Establishment of Maternal and Perinatal Database for Quality, Equity and Dignity programme: Processes, challenges, lessons and prospects

Abstract

Many low- and middle-income countries (LMICs) with high maternal mortality rates lack reliable routinely collected data on the quality of care provided to women giving birth in facilities nationwide and information when a death occurs. Although global efforts have been made to address the high numbers of maternal and perinatal deaths (e.g. by increasing the number of women who access health facilities for pregnancy care and delivery), this has not translated into reducing mortality.1, 2 Despite reaching health facilities, many women and their babies still die because of delays in receiving quality care.3, 4 Health information systems designed to assess quality of care and outcomes for women and their babies can play a crucial role in facilitating timely interventions, as part of a continuing quality improvement process to enhance maternal and newborn care. Nigeria is one of the highest contributors to maternal and perinatal mortality worldwide, with approximately 67 000 annual maternal deaths and stillbirth rates of 43 per 1000 total births.5, 6 However, as a tool for improving maternity care, the collation of actionable routine healthcare data is hindered by the lack of harmonised data and by the poor quality of data available in Nigeria. This limitation impedes the ability of researchers to pinpoint and address gaps in the quality of care at regional and national levels. Recognising this limitation, the World Health Organization (WHO) in partnership with the Nigerian Federal Ministry of Health (FMOH) established a nationally representative electronic database, the Maternal and Perinatal Database for Quality, Equity and Dignity (MPD-4-QED), in referral-level hospitals across Nigeria, with the aim of monitoring the quality of maternal and early newborn care and outcomes for women and their babies. In this article, we present the process for setting up the database, challenges encountered in its implementation, lessons learned and prospects of the programme to assist other countries that share similar health system characteristics in establishing nationally representative facility-based data collection systems. In the planning stages, the WHO engaged with the Nigerian FMOH, including the Minister of Health, to establish the MPD-4-QED programme. This successful engagement led to a partnership with the FMOH for the programme and facilitated the participation of referral-level public health facilities under the jurisdiction of the FMOH. In addition, the WHO met with key in-country stakeholders in maternal and child health, including professional medical associations, to gain their support for the programme and enable the participation of health facilities run by the State Ministry of Health as well as health facilities that are privately owned. Consequently, the MPD-4-QED programme was established in 54 consenting referral-level hospitals (48 publicly funded and six privately funded) in Nigeria. These hospitals receive patient referrals from lower levels of care and manage both high-risk and low-risk pregnancies. The hospitals were spread across 36 states of the country (Appendix 1). Each hospital had a dedicated team for the MPD-4-QED programme, consisting of two medical record officers (MROs) conducting data entry and two hospital coordinators (an obstetrician and a neonatologist) overseeing the data collection. The hospitals in each of the six geopolitical regions of Nigeria were supervised by a regional coordinator, and a national coordinating unit for the programme was set up, comprising a national coordinator, a data manager and an administrative team. The study population comprised all women and their babies who were admitted for delivery or for complications within 42 days of delivery or termination of pregnancy. Babies who were delivered outside the participating hospitals but were admitted to the hospitals within the first week of life were also included. Data collection was designed in line with the WHO Quality, Equity and Dignity (WHO QED) strategic objectives for improving maternal, newborn and child health.1 The WHO QED strategy utilises a set of common indicators, the ‘QED indicators’, to monitor the key aspects of quality of care (Appendix 2). We developed an individual-level data collection template using a participatory approach with the Nigerian FMOH and in-country stakeholders, including obstetricians and neonatologists. The wording of the questions and the selection of variables reflected standard case definitions and context-specific practices regarding clinical documentation in Nigeria. The information included biodemographic data, antenatal and past medical history, clinical conditions during pregnancy and at admission for delivery, labour interventions and complications (if any), time-related events in clinical management (such as labour augmentation and decision-to-delivery interval), mode of delivery, immediate postpartum condition, and maternal and newborn outcomes (Appendix 3). These data were collected through an electronic platform developed by customising the open-source District Health Information Software (DHIS-2).7 Routine data entry was primarily performed by trained MROs using internet-enabled tablet devices and synchronised in real time to a secured central cloud-based server. Data entry was initiated at the time of admission and updated until discharge or death (whichever occurred earlier). For newborns, information was collected up to 7 days after birth, at discharge or upon death (whichever happened first). In the event of maternal or perinatal death, the local mortality audit team (led by an obstetrician and neonatologist) analysed and documented the primary cause of death (using the International Classification of Diseases for Maternal Mortality, ICD-MM, and the International Classification of Diseases for Perinatal Mortality, ICD-PM) and the avoidable contributing factors. A unique identifier was used to link the data between a woman and her newborn. In instances of poor internet connectivity, it was possible to enter data offline and then automatically synchronise once internet connectivity was restored. Furthermore, facility-level audits were conducted quarterly to evaluate personnel, conditions of the facility, quality management, basic hygiene and sanitation, equipment and the commodity inventory. The model of the MPD-4-QED electronic platform is shown in Figure 1. To minimise heterogeneity in data collection, a standard operating procedure (SOP) manual detailing how cases should be defined and how electronic data collection should be completed was developed. The SOP manual was used to train the project team members from all participating facilities at the regional level. The electronic platform had built-in validation rules, reflecting demographic, biological and medical plausibility, which were applied to minimise data entry errors and ensure data completeness and internal consistency. Before deployment, the electronic platform was pretested for 4 weeks and the problems identified were fixed iteratively. The data entry error was less than 1% in all facilities before the formal launch of the platform on 1 September 2019. The programme continues to collect data to date. To ensure that all eligible data were captured, a hospital coordinator compared weekly entries with admission registers and randomly sampled 5% of cases to review for accuracy. Any observed missing cases or errors in the data entries were resolved before closure of the data record for each participant. At 6-monthly intervals, regional coordinators paid unscheduled visits to participating institutions for independent external data monitoring for completeness and internal consistency. The national coordinating unit conducted periodic analyses of available data at the facility, regional and national levels, and shared the results monthly with collaborating institutions and the FMOH. Each hospital coordinator had access to data visualisation specific to their respective hospital, whereas regional coordinators were provided with data visualisation for their designated regions. There were several external challenges in implementing the MPD-4-QED programme. The high turnover of hospital staff who were involved in the programme as hospital coordinators or MROs posed a significant challenge. The need to continually train new personnel in the programme arose at different intervals and across diverse hospitals and regions. To navigate this challenge, a trainer model was implemented. This approach empowered regional coordinators to conduct regular training sessions for new hospital coordinators and, in turn, enabled hospital coordinators to provide training for new MROs as required. Second, the additional workload for hospital coordinators and MROs was an issue initially, but over time became more seamlessly integrated into routine practice. Motivation increased when the first results became available, and the data could be visualised regularly and used in local quality improvement meetings and for research purposes. Another challenge was the variability in the strength of internet networks among providers across different regions. Consequently, multiple network providers were needed to ensure unlimited strong internet access for the data-capturing tablets. However, the system also enabled offline data entry and synchronisation when an internet connection was available. In general, it is feasible to establish a routine electronic health database in a low -to middle-income country such as Nigeria. The past decade has witnessed a rapid surge in mobile and internet penetration in sub-Saharan Africa, enhancing the prospects for digital health interventions.8 Although the likelihood of poor internet services remains, designing the electronic platform to capture data offline and synchronise when the internet network is available was successful in minimising disruptions. The success of the programme was driven by the meaningful engagement of in-country stakeholders, including the FMOH, hospital authorities and healthcare staff. The motivation of staff played a pivotal role in achieving the project's goals. Regularly disseminated newsletters and infographics, illustrating care performance, areas requiring improvement, and advancements in maternal and perinatal care, proved effective in sharing outcomes with the team and contributed to sustaining motivation. The availability of the SOP manual and the standardisation and uniformity of the data collection template resulted in high-quality data that could be aggregated at regional and national levels to monitor care performance and improve quality. The close monitoring and maintenance of an effective two-way communication system for raising and addressing issues also contributed to the quality of the data collected. We collected high-quality data on maternal and perinatal death audits and found that timely death audits, conducted within 3 days, were important. Our experience was that it was very difficult to conduct retrospective death audits after a prolonged period. On the whole, the MPD-4-QED programme shows the future potential for monitoring key maternal, newborn and child health indices in Nigeria (and by extension in other LMICs), rather than solely depending on estimates. The success of the MPD-4-QED programme demonstrated the feasibility of implementing a harmonised and unified routine data collection system for maternal and perinatal outcomes, as well as the quality of care, in Nigeria. The rich data collected could be used to identify gaps in the quality of care and improve clinical practice at individual hospitals. To enhance the translation of recommendations resulting from the analyses of the database into actionable steps, the establishment of a national technical working group is required. To avoid drops in data quality, continual training and retraining in clinical documentation and data capture will be necessary, as well as the continuous monitoring and evaluation of the data collection processes. The MPD-4-QED database yielded high-quality data, proving valuable for assessing the quality of maternal and perinatal care. The successful development of this database suggests its potential for scalability to other LMICs and demonstrates that it can be implemented elsewhere. The idea for the commentary was conceived by AA, TL, JT and OTO. AA wrote the first draft, with substantial input from TL, JT and OTO. All authors (AA, JT, TL, BE, IA, PA, OA, CC, SE, HG, JI and OTO) revised the article for intellectual content and approved the article for publication. The article represents the views of the named authors only. We acknowledge the contributions of academic staff and hospital personnel in all hospitals across the MPD-4-QED network. We offer special thanks to the MROs for their dedication to meticulous data collection and their contributions to the success of the programme. We acknowledge the Nigeria FMOH for their collaboration and support in establishing the programme. This work was funded by MSD for Mothers; and the 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). The funders did not play any role in the design of the program, in the data collection, analysis, and interpretation, in the writing of this report and the decision to submit the paper for publication. All authors declare that they have no competing interests associated with this work. The scientific content of the programme was approved by the WHO HRP Research Project Review Panel (A65930, 6 May 2018). The WHO Ethics Review Committee (A65930, 5 June 2018) and the Nigerian National Health Research and Ethics Committee (NHREC/01/01/2007, 5 September 2018) approved the study. Authorities of all participating hospitals granted written institutional approval to participate in data collection, periodic analyses and reporting. Number of women (obstetric admissions) who had a companion present, as recorded in medical records. In the case report forms, three responses were available: yes (medical record reported that the woman had a companion during labour); no (medical record reported that the woman did not have a companion during labour); unknown (medical record did not report whether the woman had a companion in labour or not). If this information was left blank in the case report form, it was considered missing data For women who had a companion in labour, the type of companion was also recorded (spouse, family member or another person) This is a commentary article with no data to share.