Pilot Study for Postoperative Machine Learning

NCT04877535 | Completed Results FDA Device

• 2 other identifiers: 202103210KL2TR002346
• Study Type: interventional
• Target: 222
• Locations: 1 country
• Sites: 1

Brief Summary

The objectives of the study are to determine the interpretability, workflow role, and effect on communications of showing report cards containing Machine Learning (ML)-based risk profiles based on pre- and intra-operative data to postoperative providers.

Conditions

Surgery--Complications

Keywords

Telemedicine; Anesthesia Control Tower; Machine Learning; Forecasting Algorithms; Handoff effectiveness

Outcome Measures

Primary Outcomes (1)

• Overall Handoff Effectiveness

  After handoff was completed, receiving nurses were asked: Globally, how effective was the handover 1. Not at all effective 2. Somewhat effective 3. Moderately effective 4. Very effective 5. Extremely effective The item is taken from PMID:25806398 but has no name

  8 hours postop

Secondary Outcomes (3)

• Number of Participants With ML Topics Discussed During Handoff

  8 hours postop

• Number of Participants With Anticipatory Guidance During Handoff

  8 hours postop

• Number of Handoff Receivers Agreeing That They Received All Needed Information

  8 hours postop

Other Outcomes (1)

• Number of Handoff Recipients Self-reporting Referring to Report Card OR Report Card Observed Directly Referred to During Handoff

  8 hours postop

Study Arms (3)

Stage 2: Intervention

EXPERIMENTAL

ML will be used to create a report card for each patient that summarizes the preoperative assessment and intraoperative data. Report card data will be made available to providers through multiple methods: integration into electronic health records workflows, electronic health records notifications, mobile device notifications, and print outs in the paper chart Patients who have a previous assignment (from another day) are not eligible.

Device: ML-based report card

Stage 2: Usual Care

NO INTERVENTION

The standard of care. The report card will be electronically generated (to determine eligibility) but it will not be visible to clinicians. Patients who have a previous assignment (from another day) are not eligible.

Stage 1: Usual Care

NO INTERVENTION

The standard of care. The report card will be electronically generated (to determine eligibility) but it will not be visible to clinicians.

Interventions

ML-based report card DEVICE

PACU and ward providers caring for participants will be notified by Anesthesia Control Tower clinicians before arrival if the patient's report card. The notification will contain a report card of the patient's forecast risk of major adverse events, explanatory machine-learning outputs, most influential pre- and intraoperative data, and predicted treatments.The ML risk profile generated for each patient will include risk of 30 day mortality, risk of respiratory failure, risk of acute kidney injury, and risk of postoperative delirium

Stage 2: Intervention

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• ODIN-Pilot will intervene on a subset of TECTONICS participants meeting all the following criteria:
• Within the TECTONICS contact arm (adults undergoing OR procedures)
• Operating room at BJH South campus (including all of "Pod 2", "Pod 3", "Pod 5") (excluding all procedure suites such as Interventional Radiology, Parkview Tower "Pod 1", Center for Advanced Medicine "Pod 4", Labor and Delivery suites)
• Surgeon is a member of the Acute and Critical Care Surgery division or the postoperative bed is 16300 observation unit.
• Planned non-ICU disposition ("floor" and "observation unit" collectively "ward" patients).

You may not qualify if:

• Not enrolled in TECTONICS Study
• Randomized to the observation arm in TECTONICS study
• Planned ICU admission
• Patients are only included once; if previously included a subsequent surgery is not eligible

Sponsors & Collaborators

• Washington University School of Medicine: lead
• National Center for Advancing Translational Sciences (NCATS): collaborator

Study Sites (1)

Barnes-Jewish Hospital

St Louis, Missouri, 63110, United States

Location

Limitations and Caveats

During the intervention, we observed that despite education and encouragement, few OR nurses included the report card with the paperwork for the PACU nurse. We determined that the study was unlikely to meet its goals with poor intervention uptake and stopped the trial. In 3 interviews with staff, we found that the workflow was difficult to remember. Because only 2-5 patient per day were eligible, and because of "control" days, staff forget what they were supposed to do with the report card.

Results Point of Contact

• Title: Christopher King
• Organization: Washington University School of Medicine

christopherking@wustl.edu

(314) 362-8649

Study Officials

• Christopher R King, MD, PhD

  Washington Univeristy School of Medicine

  PRINCIPAL INVESTIGATOR

Publication Agreements

• PI is Sponsor Employee: Yes

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: NON RANDOMIZED
• Masking: NONE
• Purpose: OTHER
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Model Details: Blended pre-post and parallel design. Participants will be allocated to up to two groups in two time periods.

Study Record Dates

• First Submitted: May 3, 2021
• First Posted: May 7, 2021
• Study Start: June 3, 2021
• Primary Completion: May 11, 2023
• Study Completion: May 11, 2023
• Last Updated: April 4, 2025
• Results First Posted: April 4, 2025

Record last verified: 2025-03

Data Sharing

• IPD Sharing: Will not share

Locations

US (1)