Forecast trial-ready effect sizes with patient stratification before a phase-advancement decision.
Enter your program directly
The OPTIMIZATION mission is the workflow for an asset you already have in hand — a candidate already in or approaching the clinic — where the question is no longer "is this target real?" but "how large an effect can this program deliver, and in which patients?" Unlike the FIC Targets mission, you do not pass through the Generator first. You enter the Validator directly with the program parameters you already know.
When to use this mission: You have a defined drug and indication and you need a forecasted, trial-ready effect size — stratified by clinical parameters — to support a phase-advancement or program-prioritization decision.
1
Enter your program directly
From the Hub, click "Start" on the Programs card. The Validator opens directly — there is no Generator step for this mission. Set the program parameters:
Disease / indication — the condition the program targets.
Therapy — your candidate drug or modality.
BVCT Type — set to ONGOING if the trial is active, or PLANNING if you are designing the next study.
NCT ID (optional) — paste the registry ID to align the simulation with the real protocol.
Tip: For a program in flight, BVCT Type = ONGOING plus the NCT ID gives you the tightest alignment between the virtual trial and the protocol your team is actually running.
Why Programs skips the Generator
The Generator exists to build and annotate a target-disease hypothesis from scratch. In the OPTIMIZATION mission the asset is already defined, so the hypothesis is implicit in the program parameters you enter. You go straight to stratification and effect-size forecasting.
Stratify by clinical parameters
2
Stratify by clinical parameters
The platform generates clinical stratification parameters specific to your indication, and assigns each subgroup its own predicted effect size. The parameters are chosen for the disease you entered — for example:
Indication
Stratification parameter
Type 2 diabetes
HbA1c baseline
Lung cancer
EGFR mutation status
Immunotherapy programs
PD-L1 expression
Each stratum receives an independent forecasted effect size, so you can see where the program's signal concentrates rather than reading a single blended average.
Tip: Notice that renal function appears as a stratification parameter, not just an eligibility gate. The platform predicts different effect sizes for patients with normal vs impaired renal function — a distinction a simple inclusion/exclusion criterion would hide.
Stratification vs eligibility
An eligibility criterion answers a yes/no question: is this patient allowed into the trial? A stratification parameter answers a quantitative one: how large is the predicted effect within this subgroup? OPTIMIZATION treats clinically meaningful axes — renal function, biomarker status, baseline severity — as stratification axes so the forecast tells you not just who can enroll, but where the program performs best.
Per-stratum forecasting: Because every subgroup carries its own predicted effect size, the OPTIMIZATION mission surfaces enrichment opportunities — the subpopulations that, if prioritized in enrollment, would maximize the trial's measured effect.
Generate and approve protocols
3
Generate and approve protocols
Protocol cards display the per-stratum predictions side by side. Work through them to:
Identify the driving subgroup — which stratum carries the strongest predicted signal.
Compare GO / NO-GO across strata — a program can be NO-GO on the overall population yet GO in an enriched subgroup.
Download PDFs for your statistician and CRO.
Verify and approve the protocols you want to advance.
Tip: When the overall result is borderline, read the per-stratum cards before deciding. A CONDITIONAL GO at the population level often resolves cleanly once you see which subgroup is driving — or diluting — the signal.
Reading the recommendation
Badge
Meaning
Action
GO
Predicted effect size meets or exceeds your target threshold.
Advance the program to the next stage.
CONDITIONAL GO
Borderline overall — the answer depends on specific subgroups.
Review the per-stratum cards before deciding.
NO-GO
Predicted effect size falls below threshold.
Reassess the design, enrich enrollment, or reprioritize.
Once approved, the downloaded protocol package is ready to hand to your biostatistics and clinical-operations teams. The forecast is a model-based decision-support analysis — pair it with your own clinical judgement before committing trial budget.
ONGOING type and multi-NCT alignment
OPTIMIZATION is where stratification and clinical-trial alignment meet, so it is also where the ONGOING BVCT Type does the most work. Setting BVCT Type to ONGOING tells the platform that a real trial is already running, and that the virtual trial should mirror that protocol rather than design one from scratch.
Aligning to a single ongoing trial
Paste the NCT ID of the active study in the program-entry step. The platform fetches the registered protocol so the simulated arms, dosing, and endpoints track the real trial. The per-stratum forecast then reads as a prediction for the trial your team is currently enrolling — useful for anticipating interim readouts and pressure-testing your powering assumptions before data locks.
Why ONGOING + NCT matters: A forecast that is anchored to the actual protocol is directly comparable to the trial's own results as they emerge. That alignment is what makes the prediction actionable for a live program rather than a hypothetical one.
Multiple ongoing trials in one program
Mature programs rarely run a single study. When several registered trials cover the same asset — different lines of therapy, geographies, or combination arms — align each relevant NCT so the stratified forecast reflects the full set of ongoing studies rather than one in isolation. This is especially valuable when you are deciding which of several active trials is most likely to deliver the program-defining result.
Setting
Use when
What the forecast tells you
ONGOING + NCT ID
An active trial you want to track.
Predicted effect for the running protocol, per stratum.
ONGOING + multiple NCTs
Several active trials on the same asset.
Where the program's strongest signal sits across all active studies.
PLANNING (no NCT)
Designing the next study.
Forecast for a trial you are still shaping.
Tip: If you are not sure which active trial best represents the program, align each one and compare the driving subgroups across the protocol cards. The trial whose enriched stratum shows the strongest, most robust forecast is usually the one to prioritize.
For the platform's prospective track record on real, registered trials, see data.bioinvestgpt.com. To open the Validator and start a Programs run, use the BVCT Validator in a new tab.
BioinvestGPT BVCT Platform User Guide — Clinically Derisk Programs (OPTIMIZATION). BVCT outputs are model-based decision-support analyses, not investment advice. Prospective track record: data.bioinvestgpt.com.
