Case Study: Drexel University

Drexel investigated new treatment protocols with a mobile app and integrated algorithm

When the Drexel research team first came to Promptworks, they had developed a hypothesis about eating disorder treatment. By incorporating glucose monitoring and patient reminders, the Drexel team hoped to improve outcomes among patients, build a case for their larger hypothesis, and win a grant to help move their research forward. They needed help building a tool to test this hypothesis and to utilize machine learning to help gauge eating disorder behavior as part of an expanded treatment protocol.

Here was our challenge:

The Drexel team had a clear vision for their research but were reliant on a number of integrations to build a product that could help support their hypothesis about eating disorder treatment. They also needed an interface that would incorporate current treatment protocols around food and behavior tracking.

In order to create a solution, Promptworks needed to incorporate the machine learning framework that had been provided by a University of Pennsylvania professor, integrate with a consumer glucose monitor, and transition real-world meal tracking solutions from hard copy to digital. The product needed to be easy to use by patients and to provide guidance and support without being invasive.

Here's how we solved it:

Built for Research Where most technical products come with an external user in mind, Drexel’s project required a slightly different focus when it came to measuring efficacy of the application and success of the overall project. From day one, the goal of our engagement was to facilitate research by way of clinical trials but also provide a workable application for trial participants. The Promptworks team conducted interviews with Drexel researchers and investigated traditional treatment protocols. We delivered an app that addressed current patient and clinician needs while also offering innovative integrations to test the overall research hypothesis.

Third Party API Integrations Because Drexel’s research hypothesis was directly built around the inclusion of glucose monitoring in eating disorder treatment, it was required that an API from a consumer facing glucose monitor be integrated into the application. After experimenting with this API, the team realized it did not deliver data in a timely manner and they needed to implement a backup solution. Instead, the team needed to connect to the glucose monitor directly via bluetooth in order to analyze the data in real-time.

Promptworks also needed to incorporate a variety of technical skills to ensure that the final application was stable and usable by trial participants. The glucose algorithm provided to Promptworks was written in C. In order to communicate with the algorithm, Promptworks created a command-line interface using Go. The command-line program accepted JSON as input and returned JSON as output and allowed Promptworks to incorporate the algorithm into the Sense Support API, written in Ruby. Clinicians and patients used a React application to communicate with the Sense Support API. In order to communicate with the glucose monitor via bluetooth, some parts of this application had to be written in Swift or Kotlin.

By remaining nimble and technically creative, Promptworks was able to help Drexel and our third party partners navigate any issues that came up during integration and deliver a clinical-trial ready application.

See the results:

In spite of challenges with API integration and third party access, Promptworks was able to deliver a product that provided Drexel with a valuable clinical trial tool. The final application was launched privately and is currently being used as part of a clinical trial protocol to test the long term viability of Drexel’s eating disorder therapy hypothesis. Without the delivered application, it would be impossible for Drexel to test its theories and work toward future research opportunities.

Interested in more examples of our work?

See how we helped IQVIA deliver targeted recruitment data with intelligent automation or how Aero2Max put laboratory physiology into athletes’ hands with a machine learning based mobile app.

Ready to get started?

Contact us to discuss your next project.


We'd love to keep in touch!

Opt in for occasional updates from us. Privacy Policy