Close
Enter your
text here

Can Artificial Intelligence (AI) Improve Access to Social Resources for Better Patient Outcomes?

Can Artificial Intelligence (AI) Improve Access to Social Resources for Better Patient Outcomes? 1024 540 Pam Georgiana

A recent study tests the ability of the DAPHNE© Chatbot to do just that.

Health care providers and organizations recognize the impact of social factors on health and are increasingly addressing social determinants to improve health outcomes and equity. However, integrating social care into health practices remains challenging due to structural barriers such as staffing, time and space constraints.

Emre Sezgin, PhD, principal investigator in the Center for Biobehavioral Health at Nationwide Children’s Hospital, and his team want to address this problem by using a chatbot called DAPHNE© (Dialog-Based Assistant Platform for Healthcare and Needs Ecosystem) to screen for unmet social needs and link patients and families to community resources. The innovative chatbot they created uses AI and natural language processing to better understand families with unmet social needs, while allowing health care providers to learn more about patient’s needs out of clinic and integrate the knowledge in medical care.

“Social service resources are available for those who need them, but initial screening and/or follow-up often fall through the cracks in health care settings. DAPHNE© links families to social services such as food pantries, transportation, housing, and more,” says Dr. Sezgin, who leads the Intelligent Futures Research Lab at Nationwide Children’s. “In essence, it aims to help lowering the barriers to much-needed support.”

Recently, Dr. Sezgin and team published a paper on the iterative design and evaluation study of DAPHNE© in JMIR Human Factors.

For the study, the team used a three-stage approach. The first stage included a web-based end-user survey sent to more than 200 low-income U.S. households. The goal was to understand the common themes that families use to express their needs and which needs are routinely unmet. The survey also measured end-user comfort level with technology. Among the survey participants, employed and younger individuals reported a higher likelihood of using a chatbot.

The second stage included web-based sessions with an interdisciplinary group including researchers, providers, public health scientists and family advocates, during which they explored the content and design of DAPHNE. These stakeholders emphasized the importance of provider-technology collaboration.

“Stakeholders do not want to eliminate the patient/provider interaction. DAPHNE© should be a complementary resource, not an alternative,” Dr. Sezgin stresses.

The stakeholders were also concerned that the chatbot’s technology deployed an inclusive conversational design that addresses multifaceted family needs.

“They also didn’t want to create frustration because the chatbot can’t follow a conversation,” Dr. Sezgin says. “Large language models have gotten so much better at understanding human language, but we learned there’s still more refining to do.”

The last stage involved a mix of surveys and focus group interviews with community health workers and social workers, focusing on usability testing. The participants interacted with a simple interface that guided them through common patient scenarios. They shared that the chatbot’s capabilities met their expectations and that it was easy to use. However, there were concerns about the accuracy of the output, electronic health record integration, and trust in a chatbot.

The insights learned in this three-part study will inform more refinement and hands-on testing of DAPHNE©. Future research will examine the chatbot’s ease of use, cost-effectiveness, and scalability in addressing social needs. Dr. Sezgin is optimistic that DAPHNE© will be valuable for two important reasons.

“Not only can DAPHNE© be a supplementary tool for health professionals to ease the screening process and provide timely social needs support to families, but it can also collect data for future resource allocation and understanding trends in community needs,” Dr. Sezgin says.

Dr. Sezgin is currently working on building a scalable version of DAPHNE with his research team, which will be used for real-world implementation and testing with a large cohort. For this effort, he received intramural funding from OSU TDAI, as well as Amazon Web Services (AWS) Health Equity Initiative grant and a fellowship from FindHelp (A social care network across U.S.)

 

Reference:

Sezgin E, Kocaballi AB, Dolce M, Skeens M, Militello L, Huang Y, Stevens J, Kemper AR. Chatbot for Social Need Screening and Resource Sharing With Vulnerable Families: Iterative Design and Evaluation Study. JMIR Human Factors. 2024;11:e57114.

About the author

Pam Georgiana is a brand marketing professional and writer located in Bexley, Ohio. She believes that words bind us together as humans and that the best stories remind us of our humanity. She specialized in telling engaging stories for healthcare, B2B services, and nonprofits using classic storytelling techniques. Pam has earned an MBA in Marketing from Capital University in Columbus, Ohio.

You might also like