The Problem
The pulmonary function test (PFT) is the EKG (electrocardiogram) analogue for pulmonary physicians. PFTs represent the mainstay of categorizing pulmonary physiology pattern and aid with diagnosis. Currently, the PFTs are performed in pulmonary clinics and are interpreted by pulmonary physicians only. However, when the pulmonary patient goes home, or if it is a different service that requests a PFT for their own patient, there is a delay in being able to view the official interpretation.
Our Solution
The project is to develop an ML model that will be able to recognize the graph of the PFT and produce interpretative statements of high-level quality when compared to the gold-standard of experienced pulmonary physicians.
Anticipated Impact
With ML, the interpretation of the PFT can be automated (just as current EKGs). This will decrease pulmonary physician burden and allow patients and consulting providers to see the results faster, improve diagnostic accuracy, and has the potential to change the current paradigm.
