Review months of diagnostic and treatment processes against current guidelines in a fraction of the time.
Built for different specialties. Your clinicians choose the question, the guidelines and the clinical context.
The origin of GLIDE
Barbora operates on patients with head and neck cancer, and cares about what happens throughout their entire care journey. She wanted to see how every step, from diagnostics and treatment decisions to radiotherapy and surveillance, held up against the latest clinical guidelines. Doing that across an entire cohort would normally mean reviewing every chart by hand.
So she ran GLIDE on 40 patients. Within weeks, the department had acted on one of the findings and changed its clinical practice, improving the patient journey.
Clinical guidelines evolve continuously: staging criteria are revised, treatment windows narrow, and new evidence redefines the standard of care. Even highly experienced clinicians can miss a step under everyday clinical pressure, and each instance tends to differ from the last. That is where GLIDE steps in: a structured clinical audit that identifies these drifts and returns the findings to clinical practice, so they are not repeated.
Reviewing every case by hand takes more time than most departments can spare. GLIDE reviews an entire cohort in a fraction of that time, giving your department a fast, clear view of how care compares to the evidence base.
Guidelines cannot anticipate every comorbidity or clinical circumstance. Departure from a guideline may, at times, be clinically appropriate; clinical context, not the record alone, distinguishes between the two.
How it works
You select a cohort to compare against current guidelines. GLIDE works best with a clearly defined diagnosis or pathway, such as acute myocardial infarction, stroke, head and neck cancer, or breast cancer.
Anonymized records: discharge summaries, imaging, pathology and tumor board notes.
A three-layer pipeline reads each record, flags candidate deviations and clusters them into themes instead of a long, unsorted list.
Real or not, justified or not, and how severe. AI never makes the clinical call alone.
Confirmed findings become a department presentation and tailored teaching, built around what was actually found.
What clinicians review
The reviewer sees the source record, relevant guideline, clinical context and severity together.
Select a specialty to view a different finding.
More than four years after radiotherapy including the neck, no thyroid function (TSH) testing had been performed, despite imaging noting a post-radiation change in the thyroid gland and repeated reports of fatigue and weight change.
Should TSH have been included in routine surveillance after neck irradiation, given these findings?
NCCN Head and Neck Cancers, Follow-up/Surveillance: “TSH should be checked every 6–12 months in patients whose neck has been irradiated.”
Atul Gawande, MD, MPH“Failure is central to medicine. What matters is how we respond to it.”
Projects
Clinical teams are applying one audit method to clearly defined departments, diagnoses and care pathways.
Banská Bystrica
Bratislava
Prague
Ružomberok
Košice
Liptovský Mikuláš
Why GLIDE
Oncologist
First Faculty of Medicine, Charles UniversityDuring internships at Harvard Medical School, Sorbonne University Hospitals and Great Ormond Street Hospital, I worked alongside clinicians who were not only following guidelines, but producing the research and writing the standards behind them. It made me ask how more of that rigor could reach everyday practice at scale.
We already have the LLM technology to extract structured information from free-text medical records and compare it with current guidelines. So we built GLIDE.
Through the Young Doctors Association in Prague, I also work to improve postgraduate education. GLIDE findings could become teaching seminars where young doctors revisit real decisions and the evidence behind them. Often, improvement needs no new drug or device, only more consistent use of the best available evidence.
Who is behind GLIDE
GLIDE was started by analysts who wanted their work to change practice, and clinicians looking for a practical way to improve care.
Maxillofacial surgeon
F. D. Roosevelt University Hospital,
Oncologist
First Faculty of Medicine, Charles University &
AI & Data researcher
Institute for Healthcare Analyses
Health analyst
Health Economics, Policy & Innovation InstituteFAQ
Anonymized documentation for the cohort you want reviewed: discharge summaries, imaging, pathology and tumour board notes, plus some review time from your own clinicians.
Documentation is anonymized before it ever reaches the pipeline, and anything your team reviews already has dates, names and identifying detail stripped out.
No. Guidelines need to be interpreted within the patient’s clinical context. GLIDE only flags possible deviations; your own multidisciplinary panel decides whether each one is real, clinically justified and how severe it is.
A pilot audit starts with collecting the relevant clinical documentation. Some hospitals can export anonymized records directly from their systems, which makes this step very fast. Once the documentation is collected, you can expect the GLIDE analysis within two weeks. Within one month of starting, your team can have both a set of systematic recommendations for improving care and material for an internal clinical seminar based on the most relevant findings.
The pilot itself is free. For subsequent audits, pricing starts at €30 per patient, depending on cohort size and complexity. A typical GLIDE analysis represents roughly 50 hours of expert physician review, condensed into a structured, reproducible audit.
Contact
Speak directly with our clinical team. Tell us your specialty and the question your department wants to answer.
Email Soňa Hudecová