Introduction Nigeria contributes more obstetric, postpartum and neonatal deaths and stillbirths globally than any other country. The Clinton Health Access Initiative in partnership with the Nigerian Federal Ministry of Health and the state Governments of Kano, Katsina, and Kaduna implemented an integrated Maternal and Neonatal Health program from July 2014. Up to 90% women deliver at home in Northern Nigeria, where maternal mortality ratio and neonatal mortality rates (MMR and NMR) are high and severe challenges to improving survival exist. Methods Community-based leaders (“key informants”) reported monthly vital events. Pre-post comparisons of later (months 16–18) with conservative baseline (months 7–9) rates were used to assess change in MMR, NMR, perinatal mortality (PMR) and stillbirth. Two-tailed cross-tabulations and unadjusted and adjusted logistic regression analyses were conducted. Results Data on 147,455 births (144,641 livebirths and 4275 stillbirths) were analyzed. At endline (months 16–18), MMR declined 37% (OR 0.629, 95% CI 0.490–0.806, p ≤ 0.0003) vs. baseline 440/100,000 births (months 7–9). NMR declined 43% (OR 0.574, 95% CI 0.503–0.655, p < 0.0001 vs. baseline 15.2/1000 livebirths. Stillbirth rates declined 15% (OR 0.850, 95% CI 0.768–0.941, p = 0.0018) vs. baseline 21.1/1000 births. PMR declined 27% (OR 0.733, 95% CI 0.676–0.795, p < 0.0001) vs. baseline 36.0/1000 births. Adjusted results were similar. Discussion The findings are similar to the Cochrane Review effects of community-based interventions and indicate large survival improvements compared to much slower global and flat national trends. Key informant data have limitations, however, their limitations would have little effect on the results magnitude or significance. Keywords Maternal mortality ratio · Neonatal mortality rate · Stillbirth rate · Perinatal mortality rate · Nigeria Significance to provide immediate quality community-based lifesaving services and increased interaction with the formal health Most obstetric and neonatal mortality continues to occur system. where access to care is limited. In July 2014, an integrated Compared to baseline, there were large, highly significant Maternal and Neonatal Health program was implemented increases in women’s 37%, neonatal 43% and perinatal 27% in Northern Nigeria, where ~90% women deliver at home, survival and a 15% stillbirth reduction. The findings are bio- logically plausible and consistent with the recent Cochrane Review effects of community-based interventions and indi- cate large survival improvements compared to global and national trends. * Nancy L. Sloan New York, USA Introduction Clinton Health Access Initiative, 383 Dorcester Avenue, Suite 400, Boston, MA 02127, USA Nigeria contributes the largest number of global obstetric Clinton Health Access Initiative, 7B Ganges St. Maitama, and postpartum deaths (58,000, 19.1% of 303,000) (World Abuja, Nigeria Health Organization 2015; World Health Organization Norwegian Agency for Development Cooperation, Bygdøy 2016), neonatal deaths (261,549, 12.5% of 2,100,000) Allé 2, 0257 Oslo, Norway Vol:.(1234567890) 1 3 Maternal and Child Health Journal (2018) 22:986–997 987 (World Health Organization 2016; Liu et al. 2012) and practice) training session in January 2015. The TBA train- stillbirths (301,267, 11.6% of 2,600,000) of any coun- ing covered early identification of women’s and neonatal try (Blencowe et al. 2016). In lower and middle income danger signs, oral misoprostol for postpartum hemorrhage countries (LMIC), hemorrhage, sepsis and eclampsia (PPH) prevention with, clearing neonate’s nose and mouth, account for half of all women’s deaths and over 80% of chlorhexidine for cord care, and first-aid management of neonatal deaths and stillbirths are caused by complica- complications, including identifying PPH as postpartum tions of preterm birth, intrapartum events, and infections bleeding which soaked two wrappers (sarongs) and appli- (Liu et al. 2012). In Nigeria, the national maternal mor- cation of Non-Pneumatic Anti-Shock Garments (NASG) for tality ratio (MMR) and neonatal mortality rate (NMR) PPH, and manual neonatal resuscitator (NNR) to stabilize are 576/100,000 and 37/1,000 livebirths, respectively patients and timely emergency referral and transport to the (National Population Commission (NPC) [Nigeria] and nearest health care facility or hospital. State certified senior ICF International 2014), with a stillbirth rate of 43/1000 nurses and midwives conducted the training. All TBAs were births (Blencowe et al. 2016). The highest rates occur in linked to their nearest health care facility whose staff men- the impoverished Northwest region, where severe chal- tored them for 6 months. One focal TBA was selected with lenges to improving health care and survival exist, includ- whom eight to ten TBAs from the same geographic ward ing poor access and transportation to reach health care, would meet with on a weekly basis to practice their skills, and where up to 90% of women deliver at home (National discuss and resolve issues and to replenish their supplies. Population Commission (NPC) [Nigeria] and ICF Inter- Between January and March, 2015, certified community national 2014). health extension workers (CHEWs), midwives and nurses at primary care facilities providing antenatal care and/or labor and delivery in the program Local Government Authori- Intervention ties (LGAs), were trained in Life Saving Skills (LSS) and Essential Newborn Care (ENC), and equipped to deliver In response to the challenges, the Clinton Health Access Basic Emergency Obstetric and Newborn Care (BEmONC). Initiative (CHAI) reviewed previous efforts in Nigeria and To ensure skills breadth and transfer in the case of staff concluded that a holistic, integrated approach was needed turnover, all the antenatal and labor and delivery care staff to efficiently use scarce donor and government resources received 2 weeks training (6 days didactic, 6 days hands-on to improve women’s and neonatal health outcomes. In July practice) conducted by government certified master train- 2014, the Government of Norway provided financial support ers, mostly doctors, including obstetricians/gynecologists. to the Nigerian Federal Ministry of Health for CHAI to work In March 2015, doctors and senior nurses from secondary in partnership with the state Governments of Kano, Katsina, level facilities were trained on NASG application and in neo- and Kaduna to develop and implement an integrated Mater- natal resuscitation with mannequins, and were equipped to nal and Neonatal Health (MNH) program. Kano, Katsina, deliver Comprehensive Emergency Obstetric and Newborn and Kaduna have among the highest MMR and NMR in the Care (CEmONC). The program distributed 3800 NASGs country, accounting for approximately 20% of women’s and and 3000 NNRs to TBAs and health facilities. In May 2015, neonatal deaths in Nigeria as estimated from Demographic a 3 months mentoring program was implemented for the Health Survey (DHS) data (National Population Commis- primary and secondary level providers. In all, 1645 nurses, sion (NPC) [Nigeria] and ICF International 2014). Con- midwives and CHEWs were trained and 1450 were men- sistent with World Health Organization (WHO) guidelines tored. To ensure timely and prompt referral to the right level (World Health Organization 2013), including averting pre- of care, a functional voluntary emergency transport and ventable deaths occurring within 48 h of birth, the approach communications system was established with the National addresses critical gaps in care by linking the health system, Union of Road Transport Workers (NURTW). Two hundred from household to hospital. Because traditional birth attend- and fifty motorbike ambulances (MBAs) were provided in ants (TBAs) had little education and training, TBAs had only June-July, 2015with 1338 drivers identified and trained been allowed to refer complications. As most complications by October, 2015. Community leaders usually retained and death occur in the community, TBAs and facility-based the MBAs to facilitate community member access, TBAs Skilled Birth Attendants (SBAs) were trained, equipped and community members were trained to call the MBA and mentored to be effective first responders. TBAs 35 to or ambulance driver who would call the nearest facility to 60 years old, with at least 8 years experience and reasonable ensure readiness for patient reception or to be re-directed vision known to facility in-charges from previous commu- to another prepared facility. If re-directed, the facility focal nity engagement activities were selected. The 2791 selected person would call the higher level facility to prepare them- TBAs were formally incorporated into the health care sys- selves to receive the emergency patient. Aggregate program tem and received a 4 days (3 days didactic, 1 day hands-on data indicate ~ 34,000 referrals and ~ 20,000 transfers were 1 3 988 Maternal and Child Health Journal (2018) 22:986–997 Fig. 1 The three state MNH integrated approach to ensure a continuum of care from the community through to the hospital level made. Figures 1 and 2 summarize the project interventions. Methods The program approach was designed to provide immediate quality community-based lifesaving services and increased Study Design interaction with the formal health system, creating a con- tinuum of care, fomenting trust and demand while facilities A pre-post design was used to evaluate change in mortality become more prepared to provide obstetric, postpartum and by comparing later with baseline incidence rates or ratios. In neonatal care. The program is ongoing in 30 (of 101) of November, 2014, the MNH program revitalized, upgraded the highest mortality LGAs without other known systematic and expanded a dormant Community Based Health Manage- MNH programs in the three focal states, targeting a popula- ment Information System (CBHMIS) in which key inform- tion of approximately ten million. ants (KIs) reported monthly births, women’s and neonatal deaths and stillbirths. The KIs were mostly respected tra- ditional community leaders with basic primary or higher Objective education. In the case of death or disability, the new com- munity leader became the KI. All traditional leaders had a The objective of this assessment was to evaluate the MNH scribe with basic primary or higher education who could program impact on reducing women’s, neonatal and perina- read and write English or Hausa. As with the interventions, tal mortality and stillbirth. the CBHMIS was established to promote local ownership. 1 3 Maternal and Child Health Journal (2018) 22:986–997 989 30 LGAs Covered 326 Political Wards Pregnant Women (10 per State) Covered Births, Newborns Mentoring system: Supply system: Knowledge improvement and competency- Support large-scale roll out of low-cost, life- based training for health care workers saving devices and commodities – NASGs, NNRs, Chlorhexidine, Uterotonics etc. Mentor and equip TBAs Support states to fill gaps with stock Continuous on-the job mentoring Develop sustainable distribution systems to Availability of protocols and clear guides on ensure continuous commodity availability management of clinical emergencies Referral system: CBHMIS: Expand Emergency Transport Scheme using Re-introduce and strengthen CBHMIS volunteer taxi drivers (NURTW) Review reporting, and program successes and Strengthen referral linkages and communication failures to improve tracking & outcomes at the community level Introduce Motorbike Ambulances (MBAs) for community level use Improve facility reporting rates for robust data availability and applicability Develop state ambulance strategy Fig. 2 Implementation of the integrated three state MNH integrated approach establishing and reinforcing linkages across levels of care to the LECs on a daily to weekly basis, depending on the Procedures geographic span of the LEC coverage area. From November 2014, over 1,500 KIs reported vital A monitoring and evaluation (M&E) officer for each state, events for settlements (groups of 5–10 households) in the and various Local Engagement Consultants (LECs), who program LGAs. Data were collected with pen on paper, were communities/LGA members, mostly with previous recording a ‘1’ as a stillbirth, livebirth or neonatal death similar data collection experience and CBHMIS familiar- and the women’s age (years) at delivery or at neonatal ity, were engaged and trained by CHAI staff to oversee data death (days) in Kano and Katsina. Kaduna reported gender collection. In November, 2014, the M&E officers led their only for deaths. Because gender influences survival (Saw - state’s initial 2 days LGA level training sessions of the dis- yer 2012), the team retroactively classified missing gender trict heads, KIs and their scribes on the use of the program using neonates’ (livebirths who survived or died ≤ 28 days) tools, forms, data templates and on the reporting structure. first names. Facility-based births and deaths were reported The training included information on MMR, NMR, SBR by families to the KIs. To maintain data collection sim- and their major causes and means of prevention. As a result plicity, multiple gestation (which increases mortality risk) of a CHAI site data audit in October 2015 that identified and individual exposure to specific interventions were not data irregularities and under-reporting, LGA level review recorded, thus their effects could not be directly assessed. meetings were held monthly through July 2015 and quar- State was systematically reported. Data were collected at terly thereafter. Each M&E officer created a report of trends the lowest(settlement)level and reported up to the traditional and gaps shared by the LECs at the meetings to resolve ward, then the village, the District LGA (where data entry data quality issues. LECs also reviewed survival trends to in Excel occurred), then to the LEC, state M&E officer, and identify and resolve program implementation limitations finally to the CHAI national M&E officer where the data and share successes achieved to continue motivated and files were aggregated. Hard copy forms for out-of-range or accurate data collection. The meetings produced a dramatic inconsistent data reported between January 2015 and June rise in reported events and stable reporting from May 2015 2016 were reviewed to correct electronic records. The Excel onwards. The M&E officers provided ongoing supervision 1 3 990 Maternal and Child Health Journal (2018) 22:986–997 Table 1 Evaluation outcomes Definition Primary outcomes Maternal mortality ratio (MMR) # Women’s deaths (age 14–45)/100,000 all births Neonatal mortality rate (NMR) # Neonatal deaths (births ≥ 28 weeks estimated by last menstrual period or, if unavailable, women’s self-reported gestation with any sign of life or that cried before dying < 29 days)/1000 live- births Stillbirth rate # Stillbirths (births ≥ 28 weeks estimated by last menstrual period or, if unavailable, women’s self-reported gestation without signs of life (not breathing or no heartbeat at birth) even after attempted resuscitation)/1000 all births Secondary outcomes Perinatal mortality rate (PMR) # Stillbirths + # early neonatal deaths (age < 8 days)/1000 all births Table 2 Numbers of Births State Quarters Total LGAs (Live and Stillbirths) and LGAs by State and Quarter Q1 Q2 Q3 Q4 Q5 Q6 1–3/15 4–6/15 7–9/15 10–12/15 1–3/16 4–6/16 Kano 5702 9073 12,930 12,498 13,351 19,291 72,845 10 (of 44) Katsina 3448 6616 10,976 11,632 12,943 14,671 60,286 10 (of 34) Kaduna 5542 7686 9027 8123 9358 12,655 52,391 10 (of 23) Total 14,692 23,375 32,933 32,253 35,652 46,617 185,522 30 (of 101) spreadsheets were securely transferred and imported into similarity of Q5 and Q6 results (described below) suggests SPSS for Windows version 23 for analysis. little (e.g., parallel if any) bias. Cross-tabulations with Chi square statistics and unad- Outcomes justed and adjusted logistic regression analyses were con- ducted for all outcomes. Where inconsistent or out-of- The primary outcomes were women’s and neonatal mortality range data correction was not possible, missing values were and stillbirth; perinatal mortality was a secondary outcome assigned for analysis. The regression Odds Ratios (OR), the (see definitions in Table 1). 95% confidence intervals (CI), and two-tailed significance values (p) are presented. Neonatal mortality regression mod- Statistical Analysis els were adjusted for gender and to adjust geographic differ - ences in socio-economic status and program implementa- The evaluation (report to be posted on http://www.clint tion state. All remaining outcomes were adjusted for state onhea lthac cess.org along with an independent qualitative only. For adjusted analyses, missing gender was re-coded to evaluation conducted by KPMG) analyzed CBHMIS vital the mean value. Data were analyzed in quarterly periods to events data and reviewed information from CHAI staff and minimize variability in event reporting by settlement. All 30 program documents. The KIs were becoming familiar with program LGAs contributed to each quarter. Variability in the data reporting between January and May 2015. Half as number of monthly reporting settlements might influence the many births were reported in quarter 1 (Q1) and two-thirds effects, however data were not adjusted for clustering given in Q2 compared with Q3–Q5 (Table 2). Further training and the large number of settlements and the small number of monitoring stabilized the number of births reported from quarterly births and deaths per settlement, which nearly rep- June 2015 onwards. As under-estimation of births would resents individual-level reporting. In the absence of extreme spuriously inflate mortality rates and over-estimate pro- reporting bias, with the very large numbers reported upon, gram effect, the evaluation conservatively used quarter 3 and large and highly significant results, adjustment for such data (July–September 2015), when all the interventions were clustering would have little effect on the findings magnitude fully rolled out, as the baseline comparison group. Quarter or significance. 6 was the endline. More births were reported in Q6, with an As the vital events system reported all deaths, MMR and increase in reporting villages and their population, but the NMR analyses assume survival through 42 days postpar- tum and 28 days of life, respectively. Assuming neonates 1 3 Maternal and Child Health Journal (2018) 22:986–997 991 Merge states' databases (n=199,047) Create neonatal death and women'sdeath variables for Kaduna Identify & exclude: infants >28 days old (n=153) unknown record types (n=13) ↙ ↘ NMR, Stillbirth, PMR MMR Identify livebirths, stillbirths & Identify unlinked women'srecords, neonatal deaths & save neonatal separate and save independent women's data file (n=198015) data file (n=867) Exclude Exclude Duplicate neonatal records (n=1) Records <1/1/15 or >6/30/16 (n=9) Invalid event dates (n=5) Already linked in the neonatal records Missing event date (n=3) (n=21) Records <1/1/15 or >6/30/16 (n=12482) Reported as both a livebirth and stillbirth that hard copy review did not resolve (n=4) Reported as both a livebirth and neonatal death that hard copy review did not resolve (n=6) ↓↓ Define NMR by event age Define MMR by women'sage 14-50 (death≤28 days, survival through 30 Exclude days or 1 month) Age ≥50 years old (n=1) Excludes Missing women'sage (n=12) Missing age at death (n=29) ↓↓ N=185,522 births N=866 women'sdeaths with known age Livebirths 181,247 ≤30 days old 14-50 years old Stillbirths 4,275 N=185,509 all (live and still) births ↓ Analyses limited to ↓ quarters 3 through 6 N=147,455 births N=866 women'sdeaths with known age Livebirths 144,641 ≤30 days old 14-50 years old Stillbirths 2.814 N=147,455all (live and still) births Neonatal deaths 1,503 Fig. 3 Flow diagram and process used to merge and analyze event registry files and women who did not die survived through 28 and 42 were reported in girls less than 14 years old. Only one days, respectively, and excluding the few infant deaths women’s death > 50 years was reported; she was 60 years beyond 28 days or women’s deaths from antepartum hem- and excluded from analysis. Women (n = 12) and neonates orrhage that did not result in births may minimally under- (n = 29) with missing age were excluded from analysis. The represent the denominators but with very large numbers of timing of women’s deaths was not reported. births, should have little effect on the results magnitude or Women’s and infants’ events were almost all reported as significance. The three women aged 14 and the five aged separate records, without unique linking identie fi rs. To attain 45–50 years were included in analysis as plausible obstet- the most accurate MMR denominator, the best woman-infant ric and postpartum deaths (Fig. 3). No deliveries or deaths links were accomplished using geographic, event date and 1 3 992 Maternal and Child Health Journal (2018) 22:986–997 name fields. With imperfect matching, all (still and live) births constitute the MMR denominator; all women’s deaths are the numerator. Ethics The program and its data collection system were approved by both the federal and program states’ governments. The program evaluation is considered exempt under 45 CFR 46.101(b) from all 45 CFR and does not require IRB approval according to the Office of Human Research Pro- tection (OHRP) guidance on exemptions at http://www.hhs. Fig. 4 Maternal mortality ratios per 100,000 births by quarter gov/ohrp/polic y/index .html#exemp t. Table 3 MMR, NMR, Stillbirth (SB) and PMR by State and quarter (Q); and comparison of outcomes for Q3 vs Q6 MMR Q3 Q4 Q5 Q6 Q6 vs. Q3compari- State N MMR N MMR N MMR N MMR son Kano 12,929 340 12,495 376 13,350 157 19,289 249 0.0074 Katsina 10,975 565 11,632 318 12,943 371 14,669 416 0.0641 Kaduna 9027 443 8123 209 9358 246 12,655 158 0.0001 Total 32,931 443 32,250 313 35,651 258 46,613 277 0.0003 NMR Q3 Q4 Q5 Q6 Q6 vs. Q3compari- State N NMR N NMR N NMR N NMR son Kano 12,662 14.6 12,238 8.6 13,104 8.7 18,933 8.0 < 0.0001 Katsina 10,723 17.4 11,363 10.4 12,690 10.6 14,374 10.8 < 0.0001 Kaduna 8854 13.3 7989 8.6 9239 7.7 12,472 7.6 < 0.0001 Total 32,239 15.2 31,590 9.2 35,033 9.1 45,779 8.8 < 0.0001 Stillbirth Q3 Q4 Q5 Q6 Q6 vs. Q3compari- State N SB N SB N SB N SB son Kano 12,930 20.7 12,498 20.8 13,351 18.5 19,291 18.6 0.2934 Katsina 10,976 23.1 11,632 23.1 12,943 19.5 14,671 20.2 0.1046 Kaduna 9027 19.2 8123 16.5 9358 12.7 12,655 14.5 0.0028 Total 32,933 21.1 32,253 20.6 35,652 17.4 46,617 18.0 0.0002 PMR Q3 Q4 Q5 Q6 Q6 vs. Q3compari- State N PMR N PMR N PMR N PMR son Kano 12,930 35.0 12,498 29.2 13,351 27.0 19,291 26.4 < 0.0001 Katsina 10,976 40.1 11,632 33.3 12,943 29.9 14,671 30.8 < 0.0001 Kaduna 9027 32.2 8123 25.0 9358 20.3 12,655 22.0 < 0.0001 Total 32,933 36.0 32,253 29.6 35,652 26.3 46,617 26.6 < 0.0001 Per 100,000 births Per 1,000 livebirths Per 1,000 births 1 3 Maternal and Child Health Journal (2018) 22:986–997 993 Table 4 Percent distribution of State Kano Katsina Kaduna Total Neonatal Gender by State Male Female Male Female Male Female Male Female N 38,133 34,702 29,418 30,863 18,099 34,041 85,650 99,606 % 52.4 47.6 48.8 51.2 34.7 65.3 46.2 53.8 Newborns with missing gender were excluded from analysis 185,522 births (181,247 livebirths and 4275 stillbirths) Findings from Q1 to Q6 were analyzed. As women’s deaths with unknown age (n = 12) or age > 50 (n = 1) were excluded, The CBHMIS reported more annual vital events than 185,509 records (including livebirths and stillbirths) were most and even multi-country household surveillance sys- analyzed for MMR (Fig. 3). The analyses of quarters 3 tems (Kirkwood et al. 2010; Goudar et al. 2012). Data on Table 5 Unadjusted and Adjusted Logistic Regressions of MMR, NMR, Stillbirth, PMR Comparing Quarters 4, 5 and 6 with Q3 OR 95% Lower CI 95% Upper CI p OR 95% Lower CI 95% Upper CI p MMR (N = 147,455) Unadjusted Adjusted Q 3 (baseline) Reference group Reference group Q 4 0.718 0.551 0.937 0.0147 0.709 0.543 0.925 0.0112 Q 5 0.606 0.462 0.794 0.0003 0.597 0.455 0.783 0.0002 Q 6 0.629 0.490 0.806 0.0003 0.634 0.494 0.812 0.0003 Katsina – – – – 1.492 1.203 1.850 0.0003 Kaduna – – – – 0.916 0.706 1.188 0.5068 NMR (N = 144,641) Unadjusted Adjusted Q 3 (baseline) Reference group Reference group Q 4 0.605 0.523 0.699 < 0.0001 0.601 0.519 0.695 < 0.0001 Q 5 0.595 0.517 0.686 < 0.0001 0.592 0.513 0.682 < 0.0001 Q 6 0.574 0.503 0.655 < 0.0001 0.576 0.505 0.658 < 0.0001 Gender – – – – 0.786 0.709 0.871 < 0.0001 Katsina – – – – 1.253 1.115 1.408 0.0001 Kaduna – – – – 0.974 0.851 1.115 0.7011 Stillbirth (N = 147,455) Unadjusted Adjusted Q 3 (baseline) Reference group Reference group Q 4 0.975 0.876 1.086 0.6440 0.968 0.869 1.078 0.5509 Q 5 0.821 0.736 0.916 0.0004 0.816 0.732 0.911 0.0003 Q 6 0.850 0.768 0.941 0.0018 0.851 0.769 0.942 0.0019 Katsina – – – – 1.096 1.007 1.192 0.0340 Kaduna – – – – 0.794 0.719 0.877 < 0.0001 PMR (N = 147,455) Unadjusted Adjusted Q 3 (baseline) Reference group Reference group Q 4 0.818 0.750 0.892 < 0.0001 0.812 0.745 0.886 < 0.0001 Q 5 0.725 0.664 0.790 < 0.0001 0.720 0.660 0.786 < 0.0001 Q 6 0.733 0.676 0.795 < 0.0001 0.734 0.677 0.796 < 0.0001 Katsina – – – – 1.146 1.07 1.227 < 0.0001 Kaduna – – – – 0.839 0.774 0.909 < 0.0001 Kano is the reference state Male is the reference gender 1 3 994 Maternal and Child Health Journal (2018) 22:986–997 through 6 include 147,455 births (144,641 livebirths and 0.503–0.655, p < 0.0001, Table 4), with declines of 39% and 2814 stillbirths). 40% in Q4 (OR 0.605, 95% CI 0.523–0.699, p < 0.0001) and Between Q3 (July–September 2015) and Q6 (April–June Q5 (OR 0.595, 95% CI 0.517–0.686, p < 0.0001). Adjust- 2016) 1291 women’s deaths were reported. Women’s mor- ment for gender and state had little effect. tality declined across all three states (Table 3; Fig. 4). Stillbirth rates declined between Q3 and Q6 (Table 3; Kano and Kaduna Q6 MMR had significantly lower MMR Fig. 5), but were only statistically significant at the state (p = 0.0074 and p = 0.0001, respectively) and Katsina’s Q6 level in Kaduna. A 15% decline in stillbirth (OR 0.850, MMR was substantially lower (p = 0.0641, marginally sig- 95% CI 0.768–0.941, p = 0.0018) was observed between nificant) than their Q3 baselines. A highly significant 37% Q3 and Q6 and an 18% decline in Q5 (OR 0.821, 95% CI decline in Q6 compared with the Q3 MMR of 440/100,000 0.736–0.916, p ≤ 0.0004), with little change between Q3 and births was observed, (OR 0.629, 95% CI 0.490–0.806, Q4 (OR 0.975, 95% CI 0.876–1.086, p = 0.6440), Table 4. p ≤ 0.0003, Table 4), with a 28% Q4 decline (OR 0.718, Adjustment for state had little effect. Progress to reduce 95% CI 0.551–0.937, p = 0.0147) and a 39% Q5 decline LMIC stillbirths has been slow (Sather et al. 2010; Little (OR 0.629, 95% CI 0.490–0.806, p ≤ 0.0003). Adjustment et al. 2010; Goldenberg et al. 2016). Most neonatal mortality for state had little effect. occurs within the first few days of life (Sankar et al. 2016), Over all quarters (Q1–Q6) between January 2015 and when NMR occurs disproportionately in very ill neonates. June 2016, 185,522 births were reported, 72,845 in Kano, A stillbirth to NMR ratio of 0.90:1.00 is expected from high 60,286 in Katsina and 52,391 in Kaduna (Table 2); 46.2% quality data, with a ratio of 0.75:100 in poor quality data were classified as male and 53.8% female (Table 5). The (Blencowe et al. 2016). Mis-reporting neonatal deaths on male-to-female birth ratio in Kano was 1.10:1 and 0.95:1.00 the day of birth as stillbirths is a common problem world- in Katsina, but was 0.53:1.00 in Kaduna and likely reflect wide (Blencowe et al. 2016; Lawn et al. 2010). The observed misclassification where most neonatal gender was assigned ratio of stillbirth to neonatal mortality was about 2.00:1.00, retroactively based on the infants’ first names. indicating that many neonatal deaths were mis-reported as Of the 183, 473 livebirths reported between Q1 and Q6, stillbirths and explains why the program had smaller still- the average and standard deviation, and the median age at birth than NMR and PMR benefit. death in days for neonates were 6.41 ± 6.29 and 5.00 (with PMR declined (p < 0.0001) between July 2015 and June an interquartile range [IQR] 8.00) days (n = 1503). Neonatal 2016 (Table 3; Fig. 5). A 27% PMR decline (OR 0.733, 95% age at death was fairly stable over time (data not shown), CI 0.676–0.795, p < 0.0001) was observed between Q3 and although fewer died in the first 2 days of life in Q5 and Q6. Q6, and a 23% Q5 decline (OR 0.725, 95% CI 0.664–0.790, Thirty-six percent of neonates’ deaths were age ≤ 1 day in p < 0.0001, Table 4). As PMR reflects stillbirth and early Q3, 42% in Q4, 32% in Q5 and 30% in Q6. Age at death was NMR, an intermediate 18% PMR change was observed 1 day greater in Q5 and Q6 than Q3 and Q4. between Q3 and Q4 (OR 0.818, 95% CI 0.750–0.892, NMR declined sharply (p ≤ 0.001) between Q3 (July–Sep- p < 0.0001). Adjustment for state had little influence. The tember 2015) and Q6 (April–June 2016; Table 3; Fig. 5). baseline PMR is almost identical (36/1000) to the most Compared with the Q3 baseline rate of 15.2/1,000 livebirths, recent DHS for the north-central Nigeria (34/1000), sup- a 43% decline in at Q6 was observed (OR 0.574, 95% CI porting the validity of the program PMR estimates (National Population Commission (NPC) [Nigeria] and ICF Interna- tional 2014). Discussion Efforts to improve survival in Nigeria have had poor to mixed results. The three state Nigeria MNH program find- ings suggest that the comprehensive, integrated program accomplished substantial, sustained and highly significant reductions in women’s mortality (37%), neonatal mortality (43%), stillbirth (15%) and perinatal mortality (27%). Much of the observed program effect may be attributable to the implementation of simple, highly effective interven- tions rooted in community-based training and involvement. Community-based neonatal stimulation without bag and Fig. 5 Neonatal mortality (Per 1000 livebirths), stillbirth (Per 1000 mask ventilation reduced stillbirth by 24% in rural India births) and perinatal (Per 1000 births) mortality rates by quarter 1 3 Maternal and Child Health Journal (2018) 22:986–997 995 (Goudar et al. 2013). Sound rural LMIC studies found oral the ratio of reproductive age women’s deaths to livebirths) misoprostol reduced postpartum hemorrhage by 36% and between 2006 and 2013 was 576 with a 95% CI of 500–652 NASG reduced PPH case-fatality by over 50% (Sloan et al. per 100,000 livebirths. Although the evaluation’s MMR 2010; Miller and Belizán 2015). In rural northern India, a denominator includes both livebirths and stillbirths, the Q1 comprehensive community-based ENC package reduced and Q2 MMRs were considerably higher, 1310/100,000 and NMR by 54%, attributing much of the effect to skin-to- 880/100,000 births (live and still). Consistent with the evalu- skin care, immediate breastfeeding and first postnatal bath- ation’s larger MMR denominator (including stillbirths), the ing ≥ 24 h after delivery (Kumar et al. 2008). Community- Q3 evaluation MMR estimate of 440/100,000 was below based application of chlorhexidine for umbilical cord care the lower DHS MMR 2006–2013 CI, and the MMR was has been shown to reduce NMR by 19% (Sinha et al. 2015). similar to the lower CI of the 2001–2008 DHS estimates The observed effects are quite similar to the results of 475/100,000 livebirths (National Population Commis- reported by the most recent Cochrane Review of commu- sion (NPC) [Nigeria] and ICF International 2014). Unlike nity-based interventions, supporting their consistency and NMR, however, DHS samples do not provide accurate point biologic plausibility (Lassi and Bhutta 2015). That review and interval estimates for MMR. Using reports on deaths found community-based interventions in various countries of women of reproductive age, however, contributes to the increased the use of simple, effective interventions, particu- strength of the evaluation as a funnel for identifying suspect larly early initiation of breastfeeding, TBA/SBA clean deliv- obstetric and postpartum deaths. With the very large sample, ery kit use and emergency transfers, as did the integrated the 95% confidence intervals indicate reliable and, except for program approach. The Cochrane Review found community stillbirth, strong effect boundaries. mobilization and ante- and post-natal visitation and home- Key informant data collection has limitations and may based treatment packages also conducted under challeng- not ensure data quality equivalent to professional prospec- ing circumstances generally reduced NMR by 37–40% with tive surveillance or periodic household surveys (Silva et al. 95% CIs from a 20% increase to a 70% NMR decrease. The 2016; Silva 2012; Joos et al. 2016). Similar to the program’s review found similar interventions reduced MMR by an KI system, facility-based integrated programs using qual- average 26–28% with 95% CIs covering a 20% increase to ity improvement systems, to provide timely information for a 50% MMR decrease. Over all community-based interven- program decisions in Niger and Mali increased attention to tions, the review found a 20% stillbirth reduction covering monitor and manage all pregnancies and thus contributed a 10% increase to a 40% decrease, and a 22% reduction in to improving survival (Boucar et al. 2014). Professionally PMR covering a 20% increase to a 45% decrease. When conducted household surveillance was initially consid- limited to well-controlled studies with low risk of bias, ered but sufficient funding was unavailable. Future profes- the average reductions were 30% NMR, 24% MMR, 25% sional assessment with household mapping that efficiently stillbirth and 27% PMR. In contrast, a recently published expands the breadth of data (Christian et al. 2013; D’souza large multi-country randomized controlled trial of a similar 1981) to evaluate receipt and use of the interventions is integrated approach to improve BEmONC and CEmONC, recommended. including SBA and TBA training, community mobiliza- The evaluation had no contemporaneous control group to tion and enhanced referral not yet included in the Cochrane help attribute survival improvement to the program (Miller reviews had little or no effect on MMR, NMR, stillbirth or et al. 2003), and to avoid unexpected potential harm. To PMR (Pasha et al. 2013). Even in that trial, diverse condi- avoid contamination, LGAs where other organizations were tions and implementation across sites may account for the conducting MNH programs were excluded from the MNH large NMR and PMR reductions observed in rural India that program. Temporal change unassociated with the program were not observed elsewhere (Goudar et al. 2015). could decrease or increase the observed effects. Nigeria has To avoid over-estimation of effect and to reflect when all experienced stable or increased neonatal and women’s mor- the interventions were fully rolled out, using Q3 as the base- tality over the past years (National Population Commission line may under-estimate the program effect as most interven- (NPC) [Nigeria] and ICF International 2014), whereas the tions were initiated in January 2015 before Q3. The most program area experienced substantial and significant mortal- recent DHS reports an NMR of 44/1000 livebirths in North- ity reductions. In the absence of extreme bias or activities west Nigeria between 2004 and 2013 and a national rate unknown to the program staff, the effects’ magnitude and of 37/1000 livebirths in 2013 (National Population Com- significance is likely attributable to the program. UNICEF mission (NPC) [Nigeria] and ICF International 2014). Even estimated that NMR reduced by 47% between 1990 and 2015 if Q1 and Q2 births were under-reported, their NMRs of (UNICEF Global databases 2016) and WHO and the United 22.6/1000 and 18.0/1000 livebirths were substantially lower, Nations Maternal Mortality Estimation Inter-Agency Group indicating that the program had some additional earlier estimates MMR reduced by 44% in that same time period effect. In contrast, the DHS maternal mortality ratio (e.g., (World Health Organization 2015; Alkema et al. 2016). The 1 3 996 Maternal and Child Health Journal (2018) 22:986–997 three state MNH Program in Nigeria observed and sustained References similar results in 12 months between July 2015 and June Alkema, L., Chou, D., & Hogan, D., et al. (2016). United Nations 2016. The capacity building of government staff instituted Maternal Mortality Estimation Inter-Agency Group collabora- at the beginning of the program supported a gradual transi- tors and technical advisory group. Global, regional, and national tion of program responsibilities. The interventions’ effects levels and trends in maternal mortality between 1990 and 2015, peaked in Q4, 9 months after the program was initiated. with scenario-based projections to 2030: A systematic analysis by the UN Maternal Mortality Estimation Inter-Agency Group. CBHMIS data not available at the time of this evaluation Lancet 387(10017), 462–474. https ://doi.or g/10.1016/S0140 indicates continued survival improvement after a period of -6736(15)00838 -7. stabilization between Q5 and Q6, when program responsi- Blencowe, H., Cousens, S., & Jassir, F. B., et al. (2016). National, bilities were successfully transitioned to the government. regional, and worldwide estimates of stillbirth rates in 2015, with trends from 2000: a systematic analysis. The Lancet In conclusion, compared to much slower global and flat Global Health, 4(2), e98–e108. https ://doi.org/10.1016/S2214 national trends, the evaluation results indicate large survival -109X(15)00275 -2. improvements are most likely attributable to the MNH pro- Boucar, M., Hill, K., Coly, A., et al. USAID Applying Science to gram and should be scaled up. In a resource constrained set- Strengthen Improve Systems University Research Co., LLC (USAID ASSIST-URC). (2014). Improving postpartum care for ting with high obstetric, postpartum, perinatal and newborn mothers and newborns in Niger and Mali: A case study of an inte- mortality, and previously disjointed MNH care, a coherent, grated maternal and newborn improvement programme. BJOG, integrated approach simultaneously addressing community, 121(Suppl 4), 127–133. primary health care and CEmONC care can greatly improve Christian, P., Klemm, R., Shamim, A. A., et al. (2013). Effects of vita- min A and β-carotene supplementation on birth size and length women’s and neonatal survival. of gestation in rural Bangladesh: A cluster-randomized trial. The American Journal of Clinical Nutrition, 97(1), 188–194. https :// Acknowledgements We thank the women and families in the MNH doi.org/10.3945/ajcn.112.04227 5. program area, the TBAs, SBAs and institutional care providers; those D’souza, S. (1981). A population laboratory for studying disease pro- providing emergency transportation, the community-based leaders and cesses and mortality—the Demographic Surveillance System, data collectors, and local and international Clinton Health Access Ini- Matlab Comilla, Bangladesh. Rural Demography, 8(1), 29–51. tiative staff who facilitated program implementation and monitoring. Goldenberg, R. L., Saleem, S., Pasha, O., Harrison, M. S., & Mcclure, We thank the State Ministries of Health and the Governors of Kaduna, E. M. (2016). Reducing stillbirths in low-income countries. Acta Kano and Katsina state for the support and collaboration during pro- Obstetricia et Gynecologica Scandinavica, 95(2), 135–143. https gram implementation, the Nigerian Federal Ministry of Health program ://doi.org/10.1111/aogs.12817. for providing guidance and support. We thank the Government of Nor- Goudar, S. S., Carlo, W. A., McClure, E. M., et al. (2012). The Mater- way for project support and guidance. We thank Cynthia Stanton for nal and Newborn Health Registry Study of the Global Network for her draft manuscript review and comments. Women’s and Children’s Health Research. Int J Gynaecol Obstet, 118(3), 190–193. https ://doi.org/10.1016/j.ijgo.2012.04.022. Author Contributions AS, OF, KM, OMW, LL and MG contributed Goudar, S. S., Derman, R. J., Honnungar, N. V., Patil, K. P., Swamy, to design of the original program and data collection and management M. K., Moore, J., Wallace, D. D., McClure, E. M., Kodkany, B. systems. AS, OF, JY were responsible for study implementation and S., Pasha, O., Sloan, N. L., Wright, L. L., & Goldenberg, R. L. data collection. NLS designed and conducted the data analysis. NLS, (2015). An intervention to enhance obstetric and newborn care AS, OF, OMW and KM drafted the manuscript. All authors contributed in India: A cluster randomized-trial. Maternal and Child Health to the interpretation of results. Journal, 19(12), 2698–2706. https ://doi.or g/10.1007/s1099 5-015-1792-0. Funding The Norwegian Agency for Development Cooperation Goudar, S. S., Somannavar, M. S., & Clark, R., et al. (2013). Still- (Norad) supported the Nigerian Federal Ministry of Health and CHAI birth and newborn mortality in India after helping babies breathe to implement and evaluate the program. Norad participated in the pro- training. Pediatrics, 131(2), e344–e352. https ://doi.org/10.1542/ gram design and co-authors from the organization contributed to the peds.2012-2112. manuscript composition and results interpretation. Joos, O., Amouzou, A., Silva, R., et al. (2016). Strengthening com- munity-based vital events reporting for real-time monitoring of under-five mortality: Lessons learned from the Balaka and Salima Compliance with Ethical Standards Districts in Malawi. PLoS ONE, 11(1), e0138406. https ://doi. org/10.1371/journ al.pone.01384 06. Conflict of interest Dr. Sloan was engaged to conduct an independ- Kirkwood, B. R., Hurt, L., Amenga-Etego, S., et al. (2010). Effect ent evaluation of the MNH Nigeria program. Authors Storey, Fasawe, of vitamin A supplementation in women of reproductive age on Yakubu, Wiwa and McCrystal are CHAI staff. Lene Jeanette Lothe and maternal survival in Ghana (ObaapaVitA): A cluster-randomised, Mari Grepstad are Norad staff. All authors have interpreted the study placebo-controlled trial. Lancet, 375(9726), 1640–1649. findings with objectivity and report no conflicts of interest. Kumar, V., Mohanty, S., Kumar, A., et al. (2008). Effect of community- based behaviour change management on neonatal mortality in Open Access This article is distributed under the terms of the Crea- Shivgarh, Uttar Pradesh, India: A cluster-randomised controlled tive Commons Attribution 4.0 International License (http://creat iveco trial. Lancet, 372(9644), 1151–1162. https ://doi.or g/10.1016/ mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- S0140 -6736(08)61483 -X. tion, and reproduction in any medium, provided you give appropriate Lassi, Z. S., & Bhutta, Z. A. (2015). Community-based intervention credit to the original author(s) and the source, provide a link to the packages for reducing maternal and neonatal morbidity and mor- Creative Commons license, and indicate if changes were made. tality and improving neonatal outcomes. Cochrane Database of 1 3 Maternal and Child Health Journal (2018) 22:986–997 997 Systematic Reviews, Issue 3. Art. No.: CD007754. https ://doi. stillbirth (5 of 7): Advocacy barriers and opportunities. BMC org/10.1002/14651 858.CD007 754.pub3. Pregnancy and Childbirth, 10(Suppl 1), S5. https ://doi. Lawn, J. E., Gravett, M. G., Nunes, T. M., Rubens, C. E., & Stan- org/10.1186/1471-2393-10-S1-S5. ton, C., GAPPS Review Group. (2010). Global report on preterm Sawyer, C. C. (2012). Child mortality estimation: estimating sex dif- birth and stillbirth (1 of 7): Definitions, description of the burden ferences in childhood mortality since the 1970s. PLoS Medicine, and opportunities to improve data. BMC Pregnancy Childbirth, 9(8), e1001287. https ://doi.org/10.1371/journ al.pmed.10012 87. 10(Suppl 1), S1. https ://doi.org/10.1186/1471-2393-10-S1-S1. Silva, R. (2012). Child mortality estimation: Consistency of under- Little, G., Niermeyer, S., Singhal, N., Lawn, J., & Keenan, W. (2010). five mortality rate estimates using full birth histories and sum- Neonatal resuscitation: A global challenge. Pediatrics, 126(5), mary birth histories. PLOS Medicine, 9(8), e1001296. https://doi. e1259–e1260.org/10.1371/journ al.pmed.10012 96. Liu, L., Johnson, H. L., & Cousens, S. et al. (2012). Child Health Silva, R., Amouzou, A., & Munos, M. et al. (2016). Can community Epidemiology Reference Group of WHO and UNICEF. Global, health workers report accurately on births and deaths? Results regional and national causes of child mortality: An updated sys- of field assessments in Ethiopia, Malawi and Mali. PLoS ONE, tematic analysis. Lancet, 379(9832), 2151–2161. 11(1), e0144662. https ://doi.org/10.1371/journ al.pone.01446 62. Miller, S., & Belizán, J. M. (2015). A promising device to save mater- Sinha, A., Sazawal, S., Pradhan, A., Ramji, S., & Opiyo, N. (2015). nal lives associated with obstetric hemorrhage: The non-pneu- Chlorhexidine skin or cord care for prevention of mortality and matic anti-shock garment (NASG). Reproductive Health, 12, 26. infections in neonates. Cochrane Database System Review, 5(3), https ://doi.org/10.1186/s1297 8-015-0019-6. CD007835. https ://doi.org/10.1002/14651 858.CD007 835.pub2. Miller, S., Sloan, N. L., Winikoff, B., Langer, A., & Fikree, F. F. Sloan, N. L., Durocher, J., Aldrich, T., Blum, J., & Winikoff, B. (2010). (2003). Where is the “E” in MCH? The need for an evidence- What measured blood loss tells us about postpartum bleeding: A based approach in safe motherhood. Journal of Midwifery and systematic review. BJOG, 117(7), 788–800. https://doi.or g/10.11 Women’s Health, 48(1), 10–18.11/j.1471-0528.2010.02567 .x. National Population Commission (NPC) [Nigeria] and ICF Inter- UNICEF Global databases. http://data.unicef.or g. Accessed November national. (2014). Nigeria Demographic and Health Survey 6, 2016. 2013. Abuja, Nigeria, and Rockville, Maryland: NPC and ICF World Health Organization. (2013). WHO recommendations on post- International. natal care of the mother and newborn. Geneva: WHO. Pasha, O., McClure, E. M., & Wright, L. L. et al. (2013). EMONC World Health Organization. (2015). Trends in maternal mortality: trial investigators. A combined community- and facility-based 1990 to 2015 Estimates by WHO, UNICEF, UNFPA, Geneva: approach to improve pregnancy outcomes in low-resource set- The World Bank and the United Nations Population Division. tings: A global network cluster randomized trial. BMC Medicine, World Health Organization. Nigeria: Neonatal and child health profile. 11, 215. https ://doi.org/10.1186/1741-7015-11-215. Accessed October 25, 2016, from http://www.who.int/mater nal_ Sankar, M. J., Natarajan, C. K., Das, R. R., Agarwal, R., Chan-child _adole scent /epide miolo gy/profi les/neona tal_child /nga.pdf. drasekaran, A., & Paul, V. K. (2016). When do newborns die? A World Health Organization. Nigeria: Maternal and perinatal health pro- systematic review of timing of overall and cause-specific neonatal file. Accessed October 25, 2016, from http://www.who.int/mater deaths in developing countries. Journal of Perinatology, 36(Suppl nal_child _adole scent /epide miolo gy/profi les/mater nal/nga.pdf. 1), S1–S11. https ://doi.org/10.1038/jp.2016.27. Sather, M., Fajon, A. V., Zaentz, R., & Rubens, C. E., GAPPS Review Group. (2010). Global report on preterm birth and 1 3
Maternal and Child Health Journal – Springer Journals
Published: Feb 9, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera