“N-of-1 trials”, by helping to better define the individual response to treatments rather than an average responses over groups of patients, may not only ensure therapeutic effectiveness at an individual level, but also promote patient safety and optimal allocation of resources. In addition, this kind of trials may stimulate patients to take on more responsibility in their own care, which improves health outcomes : actually, Nikles et al. reported that patients having completed an N-of-1 trial had greater understanding, awareness and knowledge of their condition and felt a greater sense of control when it came to decisions about their health.
Single patient studies have been performed more or less formally for centuries. The first report identified is the 1786 Caleb Parry’s crossover studies comparing in an intra-patient way the purgative effects of Turkish and English rhubarb. In 1796, Edward Jenner proved the efficacy of the smallpox vaccine using a single patient trial concept, and Louis Pasteur did the same in 1885 with the rabies vaccine. The establishment of a N-of-1 trials methodology appears progressively with Paul Martini in 1932 (blinded administration in two periods crossover), and in 1953 with Hogben and Sim (use of comparators, blinding, wash-in and wash-out effects). The modern era of N-of-1 began in the 80s with Gordon Guyatt’s establishment of an N-of-1 service in Canada. Since then, a rigorous methodology has been established for the performance and reporting of N-of-1. Even though this kind of design have been known for decades and is now considered as level-one evidence by the Oxford Centre for Evidence-Based Medicine, they are still poorly adopted throughout the medical community.
N-of-1 trials are applicable to a wide range of diseases and treatments, but certain prerequisites are required to allow reliable measures and patients’ adherence. The disease has to be chronic, stable or slowly progressive, symptomatic or affecting biomarkers that enable appropriate monitoring, and the treatment must have a rapid onset of action, a reversible effect and a short wash-out time, to avoid prolonged trials with the risk of drop-out.
Many authors recommend N-of-1 trials to support difficult treatment decisions, e.g. involving costly drugs, off-label indications, uncertain efficacy, significant inconvenience or questionable tolerability. Various articles in literature report their use in a wide range of clinical conditions and questions. N-of-1 trials are particularly interesting in certain individual situations when objective assessment is difficult (e.g. to evaluate side-effects imputability, reveal psychological dependence or overcome patients’ or physicians’ preconceptions). These kinds of tests can also be aggregated to collect data for cumulative inference when randomized controlled trials are hardly feasible (rare conditions, difficulty to recruit patients, e.g. in palliative care or pediatric contexts). It is also an excellent way to carry out pilot studies for expensive treatment.
It must be made clear that at the individual patient level, an N-of-1 trial cannot be considered as “clinical research” according to the Swiss Human Research Act, as it does not aim to produce knowledge generalizable beyond a specific patient (Art. 3 al. a). Thus, it requires no approval by an Ethics Committee.
In addition to clinical gains, the use of N-of-1 trials in routine prescribing could also lower health costs, especially by avoiding expensive treatments not confirming their efficacy, long-term prescriptions in non-responders, or costs associated with adverse effects induced by futile treatments. N-of-1 trials, as a “diagnostic test for therapeutic efficacy”, are particularly interesting for medications with unpredictable clinical response, such as typically chronic pain medications. Their cost-effectiveness has been evaluated by several trials. Larson et al., reporting 2 years of experience in planning Nof-1 trials in an academic medical center, conclude that single patient trials are feasible and that their costs compare favorably with other conventional services. Scuffham et al. concluded that N-of-1 strategy offers a realistic and viable option for increasing access to selected high cost medications. Pope et al. compared the costs to standard practice for AINS treatment of osteoarthritis: N-of-1 trials were time-consuming and more expensive, but associated with better outcomes over 6 months. However, cost-effectiveness ratios for N-of-1 trials are difficult to evaluate accurately, necessitating a control with standard care and a sufficiently long observation time to take into account the savings resulting from prevention of side effects and futile expenses in non-responders. The development of centers offering this service might participate to increase the availability and decrease the costs for Nof-1 trials.
