The broad lens of the Sustainable Development Goals (SDGs) prioritizes integrated, coordinated, and multi-sectoral approaches to the health of individuals, societies, and the planet. The UN Secretary General’s updated Global Strategy for Women’s, Children’s and Adolescents’ Health (2016-2030)  goes beyond survival to include efforts to ensure that women, families, and communities can thrive and transform to meet their full potential.
As countries move along the pathway of the obstetric transition and progress toward ending preventable maternal mortality, the causes of maternal death are also widening. Along with indirect causes of death, the more distal social and structural determinants of maternal health and survival are taking on greater significance. To understand and address the full range of causes of death, high-quality, nationally representative data on deaths are needed; however, at the beginning of the SDGs in 2015, no deaths data were available in seventy-three countries due to lack of functional civil registration and vital statistics systems .
The Strategies toward Ending Preventable Maternal Mortality (EPMM) (EPMM Strategies) , released in 2015 by the World Health Organization to provide a strategic framework for maternal health under the SDGs, likewise turns a broad lens on the determinants of maternal health and survival, both direct, and distal – including social, political, economic, and systemic determinants – highlighted in “11 Key Themes” (Table 1).
Table 1. EPMM Key Themes
EPMM – Ending Preventable Maternal Mortality
The EPMM Strategies present priority recommendations in eleven domains reflecting the full spectrum of determinants of maternal health and survival; these “11 Key Themes” emerged through consultation with global and national stakeholders through a number of channels, including an official member state consultation and public call for comments hosted by the WHO. In line with the SDG agenda, the EPMM Strategies go beyond the health sector and cover structural and societal factors that influence maternal health outcomes including chance of death. Such a multi-sectoral lens is particularly important for maternal health, where outcomes are more dependent on other sectors and upstream determinants than other more “vertical” domains of health.
The increasing complexity of the current era calls for improved approaches to monitoring maternal and newborn health. We need to broaden the range of what is monitored, at the same time that we apply more precision in choosing measures that are fit for purpose and worth the burden of collection. The principle of parsimony calls for focused attention to what matters, which varies from differing perspectives, in different contexts, and for different uses. At the advent of the SDG era, the so-called “Kirkland commentary”  identified five principles for maternal and newborn health measurement: focus through core indicator sets, relevance of measures to end-users, measurement innovation, inclusion of equity measures, and supportive global leadership. A review of these principles at the five-year mark resulted in the call for a sixth, overarching principle of country ownership through which the first five principles should be filtered . Additionally, there were calls for increased focus on policy and health system-level drivers in health measurement , as measurement efforts have traditionally focused quite narrowly on health services and interventions, neglecting measures of societal determinants that empower people to achieve full health, or systems determinants that create an enabling environment for effective service delivery and quality of care.
To enable countries to track progress toward achieving the priority recommendation in each domain highlighted in the 11 Key Themes, work was undertaken to develop a comprehensive monitoring framework for the EPMM Strategies, in two phases, described elsewhere [7,8]. Whereas Phase I focused on reaching consensus on a minimum set of core maternal health coverage and outcome measures for cross-country comparisons, Phase II focused specifically on identifying indicators to address the distal causes (social, political, and economic determinants) of maternal mortality, by seeking to answer the question, “What are the 1-3 strongest available indicators for monitoring each EPMM Key Theme?” From 2016-2017, multiple rounds of modified Delphi process were implemented to develop a menu of indicators suitable for national/sub-national monitoring. Indicators were selected by stakeholders based on pre-specified criteria focused on relevance to the EPMM thematic area, importance across a range of contexts, interpretability and utility to end users, as well as validity, feasibility and harmonization with other global measure development initiatives. A secondary outcome of this process was the emergence of a set of requests for additional indicators linked to each EPMM Key Theme, reflecting instances when participants in the process identified a need for new metrics to fill measurement gaps. Stakeholder feedback suggested that developing a small number of additional indicators would be an important contribution to a comprehensive monitoring framework for the EPMM Strategies. In total, twenty-nine additional indicators were requested to fill such gaps (Table 2).
Table 2. Additional indicators requested by EPMM Phase II stakeholders
EPMM – Ending Preventable Maternal Mortality
In 2017, the Women and Health Initiative (W&HI) at the Harvard T.H. Chan School of Public Health launched the Improving Maternal Health Measurement Capacity and Use project (IMHM Project), aimed at strengthening metrics and measurement capacity to support country efforts to drive improvement in maternal health and survival. The full body of work aimed at capturing stakeholder demand and channeling it into activities and resources to improve maternal health measurement as a core strategy for ending preventable maternal deaths is displayed in a Theory of Change diagram (Figure 1).
Figure 1. Demand-driven measurement theory of change: Ending Preventable Maternal Mortality (EPMM) – Improving Maternal Health Measurement Capacity and Use (IMHM).
- Consultation 1: Four indicators reflecting fundamental human rights principles applied to maternal health
- Consultation 2: Four indicators reflecting better measurement of births, deaths, and causes of death
The primary aim of these technical consultations was to affirm the identified measurement gaps, specify the critical components of the constructs for measurement, and fully elaborate metadata for these constructs (definition, numerator, denominator, disaggregators) to allow them to be monitored.
A secondary aim is to share the experience of adapting a technical consultation to an online platform due to COVID-19 travel restrictions, which may offer important lessons learned for others beyond the field of maternal health measurement.METHODS
Selection of constructs for measure development
This is a qualitative, iterative, structured, remote expert technical consultation. From May 11 to July 17, 2020, the IMHM Project convened two simultaneous online technical consultations focused on developing eight additional EPMM indicators from the full list of additional measures requested to fill measurement gaps that emanated from Phase II of the EPMM indicator development process. The twenty-nine additional EPMM indicators requested (Table 2) were distilled into their underlying constructs. As defined by Lavrakas , a construct is “the abstract idea, underlying theme, or subject matter that one wishes to measure”. While the original EPMM stakeholder requests were formulated with detailed (aspirational) indicator definitions, we chose to focus at the construct level to give content experts and measure developers the freedom to deliberate on the strongest formulations of the underlying concepts of interest for the purposes of measurement.
The selection of constructs to be developed into fully articulated indicators was based on inputs collected during national multi-stakeholder dialogues in five countries, along with information from a landscape analysis based on online surveys and key informant interviews with government officials and in-country UN agency staff in twenty countries on their perceived need for the proposed additional indicators. From 2018-2020, the IMHM Project supported seven national multi-stakeholder dialogues, co-sponsored with government officials in Ministries of Health and civil society organizations and focused on collecting stakeholder opinions on the priorities for ending preventable maternal mortality and maternal health measurement needs to support their achievement in each context. At dialogues held in five of the seven countries, Kenya, Cote d’Ivoire, India, Bangladesh, Pakistan and Mexico , participants took part in interactive prioritization exercises in which they ranked the importance of all twenty-nine proposed additional EPMM indicators in their context, using a four-point scale. The constructs that a majority of stakeholders across all participating countries rated “Very Important” were considered for inclusion in these consultations. These results provided additional inputs into national demand for development of the proposed additional indicators.
The final selection was made with the input of a global advisory committee including maternal newborn health and measurement experts whose organizations contributed to the development of the EPMM Strategies report and its comprehensive monitoring framework. The candidate constructs were selected from the list of those deemed of importance at country level based on the following predetermined criteria:
- The construct is conceptually clear
- The construct needs measurement
- Measurement of this construct is relevant and important
- Measurement data on this construct would be actionable
- Measurement of this construct is conducive to global monitoring and national comparisons
- There are available data of acceptable quality on which to base or compile an indicator to measure this construct
Crucial to prioritization was careful consideration of indicator development already in progress by other researchers and UN agency groups, to coordinate indicator development efforts and ensure that the constructs prioritized are not duplicative of ongoing initiatives by other organizations.
In the end, eight constructs were prioritized for development into fully articulated indicators through the current consultations (Table 3): Consultation 1 – four rights-based constructs for measurement (construct IDs: 1.1; 1.2; 4.1; 4.2); Consultation 2 – four constructs for strengthening measurement of births and deaths (construct IDs 2.2; 5.1; 5.2; 5.4).
Table 3. Constructs for consideration and development into fully articulated indicators
EPMM – Ending Preventable Maternal Mortality
We purposively identified experts with specific expertise in maternal health measure development as well as the topical areas covered by the constructs for development to participate. Sixty-one participants took part across both consultations. Thirty-one participants took part in Consultation 1, focused on developing fully articulated indicators for a set of Rights-Based constructs for measurement. Thirty participants took part in Consultation 2, focused on developing a set of fully articulated indicators for measuring births and deaths.
Due to travel restrictions imposed by the COVID-19 pandemic, measure development took place through two simultaneous, parallel ten-week online consultations stratified by focus area (see Table 3), hosted on Harvard Canvas, the university’s online education platform. The consultations used multiple modalities to collect inputs and conduct in-depth consultation with the convened groups of experts in order to achieve their intended objectives.
To orient participants to each construct for measurement and to affirm the identified measurement gaps, participants first watched presentations from subject matter experts on each construct for consideration. Twelve content experts gave presentations in Consultation 1. Ten content experts gave presentations in Consultation 2. Each presenter provided an overview of the current state of measurement of the construct they were presenting, as well as their own perspectives on the feasibility, challenges, and utility of measuring the construct. This set the stage and provided an introduction for the structured online discussion forums and technical work that followed for each construct to be developed into an indicator.
After watching presentations, participants completed online polls to identify the critical component(s) of each construct for measurement, choose the expression of the construct best capturing the specific measurement gap, and vote on whether to move forward to articulate the full metadata for an indicator to measure the construct. Next, following participation in a semi-structured online discussion, participants completed a final survey to consolidate the group’s consensus on which constructs to take forward. All eight constructs for measurement were ultimately ratified after considerable debate and brought forward for full indicator development, to include indicator definition and articulation of full metadata.
Participants then worked collaboratively in small groups to develop the full metadata for each construct. To prepare, each participant individually drafted their proposed formulation of all seven components of the metadata for each construct: definition, numerator, denominator, method of estimation, disaggregation factors, preferred data sources, and the expected frequency of data collection. These individual participant inputs then populated a group worksheet for each construct. In small groups on Zoom, organized by time zones, participants discussed the group worksheet for each construct, debating the merits and challenges related to the proposed formulations of each component of the metadata to reach agreement on the strongest versions of each draft indicator, using pre-specified criteria  (Table 4). In this way, participants in each small group developed consensus on a proposed draft indicator for their group, which a rapporteur submitted for consideration by the full plenary.
Table 4. Criteria for evaluating draft indicator metadata
The proposed draft indicators for each construct from all small groups were compiled into a final poll that included one additional version of each indicator synthesized by the consultation administrators that drew from all submissions in an attempt to reflect the “best of” all the proposals. Participants then voted on the strongest draft indicator, including the definition and full metadata, for each construct.
A final plenary discussion to review these poll results was hosted on Zoom for each consultation. All participants for each consultation came together to review the full metadata for the highest-ranking draft indicator for each construct, and to discuss any related questions and concerns posted on the Canvas discussion boards. A facilitated discussion ensued to address the outstanding issues regarding any aspect of each measure, where participants arrived at agreed solutions wherever possible. The draft indicators emanating from both consultations appear in the tables below.RESULTS
Eight new candidate indicators with relevance for monitoring the priority recommendations highlighted in the EPMM Strategies were developed through these two simultaneous technical measure development consultations. Each indicator includes a definition, numerator and denominator (if applicable), method of estimation, disaggregation factors (stratifiers), preferred data source(s), and expected periodicity (frequency).
Consultation 1: Four Rights-Based Indicators
Four indicators address gaps in measures to monitor fundamental rights-based determinants of maternal health and survival at national and subnational level, including women’s reproductive autonomy and decision making power; participatory accountability for maternal health outcomes ; and Respectful Maternity Care. These indicators address identified gaps in measures available for monitoring the following EPMM Key Themes: Theme 1. Empower women, girls, families and communities; and Theme 4. Apply a human rights framework to ensure that high-quality reproductive, maternal, and newborn health care is available, accessible and acceptable to all who need it (Table 5).
Table 5. Full metadata for four rights-based EPMM indicators (developed during Consultation 1)
EPMM – Ending Preventable Maternal Mortality
Consultation 2: Indicators for Measuring Births and Deaths
Four indicators strengthen the measurement of and ability to count, track, and link births and maternal deaths and causes of death. They address identified gaps in measures available for monitoring the following EPMM Key Themes: Theme 2. Integrate maternal and newborn care, protect and support the mother-baby dyad; and Theme 5. Improve metrics, measurement systems, and data quality (Table 6).
Table 6. Full metadata for four EPMM indicators for tracking births and maternal deaths (developed during Consultation 2)
CRVS – civil registration and vital statistics, COD – cause of death, EPMM – Ending Preventable Maternal Mortality
Implications in context of existing research
The proposed indicators contribute toward a comprehensive monitoring framework for the EPMM Strategies. As such, each proposed indicator corresponds to a specific EPMM Key Theme, and aims to fill gaps identified by stakeholders who were asked to evaluate existing maternal health metrics using pre-determined criteria to the strongest available indicators to track progress toward achieving the priority recommendations in the area of each EPMM Key Theme.
A recent commentary on measurement criticized the siloed, top-down nature of measure development, which for too long has remained an internally facing activity, controlled largely by what the authors call a “global measurement enterprise”, lacking inclusivity and transparency . In contrast, the draft indicators presented in this paper reflect the results of a process driven by end-user and country-level stakeholder demand generated through a participatory process, and include the inputs and aspirations of numerous stakeholders, gathered over time and across various platforms.
These draft indicators respond to the call for a broadened approach to measurement and for indicators useful for tracking the societal determinants and health system drivers of maternal health, in addition to those that track coverage of clinical interventions and their associated outcomes that, while critical, are not sufficient . Indicators that permit countries to track rights-based constructs such as women’s reproductive autonomy and decision-making power, participatory mechanisms for driving subnational accountability for maternal health outcomes, and the legal and regulatory frameworks to support Respectful Maternity Care are also needed. Such indicators provide a basis for monitoring the enabling environment for high-quality maternal newborn health care and a high-performing maternity care system. This enabling environment includes the promotion and protection of fundamental rights, attention to power dynamics, information systems, and organizational and sociopolitical dynamics.
A number of these indicators respond to the need to strengthen civil registration and vital statistics [CRVS] systems in many low- and middle-income countries, which are critical mechanisms to address and ensure citizen rights, improve gender equality, and enable health system planning and improvement . Indicators that track cross-system design and interoperability and monitor national health information and registration system capacity–including the ability to capture health data and link them with data on vital events, coverage of birth and death registration, and the use of CRVS information not only for the purposes of ensuring legal status but as data for decision making–provide opportunities to improve measurement, metrics, and data quality in the context of maternal health and survival .
Strengths and limitations
The iterative, discursive exploration of the concepts for measurement and the need for metrics to track them responds to recent calls for measure development to be carried out in more inclusive ways and to include a broader range of stakeholder perspectives . The process undertaken to identify measurement gaps through the EPMM indicator development consultations (2016-2017), and to prioritize the constructs for measure development through national dialogues and key informant interviews in over 20 countries (2018-2019) aligns well with a recent update and reaffirmation of the Kirkland principles for maternal health measurement, calling for measure development to be country-driven with supportive global leadership, focused for purpose, relevant to the end users, innovative in design, and sensitive to measurement of equity .
Another strength is the focus on concept-driven measure development. Recent work to improve the understanding of and the approaches to determining indicator validity in the context of maternal and newborn health measurement have focused on the critical importance of construct validity. In a definitional framework commissioned by Mother Newborn Information for Tracking Outcomes and Results (MoNITOR) , an expert advisory group to the WHO on maternal newborn health measurement , describe construct validity as the degree to which “a given operationalization (through indicator definition and its measurement) accurately reflects the phenomenon it is intended to measure.” This essential aspect of validity was highly valued by practitioners in the field who contributed to a landscape analysis by the same authors  that informed the MoNITOR definitional framework. Policy dialogue on measure development emphasizes that it should be driven by a conceptual need and framing, rather than by what data are available or what is currently feasible to measure ]. In response, the process to develop these additional EPMM indicators focused heavily on careful consideration of the underlying construct for measurement and consensus on the best expression of that construct in each case.
The virtual format of the consultation, paired with the COVID-19 pandemic and differing time zones in which the participants completed activities, led to a few limitations. The online Harvard Canvas platform posed difficulties for a small number of participants due to occasional technical glitches, poor internet connection, or user error. Additionally, participant time zones ranged from UTC/GMT – 8 hours to UTC/GMT + 6 hours, which made it difficult for all participants to participate in all activities. Finally, although the online consultations comprised the same number of hours as the planned in-person meetings, simply spread out across a longer time frame, “consultation fatigue” over the ten-week period may have led to lower rates of response to polls or participation in other activities compared to an in-person format held over four consecutive days. The impact of the COVID-19 pandemic, which achieved global magnitude just prior to the start of the consultations, on participants work and family lives must also be recognized.
On the other hand, the online format offered numerous benefits. Although Zoom meetings and online discussion boards are not a replacement for in-person collaboration, Canvas allowed all objectives of the consultation to be met during a global pandemic. The extended time frame allowed for more in-depth analysis of each construct, including rich discussions and participant polls during each step of the consultation. Additionally, the online format allowed participants to return to past objectives and review content at their leisure, such as the recorded speaker presentations and the background resources. Furthermore, the online format allowed six more participants to take part in the consultation than could have participated in person, and enabled a variety of activities tailored to each objective, including videos, discussion forums, polls, and live Zoom meetings. In the end, 83% of those who responded to the evaluations for both consultations strongly or very strongly agreed that the consultations were effective in meeting their objectives and that they learned a lot, while 67% said they would recommend this type of online consultation to others.
While recognizing that further technical and policy work remains before these draft indicators can be broadly recommended for national and subnational level monitoring and some global comparisons, they represent a solid foundation for such next steps. The inputs that went into the selection of these indicators, gathered over time, across various geographies, and from multiple stakeholder groups, ensure that the need and demand for these measures is real. The thoughtful deliberations collected during these consultations ensure that the constructs for measurement are well-defined, and that the proposed draft indicators and metadata to allow their measurement reflect the expertise of a sizeable number of measure developers and subject matter experts. Together, these indicators contribute to country capacity to improve maternal health outcomes by analyzing and acting on influential social and health system level determinants of maternal health and survival .
We gratefully acknowledge the following project advisors: Mabel Berrueta, Veronica Pingray, Sowmya Ramesh, Jennifer Requejo, Niranjan Saggurti, Marta Schaaf, Kimberly Whipkey. We thank all participants in the IMHM Additional Indicator Development Consultations for contributing their expertise and opinions. The list of participants and presenters is included as a supplemental file, and presentations are available from the corresponding author upon reasonable request.
Availability of data and materials: All data generated or analyzed during this study are included in this published article (and its supplementary information files) or at https://doi.org/10.7910/DVN/QCFZJ3. This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.