Balsam Semantica Computable clinical meaning

Clinical Meaning Integrity · From evidence to claim

Clinical Claim Review

Does the evidence support what your product claims?.

An independent review of clinical claims made by health-AI companies.The review examines whether published evidence supports each claim and whether the claim accurately reflects what the evidence shows.
It identifies claims that go beyond the available evidence, leave out important conditions, or change the meaning of the original findings.
A study may report good results in a specific population or setting. A product claim may then present those results as generally applicable. The study itself may be sound, but the claim no longer reflects what was actually established.
Clinical Claim Review identifies these differences.

What is reviewed

Each claim is examined from two perspectives:
• Is there published evidence to support it?
• Does the claim accurately represent what that evidence shows?
The review considers factors that can change the meaning of a finding, including:
• Patient population
• Clinical setting
• Comparator
• Measured outcome
• Study design
• Level of certainty
A claim may have genuine evidence behind it and still overstate what that evidence establishes.

Who it is for

This service is intended for:
Health-AI companies
Companies preparing to publish or revise clinical claims, including during:
• Product launch
• Entry into a new market

• Website or app updates

• Preparation of brochures and white papers

• Development of patient-facing materials

Claims that are already published can also be reviewed.
Investors
Investors can commission an independent review of a company's clinical claims as part of their assessment of the product and its supporting evidence.

What you provide

You provide:
• The product website and relevant materials, such as app store descriptions, brochures, white papers or patient-facing content.
• Up to ten clinical claims for review.

You may also submit the materials without selecting specific claims.
 Claims will then be selected based on their clinical importance, strength, and potential to overstate the evidence.
The review is based on published evidence.
Unpublished data supplied by the client may also be examined. However, it will be identified as unpublished in the report and cannot, on its own, establish that a claim is supported.
No patient data is required or accepted.

What you receive

You receive a written report containing a numbered register of the claims reviewed.
For each claim, the report includes:
• The original claim.
• The evidence searched for and the evidence identified.

• The review classification.

• What the available evidence actually supports.
• Any difference between the claim and the meaning of the evidence.
The report makes clear what the evidence supports, what requires qualification, and what goes beyond the available findings.

The four review classifications

1. Supported
The published evidence supports the claim as written.
2. Supported with a condition
The evidence supports the claim only under a specific condition that the claim does not mention.
This may relate to the patient population, clinical setting, comparator, workflow, outcome or study design.
3. No supporting evidence found
No published evidence supporting the claim was identified within the scope of the review.
This does not mean that the claim is necessarily false. It means that supporting evidence was not found.
4. Meaning altered
Relevant evidence exists, but the claim describes something different from what the evidence actually establishes.
The original study may be sound. The difference arises when its findings are presented in a way that changes their meaning.
Each review has a fixed scope, a fixed fee and a stated delivery period, sent in writing on request.
The engagement follows a simple process:
1. Submit your request.
2. Receive the fee and terms in writing.

3. Work begins on written acceptance of the terms and full advance payment.
4. Share the claims and relevant product materials.
The review covers up to ten claims and is delivered within seven business days of receiving the complete materials and payment.
All work is conducted asynchronously and delivered in writing. No meetings are required.
More than ten claims are reviewed as a second review, on the same terms.

Engagement and delivery

What this review is not about

Clinical Claim Review examines the accuracy of claims made about a product.
It does not include:
• Technical testing or validation of the product.
• Clinical safety assessment.
• Regulatory opinions or medical device conformity assessments.
• Legal or regulatory advice.
• Product endorsement or confirmation that a product is safe, effective, compliant or suitable for clinical use.

The review assesses the claims and their supporting evidence. It does not certify or approve the product itself.

Independence and use of the report

The review is conducted independently and prepared for the commissioning client.
To preserve that independence:
• The reviewer's name may not be used to endorse, promote or validate the product.
• The report may not be presented as regulatory approval, certification, endorsement or clinical validation.
• The report may not be published or distributed, and its findings may not be quoted in marketing materials, unless agreed in writing

See a sample review

A full sample review of claims made for Arabic-capable ambient AI scribes is published in Writing.
  
Read the sample review →

If your organisation needs an independent review of its clinical claims, submit a request.You will receive the fee, scope and delivery terms in writing.

Request the fee and terms →

[email protected]

Clinical Claim Review · up to ten claims · seven business days