Medidata: Scaling an App by Adding a Configuration Service

eConsent-Sales-Image-iPad-Mockup-2019.png

The iOS app eConsent has a complex set of site-specific configurations we needed to accommodate.

Overview

Taking a white-label app through M&A activities to scale with a new configuration service.

The Challenge

Startup company Mytrus hired me to work on their virtual trials platform as the UX lead and UI engineer on the technical team implementing custom client engagements. Shortly after joining the team, Mytrus was acquired by Medidata Solutions to strengthen their end-to-end trial design, conduct, analysis, and FDA submission process by adding eConsent. 

The challenge was to manage clients moving from unlimited customizations through the inflection point of limited tools to a robust and mature consent design service. We needed to move from custom client app instances to a single configurable instance of eConsent that could scale.

Team

I led the design of the new configuration service as UX Software Engineer, working closely with one UX/UI designer, plus a team of six engineers and two testers.

Audience

The professional services users we were designing for were well-versed in study-site setup and which parameters were required for different clients. We recognized that as soon as clients learned about the tool, they would want access to it, too.

Constraints

The original roadmap indicated six months to research, design and build, but the project evolved into a nine-month project as the complexity of integrating with Medidata’s platform revealed itself.

Process

We started by evaluating all the data points currently used in the app, an eye-opening introduction to the vast differences between studies and their wildly varying requirements. I proposed a Site Wizard to set up document templates and workflows to be used across groups of sites, then edited at the site level to customize for the country, locales, and Institutional Review Board (IRB) requirements, dictating the assets needed. After creating a product brief writing up my vision, I collaborated with the UI/UX designer on a series of task flows which we synthesized into a master user flow.

Solutions

Clinical Trial Study and Site Setup

Our product owner highlighted that this was a minimum viable product (MVP) to be used by professional services—it did not have to be perfect or configure all studies to launch. We reduced scope to launch with one static workflow serving the most significant number of studies possible, knowing we would add more flexible site workflows later. With an approved scope and feature set, we moved to divide and conquer on wireframes, with each of us taking a subset of features and flows to design.

Document Mapping

Creating the concept for the document mapping tool was an exciting challenge. During the first month of the acquisition, we experienced downtime due to the transition. One of my innovation projects in the downtime was designing a better way to capture coordinates for interactive elements in study document PDFs. The eConsent app captures patient and clinician user inputs as elements of electronic informed consent: responses to questions, initials for acknowledgments, subject name, and digital signatures. After the patient views study documents and gives consent digitally, the app knits together captured data to generate a final PDF showing subject and clinician inputs from the app.

At Mytrus, engineers integrated documents by capturing coordinates and manually entering them in the site’s configuration, using valuable engineering time on a repetitive manual task. My blue-sky idea involved using a data matrix in the document creation process and letting the tool pull coordinates automagically, with a side benefit of enabling paper loading of consents in GDPR countries in the future. My short-term solution allowed professional services to use the site import feature after setting up one instance of a form. Study forms tend to follow the same content for the core of the document with slight variations depending on the language of the text, for example, the length of the payments paragraph. When the user launches the document mapping tool on an imported form, they can quickly review coordinates and tweak those which vary from the mean

ios.png

A Five Star App

We launched eConsent in the Apple App Store.

Retrospective

As one of the most challenging projects I have completed, the payoff in accomplishment made it well worth it. We reduced site setup time by 1200% by eliminating manual setup and duplicate work and decreasing the timeline for eConsent study and site setup was reduced from months to days.

My knowledge of designing and building at enterprise scale improved dramatically from launching the configuration service for eConsent. I learned that Most Viable Product (MVP) is too vague and means wildly different things depending on your audience. In retrospect, I would refine the process further by building and launching in smaller increments. Start with Most Testable Product (MTP), allowing development work to begin with the underlying service layer first. Move to Most Usable Product (MUP) to get the product out there and being used. Iterate on the feedback to level up to the Most Loved Product (MLP).

Reacting to changing projects with an agile, dynamic, problem-solving approach can open unexpected opportunities. I started as a UX designer on the project, and then after six months, promoted to Technical Lead for UI Implementation. Halfway through the project, I stepped up to take additional tasks when a Product Manager went on leave, which helped keep team velocity steady. Based on the team’s experience working with me on the configuration tool, I was offered a permanent role and promotion to Senior Product Owner.

Previous
Previous

lululemon iOS App Checkout Redesign

Next
Next

Dumpster Fire to Dream Team