Dose Optimization with Backfill and Adaptive Randomization (BARD)

PID: 1162 ; V1.0.4.0 ; Last Updated: 05/19/2025

Yixuan Zhao, Rachael Liu, Jianchang Lin, Ying Yuan

Department of Biostatistics, MD Anderson Cancer Center

Enter design parameters in the following panels.

Doses & Sample Size



Stage 1: close a dose for backfilling if the number of patients assigned to the dose reaches \(n_{cap}\), where \(n_{cap}\) =


Stage 1 Target Toxicity Probability




Stage 1 Overdose Control

Eliminate dose \(j\) if\(\ Pr\) (\(p_{_j}\)> \(\phi\) | \(data\)) > \(p_{_E}\)



Stage 2 OBD Identification






The dose identified as OBD must satisfy the following safety and response criteria:






Simulation

Enter Simulation Scenarios




For each scenario, enter the true DLT rate and populational response rate of each dose level:








Prognostic Factors


Enter the number of levels of each prognostic factor:













Please make sure that you have set up Trial Setting and Simulation before generating the protocol.




Current Dose Escalation Cohort

1. Dose level of the current dose escalation cohort:

2. Is the current dose esclation cohort fully filled?

3. Is the DLT assessment completed on the current dose escalation cohort?

4. Is the dose escalation ended?

Enter the trial data:


Dose level
# of patients treated
# of patients completed DLT assessment
# of patients with DLT
# of patients with response


Enter trial data:


Dose level
Number of evaluable patients treated
Number of patients with dose limiting toxicity



Prognostic Factors & Dose Arms

Enter the number of levels of each prognostic factor:


Upload the Data of Enrolled Patients



Input the Data of New Patient

Enter ID and prognostic factors of the new patient:


Seed & Probability of Minimization