One size fits all? The role of social determinants of health on AI-driven CTG intrapartum decision-support tools in term babies

主要目标：分析种族和其他社会健康决定因素与足月儿分娩期间不良结局之间的相互作用;在人工智能驱动的CTG产时决策支持工具中使用种族数据和其他社会健康决定因素来预测围产期不良事件;了解来自代表性不足背景的妇女和临床医生使用人工智能驱动的决策辅助工具进行产时风险评估的经验。临床需求：分娩前或分娩期间的缺氧从其起源、检测、预防和治疗方面仍然是临床团队面临的重大挑战。此外，分娩对母亲和胎儿提出了一些挑战。长期的身体和精神挑战经常被比作一场马拉松，所需的努力使其得名“劳动”。生理上，收缩压缩血管，导致间歇性子宫肌层缺氧。然而，如果这种缺氧延长，将导致功能失调（进展缓慢）的劳动。分娩中的缺氧仍然未知，是导致妇女和婴儿不良围产期结局以及计划外手术干预的原因（Wray等人，2021）。产时缺氧缺血性损伤是全球范围内死产、新生儿脑病和死亡以及长期神经发育后遗症的主要原因（Lovers等人，2022）。此外，有证据表明，来自最贫困背景的妇女以及来自黑人、亚洲人和少数民族群体的母亲，其婴儿在子宫内或出生后不久死亡的风险更高（德雷珀等人，2018年）。因此，这些不利的围产期结果不仅对妇女、婴儿和家庭的生活产生不利影响，而且对人力、物质和经济资源产生影响。因此，预防缺氧是一个重大的公共卫生问题，也是一个重要的健康优先事项。使用心脏分娩描记术（CTG）已成为分娩期间监测胎儿状态的标准技术;然而，胎儿心率（FHR）监测的主要问题与CTG描记图的阅读和解释有关（Schiermeier et al. 2008）。目前，常规CTG监测存在争议，因为其假阳性率高，临床干预增加，但未改善新生儿结局（Alfirevic等人，2013）。然而，分娩期CTG是目前唯一获得国家协会批准的监测方法（Georgieva et al. 2014），特别是对于高风险孕妇（Ayres-de-Campos et al. 2015）。通过开展几项全国性运动和编写报告，如《拯救婴儿生命》、《国民保健服务长期计划》和《奥肯登报告》，人们认识到需要解决这一公共卫生问题。牛津分娩监测（牛津大学）是一个专家团队，开发数据驱动的心脏分娩描记（CTG）系统/软件，以持续评估分娩开始时和分娩期间的胎儿健康状况。OxSys是一种人工智能驱动的分娩决策支持工具，可作为筛选测试，以在分娩开始时识别高风险或不适合的婴儿。该工具基于一个统计预测模型，该模型考虑了产前和分娩期间的风险因素，并使用计算机化的CTG来预测围产期不良结局的风险。

Key objectives: to analyse the interaction between ethnicity and other social determinants of health and adverse outcomes during labour in term babies; to use ethnicity data and other social determinants of health in AI-driven CTG intrapartum decision-support tools to predict perinatal adverse events; to understand the experience of women from underrepresented backgrounds and clinicians using AI-driven decision aid tool for intrapartum risk assessment. Clinical need: Hypoxia, before or during labour remains a significant challenge to clinical teams, from its origin, detection, prevention, and treatment. Moreover, labour presents several challenges to mother and fetus. The long-lasting physical and mental challenge has often been compared to a marathon and the effort required giving rise to its name, 'labour'. Physiologically, contractions compress blood vessels that lead to an intermittent myometrial hypoxia. However, if this hypoxia is prolonged, will contribute to dysfunctional (slow to progress) labours. Hypoxia in labour is still unknown, and is responsible for adverse perinatal outcomes in women and babies and unplanned surgical interventions (Wray et al. 2021). Intrapartum hypoxic ischaemic injury is a major contributor to stillbirths, neonatal encephalopathy and mortality, and long term neurodevelopmental sequelae worldwide (Lovers et al. 2022). Furthermore, evidence shows that women from the poorest backgrounds and mothers from Black, Asian and Minority Ethnic groups are at higher risk of their baby dying in the womb or soon after birth (Draper et al. 2018). Consequently, these adverse perinatal outcomes have not only a detrimental impact on women's, infants and families' lives but also have an effect on human, material and economic resources. Hence, the prevention of hypoxia is a major public health issue and a significant health priority.The use of cardiotocography (CTG) has become a standard technique to monitor the state of the fetus during labour; however, the major problems with fetal heart rate (FHR) monitoring relate to the reading and interpretation of the CTG tracings (Schiermeier et al. 2008). Controversies around routine CTG monitoring exist nowadays due to the their high false-positive rate and increment in clinical interventions without improving neonatal outcomes (Alfirevic et al. 2013). Nevertheless intrapartum CTG is currently the only method of surveillance approved by national societies (Georgieva et al. 2014), particularly for high risk pregnant women ( Ayres-de-Campos et al. 2015). The need for tackling this public health issue has been acknowledged through the development of several national campaigns and reports, such as Saving Babies Lives, the NHS Long Term Plan and the Ockenden report. The Oxford Labour Monitoring (University of Oxford) is a specialist team developing data-driven cardiotocography (CTG) systems/software to continuously assess fetal wellbeing at the onset of and during term labour. The OxSys, an AI-driven decision support tool in labour serves as a screening test to identify high risk or unfit babies at the onset of labour. This tool is based on a statistical prognostic model that takes into consideration antenatal and intrapartum risk factors as well as using computerized CTG to predict the risk of experiencing perinatal adverse outcomes.