Shifts in age pattern, timing of childbearing and trend in fertility level across six regions of Nigeria: Nigeria Demographic and Health Surveys from 2003-2018

Background Nigeria’s population is projected to increase from 200 million in 2019 to 450 million in 2050 if the fertility level remains at the current level. Thus, we examined the shifts in the age pattern of fertility, timing of childbearing and trend in fertility levels from 2003 and 2018 across six regions of Nigeria. Method This study utilised the 2003, 2008, 2013, and 2018 Nigeria Demographic and Health Survey datasets. Each survey was a cross-sectional population-based design, and a two-stage cluster sampling technique was used to select women aged 15-49 years. The changes in the timing of childbearing were examined by calculating the corresponding mean ages at the birth of different birth orders for each birth order separately to adjust the Quantum effect for births. The Gompertz Relational Model was used to examine the age pattern of fertility and refined fertility level. Result In Nigeria, it was observed that there was a minimal decline in mean children ever born (CEB) between 2003 and 2018 across all maternal age groups except aged 20-24 years. The pattern of mean CEB by the age of mothers was the same across the Nigeria regions except in North West. Nigeria’s mean number of CEB to women aged 40-49 in 2003, 2008, 2013 and 2018 surveys was 6.7, 6.6, 6.3 and 6.1, respectively. The mean age (years) at first birth marginally increased from 21.3 in 2003 to 22.5 in 2018. In 2003, the mean age at first birth was highest in South East (24.3) and lowest in North East (19.4); while South West had the highest (24.4) and both North East and North West had the lowest (20.2) in 2018. Similar age patterns of fertility existed between 2003 and 2018 across the regions. Nigeria’s estimated total fertility level for 2003, 2008, 2013 and 2018 was 6.1, 6.1, 5.9 and 5.7, respectively. Conclusion The findings showed a reducing but slow fertility declines in Nigeria. The decline varied substantially across the regions. For a downward change in the level of fertility, policies that will constrict the spread of fertility distribution across the region in Nigeria must urgently be put in place. Nigeria has the largest population in Africa and the 14th largest in landmass. According to the 2006 Population and Housing Census conducted in Nigeria, the country’s population was 140,431,790 [17, 18], but the 2019 projection is based on the 2006 census figure, as the base year was above 200 million [19]. The country comprises 36 states with a Federal Capital Territory and is structured into six geopolitical zones, which are North Central (NC), North East (NE), North West (NW), South East (SE), South-South (SS) and South West (SW). The study utilized data from the 2003, 2008, 2013 and 2018 Demographic and Health Surveys (NDHS). Each survey was a cross-sectional population-based design, and a two-stage cluster sampling technique was used to select women aged 15–49 years. The 2003 NDHS programme made use of the sampling frame designed for the 1991 population census, while the sampling frame designed for the 2006 population and housing census was used for 2008, 2013 and 2018 NDHS but with modification due to expansion in the number of households between the census period and the survey years defined in all the survey rounds, the primary sampling unit (PSU) was a cluster tagged as the Enumeration Areas (EAs) from the 1991 and 2006 EA census sampling frames. Samples for the 2003 and 2008 surveys were selected using a stratified two-stage cluster design consisting of 365 clusters in 2003 NDHS and 888 clusters in 2008 NDHS. While 2013 and 2018 NDHS were conducted at three and two stages, respectively. For 2013, 893 localities were selected at the first stage with probability proportional to the size and with an independent selection from each sampling stratum. In the second stage, one EA was randomly selected from most of the selected localities. In a few larger localities, more than one EA was selected. In total, 904 EAs were selected. After selecting the EAs and before the main survey, a household listing operation was carried out in all the selected EAs. For 2018 NDHS, at the first stage, 1400 EAs were selected; and a household listing which served as a sampling frame was conducted on the selected EAs. In the second stage, 30 households were selected from each cluster by an equal probability of systematic sampling. The number of households interviewed in 2003, 2008, 2013, and 2018 was 7864, 34070, 40680 and 42000, respectively. The number of women aged 15–49 years interviewed for these year periods used in the study is given as 7620, 33385, 38948, and 41821, respectively. A detailed description of the methodology of the data set used for this study may be found in NDHS main report [15]. The changes in the timing of childbearing were examined by calculating the corresponding mean ages at the birth of different birth orders of the study periods. The mean was calculated for each birth order separately to adjust the Quantum effect for births. The level of fertility is influenced by changes in the timing of childbearing (Tempo) and children ever born (Quantum) [12]. Where μ is the mean age of childbearing that measures the timing of childbearing; x(i) is the central age-point in the age interval, and I is the age group containing the upper age limit of the childbearing span; and f(i) denotes the fertility rate experienced by women in each age group. The shifts in the age pattern of fertility can be examined by looking at observed or model age-specific fertility rates. However, due to reporting errors and the truncation effect; observed age-specific fertility may be inappropriate to describe the age pattern of fertility [20]. Thus, in an attempt to describe the fertility age pattern, several mathematical models have been proposed. These models have been used successfully to fit the age-specific fertility rates in different populations. One of such model was a relational method between a standard fertility schedule and any other schedule proposed by Brass [20]. The model is based on the assumption that the cumulative age pattern of fertility follows a Gompertz distribution function. However, it was found later that this model has two major shortcomings: first, it involves using total fertility (TF), which may be biased. The second shortcoming is the assumption that fertility has been constant. Nevertheless, Zaba’s Ratio method of 1981, which this study used, was an improved variant of the model proposed by Brass [21]. Two sets of data were used in the study. These are the women’s data set (individual recode) and children’s data set (child recode). All the variables needed in both the women’s data set and children’s data set for the study were extracted using SPSS. The variables used to estimate the numerator of TFR, the month and year of the child’s birth, were extracted from the children’s data set with the ID variable and matched with the women’s data. The observed ASFRs as presented in Fig 2 was obtained through direct estimation as described by Moultrie et al. [21]. While the estimated ASFRs were attained indirectly by following the procedures known as the “Ratio method” developed by Moultrie et al. [21]. The Gompertz parameters derived through these procedures were used to describe the age pattern of fertility and refine observed ASFRs. In the method, the average parities, 5Px, of women in each age group (x, x+5) for x = 15, 20, ——- 45, were calculated. Then the fertility standard developed by Booth [20] was chosen to fit the model as follows: Where: The plots of z(x)–e(x) against g(x) and z(i)–e(i) against g(i) (on the same set of axes) that were almost on the same line was used to fit the model. The values of α (intercept) and β (slope) are the parameters. The level of fertility (TFR) was estimated indirectly by applying the above-derived parameters (α & β) to the current fertility gompits. The parameters α & β for the periods were compared; α indicates the location of fertility, and β shows the spread in relation to the standard. Sample weights were applied to each case to adjust for differences in the probability of selection. Weighting is important to increase the sample’s extent of representativeness and reduce the errors associated with sample selection bias. Since the authors of this manuscript did not collect the data, we sought permission from the MEASURE DHS website and access to the data was provided after our intent for the request was assessed and approved on the 10th of March 2021. The DHS surveys are ethically accepted by the ORC Macro Inc. Ethics Committee and the Ethics Boards of partner organizations in different countries, such as the Ministries of Health. The women who were interviewed gave either written or verbal consent during each of the surveys.