O-148 Miscarriage prevalence and factors: a multi-source approach in France

Study question How do medical records, survey data and qualitative interviews provide complementary insights into the epidemiology of miscarriage in France? Summary answer A comprehensive understanding of miscarriage prevalence and its socioeconomic risk factors needs triangulation of data from various sources. What is known already Miscarriage prevalence and its risk factors have mainly been estimated with Nordic health (often hospital) records, showing a decline in medically managed miscarriages and suggesting the woman’s age and previous miscarriages as the main risk factors. Population-based surveys can capture miscarriages that did not lead to hospital management and allow the investigation of more factors than medical records. However, miscarriages are also deemed misreported in such data, with little knowledge of the extent, patterns and reasons. Qualitative approaches can help understand miscarriage experiences, what is reported as a miscarriage in a survey and the reasons for disclosing it or not. Study design, size, duration This study compares multiple data sources. The primary one is the French National Health Data System, which contains exhaustive medical records for 13,632,246 pregnancies, including inpatient (2009-23) and outpatient care (2013-23). This is compared to the FECOND survey (2010-11), a nationally representative sample of women of reproductive age in France (N = 7,196 pregnancies). We will also conduct around 50 in-depth interviews about miscarriage experiences, to better understand potential misreporting of miscarriages in the quantitative data sources. Participants/materials, setting, methods 1) Regression analyses estimating miscarriage risk by year and women’s characteristics in medical records (age at the start of the pregnancy, social security insurance indicating financial precarity, use of assisted reproduction, prior miscarriages). 2) Event-history analyses estimating miscarriage risk in the FECOND survey using similar information and women’s education, socioeconomic position and partnership stability at the start of the pregnancy. 3) Thematic analysis of qualitative interviews pertaining to reasons for misreporting miscarriages in surveys. Main results and the role of chance Our multi-source approach reveals significant discrepancies in miscarriage prevalence across data sources. Medical records show a decline in hospital-managed miscarriages from 6.9% to 5.1% of all clinically recognized pregnancies between 2009 and 2023. Including primary care data, these rates range from 9.9% in 2013 to 8.9% in 2023. This is below the 14% miscarriage rate estimated with self-reported pregnancy outcomes in the FECOND survey, and contrasts with the increasing reporting across women’s birth cohorts (HR = 1.419, p < 0.001 for those born in 1980–85 vs. 1961–69). Both data sources confirm established risk factors: maternal age and prior miscarriages. Interestingly, neither source shows significant socioeconomic disparities in miscarriage risk: financial precarity status in medical records has minimal effect (coeff=-0.009, p < 0.001), whereas educational attainment and self-reported financial situation in FECOND show no significant association (p > 0.05). In the qualitative material, interviewees will be presented with several typical survey questions on pregnancy outcomes, to understand how survey results can be biased by miscarriage misreporting (such as underreporting of miscarriage, or reporting ectopic pregnancies or abortions as miscarriages). Limitations, reasons for caution Health records underestimate miscarriage prevalence due to missing non-medically managed miscarriages, and likely miss very early pregnancy losses that were not administratively declared. While our forthcoming qualitative investigation will provide valuable insights into reporting behaviours in surveys and care experiences, it will not shed light on limitations of administrative data. Wider implications of the findings Our research will contribute to a better understanding of miscarriage risk factors and methodological shortcomings of various data sources. Trial registration number No