Abstract
BACKGROUND: Maternal and neonatal mortality in the UK remains high, underscoring safety concerns in maternity care. Incident reporting remains a key mechanism for identifying risks and driving improvement, yet challenges, including underreporting and limited organisational learning, persist. AIM: The primary aim of the study was to explore clinicians' preferences and behaviours in maternity patient safety reporting within a tertiary hospital. METHODS: We conducted a two-phase qualitative study in a UK tertiary teaching hospital maternity service. Phase 1 involved AI-supported Big Qualitative (Big Qual) thematic analysis (using Caplena and Infranodus) of the first 400 patient safety incident reports submitted via the local electronic reporting system over a 5-month period (June-November 2024). Phase 2 comprised semistructured interviews with 14 maternity clinicians conducted between April and June 2025 and informed by phase 1 findings. Interview data were analysed using a Rapid Assessment Procedure and framework-based thematic analysis. Findings from both phases were integrated at the interpretation stage to examine reporting practices, barriers and enablers and opportunities for organisational learning, drawing on sociotechnical systems and safety-II-informed concepts. RESULTS: Thematic analysis of incident reports identified ten recurrent topics including staffing capacity, documentation discrepancies and communication issues. Interviews highlighted barriers such as psychological safety, form complexity and limited feedback, alongside enablers including visible learning and supportive leadership. Inconsistencies in reporting behaviours, feedback mechanisms and system integration were evident, with underreporting of near misses and staff conduct identified as key gaps. CONCLUSIONS: This study offers a nuanced view on how incident reporting is enacted in practice within maternity care. By combining interview data with Big Qual incident analysis, it identifies actionable insights for improving safety and organisational learning. Recommendations include simplifying reporting systems, embedding psychological safety, standardising processes and enhancing feedback and cross-professional learning.