EHR Technical Discovery

A technical discovery in collaboration with CHLA Innovation Studio and Armenian Eye Care Project, CHLA's partner NGO. Our research focused on improving the Armenian healthcare system through digital products, specifically electronic health record (EHR) systems.

Role: Product Manager
Platform: Web

About CHLA Innovation Studio

Children's Hospital of Los Angeles (CHLA) hosts their own Innovation Studio, a small team of visionaries in healthcare that work with communities across the world, tackling problems with novel care models, digital products, and digital health equity. They work on several programs, including Open Pediatric Educational Network (O.P.E.N.), where this project falls under.
O.P.E.N. was first conceived with Armenian Eyecare Project (AECP) in 2009 with the intention of building educational tools and digital products to improve the screenings and treatment for retinopathy of prematurity (ROP). The program has since expanded to work towards leveraging technology to improve healthcare in low and middle-income countries.

The Project

In April, I started working with CHLA O.P.E.N. to assess different technical options for building out an EHR system for the Armenian pediatric care. I led a (virtual) kickoff workshop to lay out our problem statement and brainstorm a vision, and I soon realized that our task at hand was more open ended beyond EHR's. The CHLA team wanted to consider other digital products as well--perhaps a mobile app for parents to communicate with pediatricians, or a way to gather health data through wearables. We also had limited information on the state of healthcare in Armenia: we knew that only some hospitals and clinics have access to computers and internet, but nearly everyone still relies on paper health records. Our stakeholders included the CHLA team (in the US); AECP administrators (in Armenia); and pediatricians, parents, and patients (in Armenia).
Our statement of work held the Sidebench team responsible for a kickoff workshop, technical research and documentation for an EHR system, and a final deliverable with systems diagrams and recommendations. But since the possibilities for our product were truly endless, I crafted a project plan that added user interviews and market research in order to guide our technical requirements. I was able to adjust our team's hours allocation and complete our work within our agreed upon timeline and budget.
I created hourlong user interview scripts for Armenian pediatricians and NGO partners to uncover current workflows and pain points. I also validated that access to internet and computers was scarce, and that structure of health data is the highest priority, so that practitioners will be able to search for data, trends, and records. I also conducted preliminary market research to better understand the digital landscape of Armenian healthcare: there are various platforms for patients to find and discover physicians, there is a national EHR system that has been in progress for 4+ years, and NGOs run different projects that require different platforms and software.

After further discussion with the CHLA team, we also learned about a previous project they were concluding with another software vendor called Avetis, the beginnings of an EHR system for AECP. I recommended we build on top of Avetis for our current engagement, and I put together a high-level product requirements document, supplemented with our user interviews and market research for a backlog of features. By mid-May, we were finally ready to dig into our technical assessment of OpenMRS, an open source EHR platform.

The Solution

I relayed my loose PRD to my technical architect on the team, and we explored two possibilities for development: utilizing an existing open source platform (OpenMRS), or building it all custom from scratch. We measured both options on a scorecard with 21 different requirements, especially focusing on:
1. Customizability: Will we be able to control the look, feel, and user experience of the platform?
2. Functionality: Will we be able to build all the features we wanted? Will be able to build iteratively and maintain agile development? 
3. Scalability: Will we be able to process large data queries securely and reliably?
4. Distribution and adaptability: Will we be able to extend this product to other geographies and use cases (beyond Armenia and AECP)?
My technical architect put together two systems diagrams to lay out how an EHR might be developed on each option, and I cleaned up both diagrams in Figma for a final deliverable.

Under OpenMRS, we would utilize existing modules and create custom functionality and UI. We validated that querying large data sets would not significantly decrease performance, and that this option would provide easy distribution and extensibility.

With a fully custom architecture, we would work through a decoupled Node application with a React frontend, and we would be able to customize our data model to fit just what we needed. However, a custom app would not allow for easy distribution.

We concluded that both options were viable and that both would be able to support all the features we were looking for in an EHR. The main concerns were around ease of development and customization.
Our final recommendation depended on the real goals of CHLA and AECP: if they wanted to focus on AECP's single project use case, then we recommended a custom solution for streamlined development. We wouldn't be constrained with outdated modules are overly robust data models. In contrast, if the goal was to venture into global extensibility, then we recommended building on OpenMRS so that we could take advantage of their open source community.
You can view the final deliverables deck and annotated systems diagrams here.

Future

With our technical research and recommendations, we proved that we are a technical team that also puts our users first, as shown with our added user interviews and market research. After handing off our deliverables late May, we expanded our engagement with CHLA O.P.E.N. to design and build the MVP for an Armenian health record system.
I kicked off our second engagement with CHLA in mid-July, and am continuing to manage our discovery as we gather and refine product requirements in preparation for design sprints.