Generating High-Quality Clinical Evidence in Palliative Care Using n-of-1 Trials

Randomized controlled trials (RCTs) are the gold standard by which the quality of evidence about treatment effectiveness is judged. ¹ In patients receiving palliative care, many RCTs fail because it is so difficult to recruit and retain enough participants to achieve the required sample sizes. ² Alternative means of conducting trials need evaluation; n-of-1 trials may provide one such alternative mechanism.

N-of-1 trials are multiple-cycle, double-blinded, placebo-controlled crossover trials using standardized measures of effect, usually used for testing the effectiveness of medicines in individual patients. ³ The randomization order is independently generated for each patient. At the end of the trial the order is revealed, and the patient response is compared against the order of test and comparator treatments. N-of-1 trials provide the strongest evidence possible about the efficacy of a treatment in an individual patient.^3,4

It is possible to combine multiple n-of-1 trials to produce an estimate of population effect, with similar rigor to that achieved by an RCT. ⁵ This technique offers two primary advantages not currently available in palliative care. The first is that the individual patient receives the strongest evidence possible about the effectiveness of the test treatment against the comparator for them. Moreover, this result is available to the patient and clinician shortly after that patient's trial has concluded; treatment can then be tailored to the individual. These data are not available to patients in a conventional RCT until all participants in the trial have completed their intervention. In the palliative care situation, that is likely to be too long a wait for most participants. The second is that a series of such individual trials of a given treatment can yield an estimate of the population effect comparable to a full-fledged RCT, and requires far fewer participants to gather that evidence.^5,6 The participants are effectively entered into both arms of the trial, thus making the comparator groups perfectly matched. Furthermore, each repeated treatment cycle provides more information so participants can contribute different amounts of data to the trial. Data from each completed cycle of treatment and comparator can contribute to the overall result. This will allow the accumulation of high-level evidence previously very difficult to gather.

N-of-1 methods are suitable where (1) the condition remains relatively stable (to the extent that changes observed within replicate pairs cannot be attributed to disease progression); (2) the medication does not alter the pathophysiology of the condition; (3) the medication has a short half-life; (4) there is validated measurement of effect; (5) and when there is considerable patient heterogeneity in response to the treatment. ⁷ In palliative care, there are few reports of formal multiple cross-over n-of-1 trials. ⁸

There are, however, limitations of the method. Many treatments in palliative care are not suitable for this method because their half-lives are too long or the treatment effect is not immediate, causing the trial length to be impractical.

The multiple cycles require the trial to be longer than a conventional cross-over trial. Longer trials risk the possibility that the underlying condition may deteriorate within a cycle, thus potentially confounding the observed effect. As individual patient results are fed back immediately, the investigator is at risk of bias, being unblinded to the population effect of the treatment. In comparison, in a standard RCT, the results are not revealed until all participants have completed the trial. Finally, the smaller sample size required raises the question of representativeness of the participants and the applicability of the result to the general population. Careful description of the n-of-1 trial population, and a comparison with the broader population of interest is important.

A list of medicines commonly used in palliative care that have the appropriate characteristics for assessment using n-of-1 trials has been published previously by our group. ⁹ The article by Hardy et al. ¹⁰ in this journal describes a pilot study undertaken to determine the test dose of methylphenidate in improving fatigue in a subsequent trial using n-of-1 methodology. Methylphenidate is an ideal drug to use within this methodology as it meets all the criteria as described above. The n-of-1 method requires proof of concept, however, so that its place in the research armamentarium and utility in the palliative care population is more clearly understood. The availability of this information will influence research policy and change clinical practice to reflect improved availability of high-quality evidence.