Lead Institution: Dartmouth

Project Leader: David Kotz

Research Progress

• Abstract
  This goal of this project is to develop techniques to help recipients of mHealth data verify its accuracy and authenticity.
• Focus of the research/Market need for this project
  We have met several researchers and health providers who describe the need for determining the accuracy and authenticity of remotely collected health information; they suspect that information could be inaccurate because of careless or malicious patients and caretakers/health workers and malfunctioning devices. There are products that provide simple validity checks on the information they collect, which is not enough in most cases since health information accuracy could vary depending on the context in which it was collected. There is no product in the market that provides a general framework that addresses the problem of collecting context and using it to verify accuracy and authenticity of health data.
• Project Aims/Goals
  Our goal is to develop a framework for collecting and sharing contextual metadata that can be used to verify provenance properties of health data, specifically, the authenticity of the data’s origin and the accuracy of the data. We call this contextual evidence, because this metadata can be used as context in helping the recipient (whether a clinician, researcher, or family member) to determine how much confidence they should place in the data.
• Key Conclusions/Significant Findings/Milestones reached/Deliverables
  By Dec 2012, we developed the design and architecture of a framework for collecting contextual evidence.

  In January 2013, we presented a paper at NetHealth ’13 about a parser that recognizes the rule language for collecting provenance information.

  By July 2013, we developed a software platform called Ocean: Open Contextual Evidence Add-On for Mobile Applications. Ocean allows programmers to extend applications developed using this framework to collect and share context relevant to the data, helping recipients verify whether certain properties are satisfied by the data they receive. We prepared and submitted a paper to MobiSys’14, describing the Ocean framework, evaluating its impact on memory and processing power on Android phones, and detailing a user study where programmers with little Android experience used Ocean to extend an application developed by an experienced Android developer.

  In July 2013, we visited clinicians, nurses, and health workers in rural villages in India, including professionals from CoreMedical Group, which offers staffing solutions for hospitals and clinics, to observe and discuss the provenance properties needed in real-world situations. This helped us better understand the actual expectations from the provenance framework.

  In March 2014, we developed a software platform called ShareBuddy, which gives users control over sharing decisions about their information and contextual evidence. ShareBuddy allows data consumers (recipients) to specify the information and contextual evidence they need and the benefits they guarantee to the user for sharing it. ShareBuddy enables users to make informed sharing decisions by presenting a sharing controls interface that outlines the benefits and risks of sharing the information with the recipient. This project will continue under new funding from the NSF.

• Materials Available for Other Investigators/interested parties
  We have developed a prototype of the Ocean provenance middleware and the sharing architecture ShareBuddy for Android phones; we hope to share these with others in a few months. We have published technical reports and peer-reviewed papers to enable others in the research community to benefit from our work.
• Market entry strategies
  No commercial efforts, but we are working with medical groups (like the Psychiatric Research Center at Dartmouth) to make our technology available to them.

Bibliography>
A Provenance Framework for mHealth
Aarathi Prasad, Ronald Peterson, Jacob Sorber, and David Kotz
Workshop for Mobile Systems, Applications, and Services for Healthcare (mHealthSys) Poster Track, ACM Press, November 2012

A Provenance Framework for mHealth
Aarathi Prasad, Ronald Peterson, Shrirang Mare, Jacob Sorber, Kolin Paul, and David Kotz
Workshop on Networked Healthcare Systems, pages 1-6. IEEE Computer Society Press, January 2013