Background High workload among general practitioners (GPs) poses a growing challenge to clinician well-being and care quality. Ambient scribes offer a potential solution, but evidence on their effectiveness is limited and largely overlooks the patient perspective. Methods A prospective multicentre multi-perspective before-after longitudinal mixed-methods study with 12 GPs and GPs in training without prior experience using ambient scribes from the Netherlands. The intervention was an ambient scribe, which generates summaries of consultations using a microphone and large language model. The primary objective was to evaluate the effect on clinical documentation time. Secondary outcomes included total consultation time, documentation quantity and quality, patient and GP experiences, tool acceptability, and usage rate. Outcomes were observed across a two-day baseline and two day intervention period. Quantitative data were analysed using (generalized) linear mixed models, and interviews were thematically analysed. This trial was registered at ClinicalTrials.gov, NCT06691724. Findings Between December 9, 2024, and July 2, 2025, 12 GPs and 535 patients (264 baseline, 271 intervention) were included in the study. Clinical documentation time decreased by 42.7 seconds (95% CI [-56.29; -30.78], p<0.0001). No difference was found in total consultation time (-61.4 seconds, 95% CI: -131.91; 0.96; p=0.069). GPs reported reduced workload, and some patients reported improved patient-provider communication. However, potential drawbacks were inaccurate summaries, barriers for discussing sensitive topics, and possible interference with the clinician’s reasoning process. Interpretation Ambient scribing decreases workload and may improve communication. However, potential unintended consequences need further investigation to ensure that care quality and accessibility are not affected.
Uncovering Code Insights: Leveraging GitHub Artifacts for Deeper Code Understanding
arXiv:2511.03549v1 Announce Type: cross Abstract: Understanding the purpose of source code is a critical task in software maintenance, onboarding, and modernization. While large language models



