Abstract
Consumer health devices such as smartwatches, fitness bands, and smart scales collect physiological data through device-specific or platform-specific application programming interfaces (APIs). These APIs differ across vendors in their authentication mechanisms, data formats, and identifier conventions. This creates difficulty when integrating multiple device sources into a single health monitoring application without coupling device-specific code to shared application logic. This challenge is of practical relevance in clinical contexts such as remote monitoring of patients prescribed GLP-1 receptor agonists, where tracking body weight and physical function across devices from different manufacturers can be desirable.
Prior work on health data integration has addressed this problem but has often focused on backend or platform-level solutions. Mobile applications that interact directly with heterogeneous device APIs at the client layer, have received comparatively less attention in terms of formal architectural evaluation using software quality metrics. The contribution of this thesis is a two-layer architecture for integrating heterogeneous consumer health devices into a native iOS health monitoring application. The design uses device-specific API communication isolated in an Adapter Layer and shared clinical logic is defined in a Domain Layer implemented as an external package. The architecture was applied to two device integrations, Apple HealthKit and the Withings Health API. They were evaluated through coupling and cohesion analyses as well as a file delta analysis measuring the cost of adding a new device integration. This research examines whether the proposed architecture achieves localized extension, such that adding a new device source requires changes only within the Adapter Layer rather than across the existing codebase.