This was first described in the statistical literature in 1947. It is designed to estimate an ED50 in clinical trials or toxicological tests, when a quantal response is measured (see Figure 11.1). However, it should be remembered that continuous responses can be converted into quantal responses with appropriate, prospective efficacy criteria, e.g. blood pressure is a continuous variable, but a drug may be deemed effective or ineffective by stating prospectively that a desired response is quantal-positive after a 15mmHg fall in diastolic blood pressure within 60 days of commencing therapy. Theoretically, this strategy can be implemented with groups of patients treated in the same way instead of individuals. Sometimes this technique is termed an 'adaptive' trial design, because dose size is adapted according to the response of the previous patient or group of patients.
Figure 11.1 Illustration of the Dixon 'up-down' method. Patients are treated in order of randomization, and the size of each test dose is determined by the response of the previous patient. In this example, patients who do not respond (O) are followed by a patient treated at the next higher dose size; vice versa, patients who do respond (•) are followed by a patient treated at the next lower dose. When the line has changed direction at least six times, the mean effective dose is about the ECj0 (in this case, 5 u, broken line). There are many, often more sophisticated, variants of this basic technique
Was this article helpful?