Operationalising the broad definition of pre-eclampsia: a retrospective cohort study
In brief
Routine codes miss most pre-eclampsia cases captured by a broader record review
In 21,120 pregnancies, routine electronic record fields and diagnosis codes identified only 29% to 37% of cases found using a broader definition that included signs of maternal or placental dysfunction. They also captured fewer cases with severe hypertension and small-for-gestational-age birth; combining codes and record fields helped only modestly, underscoring the need for richer data to identify cases accurately.
- Journal
- The Lancet. Digital health (Q1)
- Published
- 4 November 2026
- Study design
- Retrospective cohort
- Evidence level
- Level 4, Very Low (CEBM 4)
- Authors
- Milly G Wilson, Hiten D Mistry, Laura J Slade, Jeffrey N Bone, Joel Singer, Caroline Ovadia, et al.
- PMID
- 42982697
- DOI
- 10.1016/j.landig.2026.101047
Why clinicians should know about it
- Picked for Health Informatics (top studies of the week, 8 November 2026): Compares EHR case-finding definitions and data quality for pre-eclampsia
Abstract
BACKGROUND: Identifying pre-eclampsia within electronic health records (EHRs) is challenged by its complex definition, evolving diagnostic criteria, and the nature of routinely collected data. There is a need to optimise methods for EHR-based definitions of this major cause of maternal and fetal and newborn mortality and morbidity. METHODS: We conducted a retrospective cohort study using pseudonymised data from the Early Life Cross Linkage in Research-Born in South London (eLIXIR-BiSL) cohort, linking routine maternity and neonatal records from two trusts in South London, UK, from Oct 1, 2018, to July 11, 2023. We included women with singleton pregnancies and at least one recorded blood pressure measurement from 20 weeks' gestation who gave birth at one of the contributing NHS trusts, Guy's and St Thomas' NHS Foundation Trust. A broad reference definition of pre-eclampsia, which was in line with current International Society for the Study of Hypertension in Pregnancy and UK National Institute for Health and Care Excellence guidelines and included evidence of maternal end-organ dysfunction or uteroplacental dysfunction, or both, that incorporated structured EHR data, was compared with three case-finding definitions: semi-structured EHR data using drop-down menus (EHR semi-structured definition); ICD-10 codes (ICD-10 definition); and combined semi-structured EHR data and ICD-10 codes (EHR semi-structured and ICD-10 definition). Diagnostic test properties and adverse outcome detection rates were assessed. FINDINGS: Of 33 040 pregnancies with antenatal booking information, 21 120 pregnancies with booking information, birth outcomes, and blood pressure measurements recorded from 20 weeks' gestation were included. Women at booking had a median age of 33·0 years (IQR 29·0-36·0). Pre-eclampsia prevalence varied by definition: 7·7% (n=1630; reference definition), 4·5% (n=957; EHR semi-structured definition), 3·7% (n=773; ICD-10 definition), and 5·8% (n=1224; EHR semi-structured and ICD-10 definition). Compared with the reference definition, case-finding definitions had inferior detection rates for adverse outcomes, particularly for severe hypertension (26·0% [EHR semi-structured definition], 28·1% [ICD-10 definition], 32·6% [EHR semi-structured and ICD-10 definition]; reference 68·2%) and birthweight <10th percentile (9·7%, 10·6%, 13·3%, respectively; reference 26·2%). Specificity for detection of the reference definition of pre-eclampsia was high (97% [95% CI 97-97]) for all case-finding definitions, whereas sensitivity was poor (29% [26-31] to 37% [35-40]). INTERPRETATION: Compared with a structured EHR definition, pre-eclampsia definitions relying solely on semi-structured EHR data or ICD-10 codes, or both, underperform in identifying pregnancies with adverse outcomes. The findings highlight the need to integrate structured and semi-structured EHR data to improve pre-eclampsia case ascertainment and emphasise the importance of implementing emerging technologies for mining maternity data. FUNDING: Medical Research Council, King's College London Centre for Doctoral Training in Data-Driven Health, and PRECISE Network.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.