Ask Mallard
Study design

Bayesian optimal interval design

4 min read

The Bayesian optimal interval design compares an observed toxicity proportion with prespecified boundaries to guide the next dose.

NHMRC EVIDENCE LEVEL · 2009Outside the NHMRC hierarchy
Why this level?not one of the five research areas this hierarchy covers. NHMRC 2009 source

Dose selection is a feasibility and safety question, not one of the five research areas this hierarchy covers. It does not establish comparative treatment benefit.

Fit

When to use it, and when not to

Use it when

  • You need to select a dose for later study under a prespecified safety plan.
  • The toxicity outcome and its observation window can guide the next allocation.

Not when

Before you start

What you need

Decide these first
  1. Toxicity definition and targetDefine a dose-limiting toxicity and the full assessment window before recruitment.
  2. Dose levels and safety limitsFix the starting dose, allowed moves, stopping rules and independent safety oversight.
  3. Simulated trial scenariosVary toxicity curves, accrual and delayed harm. Check selection accuracy and exposure to unsafe doses.
It rests on these
  1. Pending toxicity is not absence of toxicity.How to check: Wait for the required window or use a validated extension designed for partial follow-up.
  2. The dose ordering and safety target fit the treatment.How to check: Toxicity-focused dose selection may not identify the best benefit-risk balance or biological dose.

Plain language

How it works

Choose a target toxicity probability and define acceptable departures before recruitment. Derive escalation and de-escalation boundaries from those choices. After completed outcomes, compare the observed toxicity proportion at the current dose with these boundaries. Escalate at or below the lower boundary. Stay strictly between boundaries, and de-escalate at or above the upper boundary. Apply the overdose elimination rule too. Use the prespecified final dose-selection procedure at the end.

Local decisions and final selection

BOIN means Bayesian optimal interval. Its next-dose rule and final dose choice are different steps. Simulate both, with stopping and safety restrictions. A design for completed outcomes should not treat pending toxicity as a negative result.

The structure of the approach
Prespecified safety rules govern each level decision. Pending follow-up is not absence of toxicity. The complete design needs simulation under plausible scenarios. Selecting or expanding a level does not confirm efficacy.Prespecified safety rules govern each level decisionDefine target and safetyrulesObserve outcomes andpending follow-upApply the chosen designand safety constraintsEscalate only whenpermittedRetain or reduce levelStop when requiredPending follow-up is not absence of toxicity. The complete design needs simulation under plausiblescenarios. Selecting or expanding a level does not confirm efficacy.
Conceptual diagram. It shows the structure of the idea, not data from any study.

Worked example

A worked example

Invented for illustration
Setting
An early-phase trial follows a prespecified toxicity window.
Question
What do the observed counts imply under the stated design?
Approach
At one dose, two of eight patients have completed toxicity outcomes. Apply the protocol boundaries, not thresholds chosen from these results.
Result
Two of eight patients have toxicities: 25 in 100.
What it means
The observed share alone does not determine the action. The target, boundaries, safety rule and final selection rule must be specified.

In practice

Recent studies PubMed search · not vetted

Searching PubMed…

Watch out

Common mistakes

The one to remember

BOIN boundaries are design choices fixed before recruitment. The current proportion does not replace overdose elimination or the final selection procedure.

  • Equating dose selection with proof of efficacy.Instead: Selection is uncertain. Confirm benefit in a suitable later trial.
  • Changing allocation rules after a worrying result.Instead: Apply the prespecified safety and stopping rules with the safety team.

Inside Mallard

How Mallard uses this

Plans do not link to this design yet. Mallard has no sample-size form for the Bayesian optimal interval design. A plan sized by simulation gets no headline sample size; Mallard does not run that simulation.

No Mallard code example points to this page.
Atlas node design.boin, the id a plan cue records.

How to report it

Use CONSORT-DEFINE to report the dose-selection goal, toxicity window, allocation, safety and stopping rules, deviations and participant flow. Separate observed toxicity from model estimates. State the final dose recommendation and its uncertainty. A reporting guideline is not a design manual: use it to make the study transparent, not as a substitute for careful design. Check the target journal's own requirements as well, which this page cannot know.

Related Learn pages

Related designs and methods

Evidence

Sources & further reading

  1. Bayesian Optimal Interval Design: A Simple and Well-Performing Design for Phase I Oncology Trials (opens in a new tab) · Clin Cancer Res · doi 10.1158/1078-0432.CCR-16-0592
  2. Enhancing reporting quality and impact of early phase dose-finding clinical trials: CONSORT Dose-finding Extension (CONSORT-DEFINE) guidance (opens in a new tab) · BMJ · doi 10.1136/bmj-2023-076387
Prepared with AI assistance. PubMed records checked on 2026-11-07 UTC for title, authors, year and available DOI. Record verified, confirm it supports this use. No independent professional or methodological approval is claimed.
Last source review: 2026-11-07
AI-generated educational content
Mallard uses reputable methodological and reporting sources, but this material may contain errors or omit important context. Verify important methodological decisions with the cited sources and, when appropriate, a qualified research methodologist or statistician.