The Usability and Trustworthiness of Medical Eye Images

The majority of blindness is preventable, and islocated in developing countries. While mHealth applications forretinal imaging in combination with affordable smartphone lensadaptors are a step towards better eye care access, the expertknowledge and additional hardware needed are often unavailablein developing countries. Eye screening apps without lens adaptorsexist, but we do not know much about the experience of guidingusers to take medical eye images. Additionally, when an AIbased diagnosis is provided, trust plays an important role inensuring in the adoption. This work addresses factors thatimpact the usability and trustworthiness dimensions of mHealthapplications. We present the design, development and evaluationofEyeGuide, a mobile app that assists users in taking medicaleye images using only their smartphone camera. In a study(n=28) we observed that users of an interactive tutorial capturedimages faster compared to audible tone based guidance. In asecond study (n=40) we found out that providing disease-specificbackground information was the most effective factor to increasetrustworthiness in the AI based diagnosis. Application areasofEyeGuideare AI based disease detection and telemedicineexaminations.