Balsam Semantica Computable clinical meaning

About the practice

Independant by design

Balsam Semantica is an independent, physician-led practice dedicated to computable clinical meaning - the accurate, context-aware representation of what a clinical encounter established, including what the patient reported, what the examination showed, and what the clinician concluded.
That representation has to be in a form machines can act on without obscuring what matters clinically.
 The practice approaches clinical AI through the integrity of the information these systems depend on. It follows that meaning as it moves between clinical records, terminology systems, structured data, and computational technologies.

THE CHALLENGE 

When clinical meaning is lost

Clinical information does not keep its full meaning simply because it has been recorded.
 A statement can be transcribed accurately and still lose essential distinctions during structuring, terminology mapping, coding, or later computational use.
Negation, uncertainty, temporality, and patient context often decide how information should be read.
 When they disappear, downstream systems go on processing data that no longer represents what the encounter established. This is not a technical inconsistency to be tidied away. It shapes the reliability of documentation, the interpretation of health data, and the foundations on which clinical AI is built.
Balsam Semantica locates where these changes occur and asks whether the resulting representation remains faithful to its source.

FOUNDER

A physician’s perspective on semantic integrity

Dr. Balsam Khalil holds an MD from Tishreen University, Syria, with clinical training in pediatrics. Her training in revenue cycle management and clinical documentation gave her a close view of the point where clinical encounters become codes, claims, and structured records, and of what tends to be lost in that passage.
 In 2026, she held a clinical AI leadership role with a health AI company. There she reviewed ambient AI-generated clinical documentation and the SNOMED CT mappings behind it. That work sharpened the question at the centre of her practice: whether what a clinical encounter established remains accurate and usable once it takes structured, computational form.

SERVICES

Reviews & Assesment 

The practice currently offers independent reviews that test whether a representation preserves the distinctions clinical interpretation depends on: negation, uncertainty, temporality, and context.
• Clinical Claim Review: health-AI product claims, set against published evidence. • Clinical AI Evaluation: simulated clinical interactions and AI-generated documentation, reviewed for the preservation of clinical meaning. • Terminology Mapping Review: mappings between local clinical terms and standard terminologies, assessed against the meaning intended for their stated purpose.
CIRI™ (Clinical Intelligence Readiness Instrument) is a separate framework for assessing an organization's readiness to govern clinical AI and preserve clinical meaning. It covers governance, clinical practice, and technical-semantic foundations. It is currently in development and not yet validated.

TRAINING

Grounded in the standards

Dr. Khalil's professional development centres on clinical terminology, health information systems, and the semantic foundations of clinical AI. Her training includes:
• SNOMED CT Foundation: SNOMED International • Revenue Cycle Management specialist training: AAPC • Clinical Documentation Improvement training • Electronic Health Records coursework

WRITING & RESEARCH

Examining the foundations of clinical meaning

Writing is part of how the practice works. Dr. Khalil writes on clinical meaning, semantic integrity, health data standards, and what information loss means for clinical AI. Her developing guide, Clinical Meaning in the Age of Health Data & AI, follows clinical information through documentation, representation, exchange, and computational use. Its aim is to make the semantic foundations of clinical AI visible to clinicians, health informatics professionals, and those responsible for health data systems.

INDEPENDENCE & SCOPE

Defined Reviews. Evidence-Led Analysis

Engagements are asynchronous and conducted in writing.
Each engagement has a defined scope and is delivered remotely.
All work is carried out independently and delivered as written analysis against explicit criteria:
Scope: Each review is strictly limited to its stated purpose, source materials, and evaluation criteria.
Evidence: Findings clearly distinguish between supported conclusions, unresolved questions, and matters that cannot be established.
Transparency: Reports identify relevant limitations, assumptions, and evidentiary gaps.Professional Boundaries: The practice does not provide direct patient care, medical consultations, or comprehensive safety certifications.

THE PRINCIPLE

Making clinical meaning computable

Clinical meaning is not generated by AI. It is captured, governed, and made computable.”
The reliability of clinical AI depends on whether the information it receives still represents what the clinical encounter established. Balsam Semantica exists to test whether it does.

Written enquiries about related work are welcome.

Clinical meaning is not generated by AI. It is captured, governed, and made computable.