Skip to main content
Rediscover Dosage

Transparency

Intended use, limitations and current validation state.

Intended use: a pharmacist-facing, human-in-the-loop medication-review workspace for workflow design, stakeholder evaluation and progressive rollout.

Not intended for

Uses this product is explicitly not for.

  • Patient care, triage or clinical decision-making outside pharmacist review.
  • Autonomous or automated dosing of any kind.
  • Patient-facing use or patient self-titration.
  • Any use implying clinical or Canadian regulatory approval.
  • Any handling of real or identifiable patient information outside approved deployments.

Policies

Synthetic data, human review and source control.

Synthetic-data policy

The case library uses illustrative, non-identifiable cases and must not contain PHI.

Case library material supports workflow design, testing and review. It does not establish clinical or Canadian validation on its own.

Human-review requirement

AI assistance does not replace professional judgment, and no autonomous patient-specific dosage decision is provided.

Source control

Guidance is represented through approved source categories with citation references. Final corpus content is pending approval, and no guideline recommendation is quoted or invented in this build.

Governance documentation

Model card and data card.

Model card

Model overview

Purpose
Supports a pharmacist-led statin-related medication review by organizing case context, guidance and drafting into one traceable workflow.
Intended use
To assist a licensed pharmacist in preparing a structured, source-linked review summary that the pharmacist edits and approves.
Out of scope
Autonomous dosing, patient-specific dose selection, patient-facing use and any unsupervised clinical decision-making.

Inputs and outputs

Inputs
Structured case fields such as review objective, medication context, control context, tolerance notes and co-medication summary.
Outputs
A draft review narrative, an explanation view and a confidence state that can abstain when information is insufficient.
Human control
Every output is an editable draft until a reviewing pharmacist applies approval as a separate step.

Evaluation and governance

Evaluation program
Benchmarking, calibration and conformal validation are an active area of ongoing technical work.
Known limitations
Explanation factors describe association, not causation. Confidence states are directional rather than statistically calibrated today.
Validation status
No clinical or regulatory approval is claimed. The team is pursuing structured validation ahead of any future clinical claim.

Data card

Data overview

Data type
Illustrative, non-identifiable case material authored to represent realistic statin-related review scenarios.
Prohibited content
No names, addresses, dates of birth, health card numbers, contact details or other identifying data may appear.
Handling policy
Case material used across the public site and product preview stays non-identifiable and contains no protected health information.

Composition

Scenario coverage
Five illustrative scenarios: new review, inadequate control, possible tolerance concern, polypharmacy context and incomplete data.
Fields
Structured fields such as review objective, medication context, control context, missing information and audit events.
Guidance sources
Represented through approved source categories that anchor guidance to inspectable material.

Limitations

Representativeness
Illustrative cases are designed for workflow clarity and are not sampled from a real population.
Validation
Case material supports workflow design and product evaluation; clinical and Canadian validation studies are ongoing work.
Trend data
Trend visuals illustrate interface behaviour rather than measured outcomes.

Product truth registry

Capability-by-capability build status.

  • Guided Review Intake

    A structured intake sequence that organizes case information for pharmacist review.

  • Case Library

    Illustrative scenarios covering new review, inadequate control, tolerance concerns, polypharmacy context and incomplete data.

  • Statin-Intensity Review Summary

    A structured summary view describing the review context without producing a dose recommendation.

  • Human Approval Gate

    A separate approval control that a reviewing pharmacist operates deliberately.

  • Why This Result

    An explainability panel presenting the factors behind a review summary, with no causal claim implied.

  • Confidence and Abstention

    A panel that shows a normal review state and an explicit abstention state when information is insufficient.

  • Guideline Copilot

    A source-grounded question panel that returns answers with source-category citations.

  • AI-Assisted Draft

    A medication-review draft that remains fully editable before approval.

  • Editable Report Export

    A printable review preview built from the reviewed and approved draft.

  • Model and Data Cards

    Model card and data card describing intended use, inputs, limitations and governance.

  • Audit History

    A session record of review events such as case opened, draft generated and approval changed.

  • Product Direction

    A published roadmap describing what is available now, what is next and what is planned.

Validation state

No clinical or Canadian validation exists for this build.

Current validation state: no clinical validation, no Canadian regulatory approval, no benchmarking result and no calibrated or conformal guarantee exists for this build yet.

  • Live model inference and retrieval are being connected progressively as the platform matures.
  • No performance metric of any kind is claimed or displayed.
  • Trend visuals are illustrative and used to show interface behaviour.
  • The audit history is a session-local product record, not yet a production compliance log.
  • Positioning remains subject to ongoing MVP, business-plan and pitch-deck reconciliation.

Session-local audit history — not yet a production compliance log.

Questions about this build

This transparency page is kept current as capabilities change.