Petra Zimmermann

Background: Antibiotics administered during pregnancy and delivery can disrupt the developing infant intestinal microbiome, but their effects on infant sleep remain poorly understood. Given the critical role of sleep in early neurodevelopment, alterations in sleep during this sensitive developmental period may have lasting consequences for later cognitive and behavioral outcomes. Methods: In a prospective birth cohort study, associations between indirect antibiotic exposure (maternal antibiotic use during the third trimester of pregnancy or during delivery) and parent-reported sleep at 2 and/or 6 months were examined in 192 infants (46% females) using linear mixed-effects models. Results: Indirect antibiotic exposure during delivery was associated with a higher daytime-to-nighttime sleep ratio, reflecting relatively greater daytime sleep duration compared with nighttime sleep at both ages. Indirect antibiotic exposure during delivery was also associated with fewer nighttime awakenings, while no significant associations were observed for sleep latency. Effect sizes were small to moderate, suggesting subtle alterations in sleep behavior rather than overt sleep disturbance. Conclusion: These findings suggest that indirect perinatal antibiotic exposure may be associated with altered sleep behavior in infants and warrant further investigation into potential underlying mechanisms, including alterations of the developing microbiome. Impact: Indirect antibiotic exposure during delivery is associated with subtle differences in sleep behavior at 2 and 6 months. This study extends existing literature on perinatal antibiotic exposure by examining early sleep outcomes, an understudied domain despite the importance of sleep for neurodevelopment and infant health. Although the observed differences were small, these findings suggest that perinatal antibiotic exposure may influence early sleep regulation and highlight the need for further research into underlying mechanisms, including alterations in the infant microbiome.

Introduction Continuing medical education (CME) is essential for more than 100 million healthcare professionals worldwide, supporting enhancement of knowledge, skills, and patient outcomes. With a growing shift towards online and blended CME formats, expert-informed AI-based training (EIAT) is a novel e-training approach enabling rapid course creation and a timely content updates. However, its effectiveness remains underexplored. Methods A within-subject pre-post study was conducted with 17 medical students and professionals . Participants first completed a self-paced EIAT module, followed by live expert-led training (LET). Knowledge acquisition was measured via pre- and post-exams for each modality. Satisfaction, engagement, learning outcomes, and content quality were assessed using a 20-item CME-relevant feedback tool aligned with European Accreditation Council for CME standards. Paired t-test and McNemar’s tests were used to compare outcomes between the two training formats. Results EIAT completion was 76%, compared to 82% for LET. Pre-test scores were lower in LET (63 ± 11.4) than EIAT (73.8 ± 11.2), though not significantly different (p = 0.14). Mean knowledge gain was 17.3 ± 13.3 for EIAT (p = 0.0005) and 29.2 ± 13.2 for LET (p = 0.003), with no significant difference between modalities (mean difference 14.2 ± 21.5, p = 0.17) or in post-test scores (93.5 ± 5.5 vs. 93.3 ± 9.3, p = 0.79). Both formats achieved high satisfaction, with mean recommendation scores of 9.7 (SD = 0.67) for EIAT and 10 (SD = 0.0) for LET, respectively, and no significant differences across most evaluation domains (p > 0.05). Engagement was strong in both groups, though 71.4% of LET participants reported being ‘enthusiastically involved’ compared to 60% in EIAT, and 85.8% of LET participants gaining substantial practical knowledge versus 50% for EIAT (p = 0.08). Both formats were rated highly in terms of fairness, content quality, and absence of commercial bias. Conclusion: This pilot study provides preliminiary evidence that expert-informed AI is an effective alternative or complement to traditional live CME format. These findings offer further research into EIAT’s role in scalable, high-quality CME for healthcare professionals, laying the groundwork for future research and broader implementation.