Healthcare Shorts: Pneumonia Prediction Model
Published Nov 01 2022 11:40 AM 1,014 Views
Microsoft

Story Background

​​​​​​​Pneumonia is an infection that affects one or both lungs. It causes the air sacs, or alveoli, of the lungs to fill up with fluid or pus. Bacteria, viruses, or fungi may cause pneumonia. signs vary from mild to severe, depending on the type of germ, age, and overall health. Mild signs and symptoms often are similar to those of a cold or flu, but they last longer which includes chest pain, cough, fatigue, nausea, vomiting or diarrhea. A traditional approach to diagnosing pneumonia includes a review of the patient’s medical history, a physical exam, and ordering diagnostic tests such as a chest X-ray. This information can help determine the type of pneumonia.

 

Business Challenge 

​​​​​​​Providers could use a machine learning diagnostic process to validate and fast-track treatment without the need for a second opinion from another physician which takes longer than necessary. Diagnostic errors are minimized with the added layer of confirmatory evidence from AI model. If not detected early, it is contagious via air transmission or surface residual effects.

 

Business Outcomes 

Implementation of this solution will lead to a lower healthcare burden by reducing: 

  • Hospital-acquired pneumonia (HAP)
  • Community-acquired pneumonia (CAP).
  • Ventilator-associated pneumonia (VAP).

 

Solution Overview 

​​​​​​​Use chest X-ray images to train a machine learning algorithm in Azure ML workstation to detect pneumonia with high accuracy, currently 92%. Deploy solution using Power App with mobile capability. Used AP view of images to develop final model, any view can be adopted for this same model. 

 

 

Thanks for reading, Shelly Avery |EmailLinkedIn 

If you are interested in more content like this, then follow  https://aka.ms/HealthcareShorts 

 

Please follow aka.ms/HLSBlog for all this great content.

Co-Authors
Version history
Last update:
‎Nov 01 2022 11:41 AM
Updated by: