EVIDENCE-BASED MEDICINE

The concept of EBM was developed by clinical epidemiologists more than 20 years ago.¹ Since then, EBM has become very popular. Whereas a Medline search of EBM in 1993 yielded only six citations,² a search in early 2008 yielded more than 29,000 citations. The Equine Veterinary Journal currently dedicates a section to EBM, and the term EBM is increasingly used in presentations and publications. Recently, EBM has been described as “the integration of the best research evidence with our clinical expertise and our patient’s unique values and circumstances.”³ The emphasis of EBM is on acquiring, assessing, and utilizing evidence to improve clinical decision making. A fundamental principle of EBM is that, whenever possible, evidence for clinical activities should be derived from well-designed studies of patients with spontaneous disease (i.e., epidemiologic studies).

The practice of equine medicine has traditionally been empiric and driven by the knowledge, wisdom, and experience of experts. The field has depended largely on combining an understanding of the pathogenesis and mechanisms of disease with authority, theory, logical deduction, and intuition. The human medical profession has moved away from these traditions toward a medical practice that is based on clinically relevant evidence.^2,³ The principle of EBM is to base all clinical decisions regarding prevention, diagnosis, treatment, and prognosis of disease on the use of the best existing evidence. It is difficult to dispute the value of this approach, which aims to discourage a reliance on authority (“I treat horses with equine protozoal myeloencepalitis [EPM] this way because that is what I was told” by a textbook or expert) and promote instead a reliance on the best available evidence (“Why do I treat EPM this way?” and “What is the most successful approach to treating EPM?”). Currently, EBM refers not only to this laudable principle but also to a methodology by which the principle is realized.

Methodology

The methodology of EBM consists of five steps: (1) defining, or asking, a clinical question; (2) searching for evidence to answer the question; (3) critically appraising the evidence gathered; (4) applying the results of answering the question; and (5) auditing the outcome of applying the results (i.e., how well did the EBM-derived answer work?).² Each of these steps is considered in turn. As will be discussed later, the state of development of the methodology is highly variable among steps.

STEP 1: ASKING THE QUESTION(S)

The first step in EBM is defining the question that pertains to the clinical circumstances of the patient. This is not always as simple as it might initially seem. A question that is too broad (e.g., “How do I treat pneumonia in horses?”) may yield numerous citations that are not germane, whereas a question that is too focused (e.g., “How do I treat Pneumocystis carinii pneumonia in a foal with selective IgM deficiency?”) may yield no results at all. Typically, veterinarians ask several questions about a given patient, including questions described as background and foreground varieties. Background questions are general knowledge questions, such as “How does cervical stenosis or instability develop in horses?” and “What causes abscessing pneumonia of foals?” Foreground questions are more specific to the particular case, such as “In a mare with congestive heart failure and atrial fibrillation that developed coincident with developing pleuropneumonia, what is the best approach for treating the arrhythmia?” Generally, veterinarians pose fewer background and more foreground questions as they gain experience with a particular condition.

Any clinical encounter will generate a number of questions pertaining to signalment and history (e.g., “Does the signalment suggest that certain disorders will be more common?” “Does the herd history help eliminate certain causes?”), clinical examination findings (e.g., “How do I interpret a head tilt and signs of facial nerve paralysis in this horse?”), etiology and differential diagnosis (e.g., “What might cause these signs?”; “What is the cause of these signs?”), diagnosis (e.g., “What tests are most meaningful in a patient with this clinical complaint?”), treatment (e.g., “What is the best way to treat temporohyoid osteoarthropathy?”), prognosis (e.g., “What is the probability that this horse will return to use?”), prevention (e.g., “What can be done to prevent recurrence?”), experiences and meanings (e.g., “Have I effectively communicated with the clients?”; “Does the client perceive that I empathize?”), and improvement (e.g., “Have I learned new information about managing this case from the experience?”; “Did my patient benefit?”).

Attempting to answer all relevant questions simultaneously will be unrewarding for the clinician, client, and patient. Veterinarians must learn to prioritize their efforts. Often, first addressing certain questions of case management (i.e., history, physical examination and diagnosis) seems logical; however, the client’s needs may also influence the order of the questions (e.g., matters of prognosis or experience and meaning). Moreover, veterinarians must consider which questions may feasibly be answered within the time and setting of the clinical encounter. The veterinarian should consider recording those questions that must be postponed so that they will not be forgotten.

The process of carefully considering questions to ask, prioritizing them, and saving unanswered ones to be answered later might seem like an inconvenience and a waste of time: Veterinarians do much of this already without needing to spend a lot of time formulating and recording their clinical inquiries. The principal advantage to the process is that it will help identify evidence-based solutions for the clinical problems that veterinarians face. With experience veterinarians become more proficient and expeditious in answering questions, and the process necessitates remaining current on new developments, such as new medical and surgical approaches to temporohyoid osteoarthropathy (and their evaluation) and new diagnostic tests for EPM.

Perhaps the most important principle of EBM is that veterinarians need to ask themselves what evidence exists for their clinical activities. Breaking away from the tradition of authority and empiricism in equine medicine will not be easy, and it will happen only if veterinarians are willing to question what they really know about their clinical decisions and interpretations.

STEP 2: SEARCHING FOR EVIDENCE

Once a question has been posed, it is important to find the best available evidence using a review of the literature that is as comprehensive as possible. Relying on book chapters or review articles will not suffice for at least two important reasons. First, new knowledge emerges more quickly than it can be incorporated into textbooks. Of course, well-written textbooks remain excellent resources for many background questions (e.g., pathophysiology of diseases, etiology of diseases with well-defined causes) and often will be fairly current in a field such as equine medicine, wherein the pace at which new information is generated is fairly slow. Unfortunately, it is often difficult to determine which information in these textbooks is current and which is outdated. Furthermore, authors of review articles have their own particular prejudices, beliefs, and perceptions, and these biases influence their interpretations and recommendations. Practitioners must guard against biases that may result in selecting reports that conform to their preexisting beliefs and perceptions, which is easier said than done.

In human medicine a hierarchy of resources for EBM has been proposed (note that this is a hierarchy of resources, not of evidence).² The highest tier of this hierarchy is a computerized decision support system (CDDS) in which clinical information from a patient’s records is automatically linked to all relevant, important research findings pertaining to the patient’s circumstances. The next highest-level resource is synopses of individual studies or reviews. These synopses are designed to be concise and precise distillations of the important facts needed by busy clinicians. The next level is syntheses of reports based on exhaustive searches for evidence, implementation of explicit scientific criteria for review, and systematic assembly of the evidence. Syntheses are epitomized by the Cochrane Reviews (www.Cochranelibrary.com/). At the bottom of this hierarchy are individual scientific reports.

Equine practitioners currently operate primarily at the level of individual reports. Although a veterinary CDDS will likely be developed in the future, it is not clear when. Synopses of numerous studies on a given topic are not likely to be available soon, primarily because numerous studies for a given topic do not yet exist. The same problem arises for systematic reviews: Although there are examples,⁴ they are exceedingly rare. The advantage of relying on individual reports is that the information is generally current (whereas it can take many months or years to develop the evidence on which synopses and systematic reviews are based). The disadvantage is that the evidence is weaker from individual studies, and it puts the onus on the reader to critically appraise each study. Critical appraisal is discussed in greater detail in the next section, but first it is important to consider resources for finding information.

Those practitioners who are fortunate to work at institutions with appropriate licenses for digital publications will find that a great deal of information can be retrieved electronically with little effort, and librarians are generally available to assist. In settings without such licenses, it may still be possible to gain access to the necessary resources through colleges or universities. Regardless, veterinarians will find a considerable amount of high-quality information on the Internet through resources such as PubMed (http://wwwncbi.nlm.gov/PubMed/) and BioMed Central (http://www.biomedcentral.com). It is important to remember, however, that the Internet also offers much in the way of low-quality information. Relying on subscription journals is generally inefficient: There are simply too many journals with too many articles appearing each month that are relevant to the daily activities of equine specialists. Moreover, important articles are often published in journals to which few veterinarians subscribe. For example, not many veterinarians subscribe to the journal Genomics, but many would be interested in reading about the mutation that causes polysaccharide storage myopathy.⁵