Identification of a Responsive Subpopulation to Hydroxychloroquine in COVID-19 Patients Using Machine Learning

Purpose

The purpose of this study was to assess the performance of a machine learning algorithm which identifies patients for whom hydroxychloroquine treatment is associated with predicted survival.

Conditions

  • COVID-19
  • Coronavirus
  • Mortality

Eligibility

Eligible Ages
All ages
Eligible Genders
All
Accepts Healthy Volunteers
No

Inclusion Criteria

  • Patient admitted to covered ward and tested positive for COVID-19 - Patient had COViage applied to electronic health record data within four hours of COVID-19 test

Exclusion Criteria

  • Patient not admitted to covered ward or tested negative for COVID-19 - Patient had COViage applied to electronic health record data greater than four hours after COVID-19 test

Study Design

Phase
N/A
Study Type
Interventional
Allocation
Non-Randomized
Intervention Model
Parallel Assignment
Primary Purpose
Diagnostic
Masking
None (Open Label)

Arm Groups

ArmDescriptionAssigned Intervention
Experimental
Exposed group 		All patients were exposed to the algorithm and were characterized as being likely responders to hydroxychloroquine treatment. Treatment decisions regarding the administration of hydroxychloroquine were made independently by care providers.
  • Device: COViage
    Machine learning intervention

Recruiting Locations

More Details

NCT ID
NCT04423991
Status
Completed
Sponsor
Dascena

Detailed Description

In a multi-center pragmatic clinical trial, COVID-19 positive patients admitted to 6 United States medical centers were enrolled between March 10 and June 4, 2020. A machine learning algorithm was used to determine which patients were suitable for treatment with hydroxychloroquine.