Strengthening the Health System in Ethiopia Through Maternal Death Surveillance and Response

Session: Audit Systems for Maternal Newborn Health: Informing and Improving Practice

Presenter: Ephrem Tekle Lemango, Federal Ministry of Health, Ethiopia
Download presentation slides

Background: Ethiopia’s National Maternal Death Surveillance and Response (MDSR) system was introduced in 2013 and aims to create an ongoing cycle of data collection, analysis, and action to improve the quality of MNCH services across the country. To further accelerate efforts to reduce preventable maternal deaths, MDSR has been integrated into national public health emergency management, with mandatory weekly notification by frontline surveillance officers, followed by collection, analysis, and reporting of verbal autopsy data upwards through each level of the health system.

Methodology: To explore facilitators and barriers to the establishment of a fully functioning MDSR in different contexts, six case studies were conducted in four agrarian and two urban zones where the MDSR system was first implemented. Semi-structured interviews and focus group discussions were conducted with over 50 key stakeholders at national, regional, zonal and district level to document MDSR introduction and development, with a focus on stories of change to assess effects on the health system beyond the generation of data on maternal deaths.

Results: Several pathways through which the process of MDSR has contributed to health system strengthening were identified, including raising awareness of maternal deaths at district level leading to improved infrastructure; increasing communication between hospitals and their catchment health centres resulting in more efficient referral mechanisms; strengthening relationships between health institutions and the communities they serve; and enhancing skills within facilities to recognise and act on obstetric emergencies.

Conclusion: The national MDSR in Ethiopia is catalysing positive changes at different levels of the health system as it gains strength and coverage, even before full implementation is achieved. Lessons learned from the Ethiopian experience could help galvanise nascent efforts to establish MDSR in other settings, particularly those experiencing significant challenges in the early phases, by demonstrating the immediate and medium term benefits.

Dead Women Talking: Community Led Social Autopsies of Maternal Deaths

Session: Audit Systems for Maternal Newborn Health: Informing and Improving Practice

Presenter: Subha Sri B, CommonHealth
Download presentation slides

Background: Existing methods for measurement of maternal death in India give only aggregate maternal mortality ratios and do not provide information on who the women dying are or the causes and contributors in the pathways to their death. Verbal autopsies of maternal deaths by the government are infrequent and largely focus only on medical cause of death.

Methodology: Over two years, in a collaborative process called Dead Women Talking, several civil society groups in India documented maternal deaths through the use of a specially developed social autopsy tool. Community representatives documented the lived experiences of deceased women and their families to identify causes and contributors to deaths that went beyond the traditional bio-medical paradigm. An inductively developed framework called SSSR framework analyzed the different domains contributing to maternal deaths including science, health systems, and social issues and from a rights perspective.

Results: 124 maternal deaths were documented across ten states of India over two years. The documentations highlighted that the deaths happened in women with multiple vulnerabilities including very young women, those belonging to marginalized caste groups, poor women, migrant women. This provides crucial information on who the women dying are. Use of the social autopsy tool and the SSSR framework also revealed the complex pathways that led to women’s deaths highlighting both health systems issues like lack of emergency obstetric care and poor referral services, and social determinants including gender inequities as contributors. This has implications for interventions required to reduce maternal mortality. The analyzed information was used for local level action by communities themselves in addition to contributing to state and national level advocacy to demand more accountability for maternal health.

Conclusion: Community led social autopsies of maternal deaths can broaden the understanding of contributors to maternal deaths and lead to more accountability for maternal health.

Innovative Software Solution to Fast Track Maternal Death Review (MDR) – Lessons from India

Session: Audit Systems for Maternal Newborn Health: Informing and Improving Practice

Presenter: Dhawal Naik, Avni Health Foundation
Download presentation slides

In India, maternal death review provides detailed information on various factors at community and facility levels that influence maternal health outcomes. A key challenge is to analyse the large volume of paper based data collected through the tools. Further challenges include the expansion of the database of records of maternal deaths on a daily basis, data elements are repeated across tools, tracking pregnant women who move between mothers and husbands houses during and after delivery, gaps in linking MDR with an existing online health management information database makes tracking, reporting, and timely analysis of data difficult to inform policy, program implementation and monitoring towards results. A national MDR software to address the above-mentioned concerns and strengthening data management systems was developed, deployed and scaled up in India. Till date over 2100 health personnel have been trained across ten states and 2200 maternal death records are available in the system. One set of data analysis shows that majority of deaths occurred in women aged between 18-25 years, with at least 2-3 antenatal visits and no postnatal checkups, with causes linked with postpartum hemorrhage, eclampsia, anemia, sepsis, ruptured uterus. Inability of the women to recognize the danger signs, coupled with delay in seeking care, arranging for transport, initiating treatment at the facility, and lack of blood contributed to the maternal death. The timely information available at all districts and blocks has enabled health teams to better plan and act on solutions to improve maternal health outcomes. Key challenges, lessons learnt, and demonstration of the use of the software to track maternal deaths, understanding related causes, gaps in services delivery etc. will be highlighted. The lessons learnt from India’s experience has implications for other countries in using innovative technologies to address large scale public health issues towards improving maternal health outcomes.

Availability of Maternal Newborn and Child Health Quality of Care Indicators in Faith Based Hospitals in Africa and Asia

Session: Quality of Facility-Based Maternal and Newborn Care: Development and Application of Improved Metrics, Tools and Analyses

Presenter: Matthews Mathai, World Health Organization
Download presentation slides

Background: Experts at a WHO technical consultation in 2013 proposed 19 core indicators for reporting on quality of maternal, neonatal and paediatric care in health facilities. Initial reports from countries indicate that not all proposed indicators are currently being reported from health facilities. Faith-based hospitals play a major role in health service delivery in Africa and Asia. In collaboration with a network of faith based hospitals in Africa and Asia, we are currently studying the availability of information and challenges to obtaining information on the 19 WHO indicators of quality of care.

Methods: This study is being conducted in 25 faith based hospitals in Chad, Democratic Republic of Congo, Tanzania, Zambia and Zimbabwe, and 15 faith based hospitals in India. A standardized data collection questionnaire was developed for the study. Data collectors in each facility are oriented through written instructions and teleconferences. Completed data collection sheets will be scanned and returned to Geneva. Clarifications on the data, if required, will be obtained by email and/or telephone. Data will be analyzed centrally and shared with individual hospitals. In addition, a follow up call is planned with each data collector to understand better challenges with data collection.

Results: The study commenced in April 2015 and is expected to be completed by end of May 2015. Results will be available by end of June 2015.

Conclusion: This study will provide information on quality of maternal, neonatal and paediatric care in selected faith based hospitals in Africa and Asia, and to identify gaps in information and challenges of data collection required for global reporting on quality of care. The information will, if required, contribute to further refinement of the list of quality of care indicators.

Development and Validation of Short Indices to Measure the Quality of Labor and Delivery Care Processes in sub-Saharan Africa

Session: Quality of Facility-Based Maternal and Newborn Care: Development and Application of Improved Metrics, Tools and Analyses

Presenter: Vandana Tripathi, EngenderHealth
Download presentation slides

Background: There is global recognition that quality of care (QoC) must improve to achieve further reductions in maternal and newborn mortality. There is a need for valid, reliable, and efficient labor & delivery (L&D) quality assessment tools. Observation-based measures of obstetric care quality are infrequently used and often lengthy and difficult to administer. This study developed and validated two measures to address these gaps. The study focused on the intrapartum and immediate postpartum periods when most maternal and newborn deaths occur and quality may have the greatest impact.

Methodology: A group of global maternal and newborn care (MNC) experts participated in a modified Delphi process to identify key dimensions of L&D care quality. Experts rated >130 items used to assess L&D care in a series of facility Maternal and Newborn Quality of Care Surveys. Potential QoC indices were developed from highly-rated indicators. Face, content, and criterion validation of these indices used data from 1,145 deliveries observed in Kenya, Madagascar, and Tanzania (including Zanzibar).

Results: An index that performed best on validation benchmarks was identified, including 20 indicators of intrapartum/immediate postpartum/essential newborn care. This index represented most key dimensions of L&D QoC, effectively discriminated between poorly and well-performed deliveries, and appears to be a strong proxy for overall care quality. Responding to concerns about time required to observe full L&D care, a shorter version of the index was created, containing just 13 items that can be assessed at delivery and in the first hour after birth. The comprehensive index is preferred, as it provides a more complete picture of L&D QoC. However, the “delivery-only” index may be a robust alternative in resource-constrained settings.

Conclusions: These tools complement existing MNC quality assessment approaches. Following further validation and piloting, they may be useful in ongoing supervision processes as well as quality improvement research.

Effects of a Quality Improvement Program on Antenatal Care in 12 Regions of Tanzania

Session: Quality of Facility-Based Maternal and Newborn Care: Development and Application of Improved Metrics, Tools and Analyses

Presenter: Scholastica Chibehe, Jhpiego
Download presentation slides

Background: World Health Organization guidelines for focused antenatal care (ANC) recommend a minimum of four ANC visits during pregnancy. From 2008-2014, the Mothers and Infants, Safe, Healthy, Alive (MAISHA) program collaborated with the Tanzanian Ministry of Health to improve quality of ANC to reduce maternal and newborn mortality. ANC providers and supervisors received clinical and supervision skills training and a quality improvement approach with quarterly monitoring and planning to bridge gaps in quality of care was institutionalized.

Methodology: A facility survey was conducted in 12 regions to assess the quality of ANC services using direct observations, facility inventories, record reviews, and health worker knowledge assessments. It employed a pre/post design with data collected in 2010 and 2012.

Results: In 2012, improvements were evident in the provision of preventive care: 83% of women received iron or folic acid at their first visit, 77% received tetanus toxoid injections, and 76% received vouchers for insecticide-treated nets. Nominal increases were noted in the proportion of clients who received counseling and testing for HIV at their first ANC visit (6%), blood pressure screening (5%), and intermittent preventive treatment for malaria (two doses, 3%). Counseling declined for iron/folic acid, primarily at regional hospitals, with only 10–18% of ANC clients receiving such counseling. Slight improvements in counseling on tetanus toxoid, anti-malarials and danger signs were noted, including: severe headaches /blurred vision (37% increase), vaginal bleeding (23% increase) and severe abdominal pain (18% increase).

Conclusion: Findings suggest the MAISHA interventions contributed to improvements in key evidence-based components of ANC, while a few gaps in quality of counseling remained. Greater improvements in quality at lower-level health facilities were achieved than regional hospitals, suggesting a need for renewed emphasis on counseling in pre-service and in-service training and during supervision/mentorship visits at those hospitals.

An Update on the Status of Assisted Vaginal Delivery in Lower and Middle Income Countries

Session: Quality of Facility-Based Maternal and Newborn Care: Development and Application of Improved Metrics, Tools and Analyses

Presenter: Patsy Bailey, FHI 360 & Averting Maternal Death & Disability
Download presentation slides

Background: Assisted vaginal delivery with vacuum extraction or obstetric forceps is used to expedite a birth when maternal and fetal conditions are indicated. This presentation aims to show to what extent assisted vaginal delivery (AVD) is available at different levels of the health system and explores systemic reasons why it is not practiced.

Methods: National and subnational needs assessments in emergency obstetric and newborn care (EmONC) are health facility surveys that include detailed information on the provision of AVD. Of 40+ countries, at least 10 have data from two points in time. Assessments include information on whether AVD was performed in the past 3 months, and if not, why not; what health worker cadres perform the procedure(s); whether equipment is present; and recent assessments include the number of institutional births delivered with vacuum extraction/forceps as well as by cesarean. Where data are available changes over time are described.

Results: Over the past 15 years, AVD is the EmONC signal function least likely to have been performed, however, between the two instruments vacuum extraction is more common than forceps. The majority of African and Asian hospitals provide AVD while only a minority of health centers do so. The primary reasons reported for not providing AVD are lack of trained human resources and lack of equipment. Regional differences in the provision of AVD and which professional health workers actually provide it will be discussed, as will changes over time. Finally, we will include the percentage of institutional deliveries delivered by AVD and cesarean.

Conclusions: Policies to limit the use of AVD in Latin America are born out in the limited data from that region but its use continues in much of Africa, and furthermore, is practiced widely by non-specialists. AVD is underutilized in the face of rising cesarean delivery rates.

A Regional Approach to Standardize Neonatal Deaths Surveillance in Latin America and the Caribbean

Session: Innovative and Collaborative Regional Response to Improved Newborn Surveillance in Latin America and the Caribbean

Presenter: Goldy Mazia, Steering Committee Member of the LAC Neonatal Alliance / Newborn Advisor, Maternal Child Survival Program
Download presentation slides

Background: The gold standard for neonatal mortality data collection is a good civil registration system. Nevertheless, about 30% of babies lack birth certificates by their first birthday, with more than 1 million children unregistered in LAC. The most common data source for neonatal mortality is household surveys, which rely on birth histories without reporting neonatal outcomes. Evidence indicates this process results in under-reporting of neonatal deaths and misclassification of stillbirths. Important information for policymaking, planning, and evaluation for newborn health is lacking.   In 1990, a PAHO resolution recommended the establishment of a death surveillance system in LAC; some countries have developed various approaches since, but few have such systems in place. One priority activity of the LAC Regional Neonatal Alliance’s Action Plan is the promotion of peri-neonatal morbidity and mortality surveillance initiatives.

Methodology: Building upon maternal mortality surveillance systems well established in the region, LAC is enhancing surveillance regionally, by providing: open access to new software packages and implementation manuals in Spanish and English; south-to-south technical assistance in updating surveillance systems; and communications through collaboration with national newborn networks. The LAC Neonatal Alliance is working to enhance newborn deaths surveillance through a) standardization of definitions and their use to improve data robustness and comparability; b) improvement of user-friendly software packages adaptable to paper and computer systems; and, c) linking analysis to action for real-time quality of care improvement.

Results and Conclusions: Activities initiated and led by the LAC Neonatal Alliance include: literature review of systems in developing countries; partnership with the Regional Task Force on Maternal Mortality Reduction to understand current status of maternal and newborn surveillance in countries in the LAC region; two-day experts consultation to share success stories and discuss the way forward; and formation of a working group to address definition consensus and strengthening of national surveillance systems.

Epidemiological Surveillance of Fetal and Neonatal Mortality in Peru

Session: Innovative and Collaborative Regional Response to Improved Newborn Surveillance in Latin America and the Caribbean

Presenter: Jeannette Avila Vargas-Machuca, Ministry of Health of Peru
Download presentation slides

Background: Beginning in 2010, Peru initiated the implementation of epidemiological surveillance of fetal-neonatal mortality with the objectives of determining the magnitude, trends, and causes of fetal-neonatal death and in order to address and improve quality of care.

Methodology: The surveillance program includes i) weekly mandatory reporting of fetal-neonatal deaths in all health establishments; ii) investigation of preventable hospital deaths; and iii) notification of sentinel hospital cases (asphyxia, respiratory difficulty, and sepsis).

Results: Using this surveillance information 1) perinatal-neonatal deaths are epidemiologically categorized, and risk situations have been identified; today the greatest concentrations of preventable newborn deaths are known to be in the rural mountain districts, marginally poor urban, and rural jungle districts. 2) The causes of neonatal death are differentiated by direct and indirect factors. 3) Estimation of the national neonatal mortality rate (12.8 per 1000 live births); and 4) Contribution to the development of two important national documents, the National Plan to Reduce Neonatal Deaths and the Plan Welcome to Life, addressing late neonatal mortality in poor districts. After five years of experience, surveillance systems need ongoing improvement. Case definitions will be incorporated at the community level to facilitate the differentiation between antepartum and intrapartum deaths, and neonatal deaths caused by asphyxia, infections, prematurity, and lethal congenital anomalies will be classified. The quality of care indicators have been updated and the system for immediate notifications has been improved.

Conclusion: Surveillance has permitted a deeper understanding of the fetal-neonatal mortality situation in Peru and has allowed for the redirection of intervention strategies for its reduction.

Surveillance of Fetal and Child Mortality in Brazil

Session: Innovative and Collaborative Regional Response to Improved Newborn Surveillance in Latin America and the Caribbean

Presenter: Juan José Cortez-Escalante, Pan-American Health Organization, World Health Organization

Background: In recent years, Brazil has strengthened the Surveillance of Fetal and Child Mortality (SFCM) as part of an effort to reduce infant mortality. IMR was 14.9 per 1000 live births (2012).Three objectives were established for this initiative: i) increase in case notification; ii) improvement in establishing correctly the underlying cause of death; and iii) identification of possible factors that may have led to the death.

Methodology: Since 2009, SFCM has been implemented nationally. This process was coordinated by the MoH Secretariat of Health Surveillance, and included: establishment of guidelines, decentralization at the local level, setting up of procedures at all levels (state, municipality and special districts of indigenous health), priority setting for physical, human, and financial resources, and timely reporting. The following actions were carried out: 1) improved Information System Regulation 2) implementation of a Mortality Research Subsystem; 3) provision of information “dashboard” to facilitate monitoring of individual cases; 4) development of Mortality Surveillance guidelines; 5) technical/ financial support; 6) periodic assessment workshops; 7) training of health professionals; and 8) partnership with health authorities (Secretariat of Health Care and Special Secretariat for Indigenous Health).

Results: In the decade prior to the implementation of SFCM, fewer than 2000 infant deaths were recorded annually. Following, the number of recorded infant deaths increased (also approximately 2000 annually) as both deaths declined and identification of deaths improved. An impressive increase in the investigation of fetal and child deaths has resulted, rising from 12% (8751/74475) in 2009 to 72% (51106/71029) in 2013. Also the time taken for the investigation to occur has been reduced: from 570 days on average in 2009, to 128 days in 2013.

Conclusion: SFCM has been very effective in strengthening surveillance, providing strong support for the development of better health actions at local and federal levels.