Charts that read themselves

For ambulatory surgery centers, the most expensive moment of the day isn't the case that runs long — it's the case that never runs at all. Industry data from HST Pathways' State of the Industry Report shows the average ASC carried a 21% case cancellation rate in 2024, up from 20% in 2023.¹ Even a small share of those that hit the day of surgery carries an enormous price tag. Peer-reviewed work in JAMA Surgery pegs the cost of an ambulatory OR minute at $36.14², which means a single 90-minute case that doesn't go puts a hole of more than $3,200 in the daily schedule before you count cascading effects on staffing, anesthesia, and downstream cases.³

What separates the centers running at single-digit same-day cancellation rates from those running at 5%–10% isn't luck. It's pre-operative screening done early, completely, and against the actual chart — not the scheduling fax.

In most ASCs, the chart for an upcoming case lives in pieces across many systems. The H&P is in the surgeon's EMR. The cardiology clearance is in the cardiologist's EMR. The medication list is in the patient's primary care portal. The imaging report is in PACS. The pre-op nurse's job is to chase all of it down — often by phone, fax, and patient self-report — and translate it into a single risk picture before the patient ever shows up.

Without that infrastructure, the most common reason for day-of cancellation is what an anesthesiologist or surgeon discovers in pre-op holding that should have been flagged days earlier — a new medication, an unmanaged comorbidity, a missed clearance, an NPO violation rooted in unclear instructions.

Lily's Pre-Op Screening agent pulls every chart available — from the connected EMR and from the HIE — for each case on the schedule. Through Health Information Exchange networks like Carequality and CommonWell, that pull can reach hospitals, primary care offices, specialty clinics, and labs the patient has touched in the last several years, not just the records inside the ASC's four walls.⁴

Once the chart is assembled, Lily reads it. She reconciles medications, surfaces anticoagulants and GLP-1 agonists that require specific stop dates, checks the most recent A1C and BMI against the facility's protocols, flags missing cardiac or pulmonary clearances, and compares everything against the surgeon's and anesthesiologist's risk thresholds for that procedure. When something doesn't line up — a clearance that expired, an EF reading below threshold, an unaddressed sleep apnea diagnosis — the case is flagged before it stays on the schedule.

The literature is clear that better screening drives down cancellations. A Mayo Clinic study of more than 41,000 ambulatory surgical procedures reported a same-day cancellation rate of just 0.5%, made possible by a mature pre-op clinic and structured assessment.⁵

Chart review is a pattern-recognition task with a high cost of missing things. It rewards completeness, consistency, and the ability to read across many documents in many formats. That is exactly the work that modern language models, paired with structured EMR data, do well — and exactly the work that pulls a pre-op RN away from the patient-facing conversations that move the needle on outcomes and experience.

The point isn't to remove clinical judgment from screening. It's to put a fully assembled, fully annotated chart in front of the nurse or anesthesiologist before they sit down. When the human reviewer's first question stops being "where's the H&P?" and starts being "should we move forward?", the screening process gets faster, safer, and far less dependent on whether the right person happened to make the right phone call on the right day.