Strengths and limitations
To the best of our knowledge, this is the first systematic review that synthesised data on the ethnic variation of the causes of stillbirth in high income nations and to standardise reporting of outcomes using the ICD-10-PM classification. To increase the accuracy of identified causes, we included only those studies that reported foetal autopsy and/or placental pathology data to determine the actual cause of stillbirth38 rather than associated risk factors, and excluded studies that used surrogate markers of ethnic identity such as mother tongue or language spoken at home. Mapping disparate causation data from individual studies to the ICD-10-PM and subsequently pooling results by meta-analysis allowed improved interpretation of findings and easier comparison between ethnic groups.
Results from this study should be interpreted in the context of some limitations. First, exclusion of studies reporting only associations between various risk factors and stillbirth may have resulted in loss of information and limited our ability to assess the full value of the ICD-PM. Second, use of crude categories to define ethnicity may be limited by being only proximal guides to experiences, practices, and beliefs of individuals and may not capture the associated complex and interrelated socioeconomic determinants such as education and employment that may be driving disparities in stillbirth rates.27, 39-44 Inequalities in stillbirth between various ethnic groups have been shown to reduce after acculturation and adoption of host nationality,45 highlighting the limitation in interpreting stillbirth data segregated by ethnicity. Third, differences in causes of stillbirth between studies may be a consequence of differences in classification rather than any important differences in pathological processes per se. This is especially pertinent in the context of the myriad classification tools available and used in various studies, making regional and global comparisons difficult. The ICD-PM itself faces challenges including insufficient differentiation of causes from associated conditions and insufficient detail on maternal conditions. Fourth, most data used in our analyses were collected more than 10 years ago and may not reflect recent migration patterns and demographic changes, and current socioeconomic and health care environments. Finally, despite including studies investigating thecauses of stillbirths, we found considerable variation in autopsy rate between studies with some studies reporting rates as low as 38%. Previous studies have shown that ethnic minority groups have lower rates of being asked for consent for a post-mortem study, and also lower rates of consent when asked.46, 47 This may be due to assumptions and/or actual differences in religious or cultural observances and may reflect an inherent limitation in studies involving minority ethnic groups and causes of stillbirth. Due to this limitation it is also worth considering alternatives to post-mortem examination which may be accessible in high income counties, such as MRI,48 to assist in assigning cause of death when post-mortem examination is declined.