Department of Sexual and Reproductive Health and Research, WHO, Geneva, Switzerland.
WHO Jordan Country Office, Amman, Jordan.
Lancet Glob Health. 2023 Jun;11(6):e854-e861. doi: 10.1016/S2214-109X(23)00125-0. Epub 2023 May 8.
In most low-income and middle-income countries (LMICs), national surveys are the main data source for stillbirths and perinatal mortality. Data quality issues such as under-reporting and misreporting have greatly limited the usefulness of such data. We aimed to enhance the use of mortality data in surveys by proposing data quality metrics and exploring adjustment procedures to obtain the best possible measure of perinatal mortality.
We performed a population-based analysis of data from 157 demographic and health surveys (DHSs) from 1990 to 2020, with reproductive calendar and birth history data from 53 LMICs. Pregnancies terminated before 7 months' gestation were excluded. We examined data quality and compared survey values with reference values obtained from a literature review to assess misreporting of the age at early neonatal death, omission and transference of stillbirths, and very early neonatal deaths. Real cohort life-table rates of stillbirth, early neonatal, and perinatal mortality per 1000 births were calculated. The underlying risks of stillbirth and daily deaths were modelled using modified Gompertz-Makeham models.
Data for 2 008 807 pregnancies of ≥7 months' gestational age were extracted from the reproductive calendar for the analysis period. Age heaping at day 7 occurred in most surveys. The median value for the heaping index of deaths at day 7 was 2·05 (IQR 1·36-2·87). The median ratio of stillbirths to deaths on days 0-1 was 1·15 (0·86-1·51). Of the 157 surveys, 23 (15%) were considered to have plausible ratios, 71 (45%) had probable ratios, and 63 (40%) had improbable ratios. The ratio of deaths on days 0-1 to deaths on days 2-6 varied considerably between surveys and 119 surveys (76%) had ratios of less than 2·4, indicative of under-reporting of very early neonatal deaths in most surveys. The fully adjusted model increased the median stillbirth rates from 12·2 (9·4-15·9) to 25·6 (18·0-33·4) per 1000 births, with a median relative increase of 95·0% (56·6-136·6). The median perinatal mortality rate also increased from 32·6 (23·6-38·3) to 44·8 (32·8-58·0) per 1000 births, with a median relative increase of 47·8% (6·9-61·0).
A simultaneous focus on stillbirths and early neonatal mortality facilitates a comprehensive assessment of inaccurate reporting in household surveys and allows for better use of surveys in planning and monitoring of efforts to reduce stillbirths and early neonatal mortality.
None.
在大多数低收入和中等收入国家（LMICs）中，国家调查是死产和围产儿死亡的主要数据来源。数据质量问题，如漏报和错报，极大地限制了此类数据的使用。我们旨在通过提出数据质量指标并探索调整程序来提高调查中死亡率数据的使用，以获得围产儿死亡率的最佳衡量标准。
我们对 1990 年至 2020 年期间来自 53 个 LMIC 的 157 项人口普查和健康调查（DHS）的生殖日历和生育史数据进行了基于人群的分析。排除妊娠 7 个月前终止的妊娠。我们检查了数据质量，并将调查值与文献综述中获得的参考值进行了比较，以评估早期新生儿死亡年龄、死产和死胎的漏报和转移以及极早期新生儿死亡的报告错误。计算了每 1000 例活产的死产、早期新生儿和围产儿死亡率的实际队列生命表率。使用改良的 Gompertz-Makeham 模型对死产和每日死亡的潜在风险进行建模。
从生殖日历中提取了 157 项分析期间≥7 个月妊娠的 2008807 例妊娠数据。大多数调查都存在第 7 天死亡的年龄堆积现象。第 7 天死亡堆积指数的中位数为 2.05（IQR 1.36-2.87）。第 0-1 天死产与死亡的中位数比值为 1.15（0.86-1.51）。在 157 项调查中，有 23 项（15%）被认为具有合理的比值，71 项（45%）具有可能的比值，63 项（40%）具有不太可能的比值。第 0-1 天的死亡与第 2-6 天的死亡比值在调查之间差异很大，119 项调查（76%）的比值小于 2.4，表明大多数调查中极早期新生儿死亡报告不足。完全调整后的模型将每 1000 例活产的死产率中位数从 12.2（9.4-15.9）增加到 25.6（18.0-33.4），中位数相对增加 95.0%（56.6-136.6）。围产儿死亡率中位数也从 32.6（23.6-38.3）增加到 44.8（32.8-58.0），中位数相对增加 47.8%（6.9-61.0）。
同时关注死产和早期新生儿死亡有助于全面评估家庭调查中不准确的报告，并允许更好地利用调查来规划和监测减少死产和早期新生儿死亡的工作。
无。
