Glaucoma is a disease that affects 60M people worldwide. It accounts for $2.5B in health care spend and 10M physician visits each year in the US alone.
Unfortunately, 1 in 8 patients who are undergoing treatment still go blind. Glaucoma has no cure and the only treatment option is to minimize its progression.
Early detection and accurate monitoring of progression is critical to ensure vision loss is minimized. One of the most widely used parts of a patient's vision assessment
is a visual field test that detects the patient's sensitivity to light at 54 distinct points. We obtained access to one of the biggest datasets available in this field, with
more than 177k unique patients across five US based sites and a total of over 800K visual fields. There have been several attempts to use machine learning algorithms
to try to classify how glaucoma is progressing, but this is the first time that it has been implemented on such a big data set.