A civil registration system that continuously records all births and deaths in a population is the preferred source of high-quality underlying data on under-five mortality but these systems are not well developed in many low- and middle-income countries. Instead, household surveys and population censuses are the primary sources of underlying data in these countries.

The reliance on multiple data sources, i.e. surveys and census conducted several years apart and producing retrospective time series, can result in disparate mortality rates from different sources, sometimes referring to the same time period. Available data also suffer from sampling and nonsampling errors, including misreporting of age and sex, survivor selection bias, underreporting of child deaths, and recall errors as data are collected retrospectively. Further misclassifications can also impact the accuracy of data, for example, early neonatal deaths may be classified as stillbirths. Thus, simply comparing two country data points from different sources and drawing a line between them is not a technically sound way to assess levels and trends. Given varying levels of data quality across different sources, this sort of trend assessment will provide misleading results. Hence, the United Nations Inter-agency Group for Child Mortality Estimation (UN IGME) fits a statistical model to these data that takes into account these various data sources to produce annualized estimates.

It is important to keep these challenges in mind when looking at available country data and also when discrepancies between country data and the UN IGME estimates are being discussed. The following

points are important to highlight:

The UN IGME aims to minimize the errors for each estimate, harmonize trends over time and produce up-to-date and properly assessed estimates of child mortality. Thus, UN IGME estimates are derived from country data. Notably, UN IGME assesses the quality of underlying data sources and adjusts data when necessary.
National estimates may refer to an earlier calendar year than the UN IGME estimates. This is particularly the case where estimates from the most recent national survey are used as the national estimate, since the survey estimates derived from a birth history are retrospective and typically refer to a period before the year of the survey, which may be several years behind the target year for the UN IGME estimates. National estimates may also use a different combination of data sources, or different projection or calculation methods.
In the absence of error-free data, there will always be uncertainty around data and estimates. To allow for added comparability, the UN IGME generates estimates with uncertainty bounds. When discussing the UN IGME estimates, it’s important to look at the uncertainty ranges, which might be fairly wide in the case of some countries.

Hesin internet-kagin er ikki longur stuðlaður

Rætta uppseting: 3.2.2