Background: Nigeria’s high perinatal mortality rate (PNMR) could be most effectively reduced by targeting factorsthat are associated with increased newborn deaths. Low access to skilled birth attendants (SBAs) and weak healthsystem are recognized factors associated with high PNMR but other socio-demographic and reproductive factorscould have significant influences as well. Identification of the major factors associated with high PNMR would berequired in designing interventions to improve perinatal outcomes.Methods: For this cross-sectional study, data from the Nigeria Demographic and Health Survey 2008 were used toestimate the PNMR of non-hospital births in identified socio-demographic and reproductive situations that areknown to influence PNMR. The estimated PNMR were compared using logistic regression analysis.Results: The PNMR was 36 per 1000 live births. North central region had the lowest PNMR while the south eastregion had the highest rate (odds ratio 1.59; 95% CI: 1.03, 2.45). Other correlates of high PNMR were belonging tothe poorest wealth quintile (odds ratio 1.87; 95% CI: 1.30, 2.70), maternal age group 15-19 years (odds ratio 1.59;95% CI: 1.05, 2.22), multiple birth (odds ratio 3.12; 95% CI: 2.11, 4.59), history of previous perinatal death (odds ratio3.31; 95% CI: 2.73, 4.02), birth interval shorter than 18 months (odds ratio 1.65; 95% CI: 1.26, 2.17) and having a smallbirth size (odds ratio 2.56; 95% CI 1.79, 3.69). Birth attendant, place of birth, parity, maternal education and rural/urban residence had no association with PNMR.Conclusions: Reproductive factors that require midwifery skills were found to contribute most to PNMR. Werecommend general strengthening of the health system, recruitment of SBAs and retraining of available birthattendants with emphasis on identification and referral of complicated cases. Family planning should be a coreMCH activity to address the issues of teenage pregnancy and short pregnancy intervals.
The study was based on an analysis of data from the Nigeria Demographic and Health Survey 2008 (Nigeria DHS 2008) which took place from June to October 2008 [16].The Nigeria DHS 2008 was a face-to-face nationally representative cross-sectional survey of women of reproductive age (15-49 yrs). Using the 2006 census enumeration area (EA) list as a sample frame, 888 (286 urban and 602 rural) EAs were selected from the 36 states and Federal Capital Territory (FCT) with each EA consisting of about 41 households. The target of the survey was to get 36,800 completed interviews. Based on the non-response rate of 2003 DHS, to achieve the sample size, 36, 800 households were selected and all age-eligible women were interviewed. Information was obtained from eligible respondents on a number of demographic and reproductive health issues including a detailed history of all children ever born alive, whether they were alive or dead at the time of interview and if dead, at what age they died. Information on place of birth and who assisted each birth was also obtained. They were also asked if they had ever had a previous pregnancy that did not result in live birth and how many months the pregnancy was when it terminated. The analysis for perinatal mortality in this study was based on the birth histories and on pregnancies that terminated at 28 weeks or older. The power for the survey was calculated to detect prevalence and effect estimates of key health indices at rural/urban residence, six regions and 36 states plus the FCT. It also has precision to detect differences in the estimates of the selected health indices including PNMR at the 5% level. The main outcome measure for this study was the perinatal mortality rate. This was estimated from early neonatal deaths of births from 2003–2008; and stillbirths (pregnancies that lasted for 28 weeks or more but did not result in live birth) from 2003–2008. Early neonatal deaths and stillbirths were in turn respectively derived using the variables for year of birth ( b2) and age at death (b6 ) for early neonatal death, and year of non-viable pregnancy (v230 ) and its duration before it terminated ( v233) for stillbirth. Birth attendant was the main exposure variable. Other a priori exposure variables that are known or thought to affect perinatal mortality include the following: Demographic factors: region, residence (rural/urban), wealth index, mother’s age and mother’s education. Reproductive factors: mother’s parity, previous mortality experience, place of delivery, number of babies (singleton or multiple) length of the birth intervals and size of baby at birth. Births at PHC centers, health posts, other non-hospital public and private places, respondents’ homes and other homes were included in the main analyses. Supplementary estimation of the PNMR of hospital births (births at government hospitals and private hospitals combined) was done for the purpose of comparison with the PNMR of non-hospital births where appropriate. We reported the PNMR of hospital births only in those circumstances where the pattern of perinatal death in hospital births was different from that of non-hospital births. The dataset obtained online from Measure EvaluationR was already cleaned and recoded. Missing dates were not allowed as dates were calculated and imputed for them. Missing values, inconsistent and impossible values and “I don’t know” responses were assigned special value. Such values were identified and recoded as missing values for purpose of the current analysis. In order to answer the research question, we generated a number of new variables from existing variables and recoded some variables. The data analysis was done with StataR statistical package version 12. Descriptive and logistic analyses were used to estimate and compare the PNMR across identified demographic and reproductive characteristics. Observed differences were considered significant at the p value of <0.05; 95% confidence interval. The “gen weight” and “svyset” command functions of the stata statistical software were used to account for the complex survey features of the HDS dataset. The ethical considerations and approval for the collection of the primary data has been described [16]. Permission for the use of the data for this study was granted by Measure Evaluation®, the copyright holder of the dataset.
N/A