# Application of self-organised learning environments integrated with generative AI in standardised training for residents

## TL;DR

This study found that combining self-organized learning with generative AI improves medical residents' training outcomes, digital skills, and satisfaction.

## Contribution

The study introduces GAiSOLEs, a novel teaching model integrating self-organized learning with generative AI for resident training.

## Key findings

- GAiSOLEs significantly improved teaching effectiveness across all measured dimensions compared to traditional methods.
- Residents in the GAiSOLEs group showed higher digital literacy and teaching satisfaction scores.
- Regression analysis confirmed GAiSOLEs as a strong predictor of improved training outcomes.

## Abstract

This study aimed to evaluate the effectiveness of the GAiSOLEs teaching model, which integrates Self-Organised Learning Environments (SOLEs) with Generative Artificial Intelligence (GAI), in improving teaching effectiveness, digital literacy and teaching satisfaction in the standardized training of residents.

A single-center, double-blind randomized controlled trial was conducted in this study. A total of 114 standardized training participants for rotating residents were enrolled and randomly divided into two groups: the intervention group (GAiSOLEs teaching model, n = 57) and the control group (traditional teaching model, n = 57). The primary outcome was the teaching effectiveness score before and after intervention. Secondary outcomes included digital literacy score and teaching satisfaction score, both evaluated by standardized scales. Data analysis was performed using χ2 test. Paired and independent samples t-tests, Mann–Whitney U test, and hierarchical multiple linear regression.

Post-intervention teaching effectiveness scores were significantly elevated relative to baseline levels in both the intervention and control groups (all p < 0.05). For all dimensions of teaching effectiveness, the gain scores in the intervention group were markedly higher than those in the control group, with large effect sizes observed across all metrics (theoretical knowledge test: p < 0.001, Hedges’ g = 1.0; clinical skill operation: p < 0.001, Hedges’ g = 1.87; standardized medical record writing: p < 0.001, Hedges g = 1.16; clinical thinking ability: p < 0.001, Hedges’ g = 1.21; doctor-patient communication skills: p < 0.001, Hedges’ g = 1.27). Hierarchical linear regression analysis, after adjusting for covariates including gender, educational background, professional distribution and pre-test scores, revealed that the GAiSOLE intervention (versus the control condition) was a significant positive predictor of higher scores for all five training outcomes (all p < 0.05): theoretical examination (β = 2.329, ΔR2 = 0.294), clinical skill operation (β = 2.947, ΔR2 = 0.499), standardized medical record writing (β = 3.366, ΔR2 = 0.331), clinical thinking ability (β = 3.391, ΔR2 = 0.412), and doctor-patient communication competence (β = 3.157, ΔR2 = 0.355). Moreover, the experimental group achieved significantly higher scores than the control group across all dimensions of digital literacy (all p < 0.05). Additionally, the group also reported significantly higher satisfaction ratings for all teaching parameters relative to the control group (all p < 0.05).

The GAiSOLE teaching model was associated with improved teaching effectiveness, digital literacy and teaching satisfaction among residents, and thus holds promising potential as a novel approach for the standardized training of residents.