Defining, Recognising and Escalating Maternal Early Deterioration (DREaMED):Decreasing inequality through improved outcomes

RESEARCH QUESTION Can an electronically-embedded, data-enhanced Modified Obstetric Early Warning Score (eMOEWS) with behaviourally-informed escalation pathways prevent maternal near-miss and severe morbidity events and reduce outcome inequalities? BACKGROUND Severe avoidable morbidity affects 8700 UK women giving birth annually, disproportionately affecting particular ethnic and deprived groups. Current definitions represent near-end-stage morbidity and are not suitable for quality assurance or prediction tool development. Encoding only the current vital sign set, clinically used MOEWS are expert opinion-based and perform poorly. An eMOEWS, incorporating additional risk factors such as ethnicity and blood tests, combined with effective implementation guidance is needed. AIM To develop, validate and test an eMOEWS and protocolised escalation pathways. OBJECTIVES To: define near-miss and severe morbidity criteria derivable from routine data develop a unique representative eight-maternity unit data cohort define existing MOEWS performance develop and validate optimised MOEWS (vital-signs-only) and eMOEWS scores design and implement an eMOEWS interface develop treatment escalation pathways implement and evaluate the eMOEWS and escalation pathways Assess the health economic consequences Report a refined programme theory METHODS Workstream 1 We will develop near-miss and pre-near miss/significant morbidity definitions derivable from routine data by literature review informing an international Delphi process. Identification performance of these definitions will be assessed using data from MBRRACE-UK and three NHS hospitals. Workstream 2 With the NIHR Health Informatics Collaborative we will develop an eight-maternity unit dataset. We will assess current MOEWS ability to identify new near miss and significant morbidity events. Workstream 3 We will develop an eMOEWS candidate variable list by literature review informing a Delphi process. Using augmented WS2 datasets we will develop and validate an optimal MOEWS and eMOEWS, exploring standard statistical models and machine learning (ML) approaches. Workstream 4 We will build working eMOEWS web interfaces following user-centred design principles. We will develop escalation and response pathways for women identified as at risk by the eMOEWS, informed by published response pathways, analysis of interviews with women who experienced near-miss morbidity and focus groups. We will optimise these pathways using simulated scenarios in managing Medical and Obstetric Emergencies and Trauma (mMOET) centres. Workstream 5 We will implement the eMOEWS and escalation pathways into 4 hospitals, with four additional hospitals as control sites in a controlled before-and-after study including process and health economic evaluations. The primary outcome will be the combined near-miss, significant morbidity, or death rate per 1000 births. Workstream 6 With our patient advocates we will disseminate findings through patient groups, journals, conferences, a project-specific website and social media, a key policy maker dissemination event, through our integration with the 'maternity and neonatal safety improvement programme' (NHS England/Improvement) and the 15 patient safety collaboratives. TIMELINES – 72-month programme IMPACT New near-miss and significant maternal morbidity criteria derivable from routinely-collected data will allow routine quality assurance reporting, driving women s safety A validated vital-signs-only MOEWS will allow safer recognition of deteriorating women where healthcare is not digitally mature An eMOEWS will deliver optimal early maternal deterioration detection in digitally mature environments Protocolised mMOET-standard escalation pathways, will deliver structured, safer care