We have been developing mobile health technology to address the accessibility, disparity, and equality issues in vision care. In this paper, we demonstrate how to use our mobile app to measure inter-pupillary distance, spherical equivalent refraction, and stigmatism. Uncorrected refractive error is a major cause of vision impairment in the world.
In underserved areas, rural areas, low-income areas, many people do not have access to routine eye care. With our app, we want to empower people who are not eye care professionals to help address some of the basic vision care needs. Our vision screening app can be used once installed, in general, smartphone.
There is no need to purchase any special phone attachments. This feature makes the screening tool easily accessible, even from many rural areas. To begin, launch the app and tap the IPD button on the homepage.
Place the phone approximately 40 centimeters away from the patient at eye level and ask the patient to look at the flashlight on the phone screen. Tap the round button on the right to take a picture of the patient's face. The app will then start processing the image automatically.
When two green crosshairs appear on the screen, check if they are aligned with the center of the patient's eyes. If the crosshairs are correctly positioned, tap the SAVE button. Launch the app and tap the REFRACTION button on the homepage.
Place the phone at least two meters away from the patient. Select E for tumbling E stimuli. Then select the eye to be measured and ask the patient to cover the other eye.
Now, hold the phone with the screen facing the patient and tap the START button. Ask the patient if they can identify the orientation of the letters displayed on the phone screen. If the patient cannot identify the letter orientations, gradually move the phone towards them while asking if they can see the letter orientations.
Then tap the VERIFY button to display the correct letter orientations on the screen. Compare the correct letter orientations with the patient's responses. If all three answers match, tap the CORRECT button to conclude the test.
Launch the app and tap the REFRACTION button on the homepage. Start by placing the phone at least two meters from the patient and select the Astigmatism1 stimulus. Select the eye to be measured and ask the patient to cover the other eye.
Hold the phone with the screen facing the patient and tap the START button. Ask the patient if the line group appears as three separate lines in any direction. Stop as soon as they can see the separated lines in at least one direction.
Now, select the Astigmatism2 stimulus. Rotate the phone around the axis perpendicular to the screen until the grading aligns with the clearest line group of Astigmatism1. Fine-tune the phone's rotation to find the best angle where the patient can see the red and green grading patches equally clearly.
Once found, tap the POINT 1 button to record the first far point. Note that POINT 1 is recorded, the astigmatism two stimulus is rotated 90 degrees automatically. From the first far point position, maintain the phone orientation and gradually move it closer to the patient until they can see the red and green grading patches equally clearly and identify the wider end.
Then tap the POINT 2 button to record the second far point. Based on the two far points, the app will calculate the spherical refractive error and astigmatism and display the results on the screen. On the result message box, tap the SAVE button to save the results.
The app measurement of spherical equivalent was highly consistent with clinical auto-refractor measurements with a regression line slope of 1.018 and an R squared value of 0.965. The Bland-Altman plot for the spherical equivalent shows 95%limits of agreement. The app measurement of cylinder power was consistent with standard subjective refraction with a mean error of 0.08 diopter and a standard deviation of 0.41 diopter.
This Bland-Altman plot for astigmatism measurement shows 95%limits of agreement. The error of inter-pupillary distance measurements with the app had a mean error of minus 0.55 millimeters and a standard deviation of 1.4 millimeters.
