been used as a justification for many agendas in health care, ranging from crude cost-cutting to the promotion of extremely expensive technologies with minimal marginal returns.

In the policy arena, dealing with differing values poses even more challenges than in the arena of individual patient care. Should we restrict ourselves to alternative resource allocation within a fixed pool of health care resources, or should we be trading off health care services against, for instance, lower tax rates for individuals or corporations? How should we deal with the large body of observational studies suggesting that social and economic factors may have a larger impact on the health of populations than health care delivery? How can we resolve the tension between what may be best for a person and what may be optimal for the society of which that person is a member? The debate about such issues is at the heart of evidence-based health policy-making, but, inevitably, it has implications for decision-making at the individual patient level.


  1. Haynes RB, Sackett RB, Gray JMA, Cook DC, Guyatt GH. Transferring evidence from research into practice: 1. The role of clinical care research evidence in clinical decisions. ACP Journal Club. 1996 Nov-Dec;125:A-14-15.
  2. Napodano RJ. Values in Medical Practice. New York, NY: Human Sciences Press; 1986.
  3. Drummond MF, Richardson WS, O'Brien B, Levine M, Heyland DK, for the Evidence-Based Medicine Working Group. Users' Guides to the Medical Literature XIII. How to use an article on economic analysis of clinical practice. A. Are the results of the study valid? JAMA. 1997;277:1552-1557.
  4. Feeny DH, Furlong W, Boyle M, Torrance GW. Multi-attribute health status classification systems: health utilities index. Pharmacoeconomics. 1995;7:490-502.
  5. O’Connor AM, Rostom A, Fiset V, Tetroe J, Entwhistle V, Llewellyn-Thomas H, Holmes-Rovner M, Barry M, Jones J. Decision aids for patients facing health treatment or screening decisions: systematic review. BMJ. 1999;319:731-734.
  6. Nisbett R, Ross L. Human Inference. Englewood Cliffs, NJ: Prentice-Hall; 1980.
  7. Guyatt GH, Sackett DL, Taylor DW, et. al. Determining optimal therapy - randomized trials in individual patients. N Engl J Med. 1986;314:889-892.
  8. Guyatt GH, Keller JL, Jaeschke R, Rosenbloom D, Adachi JD, Newhouse MT. The n-of-1 randomized controlled trial: clinical usefulness. Our three-year experience. Ann Int Med. 1990;112:293-299.
  9. Larson EB, Ellsworth AJ, Oas J. Randomized clinical trials in single patients during a 2-year period. JAMA. 1993;270:2708-2712
  10. Mahon J, Laupacis A, Donner A, Wood T. Randomised study of n of 1 trials versus standard practice. BMJ. 1996;312:1069-1074.
  11. Guyatt GH, DiCenso A, Farewell V, Willan A, Griffith L. Randomized trials versus observational studies in adolescent pregnancy prevention. J Clin Epidemiol . 2000;53:167-174.
  12. Kunz R. Oxman AD. The unpredictability paradox: review of empirical comparisons of randomised and non-randomised clinical trials. BMJ. 1998;317:1185-1190.
  13. S utherland HJ, Llewellyn-Thomas HA, Lockwood GA, Tritchler DL, Till JE. Cancer patients: their desire for information and participation in treatment decisions. J Royal Soc Med. 1989;82:260-263.
  14. Greenhalgh T. Narrative based medicine: narrative based medicine in an evidence based world. BMJ. 1999;318:323-325.
  15. Greenhalgh T, Hurwitz B. Narrative based medicine: why study narrative? BMJ. 1999;318:48-50.
  16. Muir Gray FA, Haynes RB, Sackett DL, Cook DJ, Guyatt GH. Transferring Evidencefrom Research into Practice: III. Developing Evidence-based Clinical Policy. ACP Journal Club. 1997;A14.