Two major concerns in therapeutics are the efficacy (E) and adverse effects (Ae) exhibited by a pharmacological agent. Although these elements are studied routinely in both animals and humans during the drug development processes, the data are usually considered independently and in different contexts. Since E and Ae are different yet inseparable components in drug therapy, E should not be used as the sole descriptor for drug responses. To accommodate the influence of Ae on E, the present approach requires the observed Ae data to be transformed to the equivalencies of E data (Ae'). Using appropriate pharmacokinetic modeling techniques, this approach permits the prediction of the adjusted therapeutic effect (ATE) (i.e. , E minus Ae') as a function of time. Effects of pharmacodynamic variability of Ae due to variations in Hill's parameters (i.e., Ae (max), AeC (50), and n (Ae) ) on ATE were studied by computer simulations for a hypothetical one-compartment model drug that displays simple first-order absorption and elimination with central sites for E and Ae. An increase in Ae' (max) and a decrease in AeC (50) and n (Ae) cause a downward shift and peak inversion on the ATE versus time curves coupling with a longer duration of influence of Ae on E. Results also showed that the downward shift of these curves was more apparent with decreasing n (Ae) values and that peak inversion became less noticeable for n (Ae) values <1.5. Subsequent analyses established the optimal dose for the hypothetical drug studied. This approach allows a more comprehensive description of the time course of the overall drug responses and is potentially useful for therapeutic drug monitoring and dose selection during the drug development process.