Background
Preterm birth (PTB) remains a leading cause of neonatal morbidity and mortality, with long-term consequences for children and families and substantial costs to the NHS [1]. Current prediction methods are limited to women with known risk factors and rely on static assessments such as cervical length or infection markers taken at fixed time points [1]. This approach overlooks dynamic, everyday signals that could indicate emerging risk and enable earlier, more tailored interventions.
Novelty & Importance
This PhD project pioneers the use of digital phenotyping [2] to capture real-world, longitudinal behavioural and physiological data from pregnant women such as sleep/wake cycles, mobility, stress indicators inferred from phone usage, and short in-app wellbeing assessments. By integrating these signals with routine maternity records and, for a subset, deep phenotyping and multiomics biomarker data [3], the project aims to develop dynamic, personalised risk forecasts for preterm birth. It is among the first to combine digital health data with clinical records to produce interpretable, actionable predictions for PTB, potentially transforming antenatal care by identifying ‘near to preterm labour’ risk and supporting timely, proportionate responses.
Aims & Objectives
1. To collect and integrate smartphone-derived digital phenotypes and ecological momentary assessments focused on high-risk women with deep phenotyping and paired multiomics biomarker data for risk modelling and collection.
2. Develop and test machine learning models for forecasting preterm birth risk, including ‘near to preterm labour’ (e.g. “next 7 days”) to provide personalised precision support.
3. Co-design and prototype clinician and patient-facing risk summary tools that are actionable in antenatal care.
The student will receive interdisciplinary training in data science, maternal health, digital health governance, and co-design, supported by the DRIVE-Health CDT programme. The project will underpin future feasibility studies to evaluate clinical utility, equity, and acceptability within NHS maternity services.
[1]. Gravett MG, Menon R, Tribe RM, Hezelgrave NL, Kacerovsky M, Soma-Pillay P, Jacobsson B, McElrath TF. Assessment of current biomarkers and interventions to identify and treat women at risk of preterm birth. Front Med (Lausanne). 2024 Jul 26;11:1414428. doi: 10.3389/fmed.2024.1414428.
[2] Leightley D, Dilkina B, Pedersen ER, Dworkin E, Saba S, Howe E, Thota P, Nuthi S, Sedano A, Davis JP. A remote measurement study of PTSD and cannabis use among veterans: Recruitment, retention, and data availability. PLoS One. 2025 Sep 29;20(9):e0332239. doi: 10.1371/journal.pone.0332239.
[3]. Valensin, C., Côté, E.J.M., Pereira-Carvalho, D. et al. INSIGHT-2: mechanistic studies into pregnancy complications and their impact on maternal and child health—study protocol. Reprod Health 21, 177 (2024). https://doi.org/10.1186/s12978-024-01911-0