BOIN Design Calculator & Decision Table

Bayesian Optimal Interval (BOIN) design for phase I clinical trials

This tool implements the Bayesian Optimal Interval (BOIN) design for Phase I dose-finding: it generates the escalation/de-escalation decision table for a full trial, or a single dose decision for an observed cohort, using boundaries derived from a closed-form posterior-probability calculation (Yuan, Hess, Liu, et al.).

Design Parameters

Number of Doses

Starting Dose Level

Cohort Size

Number of Cohorts

Maximum patients per dose (m)

Target Toxicity & Boundaries

Target Toxicity Probability (φ)

Lower Toxicity Bound (φ₁)

Upper Toxicity Bound (φ₂)

Modify decision from de-escalation to stay when observing 1 DLT out of 3 patients

Modify decision from stay to de-escalation when observing 2 DLTs out of 6 patients

Perform accelerated titration

Dose Decision Calculation

Number of Patients at Current Dose (n)

Number of Observed DLTs (x)

Decision Result

Decision Table

Generate a decision table with columns representing the number of evaluable subjects (from the cohort size to m) and rows indicating:

Number of evaluable subjects
Escalate if # of DLT ≤
De-escalate if # of DLT ≥

Decision Table

Frequently asked questions about this calculator →

Related Tools