Estimated no-show rate from the electronic record strongly correlates with endoscopy no-show.

BACKGROUND

High demand for endoscopic procedures contributes to prolonged wait times and limited access to care, particularly in safety-net health systems. Missed appointments, including no-shows and late reschedules, further strain resources and delay diagnosis and treatment. Epic Systems provides a proprietary Risk of Patient No-Show Model, previously validated in primary care settings, but its performance in predicting attendance for endoscopy appointments has not been evaluated.

METHODS

A retrospective cohort study was conducted among adults scheduled for outpatient endoscopy at a single safety-net hospital between January 1, 2023, and August 1, 2024. Patients aged ≥ 18 years with at least one prior encounter in the electronic medical record (EMR) were included; inpatient procedures and early cancellations (> 2 days before the appointment) were excluded. The primary outcome was appointment completion versus no-show. Epic's predicted no-show risk, demographic characteristics, procedure type, and patient portal activation status were extracted. The relationship between predicted and observed no-show rates was assessed using the coefficient of determination (R²) and linear regression. Secondary analyses stratified results by patient portal activation.

RESULTS

Among 4,658 unique patients, the median age was 59 years, 52.2% were male, and the cohort was racially and linguistically diverse. Overall, 1,493 patients (32.1%) did not attend their scheduled endoscopy. Epic's predicted no-show risk demonstrated a strong linear correlation with actual no-show rates (R² = 0.87). Observed missed appointment rates followed the equation: Missed Appointment Rate = 1.30 × (Epic Risk) + 0.16, indicating a baseline no-show rate of 16%. Each percentage-point increase in Epic's predicted risk corresponded to a 1.3-point increase in observed no-show rate. Patients without an activated patient portal (MyChart) had approximately 5-percentage-point higher no-show rates across the risk spectrum.

CONCLUSIONS

Epic's Risk of Patient No-Show model shows strong correlation with real-world endoscopy attendance and may support predictive overbooking and targeted outreach to improve endoscopy unit efficiency. Given its integration within the EMR, this tool offers a practical framework for operational interventions, though further validation across diverse health systems is warranted.