John T. E. Timm

Coordinating analytics methods for mobile healthcare applications

Mobile healthcare applications enjoy increasing popularity and provide significant benets to users who can take advantage of the general availability and the increasing set of sensors in mobile devices. Healthcare analytics services developed for a variety of conditions to provide insights and guidance to patients can augment the mobile applications with computationally intensive analysis at the server-side environment. Creating applications that are supported by analytics is challenging for dierent reasons: We must deal with the limited information to get relevant recommendations, or not enough to provide appropriate predictions (patients trajectory to reach goal such as weight loss). Moreover building an analytical model in the health domain brings its own challenges such as data ingestion, data curation, and service levels, etc.This paper proposes a software architecture framework that eases the support of analytics in the mobile health applications. The paper also discusses the various components that make up the framework by taking one analytical model as a use case.

Taming complex healthcare data models with dictionary tooling

Information models used in the healthcare domain tend to be complex, in part because they were designed to be as flexible and generic as possible. This complexity presents a steep learning curve for implementers, which can lead to partial or poorly-implemented solutions. In this paper, we present a tool that facilitates the creation of sets of modular and composable clinical data abstractions. Using these, implementers can produce and consume standards-compliant clinical data correctly and efficiently without being experts in the underlying information models or the medical terminologies they reference.