Optimizing medical interpreters assignments and deployments in the hospital


for MIT Hacking Health (2018)




2 days


Angela Wang  | Chris Kwan | Jonathan Chen | Julie Berez | Ryan Graue | Serena Li | Suzanne Choi   

My Role

UX/UI Design


Best Use of Athena Health API


The Challenge

While in high demand, in-person interpreters are poorly utilized due to inefficient dispatching and slow response times. Most paging systems today rely on an admin to deploy interpreters using pagers, landlines, and printed schedules.


The Outcome

An on-demand solution that streamlines access to medical interpreters by optimizing interpreter assignments. Medilingo's system can more efficiently utilize on-staff interpreters and loop in video interpreters when demand exceeds supply.

>> Hackathon Final Presentation Deck <<


An On-demand System

Similar to Uber and many on-demand digital services, Medilingo automatically adds interpretation assignments to on-duty interpreters based on service request sent by medical staff and patients.

Toggle below to compare proposed vs. existing dispatch flows from the perspective of patients.


A Service Built for Multiple Users

The Medilingo system is designed for not only the interpreters but also the patients, the doctors, the supervisors, and the patient’s families.


1. Patient – "more participation & power"

Current Pain Points:

  • understand little about the system

  • long wait time

  • no control over quality of service

Medilingo's Solutions:

  • natural language processing & multi-lingual options

  • information about interpreters & make choices

  • transparent request process

  • experience rating system (English or native language)

2. Medical Staff – "faster & more reliable"

Current Pain Points:

  • Factor in wait time to appointment

  • No feedback channel to interpreters

Medilingo's Solutions:

  • conversation content option for time estimate

  • Feedback portal for intepretation quality

3. On-site Interpreter – "seamless flow"

Current Pain Points:

  • idle and slow time between assignment

  • have little context of the situation prior to the assignment

  • no crisis management or reporting system for hostile situation

Medilingo's Solutions:

  • real-time location and availability tracking

  • auto-queue to enable deployment at all time

  • feedback system for doctors and patients

4. Remote Interpreter – "increase value"

Current Pain Points:

  • remote interpreters are cheaper but less responsive and generally considered a less humanistic experience

  • full transition to remote interpreters is not preferable from both the business and service quality standpoint

Medilingo's Solutions:

  • include remote interpreter as a secondary option of the same system

  • leverage video chat and text-to-speech technology to augment the experience

5. Supervisor – "empower management"

Current Pain Points:

  • significant energy spent on manual scheduling that can be easily automated

  • role is limited to logistics management, leaving little capacity for service quality control

Medilingo's Solutions:

  • automate basic scheduling and let supervisors arrange more complex situations

  • create a feedback system as a basis for service quality control and management


A Profitable Financial Model

In additional to customer satisfaction, Medilingo brings significant reduction – a 15% saving in operational cost – for the hospital along with increased efficiency for on-site interpreters.

cost analysis.png

Redistribute Control through Streamlining

When designing the Medilingo, we want to make sure that we are truly simplifying the process by reducing uncertainty for patients and offload decision-making power from supervisors. With the support of technology, the improved system clarifies individuals’ roles at specific times so everyone is on the same page.

Toggle below to compare proposed vs. existing service blueprints from the perspective of patients.


The Design Process

  1. In the first day of the hackathon, our team lead Serena pitched the problem space of medical interpretation. Shortly after the team was formed.

  2. Given the short time frame, we did a very brief ideation session and quickly went into high-fidelity prototyping.

  3. We also took advantage of the medical expert resources in the hackathon to speed-date our ideas through online forums and quickly pivoted our project the night before our presentation.

  4. In addition to the service and interaction design, we also did market research and cost analysis to narrow down our customer segment and create a financial model.



One of the biggest shortcoming for this project is the fact that most of us came in with a solution in mind — an on-demand service via mobile app. It fit the making-oriented purpose of hackathon well, but the proposed design can very likely fail to address pain points we didn’t get to identify.

Some of our teammates have since built on this product and is hoping to bring this product to accelerators. Myself and Suzanne (the other designer) have remained close contact with the team to provide UX design support.