Emerging evidence
More
Caregivers of children with asthma rated an AI interviewer easy to use in a 28-person study
Study title: Evaluating a large language model-powered system for semi-structured interview in health research
- Evidence type
- Interviews or qualitative study
- Sample
- n = 28
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2945cite or check it
Study title: Evaluating a large language model-powered system for semi-structured interview in health research
Researchers built a web-based interview system powered by a large language model and tested it with 28 caregivers of children with asthma who live in subsidized housing. The aim was to see whether an AI interviewer could stand in for traditional semi-structured qualitative interviews with marginalized groups.
Read the study abstract
Objectives: Conducting qualitative health research with marginalized populations presents persistent challenges related to trust, disclosure, and resource constraints. Traditional in-person interviews can be costly and difficult to scale, while participants may hesitate to share sensitive experiences openly. This study assessed the feasibility and acceptability of a large language model (LLM)-powered conversational interview system as an alternative to semi-structured qualitative interviews with marginalized pop…
49Gain side
How the two sides scoreClose call
48Problem side
The gain
In a feasibility study of 28 caregivers of children with asthma living in subsidized housing, participants said an AI-powered interview system was easy to use and rated it positively for clarity, comfort and accessibility.
The problem
In the same 28-caregiver asthma interview study, researchers recorded intermittent technical errors, limited contextual adaptability and uneven comfort with digital technology among participants using the AI interviewer.
A fixed formula sets both scores from the evidence and its sources. Reader votes don’t move them.
What the scores mean
G is the gain side, out of 100. It goes up when the gain is specific, when it measurably helped people and when the sources behind it are strong.
P is the problem side, out of 100. It goes up when the problem is specific, when it measurably hurt people and when the sources behind it are strong.
In brief
Participants said the system was easy to use and liked its empathetic, nonjudgmental tone. They rated it positively for clarity, comfort and accessibility. Researchers also logged intermittent technical errors, limited ability to adapt to context and differences in how comfortable people were with digital tools.
The authors argue that hybrid human-AI approaches could widen participation in qualitative health research and support health equity. The study measured feasibility and acceptability, not the quality or depth of the data collected compared with human-led interviews.
Main points
- The system was a web-based interview tool powered by a large language model.
- Its design drew on a community-based participatory research framework.
- Twenty-eight caregivers of children with asthma living in subsidized housing tested it.
- Evaluation relied on postinterview sessions and researcher observation notes.
- Participants described the tool's tone as empathetic and nonjudgmental.
- Problems included intermittent technical errors and limited contextual adaptability.
- Comfort with digital technology varied among participants.
The full story
Qualitative research with marginalized populations faces challenges of trust, disclosure and cost. In-person interviews are expensive and hard to scale, and participants may hesitate to share sensitive experiences. This study tested an LLM-powered conversational system as an alternative in a community-engaged setting.
The team assessed feasibility and acceptability using postinterview evaluation sessions and researcher observation notes from 28 caregivers. They examined usability, accessibility, technical and implementation issues.
The authors conclude that such systems, guided by community engagement, ethical design and human support, are feasible and acceptable. They call for ongoing technical improvement, ethical safeguards and participatory design before wider use.
What this doesn’t fix
The study assessed only feasibility and acceptability with 28 caregivers, and participants met technical errors, limited contextual adaptability and varying comfort with digital technology.
Sources
1 on file- JAMIA Opendoi:10.1093/jamiaopen/ooag218
Talk it over
Caregivers of children with asthma in subsidized housing rated the AI interviewer clear and comfortable, yet researchers logged technical errors and limited contextual adaptability. How do you weigh those two sides?
Caregivers of children with asthma rated an AI interviewer easy to use in a 28-person studyhttps://truvace.com/item/evaluating-a-large-language-model-powered-system-for-semi-structured-interview-i
Caregivers of children with asthma in subsidized housing rated the AI interviewer clear and comfortable, yet researchers logged technical errors and limited contextual adaptability. How do you weigh those two sides?
Answer in the threadThe evaluation rested on postinterview sessions and researcher observation notes from 28 caregivers. What can that tell us about whether people shared as much as they would with a human interviewer?
Answer in the threadIf you were a caregiver in this study with little comfort using digital tools, what would you need from the research team to take part in an AI interview?
Answer in the thread
If you have more time
Which findings would you want from a comparison of AI-led and human-led interviews on the depth of the data collected before using this tool more widely?
Answer in the threadThe design drew on a community-based participatory research framework. How could caregivers themselves help decide when a human should step in during an interview?
Answer in the thread
ace
What readers say