Identification of the minimum effective dose for binary endpoints

Abstract

In response to an open topic posed by Tamhane and Dunnett (1999), this article suggests a confidence interval approach for determining th minimum effective dose of medicine with binary data. To address this issue, we employ a partitioning strategy together with a confidence interval technique. In medical research, binary data typically appear about dichotomous outcomes, such as the onset of a disease or the effectiveness of a medication. The suggested method not only determines the lowest dose of a medicine that would work, but it also offers an estimate of the therapeutic impact of the nearest unsuccessful amount. Those details are useful for further research in clinical trials. The procedure controls the family-wise error rate. The acetaminophen data example in this article was used to calculate the minimum effective dose, which was determined to be 151 $\mu$ g. The findings revealed that Wilson's score showed a strong coverage probability. The findings further revealed that the Wald and Binomial Distribution have poor coverage probability. We recommend that Wilson's score with a stepwise confidence-based procedure is more suitable for dose discovery studies with binary outcomes