Application for evaluating Phase I methods using a non-parametric optimal benchmark

Nolan A. Wages and Nikole Varhegyi

Division of Translational Research & Applied Statistics, University of Virginia;

1. Enter a set of assumed true DLT probabilities, separated by commas. The length of this set should be equal to the number of dose levels.

2. Enter the target DLT rate for the study.

3. Enter the sample size for the study.

4. Enter the number of simulated trials to be generated. A minimum of 1000 is recommended.

5. Set the seed of the random number generator.


This application provides operating characteristics for accuracy of the non-parametric optimal benchmark [1].

The Accuracy Index is given in Equation 6.1 in Cheung [2].

Click here to view the R Code for the benchmark.


[1] O'Quigley J, Paoletti X, Maccario J. (2002). Non-parametric optimal design in dose-finding studies, Biostatistics; 3 (1): 51-56.

[2] Cheung YK (2011). Dose-finding by the continual reassessment method. Chapman and Hall/CRC press: New York.