The so-called n-of-1 trial provides data on individual responses to treatment options, and this study design can supply rigorous information about the efficacy of a treatment and permit the optimal treatment to be identified for the individual patient, according to a UK medical statistician.
The traditional randomized controlled superiority trial has become the standard for assessing the efficacy of treatments in clinical medicine, the assumption being that the observed treatment effects can be generalized to all patients in the population, but this assumption is unlikely to be true, according to Philip Sedgwick, PhD, a reader in medical statistics and medical education at the Centre for Medical and Healthcare Education, St. George’s, University of London.
“Patients in the population will differ in their response to a treatment and may even benefit from the treatment shown to be inferior in a trial. Therefore, the uniqueness of patients may need to be accounted for in clinical practice,” he noted in an article published online on April 10th by bmj.com. “Nonetheless, the n-of-1 trial is not always a rational and effective study design for all conditions, and its usefulness in identifying the optimal treatment for individual patients might warrant investigation.”
Sedgwick described the case of researchers who assessed the effectiveness of “n-of-1” trials for the short-term choice of drugs for osteoarthritis. The efficacy of sustained release paracetamol was compared with celecoxib in the management of symptoms associated with osteoarthritis. A series of double-blind randomized n-of-1 controlled trials using a double dummy design was performed. The intervention was sustained release paracetamol (two 665 mg tablets, three times a day), or celecoxib (200 mg daily, or 200 mg twice a day for those who were already using this dose). Each treatment regimen was taken for two weeks, administered for three treatment cycles. The primary outcome measures included pain, stiffness, and functional limitation scores, as well as preferred treatment and adverse effects.
Participants were eligible if they had osteoarthritis in multiple sites, with pain for at least one month. In addition, their pain was severe enough to warrant consideration of the long term use of celecoxib, but the efficacy of this drug was in doubt. In total, 59 patients were recruited, with 41 completing an n-of-1 trial. It was reported that 33 of the 41 patients could not identify a difference between sustained release paracetamol and celecoxib in terms of overall symptom relief. Of the eight patients who were able to identify differences, seven reported better relief with celecoxib and one with sustained release
“It was concluded that n-of-1 trials may provide a rational and effective method to help choose the most effective drug for patients with osteoarthritis,” stated Sedgwick.
“Advisory guidelines based on results from clinical trials recommend paracetamol as the agent of first choice in managing the symptoms associated with osteoarthritis. Nonetheless, some patients, particularly those with moderate to severe pain or those whose pain is unresponsive to paracetamol, prefer non-steroidal anti-inflammatory drugs (NSAIDs).”
The n-of-1 trial, sometimes referred to as a single patient trial, receives its name by virtue of its sample size: n is equal to one, he explained. It is a randomized controlled crossover trial in a single patient, and provides a pragmatic approach to individual patient care. The patient receives each treatment, with the treatment order decided at random. Each trial is analyzed separately, with the outcomes compared within the patient to establish the optimal treatment for that patient.
“The n-of-1 trial has a within subject design, with outcomes for treatments compared within patients so that the optimal treatment for the patient can be established. Therefore, each patient acted as his or her own control,” wrote Sedgwick. “The n-of-1 trial is similar in design to the traditional randomized crossover trial... In a randomized crossover trial, patients are randomized to one of two interventions and then they receive the other intervention after a washout period.”
Both study designs compare outcomes within subjects, but there are noticeable differences. For instance, in an n-of-1 trial, the unit of randomization is the treatment order within a treatment cycle for a patient. This is in contrast to the randomized crossover trial, where the patient is the unit of randomization. In a randomized crossover trial, participants typically receive only one cycle of treatment, whereas in an n-of-1 trial they receive at least two cycles. In particular, the n-of-1 trial establishes which treatment is best for a patient. In comparison, the traditional crossover trial design estimates which treatment is best for the population.
Despite the dissimilarities between the two study designs, they are both suitable only for studying chronic stable conditions that are not resolved by treatment, he concluded.
Reference: BMJ 2014;348:g2674 doi: 10.1136/bmj.g2674