1. Purpose of this page
BayesPharma Innovator organizes evidence, models, assumptions, and review history around development decisions — including candidate selection, translational/first-in-human reasoning, dose and trial-design strategy, formulation bridging, and regulatory-evidence preparation. Several workspaces (for example, First-in-Human Dose, Dose Optimus Studio, the Pediatric Dose Calculator, PopPK/PBPK/PK-PD modeling, the Clinical Trial Simulator, and IND/NDA regulatory authoring) sit close to consequential scientific and regulatory questions. This page sets out the boundaries of what those tools do, so they are used appropriately and safely.
2. Not a medical device
BayesPharma Innovator is not a medical device, is not intended to diagnose, treat, cure, or prevent any disease or condition, and has not been cleared, approved, or classified as a medical device by the FDA, EMA, or any other regulatory authority. No content, calculation, model output, or workspace in BayesPharma Innovator should be interpreted as a medical-device function or as software that directs real-time patient care.
3. Not medical advice
Nothing produced by BayesPharma Innovator constitutes medical, clinical, dosing, or treatment advice for any individual patient. BayesPharma Innovator is built for use by pharmaceutical and life-sciences development teams working on investigational products, formulations, and evidence packages — it is not designed, validated, or intended for bedside use, direct patient care, or point-of-care clinical decision-making. If you are looking for patient-level dosing support, that is a distinct use case outside the scope of this product and this disclaimer.
4. Decision support, not a decision-maker
Every workspace in BayesPharma Innovator is designed to make the scientific basis of a decision easier to inspect: the question being asked, source evidence, assumptions, model diagnostics, uncertainty, uncertainty limitations, and reviewer interpretation. The platform does not make development, safety, regulatory, or business decisions on its own, and outputs should never be treated as an autonomous final answer. A qualified human reviewer must exercise independent scientific judgment on every consequential output.
5. Qualified personnel required
BayesPharma Innovator is intended for use by trained scientific, clinical, pharmacometric, regulatory, and quality personnel who are qualified — by education, training, and organizational authorization — to interpret the relevant scientific domain. It is not a substitute for:
- Qualified pharmacometricians reviewing PopPK, PBPK/PBBM, PK/PD, and QSP model assumptions, diagnostics, and limitations.
- Qualified toxicologists and translational scientists reviewing preclinical, IVIVE, and first-in-human safety-margin reasoning.
- Qualified regulatory affairs professionals reviewing submission strategy, CTD assembly, and agency-facing claims.
- Qualified CMC, formulation, and quality personnel reviewing developability, bridging, and control-strategy conclusions.
- Institutional review boards, safety committees, and other governance bodies your organization requires for the decision in question.
6. No guarantee of regulatory acceptance
BayesPharma Innovator can help organize, structure, and trace regulatory-relevant evidence — including for IND/NDA authoring, submission evidence mapping, and CTD assembly — but it does not guarantee that any agency (including the FDA, EMA, PMDA, or any other health authority) will accept, clear, or approve any submission, claim, endpoint, or evidence package prepared with its assistance. Regulatory acceptance depends on the underlying science, data quality, study conduct, applicable guidance, and the reviewing authority's own judgment, none of which Bayes Pharma.ai controls or can predict with certainty.
7. Model & data limitations
Models, calculators, and AI-assisted features in BayesPharma Innovator rely on assumptions, input data quality, and methodological choices that carry inherent uncertainty. Common sources of limitation include:
- Incomplete, outdated, or mis-specified input data supplied by the user or customer organization.
- Model structure, covariate, and parameter assumptions that may not generalize to every population, dose range, or clinical context.
- Extrapolation beyond the range of observed data (for example, first-in-human dose projection from preclinical species).
- Evolving scientific understanding, endpoint definitions, or regulatory expectations that a static model or evidence package may not reflect.
Diagnostics, confidence intervals, sensitivity analyses, and stated limitations shown in a workspace are there to be read and interpreted — not skipped.
8. Dosing & simulation workspace notices
First-in-Human Dose & Dose Optimus Studio
These workspaces support building a reviewable starting-dose or active-dose rationale from translational and safety evidence. They do not determine an actual first-in-human or therapeutic dose on their own. Final dose selection requires sponsor, medical, toxicology, and regulatory sign-off consistent with applicable guidance (for example, ICH M3(R2)/S9-type considerations) and institutional safety review.
Pediatric Dose Calculator
This workspace explores pediatric dose scenarios with explicit assumptions for development and planning purposes. It is not intended for use in prescribing, dispensing, or administering medication to an actual pediatric patient, and must not be used as a bedside or point-of-care dosing tool.
PopPK, PBPK/PBBM, PK/PD, QSP, and the Clinical Trial Simulator
These workspaces are modeling and simulation environments for development planning. Outputs reflect the model structure and inputs selected by the user and require qualified pharmacometric review, including assessment of model qualification and context of use, before informing a dose, trial-design, or regulatory decision.
Bioequivalence & Formulation Bridging
Outputs support internal evaluation of formulation, strength, or manufacturing-change bridging questions. They do not substitute for the bioequivalence and comparability studies, statistical analyses, and regulatory pathways required by applicable law and guidance.
9. Real-world evidence & HEOR notices
Workspaces in the Evidence Hub (RWD intake, RWE analysis, HEOR model building, and payer evidence) are designed to work with data that the customer organization has the rights to use, appropriately de-identified or governed by applicable consent, ethics, and privacy requirements. Bayes Pharma.ai does not independently verify the provenance, quality, or lawful basis of real-world or health-economic data uploaded by a customer, and the customer organization remains responsible for ensuring its use complies with applicable health-privacy, human-subjects, and data-protection law. See our Privacy Policy for how Bayes Pharma.ai treats health-related data generally.
10. Your responsibility
- Independently validate any model, calculation, claim, or AI-assisted output before relying on it for a scientific, safety, regulatory, or business decision.
- Ensure appropriate qualified personnel review and sign off on outputs used in a consequential decision, consistent with your organization's SOPs and governance requirements.
- Maintain your own model-validation, data-quality, and change-control practices independent of BayesPharma Innovator.
- Not rely on BayesPharma Innovator as a sole source of truth for any regulatory submission, safety determination, or patient-facing decision.
11. Reporting a concern
If you believe a workspace is producing incorrect, misleading, or unsafe output, or if you have a safety or quality concern about the product, please contact us promptly so we can investigate.