Development and educational evaluation of iLN, a DeepSeek-based generative AI platform for teaching lupus nephritis: a randomized controlled, single-blind (outcome assessor-blinded) study

Lupus nephritis (LN) is the most common and severe organ manifestation of systemic lupus erythematosus, and its diagnosis and treatment require the integration of clinical, serological, and renal pathological information. Few studies have evaluated generative artificial intelligence (GAI)-assisted teaching dedicated to LN training. We developed an interactive LN teaching platform (iLN) based on the DeepSeek large language model, integrating current textbooks, guidelines, a renal pathology atlas, and authentic cases across seven modules. In a randomized controlled study, 50 third-year medical students were randomized in a 1:1 ratio to iLN-based multimodal teaching (iLN group) or traditional lectures (control group). Randomization was performed using a random number table after all participants had been enrolled, and the outcome assessor was blinded to group allocation. The primary outcome was an immediate postintervention 20-item multiple-choice examination (100 points). The iLN group achieved a higher median score compared to the control group (median 90 [IQR 90–95] vs. 75 [IQR 75–80]; Hodges–Lehmann median difference of 15.0 points, 95% CI 10.0–20.0; Mann–Whitney U test P < 0.001). There was no significant difference in score dispersion between the groups, as the 95% confidence interval for the ratio of standard deviations included 1.0. The nominally smaller standard deviation does not provide evidence that the platform reduced variation in learning outcomes. Additionally, iLN students reported a high satisfaction rate of 96.0%. The iLN platform was associated with higher immediate post-intervention test scores and high student satisfaction, supporting the feasibility of integrating generative AI into a multicomponent platform for LN. Larger multicenter studies with validated instruments and long-term follow-up are required. NCT07706998 (retrospectively registered on June 30, 2026).

Authors

Institutions

Publication Details

Journal
BMC Medical Education
Published
2026-11-19
DOI
https://doi.org/10.1186/s12909-026-10856-w
Primary Topic
Artificial Intelligence in Healthcare and Education
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Development and educational evaluation of iLN, a DeepSeek-based generative AI platform for teaching lupus nephritis: a randomized controlled, single-blind (outcome assessor-blinded) study

Juan Wang, Hao Bao, Shaoshan Liang, Yuanmao Tu
BMC Medical Education
Artificial Intelligence in Healthcare and Education
article

Development and educational evaluation of iLN, a DeepSeek-based generative AI platform for teaching lupus nephritis: a randomized controlled, single-blind (outcome assessor-blinded) study

Juan Wang, Hao Bao, Shaoshan Liang, Yuanmao Tu
article en

Abstract

Lupus nephritis (LN) is the most common and severe organ manifestation of systemic lupus erythematosus, and its diagnosis and treatment require the integration of clinical, serological, and renal pathological information. Few studies have evaluated generative artificial intelligence (GAI)-assisted teaching dedicated to LN training. We developed an interactive LN teaching platform (iLN) based on the DeepSeek large language model, integrating current textbooks, guidelines, a renal pathology atlas, and authentic cases across seven modules. In a randomized controlled study, 50 third-year medical students were randomized in a 1:1 ratio to iLN-based multimodal teaching (iLN group) or traditional lectures (control group). Randomization was performed using a random number table after all participants had been enrolled, and the outcome assessor was blinded to group allocation. The primary outcome was an immediate postintervention 20-item multiple-choice examination (100 points). The iLN group achieved a higher median score compared to the control group (median 90 [IQR 90–95] vs. 75 [IQR 75–80]; Hodges–Lehmann median difference of 15.0 points, 95% CI 10.0–20.0; Mann–Whitney U test P < 0.001). There was no significant difference in score dispersion between the groups, as the 95% confidence interval for the ratio of standard deviations included 1.0. The nominally smaller standard deviation does not provide evidence that the platform reduced variation in learning outcomes. Additionally, iLN students reported a high satisfaction rate of 96.0%. The iLN platform was associated with higher immediate post-intervention test scores and high student satisfaction, supporting the feasibility of integrating generative AI into a multicomponent platform for LN. Larger multicenter studies with validated instruments and long-term follow-up are required. NCT07706998 (retrospectively registered on June 30, 2026).

BMC Medical Education
Nanjing General Hospital of Nanjing Military Command (CN), Nanjing University (CN)
Artificial Intelligence in Healthcare and Education
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon