Making Medical Assessments Available and Objective Using Smartphone Sensors

Abstract

Access to healthcare resources is a worldwide issue, but people do not always need access to such resources to discover a medical condition. Time and time again, people have been able to discover medical symptoms in themselves and others using their human senses—namely sight, touch, and hearing. However, observations with the senses are subjective, which can lead an untrained person to ignore their own symptoms and neglect treatment until their condition worsens. I propose that subjective health measures can be made objective with little additional burden using smartphone sensors. For my thesis, I provide three examples of how the smartphone camera can be used in place of visual inspection to automatically interpret diagnostic observations related to the eye; these projects cover medical conditions like glaucoma, pancreatic cancer, and traumatic brain injuries. My work in this space has lead me to uncover a number of challenges that impede progress in smartphone-based health-sensing. One of those challenges is ensuring that people make rational decisions when they are given health-screening tools despite not having formal training on diagnostic decision-making. I address this challenge by presenting a low-fidelity survey instrument that enables researchers to rapidly explore the effects of design decisions on the expected acceptability and effectiveness of a ubiquitous health-screening technology.