In this issue of Jun
In this issue of , Jun Zhu and colleagues report data retrieved from China\'s National Maternal Near Miss Surveillance System (NMNMSS), yielding a stillbirth rate of 8·8 per 1000 births for the protein kinase inhibitor 2012–14. By comprehensively incorporating datasets based on level II (county or city district) and level III (provincial and subprovincial regional) hospital deliveries, this study is to our knowledge the largest to date and fills a gap in China\'s vital statistics. One significant limitation, however, is the exclusion of level I (township) hospitals and home deliveries. According to a survey of a subprovincial complete birth population for one region in 2010, births occurring in level I hospitals accounted for 52% of the population compared with 33% and 15% in level II and III hospitals, respectively. Of note, the stillbirth rate in level II and III hospitals was 7–8 times that in level I hospitals, demonstrating highly centralised management of high-risk pregnancies and deliveries. The proportion of home deliveries ranged from 2·5% in cities to 36·1% in the least developed rural area during 1996–2008. However, the New Rural Cooperative Medical Scheme, a nationwide health-care insurance policy launched since 2010 to cover all rural residents\' hospital costs including subsidies to encourage hospital delivery for rural pregnant women, along with the suspension of level I hospitals unqualified for facility-based deliveries, might have reduced the rate of home births and caused a shift in birth distribution from low-level to high-level hospitals. Given the absence of data from level I hospitals and a lack of clarity on the distribution of births among different levels of hospitals excluded from the sampling, the capacity of Zhu and colleagues\' study to represent the national population is limited, and the stillbirth rate could be biased. Furthermore, the births sampled in this study (n=3 956 836) account for about 12% of the total national population (there were 33 million births nationwide during 2012–14). How many and which provinces have been included in the sampling is the key methodological issue and thus has a significant impact on data representation. Zhu and colleagues state that “urban populations were over-represented in the NMNMSS, particularly in central and western regions”, indicating a biased sample, large as it is. There are other important issues that also need to be addressed. First, births before 28 complete weeks are conventionally excluded from vital statistics registration and analysis by China\'s family planning system. However, as advances in neonatal intensive care have pushed the boundary of fetal viability to around 22–24 weeks in industrialised countries and 25–27 weeks in emerging regions in China, the stillbirth and perinatal period should be redefined with a threshold earlier than 28 complete gestational weeks. Additionally, there is confusion especially among medical staff in low-level hospitals over the measurement of perinatal and neonatal mortality. Neonates who were born with detectable life signs but died shortly thereafter may be misclassified as stillbirths rather than livebirths, resulting in a biased stillbirth rate and early neonatal mortality. Resuscitation during the early post-partum period followed by parents\' withdrawal of their babies\' medical treatment owing to financial and prognostic concerns may further contribute to the risk of misclassification. Second, the proportional contribution of abortion to the stillbirth rate is yet to be elucidated. In practice, abortions are registered as miscarriages rather than stillbirths and will probably contribute to an increase in stillbirth rate should they be calculated as such in future. Notably, with universal implementation of the two-child policy starting very recently, we might expect a declining stillbirth rate in years to come. Third, migrant workers (approximately 8% of the total population of 1·34 billion by 2010) shuttling between rural areas and economically advantageous coastal regions have been largely unaccounted for in birth registries. Due to the instability of socioeconomic status, the stillbirth rate among migrant populations may be high and should be addressed in future studies.