Use of sequential analysis in clinical trials of caries prophylactic agents

Using a sequential t test, the objectives of two 3-year experimental clinical trials were achieved more efficiently by using fewer subjects over a shorter time. The assumption of identical distributions of each batch of data analyzed was not apparent and further work to establish the robustness of sequential t tests is indicated. Identification of subjects with more consistent caries increments during the period of a trial would also help to satisfy this assumption.