2. Individual Patient Data (IPD) meta-analysis
This phase of the research aims to establish an international individual participant data (IPD) collaboration to share data on early pregnancy adiposity and pregnancy outcomes. Data obtained from this phase will be used as an external validation dataset. We will use IPD meta-analysis methods to determine whether the prediction model(s) developed in phase 1 (SHAPES cohort study) produce similar findings when applied to external populations.
This phase is builds upon our previous systematic reviews which identified existing published studies that measured early pregnancy adiposity (alternative to BMI) <20 weeks gestation and reported associations with adverse maternal and infant pregnancy outcomes. We are updating these searches to identify any studies which have collected early pregnancy adiposity measures and pregnancy outcomes, even if they have not reported associations between these variables. This stage involves rescreening the studies excluded from the previous reviews as well as searching for pregnancy cohort studies which may have collected, but not published, these data. The protocol for this stage has been registered on PROSPERO.
Authors of any eligible studies will be invited to join an international IPD collaboration. We will establish data sharing and co-authorship agreements with any authors who agree to collaborate in the IPD study. The IPD will be obtained in an anonymised format and stored in a secure data repository. We will also consider using DataShield to enable more collaborators to include their data in IPD meta-analysis studies if they cannot share the raw data.
We will externally validate the risk prediction model(s) using calibration and discrimination analyses. Calibration will examine the agreement between the probability of developing the outcome as estimated by the model, and the observed outcome frequencies (assessed with plots). Discrimination analysis will examine the ability of the model to distinguish between individuals who develop the outcomes or not. This will be assessed using the c-index (area under the receiver operating characteristic curve).
We will start contacting authors of identified eligible studies from May 2022. If you think you have data which might be eligible for inclusion in this international IPD collaboration, please contact Dr Gina Nguyen at Newcastle University, UK.
