Abstract:
This thesis proposes a patient's health data acquisition system for community health care through mHealth. In this system, patients gather in a community clinic where an attend• ing health worker records patient condition through a mobile app. The collected data can be essentially multimodal. They may be of different formats: numbers ( e.g., temperature reading), texts (e.g., description of symptoms), images (specific sign/scar/wound on body parts), audio (recorded interviews, health conditions such as cough/ breathing sound), and video (if required). These contents are partially synced to a backend server. They can be accessed by distant doctor now or later time to assess the patient case. Due to large size of media content and bandwidth limitation of the worker device, uploading all contents to the server is not possible (nor even required); most part of them can be stored locally and later can be delivered to the doctor's device upon request. On-demand deliv• ery delays content upload to save costly upload bandwidth at the cost of slight increase in latency to retrieve content when needed. To this end, we build an mHealth service, called dursheba, that entails mobile apps (both for doctor and health worker), a server and a cloud messaging service enabling an effective synergy between patient and doctor in a rich fashion even though they are distant apart.
