Determining Optimal Therapy — Randomized Trials in Individual Patients

TLDR

Individual patient treatment is often treated as an experiment, but its bias has led to a demand for multi‑patient, double‑blind randomized trials, which are frequently infeasible or difficult to generalize to single patients. The authors propose using single‑patient, double‑blind randomized trials with paired active and placebo or alternative treatments to determine optimal therapy. The method randomizes the order of paired active and placebo/alternative treatments for a single patient, uses.

Abstract

Abstract Although the treatment of an individual patient in routine clinical practice has been likened to an experiment, the method is so susceptible to bias that we have come to demand multi-patient, double-blind, randomized controlled trials on matters of efficacy. Unfortunately, such trials have not or cannot be carried out for many clinical disorders; even when they have been executed their results may be difficult to extrapolate to individual patients. To resolve this problem, we have begun to use double-blind randomized trials in which a single patient under-goes a series of pairs of treatments, consisting of one active and one placebo or alternative treatment per pair, with the order determined by random allocation. Appropriate treatment targets (signs, symptoms, or laboratory tests) are used as the measure of efficacy, and the trial is continued until efficacy is established or disproved. We describe such a trial, which resulted in a dramatically beneficial modification of treatment in a patient with partially reversible airflow limitation. We have established a clinical service that facilitates the widespread use of the method in our community. (N Engl J Med 1986; 314:889–92.)