cliexa's Explainable Clinical AI and Measured Fairness Gains Featured in Mayo Clinic Platform Case Study

Published evaluation reports a 76% reduction in the racial performance gap for high-risk opioid use disorder detection and documents a reasoning-first architecture's progression toward enterprise deployment.

DENVER, Sept. 15, 2026 /PRNewswire/ -- cliexa, the clinical reasoning layer for healthcare, today announced the publication of a Mayo Clinic Platform case study documenting the development, evaluation and commercial preparation of cliexaAI Opioid Use Disorder Solution. The case study details how cliexa combined evidence-grounded clinical rules with predictive AI, measured performance across patient groups, and progressed through clinical validation to Mayo Clinic Platform qualification and a pathway to enterprise deployment.

The publication comes as healthcare organizations place greater emphasis on the evidence supporting AI adoption. In its 2026 physician survey findings, the American Medical Association reported that 88% of physicians identified robust safety and efficacy validation as critical to broader AI adoption, with clinical evidence and practical implementation guidance among the most requested resources.

"Generating an answer is not the same as making a clinical decision defensible," said Mehmet Kazgan, founder and CEO of cliexa. "Our focus is the reasoning between the record and the decision: the clinical evidence, the patient context and the explanation a clinician can inspect. This case study documents the development and evaluation behind that approach, not just what the technology is intended to do."

Measured performance, including results across patient groups
The case study reports 82.3% overall test accuracy and a macro-F1 score of 79.5% for the OUD clinical decision support model on a held-out test set of 1,571 patients. The model stratifies risk into four tiers—Minimal, Low, Medium and High—using clinician-derived tier boundaries.

The reported bias-mitigation results include:

• 76% reduction in the racial performance gap for high-risk detection.
• 59% reduction in the age-based accuracy gap.
• Improvement in high-risk detection among Black patients from 58.3% to 75.0%, while maintaining the gender-based performance gap below three percentage points.

These findings describe performance in the reported OUD evaluation. The formal Mayo Clinic Platform qualification documented in the case study applies specifically to cliexaAI – Opioid Use Disorder Solution.

A clinical reasoning layer, not another system to replace
The architecture described in the case study pairs a deterministic clinical rules engine grounded in published medical evidence with predictive models trained on real clinical data. This dual-layer approach was developed to make risk intelligence explainable for point-of-care use, rather than presenting a prediction without its clinical basis.

Across cliexa's broader platform, the reasoning approach supports Admissions, Documentation Integrity, Care Coordination and Revenue, connecting patient context with the requirements surrounding entry into care, clinical documentation, transitions and claim submission. Enterprise Intelligence provides a cross-functional interface for questions spanning clinical, operational and financial data.

cliexa is designed to work with existing electronic medical records and healthcare applications. Its API-first architecture also allows technology partners to embed capabilities such as documentation scoring, medical-necessity validation, denial-risk assessment and appeals within their own products. The case study describes preparation for enterprise deployment through EMR-agnostic and on-premises-capable implementation patterns.

A documented development pathway, with work beyond OUD
The case study traces cliexa's participation from selection as one of four companies in the inaugural Mayo Clinic Platform Accelerate cohort in 2022, through subsequent access to de-identified data and additional modeling work, to qualification and commercial preparation. It also describes a pathway for distribution into Mayo Clinic Care Network health systems.

The publication also describes how the engagement strengthened cliexa's development practices, making fairness analysis, explainability, demographic auditing and pre-registered evaluation criteria standard expectations across model development.

The full case study, "From records to reason^®: How cliexa advanced explainable clinical AI with Mayo Clinic Platform," is available on Mayo Clinic Platform's website.

About cliexa
Founded in 2016, cliexa develops clinical intelligence that connects medical records, clinical workflows and revenue cycle systems. Its cliexaAI platform combines a proprietary clinical rules engine with predictive models to support clinical decision-making, patient engagement, billing intelligence and AI-assisted documentation. cliexa's approach is designed to strengthen the systems and workflows healthcare organizations already use.

SOURCE cliexa, Inc.