Clinicians should use the results of randomized clinical trials (RCTs) of groups of patients to guide their clinical practice. When deciding which management approach will be best for an individual patient, however, clinicians cannot always rely on the results of RCTs. An RCT that addresses the particular issue may not be available; for example, some conditions are so rare that randomized trials are not feasible. Furthermore, even when a relevant RCT generates a clear answer, its result may not apply to an individual patient. First, if the patient is very different from trial participants, the trial results may not be applicable to that patient (see Chapter 13.1, Applying Results to Individual Patients). Second, regardless of the overall trial results, some similar patients may benefit from a given therapy, whereas others receive no benefit. Clinicians may have particularly strong reservations about applying RCT results to individuals when results have revealed small treatment effects of questionable importance.
These considerations lead clinicians to conduct trials of therapy, in which the patient begins treatment and the subsequent clinical course determines whether treatment is continued. Many factors may, however, mislead physicians who are conducting conventional trials of therapy. The patient may have improved anyway, even without medication. Physicians' and patients' optimism may result in misinterpretation of the therapeutic trial results. Finally, people often feel better when they are taking a new medication even when it does not have any specific activity against their illness (the placebo effect); this may also result in misleading inferences regarding the value of the new treatment.
To avoid these pitfalls, clinicians must conduct trials of therapy with safeguards that minimize these biases. Potential safeguards include repeatedly administering and withdrawing the target treatment, performing quantitative measurements of the target outcomes, and keeping both patients and clinicians blind to the treatment being administered. Investigators routinely use such safeguards in RCTs that involve large numbers of patients.
To determine the best care for an individual patient, clinicians can conduct RCTs in individual patients (n-of-1 RCTs). In contrast to most of this book, which provides a guide to using the medical literature, this chapter provides an approach to applying the principles of evidence-based medicine toconduct an n-of-1 RCT in your own practice.
N-of-1 Randomized Clinical Trials: Study Design
Although there are many ways to conduct n-of-1 RCTs, the following is the method we have found to be most widely applicable:
A clinician and patient agree to test a therapy (the experimental therapy) for its ability to improve or control the symptoms, signs, or other manifestations (the treatment targets) of the patient's ailment.
The patient then undergoes pairs of treatment periods organized so that one period of each pair applies the experimental therapy and the other period applies either an alternative treatment or placebo (Figure 11.5-1). The order of these 2 periods within each pair ...