Robust and efficient estimation of effective dose

In dose-response studies, experimenters are often interested in estimating the effective dose EDp, the dose at which the probability of response is p, 0 < p < 1. For instance, in pharmacology studies one is typically interested in estimating ED0.5, whereas in toxicology studies the main interest is EDp for smaller values of p. In this context, methods based on parametric, semiparametric, and nonparametric models have been developed. Traditional estimators based on parametric models are generally efficient but are not robust to model misspecification. On the other hand, nonparametric estimators are robust to model misspecification but are less efficient. Semiparametric methods are a compromise. Two new parametric methods are presented in this paper for estimating EDp using minimum-distance techniques. It is shown that the proposed estimators are efficient under the model and simultaneously have some desirable robustness properties. The asymptotic properties such as consistency and asymptotic normality are studied. Small-sample and robustness properties of the proposed estimators are examined and are compared with traditional estimators using Monte Carlo studies. Two real-data examples are also analyzed. (C) 2015 Elsevier B.V. All rights reserved.