Mitigation of Postoperative Delirium in High-Risk Patients

Integration of Machine Learning and Clinical Decision Support to Prevent Postoperative Delirium in Patients With Cognitive Impairment

Among patients with cognitive impairment (CI) that undergo surgery, the risk for developing postoperative delirium (POD) is high (50%) and associated with further morbidity and mortality. Yet, 30-40% of POD cases are preventable with perioperative management. This randomized pragmatic clinical trial aims to assess incidence of POD in adult surgical patients with CI, as well as provider adherence to a set of 12 perioperative best practice recommendations for perioperative management. Electronic health record (EHR) data will be used to identify patients as high risk for developing POD and clinical decision support (CDS) prompts within the EHR will display best practices. Cases will be randomized to either the control group, usual care or the intervention which includes the high-risk alert and best practice prompts.

Study Overview

• Status: Terminated
• Conditions: Cognitive Impairment; Decision Support Systems, Clinical; Post-operative Delirium
• Intervention / Treatment: Other: Clinical Decision Support

• Study Type: Interventional
• Enrollment (Actual): 7412
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • New York
    • New York, New York, United States, 10029
      • Mount Sinai Health System

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years to 120 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Adult patients with a history of cognitive impairment (CI) undergoing surgery at the Mount Sinai Health System.
• Attending physicians, resident physicians, and nurse anesthetists at the Mount Sinai Health System receiving CDS alerts during EHR system implementation.

Exclusion Criteria:

• Patients will be excluded if this not their first surgery since study start, if they do not have cognitive impairment based on EHR data, or if the surgery is for organ donation.
• There will be no exclusion criteria for providers or historical controls.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Prevention
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Clinical Decision Support; Clinical decision support alerts in the electronic health record directed towards anesthesiologists caring for patients with preexisting cognitive impairment.	Other: Clinical Decision Support; The intervention will consist of clinical decision support alerts in the electronic health record directed towards anesthesiologists caring for patients with preexisting cognitive impairment. This intervention will alert towards delirium risk informed by history of cognitive impairment and promote 12 evidence based best practices during care for perioperative patients.
No Intervention: Standard of Care; No clinical decision support will appear, and standard of care procedures will take place.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
4AT Delirium Score	Postoperative Delirium measured using the 4 A's Test (4AT) delirium assessment documentation. The 4AT is a commonly used tool that identifies the presence/absence of Postoperative Delirium (POD). It consists of 4 sections; alertness, AMT4: Abbreviated Mental Test -4, attention, and acute change of fluctuating course. The 4AT is scored from 0-12, with higher score indicating poorer health outcomes. Scoring: 4 or above: possible delirium +/- cognitive impairment 1-3: possible cognitive impairment 0: delirium or severe cognitive impairment unlikely (but delirium still possible if [4] information incomplete)	Postoperative day 7

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Number of Participants Where Perioperative Best Practices for Intervention Was Performed	Adherence will be measured as a binary variable; overall protocol adherence will be defined as the number of participants where best practices was performed by the anesthesia team. The 12 perioperative best practices are grouped in 5 intervention domains including, avoid potential inappropriate medication, perioperative glycemic control, avoid hypotension, maintain normothermia, and titrate anesthetic depth. Three of 5 intervention domains measured by number of participants where the following conditions met: Within avoid potential inappropriate medication is avoid diphenhydramine, scopolamine, and midazolam. Within perioperative glycemic control is check pre-op glucose, check glucose every 2 hours, maintain glucose <200 mg/dL, and check post-anesthesia care unit glucose. Within maintain normothermia is use temperature probe and maintain temperature >36 degrees Celsius.	Day 1
Mean Arterial Pressure >65 mmHg	Avoiding hypotension is one of the 12 perioperative best practices are grouped in 5 intervention domains including, avoid potential inappropriate medication, perioperative glycemic control, avoid hypotension, maintain normothermia, and titrate anesthetic depth. Within avoid hypotension is Mean Arterial Pressure (MAP) >65 mmHg. MAP assess blood flow throughout the body.	Day 1
Age Adjusted Minimum Alveolar Concentration (MAC)	Age-adjusted Minimum Alveolar Concentration (MAC) represents the concentration of anesthetic gas required to prevent movement in response to surgical stimuli, adjusted for a patient's age. For every case, age-adjusted MAC values were calculated for each minute during anesthetic administration using standard formulas. The calculation was performed as follows: For every case and for every minute that anesthetic gas was administered to a person, the age-adjusted MAC was calculated using the formula. For each minute, compute the sum of all age adjusted MAC values for all gases administered Next, % MAC was calculated for every case as TOTAL # of mins where MAC>1 /# of minutes X 100 MAC>1 indicates that it is more anesthetic than is typical for a patient with given demographics is being administered.	Day 1

Collaborators and Investigators

• Sponsor: Icahn School of Medicine at Mount Sinai
• Collaborators: University of California, Los Angeles; National Institute on Aging (NIA)
• Investigators: Principal Investigator: Ira Hofer, MD, Icahn School of Medicine at Mount Sinai

Study record dates

Study Major Dates

• Study Start (Actual): June 20, 2023
• Primary Completion (Actual): August 31, 2024
• Study Completion (Actual): August 31, 2024

Study Registration Dates

• First Submitted: March 8, 2023
• First Submitted That Met QC Criteria: March 8, 2023
• First Posted (Actual): March 21, 2023

Study Record Updates

• Last Update Posted (Actual): February 19, 2026
• Last Update Submitted That Met QC Criteria: February 1, 2026
• Last Verified: February 1, 2026

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Neurologic Manifestations; Nervous System Diseases; Mental Disorders; Postoperative Complications; Pathologic Processes; Confusion; Neurobehavioral Manifestations; Neurocognitive Disorders; Cognition Disorders; Delirium; Pathological Conditions, Signs and Symptoms; Signs and Symptoms; Emergence Delirium; Cognitive Dysfunction
• Other Study ID Numbers: STUDY-22-00920; 22-00920 (Other Identifier: Advarra/CIRBI); 5U54AG063546 (U.S. NIH Grant/Contract)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: The research team will be collecting the data in a data warehouse, but it will be extracted by an honest broker de-identified. No participant level data will be shared outside Sinai or the study team.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No
• product manufactured in and exported from the U.S.: No