Graduate Students Propose AI-based Solutions to Support Social Service Providers

 

A spring semester capstone project challenged graduate students to understand the daily workflows of social service providers — including social workers, community health workers, and peer support specialists —  and to propose an AI-based solution to address some of the challenges associated with providing care.

The capstone project was a joint collaboration between the IC² Institute and the M.A. in Design Focused on Health, a graduate program developed by the School of Design and Creative Technologies (College of Fine Arts) and Dell Medical School.

The IC² Institute is heavily engaged in qualitative and quantitative research that will advance the mindful integration of AI into health care. As AI expands into many corners of the healthcare space, the Institute is working to ensure that underserved populations and a diverse healthcare workforce are able to reap the benefits of these technological advances. IC² designed the capstone project to meet two primary goals:

• To better understand the priorities and pain points of social service providers.
• To identify opportunities for artificial intelligence (AI) to enhance providers’ workflows within integrated care models.

 

The Research

The participating students — Isabel Alexander, Laura Long, Tanya Sasnouskaya, and Karl Sheeran — kicked off their research with several quick interviews at a local gathering of community health workers, chatted with subject matter experts about the current landscape of health-related AI, and spent a day on a “ride along” — shadowing an Austin-Travis County Community Health Paramedic team. But the centerpiece of their research was in-depth interviews conducted with 18 service providers — these included community health workers, social workers and peer support specialists.

Mostly meeting over Zoom, and sometimes sandwiching their interview between patients, the care providers fielded a range of questions from the students —  some questions centered on workflow; others probed the providers’ perceptions of the potential benefits and consequences of AI as an “assistant” in their daily work. Providers answered questions like these:

• “What is the most difficult part of your day?”
• “How does collaboration work — are you messaging through a chart or having face-to-face interactions?”
• “In your own words, how would you define artificial intelligence?”
• “On a scale of 1-5, how much do you trust AI to help you with patient care?”

When asked how the students developed their interview questions, Isabel explained, “IC² left it very open-ended, on purpose.” But the students quickly discerned that there were three core components of provider workflow: an element of patient care, an element of administrative work, an element of collaboration with other members of the care team. The students developed their questions around these three prongs.

 

Synthesis

Several themes emerged from the interviews — one was technology overload. Laura explained, “People in the medical field already have a lot of devices and plug-ins. They don’t want more technology that wasn’t really designed with them in mind.”

Another theme was that the level of trust in AI seemed to grow as the interviewees came to understand it better. Laura said, “When we would start our interviews with “Do you use AI?” they would say, “No, No!” But by the end of the interview, they were actively ideating all these different applications for how it could be used. Once people were given the space to think about it, they started to understand the potential.”

The students were careful to let the interviewee’s take the lead in talking through their understanding of, and feelings about, AI. Isabel said, “We didn’t want this to feel like we were somehow educating them as designers. We let them be the experts in the room, so to speak, and educate us.”

 

Design Intervention

For their final capstone deliverable, the students designed an AI-based solution aimed at three distinct user personas (fictional characters, based on real interviews), with each persona representing the needs of a particular subgroup:

1. Samantha, a community health worker at a small nonprofit clinic, feeling overwhelmed while entering and synthesizing notes after patient appointments.
2. Mattie, a social worker at a large teaching hospital, spending hours trying to sift through large amounts of data to understand her patients’ complex histories.
3. Greg, a peer support specialist at a private psychiatric hospital, mentoring patients in crisis, but struggling to get clinicians to engage with him in the treatment process.

 

To address Samantha’s workload, the students proposed applying natural language processing (NLP) to spoken notes in order to streamline the documentation process. To give Mattie a more efficient way to visualize and comprehend patient data, they proposed an AI-generated synthesis of post-encounter notes, translated into a dynamic dashboard that provides new insights to inform patient care. To help Greg advocate for his patient, the students suggested an expanded dashboard that would allow for more transparent communication  of things like family dynamics or access to transportation or living conditions, which could help inform treatment.

 

What’s Next?

The final capstone product was conceptual: “We were prototyping — we were pondering the feel, the emotional experience of a concept without limiting ourselves to existing technical capabilities” Isabel explained. “Hopefully, down the line, IC² can test and implement our prototype.” In addition to translating the team’s design concepts into a more fully-developed prototype, IC² plans to wrap the students’ findings into a paper for publication — with the students as co-authors.

 

…And then there’s the bit about getting jobs.

All four students graduated in May. Isabel is seeking a design strategy role within a hospital or social impact consultancy, where she can “figure out user needs and design solutions to meet those needs.” Tanya, who spent the 23/24 year juggling academic responsibilities with NCAA Division 1 tennis matches, plans to join the women’s pro tour. Laura is deciding between design research or working at the intersection of architecture and health care. Karl plans to work at a hospital, doctor’s office or clinic — somewhere he can make “positive changes in how patients are impacted by the healthcare system.”

 

Group Project Secret Sauce

All four students impressed the project sponsors with their organizational chops, empathetic approach, and quick learning. IC² Senior Research Scientist Matt Kammer-Kerwick commented, “They approached the project with curiosity, creativity, and a willingness to learn from their sponsors and from the professionals they interviewed.  Throughout a demanding project, they appeared to have fun conducting the research and ideating solutions.”

The students cite their intentional, front-end planning as a key to team success. Laura explained, “We took the time to get together and establish team norms…We also met with our client (IC²) and established what our success metrics might look like and how we should communicate with them. I think that early work paid off.”

 

Hope for the future

Having concluded their research and walked the graduation stage, do the students really think AI can play a positive role in health care?  Karl sees potential in the diagnostic realm: “AI can gather all of the information and make it easier for clinicians to digest – so they can focus more on the person. Less time summarizing; more time analyzing.”  Isabel said, “I’m feeling a lot of hope. AI can alleviate potentially menial tasks and shift the dynamic so that providers have more bandwidth, emotionally and cognitively, to care for their patients.”  Laura added, “If implemented mindfully, I wonder if AI can help these roles almost become more human. It’s kind of an interesting juxtaposition – using technology to let the human element really shine through.”

 

Notes

Capstone project coordinators were Bonnie Reese (Master of Arts in Design Focused on Health) and S. Craig Watkins, Matt Kammer-Kerwick, and Gregory Pogue (IC² Institute).