In our last post, we explored some of the limitations of traditional clinical trial designs, especially when applied to precision medicine. In recent years, a new contestant has stepped into the fray: the N-of-1 trial.

The N-of-1 Trial: Crossover Meets Precision Medicine

A more extreme sibling of the crossover design, an N-of-1 trial is a multiple crossover study of a single individual, with repeated crossovers taking the place of multiple study subjects in the statistical analysis (Figure 1). Several recent commentaries have argued that N-of-1 trials are an effective mechanism for generating truly personalized clinical evidence, since they directly estimate a treatment’s effect on a particular individual without needing to explicitly account for prior history, disease status, or demographics.

Figure 1 Figure 1: Comparison of three different study designs. The number of people shown for each is the average sample size of studies with that design from Hopewell 2010. The N-of-1 study design shown is a particular type called a “singly-counterbalanced” study. The gray arrows shown for crossover and N-of-1 studies refer to “washout” periods, which are necessary to avoid carryover effects between treatment blocks. The syringes refer to measurements/evaluations of the patient.

Despite the fact that the AHRQ recently published a fifty-page manifesto on N-of-1 trials and touted their effectiveness in patient-centered outcomes research, N-of-1 studies are not popular, at least in the published clinical literature. Between 1986 and 2010, a nearly 25-year period, only 108 N-of-1 trials were published, involving a total of only 2154 patients. To put that number into perspective, in the period since the NIH launched its public clinical trial registry in 2000, 226,987 trials have been registered (83,416 between 2000-2010).

It seems likely that part of the reason for this lack of popularity may be the absence of standardized analysis methods and software for performing N-of-1 studies. Ad-hoc analysis techniques make it difficult to generate robust conclusions from N-of-1 studies; a significant fraction use no statistical analysis whatsoever, instead relying on graphical presentations of the data to guide inference. Another issue may be the perception that the results of an N-of-1 study are applicable only to a single patient. Physicians also cite lack of funding, ethical concerns, and an unwillingness to disrupt the doctor-patient relationship as reasons for avoiding N-of-1 studies.

Some Advantages of N-of-1 Trials

We at HD2i share AHRQ’s excitement about the potential of N-of-1 studies. Each N-of-1 study is essentially a formalized “story” about a particular patient. These studies interface naturally with contemporary ideas about how patient-level data can be used to generate clinical evidence, such as the Green Button, and could be combined with observational sources like electronic medical records.

Some other advantages of N-of-1 trials:

  • An N-of-1 study formalizes the process of searching for the right treatment and makes the search replicable and applicable to other patients. If multiple patients undergo the same search process, their results can be combined to yield a population-level crossover study. This enables patients, especially those with rare diseases, to benefit others separated by time or geography with no additional effort.
  • N-of-1 studies provide a new way for physicians to communicate with patients. Evidence shows that when patients are presented with evidence from N-of-1 studies, the majority (54%) make treatment decisions consistent with study findings, compared to only 8% inconsistent (38% were ambiguous).
  • As remote monitoring technologies such as wearables, continuous glucose monitors, environmental monitoring devices, etc. become more common, we gain the ability to monitor patients with much higher resolution than ever before. This will alleviate one of the principal limitations of N-of-1 studies (the inconvenience of repeat measurements, especially those performed in a clinic).
  • Because they are relatively inexpensive compared to traditional clinical studies, N-of-1 studies open the door to systematic investigations of treatments with little government or industry backing, such as complementary and alternative medicine, dietary supplements, and biohacking.

Single-Patient Trials: Docking Research and Clinical Care

Perhaps the biggest advantage of N-of-1 studies is the resemblance they bear to standard clinical practice, especially in the management of chronic disease. A patient with major depression, for example, will often experiment with several different antidepressants before identifying one that provides relief from disease symptoms while minimizing side effects. In practice, a physician will prescribe a drug and then check back in with the patient in a few weeks or months. Based on the patient’s feedback, the physician may elect to change the drug or its dosage, or do nothing.

With a few modifications, such practices could easily be formalized into N-of-1 trials.

  1. It is unusual for a patient to revisit an earlier treatment that was deemed ineffective. In the N-of-1 paradigm, treatment sequences would be decided in advance by doctor and patient, and block durations, washout periods, and other study parameters would be designed to achieve the most robust possible evidence for a patient’s optimal treatment.
  2. Treatment decisions are based on the patient’s feelings at the time of the appointment and not on statistical analysis of a predefined outcome. In the N-of-1 paradigm, outcomes would be based on validated clinical instruments and other objective measures of success, enabling trial results to translate effectively across patients and institutions.
  3. Treatment decisions are adaptive, and do not follow a predefined protocol. We envision future methodological development around N-of-1 studies that enables more complex trial designs, such as adaptive protocols.

Building a Platform for N-of-1 Studies

At HD2i we are in the process of developing a software platform to enable the efficient design, execution, analysis, and meta-analysis of N-of-1 studies. Our platform will be implemented first in the Next-Generation Health Clinic at INGH, and later released to the public.

We are excited by the possibility of creating an entirely new vehicle for clinical research that addresses the need for high-quality evidence in the era of precision medicine. We envision a future in which a physician’s treatment decisions are supported by a comprehensive database of N-of-1 trials. With a little forethought, we should be able to standardize these trials appropriately so that physicians have access to specific, relevant data for the individual patients they are trying to treat.

All comments and suggestions are welcome and appreciated.

Notes and Further Reading

The Agency for Healthcare Research and Quality published a manifesto on N-of-1 studies in 2014 that includes everything from statistical analysis to financial considerations to ethics.

Two great summaries of the statistical and methodological considerations surrounding N-of-1 trials can be found here and here. The second offers a broader perspective on the future of N-of-1 trials and their role in clinical practice, especially with respect to precision medicine.