Step-by-step implementation

The process begins by consolidating the existing EmONC data in the country. Various sources can be used to support this process, starting with the routine health information system. If no recent data are available on the EmONC signal functions and indicators, then the country needs to conduct a baseline EmONC facility survey, using the EmONC assessment or its light version (LAT) developed by ENAP/EPMM.¹²³ When possible, the baseline assessment should be part of a broader health facility survey (e.g., Harmonized Health Facility Assessment, SARA, SPA). Health facility assessments can be conducted by local research institutions or by teams assembled by the Ministry of Health and its development partners and comprised of health providers (nurse-midwives, medical doctors, medical/nursing students) and people with health information and data collection experience.

It is critical to include all health facilities that have at least 20-50 deliveries per month, including both public and private-sector facilities (census of health facilities and not sample), when conducting this kind of assessment. Data collection will include geographical coordinates of health facilities that will help determine travel times for the population to facilities and between facilities.⁴ These data will help planners map out the existing EmONC system to identify serious gaps in its functioning, so a complete mapping of the system is necessary. Sometimes accessing data at private-sector facilities can be challenging. But, in countries with large private sectors that deliver a significant amount of maternity care, it is important to include them. (Potential strategies for engaging the private sector can be found under ‘Integrating into routine data systems’.)

When establishing a baseline looking at all facilities, it is recommended that the full set of EmONC indicators be analyzed. This helps stakeholders answer all the questions in the storyline, presented with associated indicators and supplemental studies in Table 1.

One of the first questions in the storyline asks, “How many EmONC facilities will be needed in the long-term to accommodate all birthing women and newborns?” This should be asked as part of a national strategic planning process. This Guide suggests that countries be ambitious when setting out to answer this question and estimate how many functioning EmONC facilities will be needed to ensure that every small and sick newborn can receive care and every woman everywhere in the country has the opportunity to give birth in a facility where EmONC is readily available, if needed. The Long-term calculator helps with the calculations to set this goal, and data from the baseline assessment will enable planners to compare current functioning of the EmONC system to that goal (i.e., Indicator 1a). Long-term strategic planning for infrastructure, human resources, and other system changes should follow, aligned as appropriate with the country’s planning and financing cycles, often on a 10-year time horizon.

Actions to strengthen the EmONC system more immediately, to make rapid impact, can and should be taken as well. In many countries, this has been done by designating a set of facilities to receive targeted support. UNFPA has pioneered this process in 15 countries, starting by supporting countries in West Africa but now extending to East Africa and Asia. Their implementation manual for optimizing a national network of EmONC facilities provides a detailed account of the process,⁵ which has been adapted and briefly summarized below.

The designation process is conducted by national and sub-national political and technical stakeholders, MNH decision makers from the ministry of health, professional associations and civil society organizations, and partners. Using information from the baseline health facility assessment described above in Step 1, this group selects a realistic and manageable set of facilities to be designated for strengthening or upgrading and improvement in EmONC over the medium term, e.g. a 3-5 year programme cycle. Designating a subset of facilities for implementation allows technical and other inputs to be directed more intensively and effectively. The idea is that once the initial set of designated facilities is functioning well, the EmONC system evolves with repeated programme cycles and other facilities designated for implementation support (Figure 1).

Some examples of factors that can be used to determine a designated set of EmONC facilities include:

• Aspirational number of facilities: Using the aspirational number of EmONC facilities from Indicator 1a as a reference point for determining a realistic yet ambitious number of Basic, Comprehensive and Intensive EmONC facilities to be designated for implementation support over 3-5 years. For example, if Country X’s long-term goal is to have 180 EmONC facilities in 10 years and a recent assessment showed that they already have 90 EmONC facilities, a realistic target based on existing resources could be to focus implementation on an additional 45 designated EmONC facilities and improve quality in the existing 90 EmONC facilities over the next 5 years.
• Current EmONC performance: Selecting facilities that are already performing EmONC or are almost performing EmONC is another method. With this approach, all facilities classified as EmONC are selected and implementation would focus on ensuring that they are able to sustain EmONC performance under adequate conditions and provide good-quality care. Additionally, facilities that are missing only one or two EmONC signal functions could also be selected and implementation would focus on systematically ensuring they are able to perform EmONC under adequate conditions and with good-quality care.
• Caseload: Incorporating where women are “choosing” to deliver is also important. To do this, planners should look at the distribution of deliveries across facilities and select facilities that have >1000 women giving birth per year. Because the evidence suggests that facilities that attend fewer than 600 births a year are not likely to perform the six 2009 Basic maternal signal functions, this would ensure that caseloads are large enough in facilities that are designated for at least Basic EmONC.
• Geographic distribution: Maximizing access to services and the geographical distribution of EmONC facilities is key to facility selection. With GIS capacity, planning is facilitated by estimating the proportion of the population that is within 1 hour of a Comprehensive EmONC facility (Indicator 6). With this information, an equitably distributed set of EmONC facilities based on travel time can be planned.⁶
• Inter-facility referral time: Considering the proportion of Basic EmONC facilities that are within 30 minutes of a Comprehensive EmONC facility may also be useful. The optimal configuration of the EmONC system will depend on many context-specific factors, including population density, topography, institutional delivery patterns, and human resource constraints. The World Bank has used a 30-minute standard in its discussions with countries about service delivery redesign choosing this standard based on the “decision-to-incision” literature, i.e. how much time can safely elapse, on average, between a decision to do a cesarean section and the performance of the operation.⁷ If designated Basic EmONC facilities are not within 30 minutes travel time of a functioning Comprehensive EmONC facility, planners should reassess their selection of facilities that are designated to be Basic EmONC and/or decide whether more Comprehensive EmONC facilities are needed in specific areas and/or identify alternative solutions such as maternity waiting homes.
• Health workforce: Considering the availability of skilled health personnel (e.g., cadres such as midwives, medical doctors, nurses), surgical teams and neonatal specialists is another critical factor when designating EmONC facilities to ensure availability of services 24h/7d.

Once a set of EmONC facilities has been designated, program implementation and regular monitoring with the EmONC indicators can be initiated in those EmONC facilities. Implementation of EmONC services themselves is covered elsewhere,⁸ for example in guidance on training,⁹ quality improvement,¹⁰ clinical guidelines,¹¹ supportive supervision,¹² and other aspects of service delivery and health system strengthening. This Guide focuses on the monitoring of that implementation effort.

Routine monitoring of the EmONC signal functions indicates whether designated EmONC facilities are performing and sustaining the EmONC services they are expected to provide. In addition, to ensure facilities are ready to provide good-quality EmONC 24h/7d, specific elements of EmONC service readiness that are sensitive to change, such as commodities, human resources, and inter-facility referral links, must be frequently monitored to ensure they are sustained. Therefore, to minimize the data collection burden for routine monitoring, it is recommended that only a sub-set of the EmONC indicators that relate to these elements be recorded by facilities (i.e., self-reported data) on a monthly basis and analyzed by facilities and sub-national MNH managers or supervision/mentorship teams on a quarterly or semi-annual basis.

The four indicators recommended for analysis on a quarterly or semi-annual basis are:

• 1b. Progress toward interim goal for scaling up EmONC availability
• 3. Equipment, drugs and supplies
• 4. Health workforce adequate for caseload
• 5. Emergency referral readiness

(The full set of indicators and frequency of their measurement is presented at the bottom of this page.)

Other EmONC indicators are more usefully analyzed on an annual basis (Table 2 Frequency of analysis). These are indicators that measure constructs that are less sensitive to change or that represent problems for which it may take longer to implement solutions. These are also indicators that may require data collection beyond register and chart reviews such as exit interviews with women postpartum or interviews with providers. The impact-related indicators that measure maternal and neonatal deaths and stillbirths are best analyzed on an annual basis, in line with EPMM and ENAP monitoring recommendations. On an annual basis, data should be collected and analyzed by sub-national MNH managers or supervision/mentorship teams who are external to facilities. This data collection exercise includes collecting data that are normally self-reported by facilities on a quarterly/semi-annual basis (e.g., EmONC signal function performance, equipment drugs and supplies).

The indicators recommended for analysis on an annual basis are:

• The four indicators analyzed quarterly/semi-annually plus:
• 2. Basic infrastructure
• 6. Home to Comprehensive EmONC within 1 hour
• 7. Institutional delivery rate
• 9. Cesarean section as a proportion of all expected births
• 10. Person-centered maternity care
• 11. Togetherness – 24h/7d family access to inpatient newborn care unit
• 12. Health workforce wellbeing
• 13. Institutional maternal mortality ratio
• 14. Direct obstetric case fatality rate
• 15. Intrapartum stillbirth and very early neonatal death rate
• 16. Neonatal inpatient mortality rate

(The full set of indicators and frequency of their measurement is presented at the bottom of this page.)

By analyzing data on an annual basis, implementation plans can be adjusted as needed. Are some designated facilities’ caseloads lower than expected and therefore not able to sustain EmONC service delivery? How should that be addressed? Are some designated facilities struggling to achieve or maintain the performance of a specific EmONC signal function (e.g., assisted vaginal birth, CPAP)? Is the health workforce in a certain part of the country experiencing higher than expected levels of burnout? Are cesarean section rates too high?

It is equally important on an annual basis to celebrate achievements and positive changes. For example, have person-centered maternity care scores gone up and therefore women are having more positive experiences during childbirth? Have there been reductions in the intrapartum stillbirth and very early neonatal death rate? Has the proportion of the population with access to Comprehensive EmONC within 1 hour increased?

When specific problems are identified (or flagged) with one indicator, stakeholders can drill down with other indicators in the set or by conducting supplemental studies to determine root causes and design targeted solutions. For example, the cause-specific direct obstetric case fatality rate can be used to determine the most lethal obstetric complication and then by looking at related EmONC signal functions, commodities, and health workforce competencies (QIS), specific implementation strategies and plans can be developed to better manage that complication. With ‘red-flag’ reasoning, some of the problems will be highlighted by the indicators but it may also point out the need for full clinical audits, health provider competency assessments, supportive supervision, mentorship, continuing medical education sessions and other quality improvement methodologies.

The Guide also includes additional indicators and suggestions of audits (Box 1) that can be used as needed to focus on specific aspects of clinical appropriateness of care and to improve the quality of care. For example, if the cesarean section rate is very high in a sub-set of designated Comprehensive EmONC facilities, the Robson methodology can be introduced into quality improvement audit cycles to identify how to reduce the number of medically unnecessary cesarean sections. There are also other sets of indicators that complement the EmONC indicators that can be used for quality improvement including a set of indicators developed through consultative processes led by the Re-Visioning EmONC project (Adherence to selected standards of care indicators) and by the WHO (e.g., the QED indicator catalogue).¹³¹⁴

Every 3-5 years, an assessment of the full EmONC system (i.e., all facilities with 20-50 deliveries per year, including the set of designated EmONC facilities) can be repeated using the full set of EmONC indicators (Table 2 below). This provides stakeholders the opportunity to periodically have a bird’s eye view of the full EmONC system and not just the designated facilities that are routinely monitored. This also provides stakeholders the opportunity to check progress towards the aspirational long-term availability goal (and any intermediate milestones), and evaluate and adjust, if needed, the set of designated EmONC facilities. This can be a moment when new strategies for improving and sustaining good-quality EmONC service delivery for the coming 3–5-year programme cycle can be planned.

Most of the indicators will have been calculated for designated facilities on a more regular basis (as described in Step 3), with the national assessment providing an opportunity to calculate them for all facilities. Three indicators will only be calculated during the national assessments, however, as they are most useful when analyzed for the full maternity system, and not just a subset of designated facilities. If the data needed to calculate these indicators are reported from all maternity units on a more regular basis, they could be calculated more frequently (e.g., annually).

The indicators recommended for analysis every 3-5 years are:

• The four indicators analyzed quarterly/semi-annually plus:
• The 11 indicators analyzed annually plus:
• 1a. Progress toward long-term goal for EmONC availability
• 8. Met need for emergency obstetric care

(The full set of indicators and frequency of their measurement is presented at the bottom of this page.)

With country-identified long-term aspirational EmONC goals and mortality reduction targets, such as those set by countries in their national strategies, ENAP-EPMM, and the SDGs, as reference points, this iterative cycle of re-assessment, routine monitoring, and planning responsive actions continues over time.