Methods and tools to evaluate the quality of vital statistics
Methods to assess the completeness of birth registration
Birth registration should be the key source of fertility statistics to inform government monitoring, evaluation and planning. However, poor-quality birth registration data will inhibit their full usage or lead to inaccurate fertility statistics. This section describes methods to assess the completeness of birth registration.
The completeness of birth registration is defined as:
The standard approach to estimate the completeness of birth registration is to estimate the total actual number of births from various sources of fertility data. Some common sources of birth data comparison include:
- The number of births derived from a national census that occurred within the past few year
- The number of births estimated or projected from census data
- Births registered in previous years if the national civil registration and vital statistics (CRVS) system has at least a 90 per cent registration rate
- Births, or fertility rates to derive total births, from a household survey such as the Demographic and Health Survey (DHS), or the Multiple Indicator Cluster Survey (MICS)
- Asking mothers if their children’s births were registered, or if each child has a birth certificate, as frequently conducted by the DHS or MICS
- Census questions asking if each person in the household has a birth certificate or if their birth was registered (data would then need to be filtered to analyse just children)
- Births recorded in a country’s health information system can be compared with registered births
- School enrolment data, where reliable, can also be used to compare registered births with the number of children enrolled in school
If no other source of information is available, a crude birth rate can be applied to the estimated total population to estimate the total number of births for that year.
With the availability of computers and statistical software, all countries with birth registration should be able to determine, with good precision, the completeness of birth registration in their country. Knowing the extent to which births are undercounted is important for informing all health programs and social services. Similarly, when these techniques are used at the sub-national level, the results become important guidance to inform which areas should be the focus of birth registration drives, because the level of completeness is likely to be highest in urban areas, and lowest in rural and marginalised areas with indigenous populations. Mapping the completeness of birth registration by province or other sub-national administrative area is a very effective way of identifying registration gaps.
To read more about some of the complex demographic techniques used to estimate birth registration completeness see data sources and methods section below.
Data sources and methods
Many population censuses, particularly the 2010 round of censuses, asked women aged 15–49 to report how many children they have ever given birth to, and:
- How many births they have had in the previous 12 months
- Details of their most recent birth (including date of birth)
- The number of children still alive.
Various methods use these census data to estimate age-specific fertility rates (ASFRs) that can be applied to population estimates, to determine total number of births and evaluate the completeness of birth registration:
- Brass P/F ratio method: This technique produces a factor for adjusting reported age-specific fertility rates (based on vital registration or births in the 12 months before a census or survey) to the actual level of fertility. The adjustment factor is determined by comparing data on children ever born, by age of women, with a set of cumulated age-specific fertility rates.
- Relational Gompertz model: This is a commonly used method that overcomes limitations of census fertility data, including too few births reported in the previous year and errors in age reporting by older women.1 It uses information from the fertility in the previous year and children ever born by age of the woman, based on an original method developed by Brass.2
- Reverse survival: Data on the size of the population by age and sex can be used to estimate the number of births ‘x’ years ago, assuming that the population is closed to migration and the age of the mother is reported accurately. The process of reverse survival is commonly known as ‘rejuvenation’ and is done by rejuvenating the population at young ages to calculate the number of births that occurred in the past. The reliability of the estimate depends primarily on the quality of the census data by age and on the quality of the life tables used to represent mortality during the periods before the census.
- Own-children method: This technique uses a modified reverse-survival technique that matches children to their mothers to estimate fertility in periods before the census.
- Intercensal fertility estimates derived from cohort parities: Where there are data of children ever born (parities) by age group of mother from two successive censuses 5 or 10 years apart, a method that uses the relational Gompertz model can estimate intercensal fertility.
- Preston integrated technique: Based on the age structure from two consecutive censuses, this method estimates the intercensal level of mortality, the crude birth rate and an age distribution of the population for the intercensal period.
The DHS is the most prominent household survey for producing fertility rates. The DHS questionnaire includes a detailed birth history, to be answered by women (generally 15–49 and sometimes only those who were ever married). This information contains the month and year of birth, sex of the child, age of the mother, birth order, whether the child is still alive and whether the child is from a multiple birth. ASFRs can be calculated from this information and applied to population estimates to estimate total number.
1 Moultrie T (2013). The relational Gompertz model, in Tools for demographic estimation, IUSSP, Paris, pp. 54–68.
2 Brass W (1964). Uses of census or survey data for the estimation of vital rates, paper prepared for the African seminar on vital statistics, Addis Ababa 14–19 December 1964. United Nations, New York.
Read more about these techniques in the: