N-of-1 trial

In an N-of-1 trial the patient is the whole trial. Two options (a drug and placebo, two doses, two supportive medicines) are given in alternating periods whose order is randomised, ideally with the patient and clinician blinded, and an outcome that the patient can feel and record (pain, nausea, fatigue, sleep) is measured in every period. Because the same person receives both options several times, the comparison is free of the between-patient differences that make ordinary trials need hundreds of participants. Aggregating many N-of-1 trials gives a population estimate as well as an answer for each participant.

The design works only for treatments that act quickly, wear off quickly and treat a stable, chronic condition, so that each period starts from the same baseline. That rules out most anticancer therapy: tumours are not stable, drugs have carry-over effects lasting weeks, and the outcome that matters, survival, cannot be measured repeatedly in one person. Its natural home in oncology is supportive care and symptom control, where it can settle whether a given patient's fatigue responds to a stimulant, whether a particular antiemetic beats another for that patient, or whether a low dose of an old drug helps their appetite. The double-blind olanzapine trial at Tata Memorial answered the appetite question at group level in 124 patients; an N-of-1 series could answer it for the individual.

The phrase is also used loosely for something different: precision oncology in which one patient's tumour is sequenced and treated with a drug chosen for its mutations, sometimes called an N-of-1 approach. That is personalised treatment, not a randomised experiment, and it cannot tell whether the patient would have done as well without the drug. The technology entry on N-of-1 and rapid platform trials covers the infrastructure being built for both meanings.