BACKGROUND: Women with low socioeconomic status and social risk factors are at a disproportionate risk of poor birth outcomes and experiences of maternity care. Specialist models of maternity care that offer continuity are known to improve outcomes but underlying mechanisms are not well understood.

AIM: To evaluate two UK specialist models of care that provide continuity to women with social risk factors and identify specific mechanisms that reduce, or exacerbate, health inequalities.

METHODS: Realist informed interviews were undertaken throughout pregnancy and the postnatal period with 20 women with social risk factors who experienced a specialist model of care.

FINDINGS: Experiences of stigma, discrimination and paternalistic care were reported when women were not in the presence of a known midwife during care episodes. Practical and emotional support, and evidence-based information offered by a known midwife improved disclosure of social risk factors, eased perceptions of surveillance and enabled active participation. Continuity of care offered reduced women's anxiety, enabled the development of a supportive network and improved women's ability to seek timely help. Women described how specialist model midwives knew their medical and social history and how this improved safety. Care set in the community by a team of six known midwives appeared to enhance these benefits.

CONCLUSION: The identification of specific maternity care mechanisms supports current policy initiatives to scale up continuity models and will be useful in future evaluation of services for marginalised groups. However, the specialist models of care cannot overcome all inequalities without improvements in the maternity system as a whole.

Original languageEnglish
JournalWomen and Birth
Early online date26 Nov 2022
Publication statusE-pub ahead of print - 26 Nov 2022


Dive into the research topics of 'Project20: Maternity care mechanisms that improve (or exacerbate) health inequalities. A realist evaluation'. Together they form a unique fingerprint.

Cite this