Portal de Periódicos da CAPES

Establishing a nationwide maternal death and near‐miss surveillance system in a low‐income country: lessons, prospects and challenges

2019; Wiley; Volume: 126; Issue: S3 Linguagem: Inglês

10.1111/1471-0528.15649

1471-0528

BO Okusanya, OO Adetoro, Peter A Aboyeji, BA Ekele, SJ Etuk, CM Chama, OA Dada, OT Oladapo,

Maternal and Neonatal Healthcare

Nigeria, like many other countries with a high maternal mortality ratio, lacks reliable population-based data on maternal mortality. This has made proper planning and effective implementation of strategies to reduce maternal death challenging. As a result, international organizations, including the World Health Organization (WHO), use estimates from statistical modellings to determine the burden of maternal death in Nigeria.1 Such estimates, however, have limited information on the causes of deaths and health service factors which are important for designing appropriate maternal mortality reduction strategies at a national level. In contrast, high-income countries with predominantly hospital births rely on data generated from their health facilities through centralised database systems to review maternal deaths and morbidities to continuously improve pregnancy outcomes for their obstetric populations. For instance, in the United Kingdom, Mothers and Babies: Reducing Risk through Audits and Confidential Enquiries across the UK (MBRRACE-UK) database documents information on maternal and newborn outcomes that are used for confidential enquiries into maternal death and morbidity as well as stillbirths.2 In realisation of the challenges facing the establishment of a data system that is representative of the Nigerian obstetric population, researchers across Nigeria collaborated to develop a nationwide data system to gather data on maternal near-miss and death through a network of Nigerian public tertiary hospitals.3 This article presents the processes, prospects and challenges of setting up this data system with the aim of assisting other countries that share similar health system characteristics with Nigeria to assess their burden of maternal death and disabilities, and identify gaps that can be addressed to meet the targets of the third sustainable development goal. The detailed protocol to operationalise the data system was published in 2009 prior to the commencement of the project.3 Establishment of this database was conceived during a dialogue in 2008 between reproductive health policy-makers and selected researchers across Nigeria. Acknowledging that population-based data on severe maternal complications were not within reach, a decision was taken by these stakeholders to create a network of tertiary facilities as a compromise to obtaining population-based data to guide national planning for improvement in maternal health on the medium term. The urgent need to have reliable data to assess strategies for reducing maternal mortality and morbidity towards the attainment of the Millennium Development Goals in Nigeria strongly influenced the decision. Data on maternal near-miss and death was collected by each of the participating hospitals for 1 year starting from 1 June 2012 through to 14 August 2013. Information on all women admitted for childbirth or within 42 days after giving birth or termination of pregnancy and who experienced life-threatening complications was recorded. The surveillance was conducted across 42 (87.5%) of the 48 target tertiary health facilities (university hospitals and Federal Medical Centres) in Nigeria during this period.4 The survey applied a modified version of the data collection tool of the WHO multicountry survey on maternal and newborn health (WHOMCS)5 to capture data using a combination of clinical, critical interventions and organ-system dysfunction criteria to identify women who had a maternal near-miss.6 A WHO collaborating centre, Centro Rosarino de Estudios Perinatales (CREP), based in Rosario, Argentina, developed the database at no cost to the project by customising the database that was used for WHOMCS. Nigeria is divided into six geopolitical zones: northcentral, northeast, northwest, southeast, southsouth and southwest. The project central coordinating unit was based at the Centre for Research in Reproductive Health (CRRH) in Sagamu, southwest Nigeria. Regional coordinating units were based at the Departments of Obstetrics and Gynaecology of: University of Ilorin Teaching Hospital, Ilorin (northcentral); University of Maiduguri Teaching Hospital, Maiduguri (northeast); Usmanu Dan Fodiyo University Teaching Hospital, Sokoto (northwest); University of Nigeria Teaching Hospital, Enugu (southeast); University of Calabar Teaching Hospital, Calabar (southsouth); and CRRH, Sagamu (southwest). A regional coordinator, who was a respected senior obstetrician and university professor, was selected from each geopolitical zone and invited to CRRH, Sagamu, to review the project design and data collection tools. The regional coordinators participated in train-the-trainer workshops and were empowered to support the central coordinating unit in conducting regional training workshops for hospital coordinators within their respective geopolitical zones. Data tools were pretested at two secondary health facilities (outside of the survey network) in March 2012. To ensure sustainability of the network throughout the project, mid-level career obstetricians (e.g. senior lecturer/consultant) were selected as hospital coordinators. In institutions where a senior obstetrician (who was also a university professor) was selected as a coordinator, a mid-level career obstetrician was identified to provide support on continuous data verification and transfer to the central coordinating unit. Hospital coordinators were invited to a 2-day regional training workshops, which took place at the six geopolitical zones in May 2012. These workshops were held at hospitals serving as the regional coordinating units at the respective geopolitical zones, except for northeast zone where the workshop took place at the Federal Medical Centre, Gombe. Each workshop was facilitated by the regional coordinator with the support of the project lead. The regional training workshops for hospital coordinators were participatory, with opportunities for participants to make suggestions to the data collection tools and project implementation procedures. The hospital coordinators provided different perspectives on regional and/or institutional circumstances that could hinder smooth implementation of the project, and agreed on the corresponding solutions. This approach stimulated a sense of ownership of the project by the hospital coordinators. A Google-based website was created for the project.7 The website had a short profile of all key stakeholders in the project, including the project lead, regional coordinators and hospital coordinators. Each investigator's institutional details and letter of ethics approval were also made available on the website. Electronic feedback (by email) was provided to all stakeholders within the network on total number of births, maternal near-miss and maternal death on a weekly basis throughout the period of data collection. Hospital and regional coordinators received an average of US$100 to procure internet services from mobile communication networks on a monthly basis. Midway into the project, a plenary session was held at the annual scientific conference of the Society of Gynaecology and Obstetrics of Nigeria (SOGON) in Abakaliki, Nigeria, to present the mid-project findings regarding maternal near-miss and death in the 42 hospitals. Six weeks after data collection was completed, an end-of-project review/collaborators’ meeting was held in Lagos, Nigeria, in September 2013. All hospital and regional coordinators attended this meeting and had the privilege of previewing and commenting on the main findings of the project before they were published. Data collection tools and a study procedure manual (Manual of Operation) were developed and supplied to the hospitals to be placed at the various points of recruitment of study participants. Data capture was performed with a simple two-page form, which restricted entered data to insertion of codes (e.g. 1 = No or 2 = Yes) into cells adjoining each variable (Supporting Information Appendix S1). This significantly reduced the time needed to complete the form once a woman met the inclusion criteria. Rather than conducting data entry at the hospital levels, the completed data forms were scanned and directly sent to the central coordinating unit at CRRH, Sagamu, for double-data entry by data entry staff. Scanned data forms were transferred by emails on laptop or desktop computers, and occasionally through a smart phone connected to laptops. The use of smart phones in settings with poor internet connectivity to transfer completed forms partly contributed to the overall success of the project. At the central coordinating unit, all data forms sent by hospital coordinators were scrutinised through visual checks and cleared by the project lead before they were entered into the database. Where any discrepancies or inconsistencies were found by the project lead, queries were immediately sent to the concerned hospital coordinator to review and revise the form for a second round of scrutiny until the form was cleared for data entry. The long time lag from proposal writing to project implementation resulted in some unanticipated challenges to the project. These included the need to identify new hospital coordinators to replace those who had moved on to work at another hospital. The fact that the scientific and ethics approval and fund release occurred approximately 3 years after submission of the proposal to the WHO Department of Reproductive Health and Research, by which time there had been a devaluation of the Nigerian currency against the US dollar, made some aspects of the project more expensive than anticipated. For example, air travel costs for the project became substantially increased and thus limited project oversight by the project lead. There were also telecommunication infrastructural challenges, particularly in areas of insurgency, which delayed regular email transfer of data to the central coordinating unit. In addition, there were unacceptable inaccuracies and inconsistencies in the data reported from two hospitals which led to their disqualification 2 months after starting data collection, amounting to loss of resources already expended on these hospitals. Several lessons were learned during the implementation of this project, which researchers working in similar contexts may want to consider before establishing this type of data system. First, it is important to aim for a short interval between proposal submission and award of research grant, though this is usually beyond the power of researchers. The 3-year gap between the submission of project proposal and actual implementation created important constraints on the project. This was because some hospital coordinators earlier earmarked for the project had changed jobs, and the Nigerian currency had depreciated in value. Early project implementation would have avoided the request for additional funds from WHO and budget adjustment, as well as saving the time lost to replacing hospital coordinators. The pretesting of data collection forms allowed the central coordinating unit to assess the time required to complete the forms and make them simple enough to be used by data collectors with minimal training and resources. The opportunity given to all regional and hospital coordinators to provide input into the data collection tools, in addition to the provision of biweekly activity summaries and quarterly project newsletters while data collection was ongoing, instilled a sense of ownership among collaborators. To ensure high-quality data, the central coordinating unit continuously made efforts to scrutinise all individual forms that were received through visual checks in addition to validity rules embedded in the database. Although visual assessments by the project lead were an arduous task, this was necessary to optimise data quality, given that prior to this project the majority of institutions had never participated in data collection using uniform identification criteria. It was important to identify patterns and quickly take action where efforts to improve data quality from any hospital were not yielding the desired results. In this project, two hospitals were disqualified and their data were removed from the database by the end of the second month because of poor data quality and the inability of these hospitals to live up to the project expectations. The development and supply of a project manual of operations to all hospitals allowed recruitment of eligible women and extraction of accurate data. The buy-in and support of, and collaboration with, international partners and organisations such as WHO and CREP, were critical to the success of the project. Lastly, the burden of maternal mortality (1.1% of live births) and the associated quality of care reported in this survey has triggered the development of national guidelines for Maternal and Perinatal Death Surveillance and Response (MPDSR) by the Federal Ministry of Health (FMOH) of Nigeria, which mandate all publicly funded Nigerian hospitals to routinely conduct maternal and perinatal death reviews and to transfer their findings on a regular basis to FMOH. To facilitate this effort, the WHO is currently working with FMOH, Nigeria, to establish a centralised electronic maternal and perinatal database to continuously gather relevant data across a network of 60 referral level hospitals, including the 42 hospitals that participated in our survey. Many sub-Saharan African countries rely on facility-based data for policy-making and planning of strategies to reduce maternal near-misses and deaths due to lack of population-based data.8 This is despite the significant differences in within-country health resources, including the availability and quality of the healthcare workforce. Historically, a landmark publication in the late 1980s on maternal mortality in northern Nigeria by Harrison et al.9 drew the world's attention to the unacceptable state of maternal health in Nigeria and provided a basis for constantly searching for ways to estimate accurately the burden of maternal death in Nigeria. That study collected data for 22 774 births over a period of over a year.9 The Nigeria Near-miss and Maternal Death Survey reported 97 634 births, 998 maternal deaths and 1451 near-misses over a 1-year period. The four-fold ratio in the number of births over a shorter period would not have been possible without a nationwide synergy of efforts among researchers. A similar database from a network of Brazilian hospitals recorded 83 388 births within a 1-year period.10 This suggests that low- and middle-income countries have the capacity to effectively establish hospital networks that could provide continuous surveillance of maternal health outcomes. Future surveillance projects on maternal near-miss and death should utilise information technology and telecommunication infrastructure in resource poor settings. Inexpensive smart phones and other portable electronic devices that are now available in low-income countries could make data gathering and transfer more efficient. In the same context, data collection applications (apps) may be developed from open source software (e.g. DHIS2 or OPENMRS) to capture and transmit data in real time to a centralised database. Such applications may be configured to require small internet bandwidth for smooth data transfer even where internet connectivity is problematic. The improvement in weight loss, using weight-loss applications on smartphones11 is an indication that smartphone data applications for the surveillance of maternal death and near-miss could be explored. Although this network of Nigerian hospitals involved tertiary level hospitals where resources for gathering good-quality data were guaranteed, countries may consider expanding the network to district or secondary level health facilities depending on the available human resources. The prospects of government ownership and sustainable funding of the data system may be enhanced by an early and continuous engagement of local health authorities and policy-makers. Establishing a nationwide data system on maternal near-miss and death in a resource-constrained setting is achievable. It requires harnessing local resources to form a strong and sustainable network of hospitals that could leverage support of international organisations. In low-resource countries struggling to obtain population-based data on severe maternal complications, a data system that networks health facilities could be a source of good-quality data for policy decision-making and health systems strengthening in the short- and medium-term. The authors have no conflicts of interest to declare. Completed disclosure of interests forms are available to view online as supporting information. The idea for the commentary was conceived by OTO. BOO wrote the first draft with substantial input by OTO. All authors (BOO, OOA, PA, BAE, SJE, CMC, AOD, and OTO) revised the manuscript for intellectual content and approved the manuscript for publication. The manuscript represents the views of the named authors only. No ethics approval required. The Nigeria Near-miss and Maternal Death Survey project and the publication of this commentary were supported by UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), a co-sponsored programme executed by the World Health Organization (WHO). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.