Delivery outcomes and patterns of morbidity and mortality for neonatal admissions in five Kenyan hospitals

A cross-sectional survey was conducted in neonatal and maternity units of five Kenyan district public hospitals. Data for 1 year were obtained: 3999 maternal and 1836 neonatal records plus tallies of maternal deaths, deliveries and stillbirths. There were 40 maternal deaths [maternal mortality ratio: 276 per 100 000 live births, 95% confidence interval (CI): 197–376]. Fresh stillbirths ranged from 11 to 43 per 1000 births. A fifth (19%, 263 of 1384, 95% CI: 11–30%) of the admitted neonates died. Compared with normal birth weight, odds of death were significantly higher in all of the low birth weight (LBW, <2500 g) categories, with the highest odds for the extremely LBW (<1000 g) category (odds ratio: 59, 95% CI: 21–158, p < 0.01). The observed maternal mortality, stillbirths and neonatal mortality call for implementation of the continuum of care approach to intervention delivery with particular emphasis on LBW babies.


I N T R O D U C T I O N
Global reports indicate that the highest risk of neonatal death is in Sub-Saharan Africa, with Kenya among the 10 countries contributing most deaths [1]. With high coverage of basic interventions [2], up to 71% of neonatal deaths could be averted with >82% of this effect attributable to facility-based care [3]. However, national, and thus global, reports are based on limited data on neonatal case-mix and outcomes; the available data are largely derived from episodic, limited-scale surveys [4]. Further, the national hospital information management system (HMIS) in Kenya has been shown to have poorquality data [5]. There are particularly few data exploring possible variability in neonatal case-mix V C The Author [2015]. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/ 4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. and outcomes. This study, therefore, sought to use data (collected specifically for the study separate from the national HMIS) from five Kenyan hospitals, from a continuum of care perspective [6,7], to (i) profile maternal characteristics, (ii) determine delivery outcomes (still births and maternal mortality), (iii) document the causes of neonatal admissions and (iv) examine the effect of birth weight on neonatal mortality.

M E T H O D S
This was a cross-sectional survey conducted in neonatal and maternity units of five Kenyan urban public hospitals in November and December 2013. Data were abstracted retrospectively from admission registers covering a 1 year period (October 2012-September 2013). These included data from all inborn neonatal admissions plus maternal data abstracted from a sample of 800 deliveries per hospital. In addition, for stillbirths, live births and maternal mortality, a tally of the total number of events over the 1 year period was obtained. The data were entered directly into REDCap V R electronic data capture tools. Data quality was checked in real time by checks built into REDCap. In addition, at the end of each day, these data would be transmitted to a central server in KEMRI-Wellcome Trust where a STATA version 12 (Stata Corporation, Texas, USA) check file was run and a list of potential errors generated and sent back to the sites for verification and correction. Analyses were also done in STATA version 12. Pooled results are presented with 95% confidence intervals (CI), while the association between birth weight and neonatal mortality was examined using a random effects logistic regression model. The effect of clustering at hospital level was taken into account in these analyses.

D I S C U S S I O N
The MMR of 276 per 100 000 live births in these hospitals compares with a recent population level estimate for Kenya of 277.2 per 100 000 (95% CI: 175.4-414.1) in contrast with 12.1 of 100 000 (95% CI: 10.4-13.7) in developed countries [8]. A higher risk population is expected to deliver in hospitals, which may account for the recorded high proportion (20%) of teenage (13-19-year-old) mothers, who are known to have a higher risk of adverse neonatal and maternal outcomes [9][10][11]. Stillbirths have remained largely invisible; there were an estimated 2.65 million third-trimester stillbirths globally in 2008, 98% of which occurred in low-and middleincome countries [12,13]. In the hospitals studied, there were 37 stillbirths per 1000 births with a rate of 20 per 1000 fresh stillbirths, for whom death is likely to have occurred intrapartum [14]. However, these fresh stillbirths may include early neonatal deaths, as misclassification between fresh stillbirths and early neonatal deaths is a challenge when enumerating still births [15].
The most common disease episodes were birth asphyxia (31%), prematurity/LBW (29%) and neonatal sepsis (14%) ( Table 1). The number of patients with multiple diagnoses is not reported but previous work demonstrated considerable overlap in these three diagnoses [4]. In addition to many fresh stillbirths, the high numbers of birth asphyxia cases are of concern, contributing to a mortality of 10% in newborn units for normal weight admissions. Small babies remain a vulnerable population; a 2012 estimate suggested that 80% of neonatal deaths in Sub-Saharan Africa and South Asia were of this group [16]. We have shown that, even in hospitals where care should be available, 68% of neonatal deaths are <2500 g. Perhaps even more important for planning long-term service development, 41% of deaths were of birth weight <1500 g.

C O N C L U S I O N
The burden of maternal mortality, fresh stillbirths and birth asphyxia in these facilities suggests significant opportunity to earn the triple return on investment offered by improving referral and the quality of perinatal care [3]. In addition, given the disproportionately poor outcomes among LBW babies, enhancement of capacity to offer care for this vulnerable group is required.

F U N D I N G
This work was supported by the Kenyan Ministry of Health, The Wellcome Trust, The Consortium for National Health Research (Kenya) and The University of Nairobi. JA s contribution was made possible by the Kenyan Ministry of Health and a grant from the Consortium for National Health Research (Kenya) to the SIRCLE Collaboration. ME has been supported by funds from The Wellcome Trust (076827, #097170). Additional funds from a Wellcome Trust Strategic Award (#084538) and a Wellcome Trust core grant awarded to the KEMRI-Wellcome Trust Research Programme (#092654) made this work possible. These grants supplemented salary support from the Ministry of Health (Kenya, to RN) and the University of Nairobi (to AW and FW). The Wellcome Trust and other funders had no role in developing this manuscript nor in the decision to submit for publication.

A C K N O W L E D G E M E N T S
Permission to conduct the study in government hospitals was provided by the Director of Medical Services in the Ministry of Health. We would in addition like to acknowledge the hospitals' medical superintendents for providing access and the research assistants and hospital staff who facilitated data collection. This work is published with the permission of the Director of KEMRI.