CALYPSO Pilot Study: Machine Learning Based Predictions of Surgical Complications (CALYPSO)

Pilot Study for a Clinical Analytical Platform for Surgical Outcomes

This is a pilot study of surgical patients whose postoperative care will be augmented with a personalized risk prediction platform (CALYPSO app).

Study Overview

• Status: Completed
• Conditions: Postoperative Infection; Complication of Surgical Procedure
• Intervention / Treatment: Other: CALYPSO; Behavioral: Standard Care

Detailed Description

CALYPSO is a software platform that uses machine learning techniques to create personalized risk predictions for postoperative complications. Individual risk factors are then linked to best-practice interventions tailored to the individual patient. The investigator aims to test the hypothesis that using CALYPSO-augmented postoperative care will reduce complication rates in patients undergoing elective general surgery. Subjects undergoing surgery will be assigned to the CALYPSO intervention in the postoperative period.

• Study Type: Observational
• Enrollment (Actual): 200

Contacts and Locations

Study Locations

• United States
  • North Carolina
    • Durham, North Carolina, United States, 27710
      • Duke University Medical Center

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years and older (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

Patients undergoing high risk general and oncologic abdominal surgery.

Description

Inclusion Criteria:

• >18 years of age
• Undergoing elective surgery
• Undergoing high risk colon, rectal, pancreas, gastric surgery with predicted length of stay greater or equal to 3 days

Exclusion Criteria:

• Surgeries with less than 3 days of anticipated length of stay (e.g. stoma takedown)

Study Plan

How is the study designed?

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
CALYPSO-augmented; Patients' postoperative care will be augmented with the CALYPSO platform.	Other: CALYPSO; Software platform of machine-learning based predictive models of surgical complication.; Behavioral: Standard Care; Previous standard of care based on surgeon judgement alone.
Historical Control; Patients' postoperative care was performed using standard practice, before adopting the CALYPSO platform	Behavioral: Standard Care; Previous standard of care based on surgeon judgement alone.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Time Frame
Thirty Day Overall Complications	30 days after surgery

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Thirty Day Mortality		30 days after surgery
Wound Complications	Includes superficial, deep, organ space infections, and wound dehiscence	30 days after surgery
Cardiac Complications	includes myocardial infarction, cardiac arrest	30 days after surgery
Respiratory Complications	includes unplanned reintubation, pneumonia, failure to wean from ventilator >48 hours after surgery	30 days after surgery
Thrombotic complications	includes deep vein thrombosis, pulmonary emobolism	30 days after surgery
Renal Complications	includes acute renal failure, renal failure requiring dialysis	30 days after surgery
urinary tract infections		30 days after surgery
Neurologic complications	includes stroke, peripheral nerve injury, coma	30 days after surgery
Septic complications	includes sepsis and septic shock	30 days after surgery
Unplanned reoperation		30 days after surgery
unplanned readmission		30 days after surgery
bleeding complications	needing at least 1 unit of pRBC	up to 72 hours after surgery

Collaborators and Investigators

• Sponsor: Duke University
• Investigators: Study Director: Zhifei Sun, MD, Duke University Department of Surgery

Study record dates

Study Major Dates

• Study Start (Actual): June 1, 2016
• Primary Completion (Actual): January 1, 2017
• Study Completion (Actual): February 1, 2017

Study Registration Dates

• First Submitted: July 7, 2016
• First Submitted That Met QC Criteria: July 8, 2016
• First Posted (Estimate): July 11, 2016

Study Record Updates

• Last Update Posted (Actual): July 3, 2018
• Last Update Submitted That Met QC Criteria: July 2, 2018
• Last Verified: January 1, 2018

More Information

Terms related to this study

• Other Study ID Numbers: Pro00072563

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO