My Role:

I was primarily the UX researcher, though the roles overlapped as the process progressed. I worked with the UX designer in iterating interfaces and helped the main engineer in developing the system. 

The Problem

At the beginning of the Covid-19 outbreak, hospitals and healthcare workers were overburdened by the influx of patients, as well as having to explore telemedicine options. There needed to be a way for clinicians to provide patient care while minimizing Covid-19 exposure, and without compromising both the experience of the patient and the doctor’s ability to examine the patient. 

Robots to the Rescue?

At the time, the hospital staff my team and I talked to were using an iPad, which nurses would deliver to the patients, to facilitate some patient-doctor interactions within the Emergency Department (ED). This current method still left nurses and other health workers exposed to Covid-19, and is not the most efficient solution. Robots could be used to partially automate the process of connecting doctors to their patients in the ED. However, robots are expensive and are not super user friendly to those unfamiliar with robots; telemedicine robots would currently not be a feasible option for patients at under-resourced facilities, who are also disproportionately affected by covid.

The goal of this project was to build and test a robot to provide a low cost and user friendly solution to infectious disease related problems in Emergency Departments. (My team and I also wanted to write and publish a paper to spur action in the robotics community to create more low-cost telemedicine robots using our open source code and their own equipment. Currently, our paper is still under review.)

Working on a “D.I.Y.” Solution

The solution that we came up with was to design a low cost, DIY robot that can be used for telemedicine in the Emergency Department. We envisioned that this robot can be used to protect healthcare workers and other stakeholders from being exposed to Covid-19 while still interacting with someone in the ED.

Our Research Study

After designing, we conducted a research study with one set of stakeholders, the ED physicians. The goal of this study was to figure out how our robot would fit into the dynamic ED environment, and what improvements can be made to the robot’s design. 

A difficult problem we faced was how to successfully conduct a study virtually. Because we were trying to conduct this study during the “stay at home” order, we were not able to test the robot at a real ED. We had to do our best to simulate an ED in one of my teammate’s apartments by designating the kitchen area to be the nurses’ station and having her bedroom be the patient room. The teammate that lived in the apartment pretended to be a patient, while the rest of us participated virtually.

I put together a user testing plan for the study which tried to mimic the user flow of the robot in the ED. I also created a pre-study guide for our participants to get them familiar with our robot, how to operate it, and the layout of the apartment. I realized that we couldn’t assume that all of our research participants would be comfortable using Google Meets or Zoom, so the pre-study guide also included instructions to get the participants to the same baseline regarding video communication and to improve their user experience as a participant in our study. When designing the guide, I also had to think about how much information to provide, as to not have an effect on the data we wanted to gather about the participant’s experience using the robot. 

After we collected all the participant data, I analyzed the study results by coding and sorting the qualitative data. Our qualitative results were pretty positive and we used the SUS scale to get a quantitative measure. After I grouped data created by transcripts from the interviews, I was able to compile a list of design guidelines. These guidelines outlined ways that our robot and interface needed to improve, as well as things the robotics community should keep in mind if they were to design for the ED environment.

The Results

A lot of our feedback had to do with providing more cues on the robot to make it more visible in a chaotic ED setting and making the robot more durable to unexpected ED and patient interactions. The piece of feedback that I thought was really interesting, and is something that the Healthcare Robotics lab will have to tackle in the future, was hearing that older patients might be uncomfortable talking to their doctors through the robot. Robots may not be something they are used to seeing, and it could be overwhelming to have to interact with one when you are unwell. One physician emphasized that patients really prefer having a person physically with them, rather than using any technology to mediate the interaction. A future study will have to be done to examine how we can make patients feel more comfortable with telemedicine through robotics, and how our design can more closely mimic the in-person interactions doctors have with their patients.