There is a policy drive in NHS maternity services to improve open disclosure with harmed families and limited information on how better practice can be achieved. To identify critical factors for improving open disclosure from the perspectives of families, doctors, midwives and services and to produce actionable evidence for service improvement. A three-phased, qualitative study using realist methodology. Phase 1: two literature reviews: scoping review of post-2013 NHS policy and realist synthesis of initial programme theories for improvement; an interview study with national stakeholders in NHS maternity safety and families. Phase 2: in-depth ethnographic case studies within three NHS maternity services in England. Phase 3: interpretive forums with study participants. A patient and public involvement strategy underpinned all study phases. National recruitment (study phases 1 and 3); three English maternity services (study phase 2). We completed n = 142 interviews, including 27 with families; 93 hours of ethnographic observations, including 52 service and family meetings over 9 months; and interpretive forums with approximately 69 people, including 11 families. The policy review identified a shift from viewing injured families as passive recipients to active contributors of post-incident learning, but a lack of actionable guidance for improving family involvement. The realist synthesis found weak evidence of the effectiveness of open disclosure interventions in the international maternity literature, but some improvements with organisation-wide interventions. Recent evidence was predominantly from the United Kingdom. The research identified and explored five key mechanisms for open disclosure: meaningful acknowledgement of harm; involvement of those affected in reviews/investigations; support for families' own sense-making; psychological safely of skilled clinicians (doctors and midwives); and knowing that improvements to care have happened. The need for each family to make sense of the incident in their own terms is noted. The selective initiatives of some clinicians to be more open with some families is identified. The challenges of an adversarial medicolegal landscape and limited support for meeting incentivised targets is evidenced. Research was conducted after the pandemic, with exceptional pressure on services. Case-study ethnography was of three higher performing services: generalisation from case-study findings is limited. No observations of Health Safety Investigation Branch investigations were possible without researcher access. Family recruitment did not reflect population diversity with limited representation of non-white families, families with disabilities and other socially marginalised groups and disadvantaged groups. We identify the need for service-wide systems to ensure that injured families are positioned at the centre of post-incident events, ensure appropriate training and post-incident care of clinicians, and foster ongoing engagement with families beyond the individual efforts made by some clinicians for some families. The need for legislative revisions to promote openness with families across NHS organisations, and wider changes in organisational family engagement practices, is indicated. Examination of how far the study's findings apply to different English maternity services, and a wider rethinking of how family diversity can be encouraged in maternity services research. This study is registered as PROSPERO CRD42020164061. The study has been assessed following RAMESES realist guidelines. This award was funded by the National Institute for Health and Care Research (NIHR) Health and Social Care Delivery Research Programme (NIHR award ref: 17/99/85) and is published in full in Health and Social Care Delivery Research; Vol. 12, No. 22. See the NIHR Funding and Awards website for further award information.