Introduction

We make therapeutic decisions all the time. Many of them are made with little thought: we make the same decisions all the time; we use aids (for instance, we may use Sanford to assist in selecting antibiotics for various conditions); we ask for advice from a colleague or from a consultant.

If you have bought into evidence-based medicine as a reasonable paradigm, then you probably are not comfortable or happy with the above-mentioned approaches to therapeutic decision-making. If you need additional convincing, let's review a few common approaches to therapeutic decision-making and their pitfalls.  If you are convinced, skip ahead to the end of this module.

Pathophysiologic approach

This is a consequence of our Flexnerian approach to medical education. In essence, this is a reductionist approach: if we understand the parts, we know about the whole. In theory, this makes sense. In the absence of evidence based on true patient-oriented outcomes, this may be all we have. We have many examples of where this approach has led us down the garden path. Back in the old days (in the 1980s!), we knew that ventricular arrhythmia was the most common cause of death among patients with acute myocardial infarction. Based on this, we routinely started patients on lidocaine and if they continued to have lots of premature ventricular contractions, we would start medications such as tocainide, encanide, or flecanide. Later we found out that these medications increased mortality!  Similarly, while reducing cholesterol should improve mortality, studies have shown that Lopid may increase overall mortality.  Finally, two recent randomized trials have shown that calcium channel blockers, which do a fine job of lowering blood pressure, also increase the risk of MI compared with ACE inhibitors.  Human beings are complex, non-linear beings, and simple, linear reasoning doesn't always (or even often) work.

Experience

This is a major component of the "art" of medicine. What could possibly be wrong with that? Generally, the experience of the clinician is an important and underappreciated part of practice. However, we have to be careful. Let's look at a couple elements that color our experience.

We observe a number of children with bronchitis who failed to respond to amoxicillin, but after putting them on azithromycin, they improve. Because of this experience, we decide to start treating all our children with bronchitis with azithromycin. There are at LEAST two explanations for this phenomenon. Unless we systematically follow-up all our patients, we forget how many children treated with amoxicillin did not come back because they got better! Additionally, bronchitis is generally a self-limited illness that gets better regardless of how it's treated. The subsequent improvement with the second antibiotic (ANY antibiotic) may merely reflect the natural history of bronchitis!

Expert opinion

We generally use consultants for a few limited things: technical expertise (a procedure) or advice on diagnosis or treatment. How many times have the recommendations of a consultant left you scratching your head? Quite often we choose to ignore their suggestions, most often because we believe that the consultant's perspective is inconsistent with the needs of the patient, their recommendations are not feasible, or the diagnosis is such a zebra that it makes no sense to pursue. Many of our consultants live in a world very different from our own. Additionally, some of our consultants may not be up to date on the literature in their own field. Eye patches have been used for treating corneal abrasions for years. Recently eye patches have been studied carefully, and guess what, patients treated with eye patches (compared to no patching) do not heal faster and do not experience less pain! This has been seen in several high quality studies. Nonetheless, many ophthalmologists still patch corneal abrasions! Caveat emptor!

Enterprise

In the good old days of medicine, we all knew of physicians who did tests or recommended procedures because they needed to make a boat payment. Now, we deny care so we can pay the HMO medical director a bajillion dollars a year!

Critically reading a paper on therapy

Being able to critically read an article on therapy puts the power back in your hands, freeing you from an overreliance on "experts".  Reading a paper evaluating a therapeutic intervention requires addressing the same three basic issues as in other studies: validity, results, and relevance. The following sections will go over some of the details of critically reading a therapeutic paper. Keep in mind that in many instances when you need information quickly and easily, several other resources can provide you with critically appraised summaries (Cochrane Library, Best Evidence, ACP Journal Club, Evidence-based Practice, Evidence-based Medicine, Journal of Family Practice POEMs, etc.).