Fall 2026 (Volume 36, Number 3)

AI Scribes Draft Great Notes, But Rheumatologists Also Need Help Entering Data: ROXI Is A New Tool That Does Both

By Tom Appleton, MD, PhD, FRCPC, Chair, ORA Informatics Committee

Download PDF

For decades, the rheumatology field has championed the importance of tracking basic clinical data in routine practice. After all, this is the data that we rely on to form our assessments and plans in everyday care. Embracing a treat-to-target approach depends on measurement of many things from swollen and tender joint counts to patient-reported measures, skin scores, and calculated disease activity indices, tracked visit over visit. International and Canadian data have made this point for years: Without structured clinical data, we cannot reliably know whether we are achieving the outcomes we aim for as individual specialists, let alone as an entire field.

Like most areas of health care, rheumatology carries a heavy documentation burden from note generation. And, since measurement and tracking are just as important to providing the best-possible care in rheumatology, capturing clinical data adds to that documentation burden. The recent wave of AI scribes has made it easier than ever to draft a clinical note and move on to the next patient, and that help was badly needed. But it hasn't touched the data side: AI scribes have not reduced the documentation burden of tracking and recording the measures that treat-to-target care depends on. Of course, a well-written note is truly important. It tells the story, and stories matter, especially in rheumatology. But a beautifully-written note is still just prose: the joint counts and outcome measures buried inside it are not captured as data unless someone manually re-enters them afterward. What if we had a tool that could reduce the documentation burden of both clinical data tracking and note generation, at once?

This is precisely the gap that the ORA’s new ROXI (Rheumatology Optimized Expert Intelligence) AI suite is built to close. Deployed within RheumView™ (the ORA’s digital rheumatology chart that plugs into conventional EMRs), ROXI is more than a simple AI note-generator. ROXI reviews the transcript and notes from the clinical encounter and extracts information directly into more than 600 rheumatology-specific structured data fields for more than 95 adult and pediatric rheumatology diagnoses, and delivers a high-quality, customizable, AI-generated note. Encompassing the same joint counts, patient-reported measures, and disease activity scores that treat-to-target care requires, every suggested entry is reviewed and confirmed by the treating physician before it becomes part of the record. Because ROXI was purpose-built for rheumatology, that structured data capture is native to the workflow, not an extra step layered on top.

That combination is what sets ROXI apart from the general AI-scribe market: outstanding clinical communication and a robust structured-data environment, generated from the same encounter, drastically reducing clicks. Each ROXI-empowered rheumatologist can finally examine remission rates, treatment patterns, and gaps in care; and, aggregated, de-identified data can give the ORA and the rheumatology community a real picture of how each rheumatologic disease is managed across Ontario.

None of this is meaningful unless it earns clinicians' and patients' trust. The College of Physicians and Surgeons of Ontario (CPSO), the Canadian Medical Protective Association (CMPA), and Ontario's Information and Privacy Commissioner have converged on similar expectations for AI in clinical care: accurate, clinician-reviewed output; physician accountability; data protected and used only for its intended purpose; evaluation for bias; and transparency and meaningful consent for patients. RheumView™ and ROXI were built around these expectations. Patients are prompted for explicit, documented consent before any encounter is processed. Data is encrypted in transit and at rest, handled under zero-data-retention agreements with AI model providers, and never used to train AI models. The platform has undergone a formal privacy impact assessment and threat-risk assessment, with ongoing security review built into its governance.

AI in rheumatology should be judged by more than how well it writes a note. To truly reduce administrative burden, AI tools must help us measure, track, and act on disease activity over time; not just document it. Used well, tools like RheumView™ and ROXI can finally give our specialty what it has long lacked: a structured, measurable record that treat-to-target care was always meant to run on.

Tom Appleton, MD, PhD, FRCPC
Division Head, Rheumatology, Western University
and St. Joseph’s Health Care, London
Associate Professor, Department of Medicine, and
Department of Physiology and Pharmacology,
Western University, Canada
Scientist, Lawson Research Institute,
London Health Sciences Research Institute,
and Children’s Health Research Institute
Chair, ORA Informatics Committee