亲爱的研友该休息了!由于当前在线用户较少,发布求助请尽量完整地填写文献信息,科研通机器人24小时在线,伴您度过漫漫科研夜!身体可是革命的本钱,早点休息,好梦!

Large language models for thematic analysis in healthcare research: A blinded mixed-methods comparison with human analysts

作者
Callum Hill,Arun Dahil,Glenn Simpson,David Hardisty,Jacob M Keast,Cameron Kumar Pinn,Hajira Dambha‐Miller
出处
期刊:PLOS digital health [Public Library of Science]
卷期号:5 (4): e0001189-e0001189 被引量:11
标识
DOI:10.1371/journal.pdig.0001189
摘要

Abstract Large language models (LLMs) are increasingly used for qualitative thematic analysis, yet evidence on their performance in analysing focus-group data, where polyvocality and context complicate coding, remains limited. Given the increasing role of such models in thematic analysis, there is a need for methodological frameworks that enable systematic, metric-based comparisons between human and model-based analyses. We conducted a blinded mixed-methods comparison of two general-purpose LLMs (ChatGPT-5 and Claude 4 Sonnet), an LLM-based qualitative coding application (QualiGPT), and blinded human analysts on an in-person focus-group transcript informing an AI-enabled digital health proposal. We evaluated deductive coding using a 10-code, 6-theme codebook against an expert consensus adjudication; inductive coding with a structured Likert-scale comparison to a reference-standard set of inductive themes generated by expert consensus; and manual quote verification of LLM segments to define LLM hallucination (evidence absent or non-supportive) and error rate (including partial matches and speaker-coded segments). During deductive coding against an expert consensus adjudication, large language models (LLMs) yielded a mean agreement of 93.5% (95% CI 92.5–94.5) with κ = 0.34 (95% CI 0.26–0.40); blinded human coders achieved 92.7% (95% CI 91.6–93.9) agreement with κ = 0.34 (95% CI 0.26–0.41). Mean Gwet’s AC1 was 0.92 (95% CI 0.90–0.93) for the blinded human analysis, and 0.93 (95% CI 0.92–0.94) for the LLM-assisted deductive analysis, reflecting high agreement despite the low overall code prevalence (7.8%, SD = 3.2%). Only one model achieved non-inferiority in inductive analysis of the transcript ( p = 0.043). The strict hallucination rate in inductive analysis was 1.2% (SD = 2.1%). LLMs were non-inferior to human analysts for deductive coding of the focus-group data, with variable performance in inductive analysis. Low hallucination but significant comprehensive error rates indicate that LLMs can augment qualitative analysis but require human verification. Author summary Qualitative research plays an important role in digital health, assisting in the implementation of healthcare technologies and innovations. However, analysing qualitative data in the form of focus groups is time-consuming and requires human expertise. Large Language Models (LLMs) are being increasingly used in qualitative research analysis, although evidence on their performance in analysing focus group data is limited. We compared the performance of LLMs to blinded human analysts in analysing a focus group transcript on AI implementations in healthcare. We used both qualitative and quantitative metrics to evaluate the performance of LLMs in thematic analysis. We found that the LLMs performed similarly to humans when applying pre-defined codes (deductive analysis), with a low rate of hallucination. However, in open-ended theme generation (inductive analysis) their performance was more variable, particularly in areas requiring interpretation of tone, nuance, or conversational context. These findings suggest that LLMs can be used to support interpretation of qualitative data, rather than replace human analysts. We provide a reproducible framework in analysing the performance of LLMs in qualitative analysis.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
朱志伟完成签到,获得积分10
刚刚
1秒前
高兴夏之完成签到,获得积分10
3秒前
CMJ发布了新的文献求助10
6秒前
pengua发布了新的文献求助10
7秒前
cheers完成签到,获得积分10
8秒前
Ava的应助被风趣含双采纳,获得10
8秒前
Jayzie完成签到 ,获得积分0
8秒前
11秒前
13秒前
田様的应助被soilman采纳,获得10
14秒前
16秒前
Yh发布了新的文献求助30
17秒前
17秒前
CMJ完成签到,获得积分10
17秒前
大个的应助被儒雅傲人采纳,获得10
18秒前
小马甲的应助被儒雅傲人采纳,获得10
18秒前
20秒前
Xenomorph发布了新的文献求助50
20秒前
小棉发布了新的文献求助10
21秒前
辛勤的涵菡完成签到,获得积分10
21秒前
Hello的应助被Brightfuture采纳,获得10
22秒前
风趣含双发布了新的文献求助10
23秒前
27秒前
30秒前
所所的应助被可乐采纳,获得10
32秒前
pengua发布了新的文献求助10
34秒前
yizimm完成签到 ,获得积分10
34秒前
37秒前
Lusteri完成签到 ,获得积分10
38秒前
板栗炒鸡完成签到,获得积分10
41秒前
42秒前
42秒前
畅快怜容完成签到,获得积分20
43秒前
敏感一一完成签到,获得积分10
44秒前
46秒前
46秒前
怡然的扬发布了新的文献求助10
49秒前
50秒前
Brightfuture发布了新的文献求助10
50秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
John McDowells Bildungstheorie 800
八旗制度の研究 600
plantTFDB 508
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8172954
求助须知:如何正确求助?哪些是违规求助? 9565479
关于积分的说明 21780097
捐赠科研通 7764853
什么是DOI,文献DOI怎么找? 3431513
关于科研通互助平台的介绍 2555530
邀请新用户注册赠送积分活动 2476595