Introduction to Evidence Based Medicine

Making Medical Decisions

Bill Cayley MD MDiv

June 2022

Fundamental Questions

Two fundamental questions need to be addressed in medical practice: First, what is the purpose of medicine? Second, how do I decide what to do? Just as a person planning a cross-country trip maps out the journey based on decisions about the most desirable destination and the most desirable routing, so in medicine physicians must work with patients to decide on both overarching goals, and on the best way to reach those goals. In both travel and medicine, you have to know where you’re going before deciding how to get there.

While medicine has many facets, including direct patient care, public health, and research endeavors, the common goal of all medical ventures is improving the quality of patients’ lives. This common underlying goal can serve as the “destination” to guide all medical journeys and endeavors, regardless of which facet or field of medicine is the stage for the patient’s medical journey.

In a much-quoted 1996 editorial, David Sackett, an early and prolific author on evidence-based medicine, defined evidence-based medicine (EBM) as the “conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients.” (1) (italics added) This characterization highlights the three important parts of evidence-based medical practice: the patient, the evidence, and careful application of generalized evidence to the individual patient.

Sackett goes on to argue that EBM is not old hat or just the same old medical practice, as evidenced by the wide variation that often occurs in clinical practice patterns. Neither is it intended to be “cookbook medicine” as EBM specifically advocates for individualized application of evidence to patient care, not forcing patient care to conform to generalized evidence. Further, EBM is not intended primarily as a savings tool – it is intended to guide practitioners to provide the best, not necessarily the cheapest, care. Lastly, EBM is not intended to be only concerned with randomized controlled trials, but with the best relevant evidence applicable to the situation in question.

“Evidence-based medicine” was first introduced in the mainstream medical literature in a 1992 article, “Evidence-based medicine: A new approach to teaching the practice of medicine,” which presented EBM as “a fundamentally new approach” emphasizing “question formulation, search and retrieval of the best available evidence, and critical appraisal of the study methods to ascertain the validity of results”. (2) A subsequent series of “Users' Guides to the Medical Literature” presented skills for searching for, appraising, and applying various types of published evidence to medical practice. As EBM gained prominence in the late 1990s and beyond, three streams of evidence dissemination developed: 1) systematic reviews gained increasing prominence in the medical literature, 2) knowledge search engines (including internet engines such as Google, and medline interfaces such as Ovid) became standard tools for medical literature searching, and 3) knowledge distillation and “push” services developed as a way to compile and disseminate concise reviews of evidence on specific topics or questions (eg, ACP Journal Club, InfoPoems, etc). (3)

Making Decisions

Heuristics are informal and often subconscious methods we use to arrive at decisions. While we may use many different heuristics depending on the clinical situation, awareness of the types of decision processes we are using, as well as their strengths and limitations, may help us to think more critically about the strengths and weaknesses of our diagnostic reasoning. Pattern recognition can allow rapid diagnosis for familiar conditions, but may lead us to miss an unexpected diagnosis in a seemingly familiar situation. Use of the scientific method involves hypothesis testing through the collection and analysis of data, such as when using a rapid influenza test to help determine the cause of an acute respiratory illness. Probabilities involve assessment of the likelihood of a condition, and assessing whether abnormal results on a test are likely true or false positives. Differential diagnosis overlaps with elements of pattern recognition and hypothesis testing, whereby we evaluate the range of possible causes for a condition, then use clinical data, labs or imaging to narrow the range of diagnostic possibilities. Tests allow us to measure or image things we cannot directly observe, but pitfalls of over-reliance on testing include overestimating the importance of unexpectedly abnormal results and emphasis on isolated test results rather than the whole clinical picture. Assessment of a “treatment threshold” includes both a decision as to whether a condition is serious enough to warrant treatment, and an assessment as to whether a potential condition would be serious enough to warrant further testing in the first place. (4)

How do I decide what to do? Decisions can be guided by a number of different approaches to reasoning. Some decisions are guided by dogma, or an a-priori set of beliefs (eg, “natural is best.”). Some decisions are guided by tradition (eg, “we’ve always done it that way”), relying on the collective wisdom of the ages. Some decisions are guided by convention (eg, “everyone else does it that way”), relying on the collective wisdom of colleagues. An evidence-based approach makes decisions based on critical evaluation and reflection (eg, “the evidence supports doing it this way”). Since EBM advocates that medical decisions should proceed from application of the “current best evidence,” an appreciation of how to evaluate, or grade, evidence is crucial to the application of “best evidence” in practice.

Classification of Evidence

Conceptually, evidence starts simply with what is observed. Every individual observation is an isolated piece of evidence. To generate higher quality evidence, however, it is important to compile, organize, and evaluate those individual observations in a systematic way. Thus, while an anecdotal observation constitutes evidence regarding a single event, a more organized compilation of several observed events can constitute a case series, a higher level of evidence. An even more organized way to evaluate an event or an intervention is to use systematic observation, as in an uncontrolled or controlled trial. A meta-analysis provides even higher quality evidence by systematically grouping together and synthesizing the results of multiple trials. Thus, the more systematic an approach that is taken to gathering and organizing evidence, ranging from the individual anecdote up to the meta-analysis of controlled trials, the higher quality the evidence.

While evidence-classification and rating may seem a daunting challenge, the basic premise is that the more systematic the observations that are available (eg, RCT instead of just a case series) the better the quality of evidence. And, since EBM seeks to apply the “current best evidence” it is important to see for the highest quality studies that are available to address a given clinical question.

The “Strength of Recommendation Taxonomy” is one system of evidence grading, developed by a collaboration among family medicine editors, that seeks to provide a user-friendly approach to classifying evidence in terms of both evidence quality, and the degree to which it bears on patient-oriented outcomes. (5)

Integrating Evidence and Practice

In order to decide what to do in practice, we also need to know how we’ll know when we are there – that is, what kind of outcomes do we seek? Medical outcomes can be broadly grouped into 3 categories. Some outcomes (eg, blood pressure or cholesterol levels) are merely surrogate markers of disease. We measure these surrogate markers because we think they tell us something prognostically about the expected course of a person’s disease process, but they do not directly impact how a patient feels from day to day. Others measure actual stage or extent of disease (eg, the stage of a diabetic ulcer, or the angiographic extent of disease). These may have a more direct bearing on a patient’s quality of life or extent of suffering, but still do not provide direct measures of long-term quality of life. The most important outcomes for guiding medical decisions are those that affect how patients feel and the quality of their lives – that is, patient-oriented outcomes such as mobilty, suffering, longevity, and other considerations that bear directly on how a patient experiences his or her quality of life. In short, patient oriented outcomes have primarily to do with long-term morbidity or mortality. (6)

An even simpler way to break down the types of outcomes that may be considered is into “disease oriented” outcomes such as blood sugar, blood pressure, flow rate, coronary plaque thickness, or “patient oriented outcomes” such as reduced morbidity, reduced mortality, symptom improvement, improved quality of life, or lower cost. (5)

In applying evidence regarding patient-oriented outcomes, it is also important to work collaboratively with the patient to determine which outcomes are most important to the patient, as this will guide the decision as to which interventions to pursue. Two examples illustrate this point: 1) Topical antibiotics for bacterial conjunctivitis may improve early and late resolution rates, but nearly all cases ultimately have complete remission. (7) 2) Digoxin for symptomatic heart failure provides no significant difference in mortality but is associated with lower rates of hospitalization and of clinical deterioration. (8) Thus, it is vital to discuss the desired target outcome with a patient when deciding how to apply evidence to a specific patient’s situation.

References

(1) Sackett DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn't. BMJ. 1996 Jan 13;312(7023):71-2. PMID: 8555924 (http://www.bmj.com/cgi/content/full/312/7023/71)

(2) Evidence-Based Medicine Working Group. Evidence-based medicine. A new approach to teaching the practice of medicine. JAMA. 1992 Nov 4;268(17):2420-5. PMID: 1404801 (http://jama.ama-assn.org/cgi/reprint/268/17/2420?ijkey=d3d27e0bf59a836b2ff7923ef06634c6304b1c75&keytype2=tf_ipsecsha)

(3) Montori VM, Guyatt GH. Progress in evidence-based medicine. JAMA. 2008 Oct 15;300(15):1814-6. PMID: 18854545 (http://jama.ama-assn.org/cgi/content/full/300/15/1814)

(4) Woolever DR. The art and science of clinical decision making. Fam Pract Manag. 2008 May;15(5):31-6. PubMed PMID: 18546805. https://www.aafp.org/fpm/2008/0500/p31.html

(5) Ebell MH, Siwek J, Weiss BD, Woolf SH, Susman J, Ewigman B, Bowman M. Strength of recommendation taxonomy (SORT): a patient-centered approach to grading evidence in the medical literature. Am Fam Physician. 2004 Feb 1;69(3):548-56. PMID: 14971837 (http://www.aafp.org/afp/20040201/548.html)

(6) Slawson DC, Shaughnessy AF, Ebell MH, Barry HC. Mastering medical information and the role of POEMs--Patient-Oriented Evidence that Matters. J Fam Pract. 1997 Sep;45(3):195-6. PMID: 9312554

(7) Sheikh A, Hurwitz B, van Schayck CP, McLean S, Nurmatov U. Antibiotics versus placebo for acute bacterial conjunctivitis. Cochrane Database Syst Rev. 2012 Sep 12;(9):CD001211. doi: 10.1002/14651858.CD001211.pub3. PubMed PMID: 22972049. https://www.ncbi.nlm.nih.gov/pubmed/22972049

(8) Hood WB Jr, Dans AL, Guyatt GH, Jaeschke R, McMurray JJ. Digitalis for treatment of heart failure in patients in sinus rhythm. Cochrane Database Syst Rev. 2014 Apr 28;(4):CD002901. doi: 0.1002/14651858.CD002901.pub3. PubMed PMID: 24771511. https://www.ncbi.nlm.nih.gov/pubmed/24771511

Link to presentation on Introduction to Evidence-Based Medicine