Reducing stillbirth in high burden settings using biomarkers and ultrasound technologies: protocol for the multi-centre prospective iTECH cohort study

Résumé

ABSTRACT Introduction Stillbirth prevention requires reliable detection of potential causes for timely interventions. Currently, there is no effective screening strategy to identify fetuses at risk of stillbirth. Prognostic models have been proposed as a potential solution, but there is a shortage of robust, clinically applicable models in low- and middle-income countries. Early birth is frequently initiated without proper risk stratification, leading to increased neonatal and infant morbidity and mortality. This study aims to develop and validate multi-modal multivariable prediction models for stillbirth and pathologies that lead to stillbirth (preeclampsia & fetal growth restriction) using widely accessible and cost-effective markers. Stakeholder perspectives will also be assessed. Methods and analysis This multi-center prospective cohort study is running in four high volume regional referral hospitals in Uganda: Kawempe, Hoima, Lira, and Mbale. We will enroll at least 6,075 pregnant women attending routine antenatal care (ANC), above 13 years of age, and ζ11 weeks of gestation. Data and biological samples will be collected at 11-23 weeks, 35-37 weeks and at birth in all women. In a subset of women, additional measurements will be obtained between 24-34 weeks and 38-42 weeks to allow for spread of the data across the full spectrum of pregnancy. This data will enable us to investigate the physiological changes with gestational development in healthy or unhealthy pregnancies, to guide future monitoring and management of women and establishment of reference values for novel markers. The placenta will be collected for histopathological analysis in women diagnosed with intrauterine fetal demise at ζ20 weeks of gestation, stillbirth nearmiss and their corresponding controls. Data on socio-demographics, obstetric history, current pregnancy conditions, and tests such as maternal hemodynamics, ultrasound, and biochemical markers will be collected from each participant, and used to develop regression and machine learning prediction models. Models will be validated and evaluated by comparing their calibration plots, precision and recall, F1 scores and accuracy, aiming for less complexity and reliable predictions. Emerging models will be translated into software as a medical device (SAMD), while taking into account user experiences, regulatory requirements, data pipelines in clinical workflows and user-friendly interfaces that facilitate access and the interpretation of outputs, to allow for seamless integration into existing electronic health information systems and decision support tools. To assess stakeholder perceptions, we will employ an exploratory qualitative component using focus group discussions, semi-structured and key informant interviews. The sample will include 81 purposively selected women and their partners who use maternity care services, local leaders and healthcare providers in and out of the four hospitals implementing iTECH in Uganda. Qualitative data will be audio recorded, transcribed verbatim and thematic analysis performed using Nvivo 12. Ethics and dissemination The study protocol was approved by Makerere University School of Medicine Research Ethics Committee (Mak-SOMREC-2022-535) on March 1, 2023 and the Uganda National Council for Science and Technology (HS2762ES) on April 6, 2023. We will disseminate the findings through scientific publications, stakeholder engagements, Wellcome Leap program meetings, professional bodies, reproductive maternal newborn child and adolescent health coalition platform, Ministry of Health Uganda, district and hospital meetings, webinars, and conferences. Strengths and limitations of this study large sample of women recruited from a general obstetric population to inform stillbirth and related outcomes prediction models based on multi-modality data. measurements taken to help understand physiological changes with gestational development to guide future monitoring and management of pregnant women. histopathology performed in selected cases and controls to contribute to the understanding of mechanisms that lead to stillbirth and refine model development. of active tracking mechanisms like phone calls, short message texts and locator forms to remind mothers of their scheduled appointments and mitigate participant drop-offs. care for women included in this study may differ from that received by other women attending routine ANC: this may lead to lower adverse pregnancy outcome rates in this cohort than usual in the setting.

Citer ce document

Accès au document

Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter

Auteur(s)

Statistiques

Consultations : 1

Téléchargements : 0