Ambulatory Surgery Center Model

Written By Phil Jacokes

Overview

Reviewing surgical cases for site optimization and appropriateness helps to ensure that patients are matched to the correct surgical location based on their health risk and risk associated with the planned procedure. In summer of 2020, U-M Health System implemented a process where physician assistants from the preoperative clinics performed prospective case reviews relative to institutional guidelines to determine if a patient was a successful candidate for surgery at an ambulatory surgical center (ASC). The manual review was successful in increasing the use of ASCs by reducing late surgery cancellations and identifying additional patients eligible for surgeries at ASCs vs university hospital, though the manual review process was lengthy and time consuming.

MPrOVE developed an ASC model that has been integrated into the clinical review process at U-M by organizing surgical cases into five risk groups and incorporating key data elements (BMI, Cardiovascular System, Respiratory System, Neurologic System, Anesthesia-related issues), which are used to assess complexity and clarity, along with historical decisions made by preoperative clinic physician assistants. The final process resulted in higher review process efficiency while requiring less clinician effort and reducing scheduling delays.

Detailed Results and Outcomes from Model Testing:

  • Faster surgery scheduling - Median days from case request to surgery scheduled decreased 4 days, from 7.5 down to 3.5 days

  • Improved Accuracy - >99% of model-approved cases were manually reviewed and were approved for ASC

  • No change in Case Movement from ASC to UH - only 0.7% of model assigned cases needed to be rescheduled from ASC to UH, which is consistent with the rate of manual reviews.

  • PA Time and Effort Saved - pre-model, PAs spent 178 hours over a two-month period. With the model, preoperative PAs now spend only 24 hours in ASC reviews over a two-month period since 75% of cases are now reviewed/assigned by the model.

You can read more here: Henderson J, Cuttitta A, Dossett LA. Using Machine Learning to Predict Suitability for Surgery at an Ambulatory Surgical Center. JAMA Surg. 2023;158(11):1212–1213. doi:10.1001/jamasurg.2023.1409

Additional Resources/Websites

MPrOVE Project Portfolio:

https://sites.google.com/umich.edu/ascmodel/home

ASC Review Website:
https://sites.google.com/umich.edu/mprove-project-portfolio/surgery-site-optimization-model

PI(s)

Geoffrey Barns, MD, James Henderson, PhD

Partnerships

Pre-Op Clinic

Ambulatory Surgery

Nursing

Anesthesia

AI & Digital Health Innovation Support

AI & Digital Health Innovation is supporting the MPrOVE team in evaluating ·several research questions related to the implementation of this model into clinical workflows. This project will address key usability and workflow issues with broad implications for model implementation. AI & DHI also supported the model deployment in Epic, enabling model predictors to be viewed at the time of score generation.

Phil Jacokes
Next

PICTURE: Pediatric General Floor Deterioration Prediction