Machine Learning Model for Perioperative Transfusion Prediction

Development and Interpretation of a Machine Learning Model for Perioperative Transfusion Prediction

This study aimed to develop and interpret a machine learning model to predict red blood cell (RBC) transfusion.

Study Overview

• Status: Completed
• Conditions: Surgery; Blood Transfusion
• Intervention / Treatment: Other: Perioperative blood transfusion

Detailed Description

A dataset from a multicenter study involving 6121 patients underwent elective major surgery was analysed. Data concerning patients who received inappropriate RBC transfusion were excluded. Twenty one perioperative features were used to predict RBC transfusion. The data set was randomly split into train and validation sets (70-30). Decision tree, random forest, k-nearest neighbors, logistic regression, and eXtreme garadient boosting (XGBoost) methods were used for prediction. The area under the curves (AUC) of the receiver operating characteristics curves for the machine learning models used for RBC transfusion prediction were compared.

• Study Type: Observational
• Enrollment (Actual): 6121

Contacts and Locations

Study Locations

• Turkey
  • Ankara, Turkey, 06110
    • Dilek D Unal

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years to 100 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

Patients undergoing major elective surgical procedures were considered eligible. All age groups and both genders were included. The surgical procedures were predefined according to the Classification of Diagnostic, Therapeutic, and Surgical Procedures of the National Social Insurance Institution, in which every procedure has a unique cod

Description

Inclusion Criteria:

• Adult
• Underwent major elective surgery

Exclusion Criteria:

• Pediatric patients
• Emergency cases

Study Plan

How is the study designed?

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Number of patients received Red blood cell transfusion	Number of patients received Red blood cell transfusion	Perioperative period
The area under the curve	The the area under the curve of the receiver operating characteristics curves	Perioperative period

Collaborators and Investigators

• Sponsor: Diskapi Teaching and Research Hospital
• Collaborators: Hacettepe University; Gazi University; Baskent University; Marmara University; Yuzuncu Yıl University; Istanbul University; Saglik Bilimleri Universitesi; Bulent Ecevit University; Selçuk University; TC Erciyes University; Akdeniz University; Dokuz Eylul University; Inonu University; Kahramanmaras Sutcu Imam University; Mersin University; Necmettin Erbakan University; Firat University; Ondokuz Mayıs University; T.C. ORDU ÜNİVERSİTESİ; Istanbul Medeniyet University; Uludag University; Eskisehir Osmangazi University; Ankara University; Trakya University; Celal Bayar University; Hitit University; Suleyman Demirel University; Karadeniz Technical University; Ufuk University; Balikesir University; Erzincan University; Osmaniye Government Hospital; Namik Kemal University; Tobb University of Economics and Technology
• Investigators: Principal Investigator: Dilek D Unal, Prof, UNIVERSITY OF HEALTH SCIENCES TURKEY DISKAPI YILDIRIM BEYAZIT TRAINING RESEARCH HOSPITAL ANKARA

Publications and helpful links

General Publications

• Murphy GJ, Reeves BC, Rogers CA, Rizvi SI, Culliford L, Angelini GD. Increased mortality, postoperative morbidity, and cost after red blood cell transfusion in patients having cardiac surgery. Circulation. 2007 Nov 27;116(22):2544-52. doi: 10.1161/CIRCULATIONAHA.107.698977. Epub 2007 Nov 12.
• Bernard AC, Davenport DL, Chang PK, Vaughan TB, Zwischenberger JB. Intraoperative transfusion of 1 U to 2 U packed red blood cells is associated with increased 30-day mortality, surgical-site infection, pneumonia, and sepsis in general surgery patients. J Am Coll Surg. 2009 May;208(5):931-7, 937.e1-2; discussion 938-9. doi: 10.1016/j.jamcollsurg.2008.11.019. Epub 2009 Mar 26.
• Walczak S, Velanovich V. Prediction of perioperative transfusions using an artificial neural network. PLoS One. 2020 Feb 24;15(2):e0229450. doi: 10.1371/journal.pone.0229450. eCollection 2020.

Study record dates

Study Major Dates

• Study Start (Actual): January 13, 2022
• Primary Completion (Actual): February 1, 2022
• Study Completion (Actual): February 1, 2022

Study Registration Dates

• First Submitted: January 13, 2022
• First Submitted That Met QC Criteria: January 27, 2022
• First Posted (Actual): February 8, 2022

Study Record Updates

• Last Update Posted (Actual): March 8, 2022
• Last Update Submitted That Met QC Criteria: March 7, 2022
• Last Verified: January 1, 2022

More Information

Terms related to this study

• Keywords: Surgery; Red blood cell transfusion; Prediction model; Machine learning; Decision tree; Logistic regression; eXtreme garadient boosting; k-nearest neighbors; Random forest
• Other Study ID Numbers: Machine learning DiskapiTRH

Plan for Individual participant data (IPD)

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

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No