Conceptualisation of ANC service integration models

Our conceptualisation of ANC service integration models started with a literature review on service delivery integration to identify key conceptual frameworks for defining and measuring integrated care. Integrated care was often described as a multidimensional concept. However, the number of dimensions varied across frameworks and there were inconsistencies in the number, naming and content of dimensions. Several dimensions were captured across all frameworks including type, breadth and degree of integration.9 25–33 Additional dimensions incorporated in some frameworks included process,27 level,25 28 time span, mechanism and real/virtual.28 While often formulated in different ways, the commonality among the dimensions of level of integration was a distinction between micro and meso/macro levels of integration where the micro level focused on a seamless care experience for the individual through improved continuity, coherence and cooperation in the delivery of care to individuals while the meso/macro level focused on system level improvements such as organisational and professional integration and mainstreaming of financing and regulation of the healthcare system.9 30 33 Common to all frameworks was a foundational belief in people-centred health services and a desire for integrated care to achieve improved health outcomes through improved coordination of care and continuity of care for individuals.

To further our thinking about what integrated care means for ANC in LMICs, we reviewed the myriad of definitions of integrated care and chose the definition of integrated health services based on the WHO Framework on IPCHS for the basis of conceptualisation of ANC service integration models as it is closely aligned with the goals of the ANC model for a positive pregnancy experience. The IPCHS defines integrated health services as ‘health services that are managed and delivered in a way that ensures people receive a continuum of health promotion, disease prevention, diagnosis, treatment, disease management, rehabilitation and palliative care services, at the different levels and sites of care within the health system, and according to their needs, throughout their whole life’.5

Our ANC service integration model ideally would capture multiple dimensions including the most common dimensions of integration (type, breadth and degree of integration), cover both the micro and meso/macro levels and include a patient-perspective of integrated care. Given the multidimensionality of integration, it would be difficult to capture all these attributes in a single model. Therefore, we outlined the following three models for conceptualising ANC service integration:

1. Integrated health systems—Integration of programmes within the health system; ANC integration within the broader health system including governance, financing, service delivery, information systems.

2. Continuity of care—Continuum of care across the life course; ANC integration across the pregnancy, birth and postpartum life course. This would assess if pregnant women receive interventions across the continuum of antenatal, intrapartum and postpartum care (eg, attended ANC, delivered in a facility with a skilled provider and received appropriate postpartum care).

3. Coordinated care—Coordination of care at a single visit; integration of disease-specific services such as HIV, tuberculosis (TB), malaria, sexually transmitted infection and/or immunisation into routine ANC services as defined by the minimum package based on the new WHO ANC guidelines. This would assess if a woman attending ANC receives the services bundled as needed based on her pregnancy and any other possible conditions (eg, syphilis testing, HIV testing, intermittent preventive treatment of malaria in pregnancy (IPTp), TB testing/treatment and/or immunisations).

Literature review on measurement of ANC service integration

A literature review on measurement of ANC service integration in LMICs identified indicators to measure ANC service integration using a variety of data sources including household surveys, health facility assessments and routine data (see online supplemental table 4). Notably, we found that most ANC integration measures conceptualised ANC service integration as coordinated patient care while a few studies conceptualised ANC service integration as integration of services across the continuum of care. We did not find any examples of measures of ANC service integration with the general health system; however, we did find several examples of disease specific service integration with the general health system which we reviewed to gain an understanding of health system measures.

Coordinated patient care was measured through the integration of ANC with one or more additional service areas. ANC services were most commonly integrated with infectious disease services (HIV, TB and malaria),34–36 however a recent initiative has also explored integration of ANC with family planning counselling, maternal nutrition, prevention of mother-to-child transmission of HIV and post-abortion care and family planning services.37 Various measures of integrated ANC were presented within each study with similar types of measures across studies at both the facility level and individual ANC client level. The most common types of facility-level ANC integration measures were co-location of services (availability of both ANC and an additional service at the same facility but not necessarily by the same provider), cross-training of staff (staff provides ANC and has been recently trained in the additional service area) and readiness to deliver services (availability of equipment, medicines and/or diagnostics required for the services being integrated into ANC). In addition, a composite indicator combining co-location of services, cross-training of staff and readiness to deliver services was developed as an overall measure of ANC integration capacity. At the individual-level, the most common ANC integration measure was receipt of integrated ANC as evidenced by provision of one or more specific interventions to a pregnant woman during an ANC visit (eg, receipt of insecticide treated nets (ITNs) or counselling on sleeping under an ITN, observation of consumption of antimalarial (IPTp-sulfadoxine-pyrimethamine (SP)) after 16 weeks gestation, receipt of HIV counselling and testing during first ANC visit, postpartum family planning counselling, or nutrition advice). The data used and proposed for measuring coordinated patient care includes data from health facility assessments,34–36 routine health information systems36 37 and household surveys.36 Studying ANC integration through the lens of integrating vertical disease specific programmes into broader primary healthcare platforms was the primary rationale for developing these ANC service integration measures.34–36 However, achieving UHC through people-centred, integrated primary healthcare was a motivating force for developing one set of ANC integration measures. Notably, this set of measures focused not only on integration of services during ANC but also at other service delivery entry points including delivery, PNC, infant care and community-based services.37

Continuity of care was measured through the integration of services across the continuum of care for maternal health with a focus on receipt of services from ANC to delivery care to PNC. The main measure assessed dropout along the cascade of services across the continuum of care to understand where women may be missing continuity of care. In addition, an analysis of the combination of maternal health services within the continuum of care that women are receiving was conducted to further understand places in the continuum where dropout occurs. A total of five services were assessed in the continuum—at least one ANC visit, four or more ANC visits, delivery with a skilled birth attendant, postnatal check within 24 hours for the woman and family planning counselling within one year of birth. Household survey data was the primary data source used for measuring continuity of care.38–43 Studying the continuum of care pathway was the main rationale for developing these measures. While these measures were not explicitly presented as measures of integration, a key goal of the continuum of care is to provide women with a unified, integrated approach to care across the life course and this analysis aligns with our conceptualisation of integrated care across the continuum of care for maternal health.

Health systems integration was measured in a series of country case studies that aimed to explore the scope and extent of integration of HIV/AIDS and TB interventions into the wider health system.44–48 These case studies used a methodology that built on the work of the Atun et al health systems framework which was developed to define and assess the nature and extent of integration of priority health interventions in relation to critical health system functions.31 Health systems integration was evaluated against 25 elements of integration across six functions of the health system: stewardship and governance, service delivery, demand generation, monitoring and evaluation, planning and financing (see online supplemental table 5 for a list of the elements). Each element of integration was classified as being ‘fully/predominantly integrated’, ‘partially integrated’ or ‘not or predominantly not integrated’. Data was collected through a combination of document review and qualitative key informant interviews.

Criteria for ANC service integration indicators

We developed a set of criteria that would be important for a ‘good’ indicator of ANC service integration that was based on the review of the literature and supplemented with criteria to ensure feasibility for country implementation. The goal was to develop a set of comprehensive criteria that would serve as guiding principles for developing the ANC service integration indicators while acknowledging that it may be difficult to incorporate all criteria into any single measure of ANC service integration. The criteria include the following:

• An indicator of ANC integration should incorporate current guidelines and recommendations for best practices for the content of care throughout the pregnancy and the multiple contacts with the health system.

  1. The WHO guidelines on ANC for a positive pregnancy experience and WHO framework for the quality of ANC along with country specific guidelines can serve as a reference for the content of care (Provision of care should include maternal and fetal assessment and management, provision of nutritional interventions, infectious disease testing and management and counselling and information sharing. Experience of care should include physiological symptoms assessment and management. Provision and experience of care are reliant on availability of physical and human resources. ANC services are delivered through the larger health system; the structures and functions of the larger health system directly impact the care delivered to pregnant women.).16 18

• An indicator of ANC integration should expand beyond the ANC service package and incorporate additional disease-specific services that pregnant women may need to access in addition to the ANC intervention package.

• An indicator of ANC integration should be measurable with extant data in LMICs. This data can come from household surveys, health facility assessments and/or routine health information systems. There is a preference for at least one integrated ANC indicator that can be measured on a routine basis so that countries can assess performance at more regular intervals.

• An indicator of ANC integration should account for differences in health system contexts within and between countries.

  1. Within countries, different levels of health facilities (ie, primary vs secondary/tertiary health facilities) are expected to have the capacity to deliver different packages of health services.

  2. Across countries, there are differing disease burdens and differing service packages offered through health facilities (ie, not all countries are malaria endemic or have a high prevalence of certain communicable diseases such as HIV or TB). A measure of ANC service integration should account for the disease burden and services offered and required in a country.

  3. The extent to which services can be integrated may be dependent on health system specific characteristics including, but not limited to:

     • The ability and willingness of healthcare workers to provide integrated services; integration may be dependent on the skills/training of health workers.

     • Service specific user fees, for example, ANC services may be free, but TB services may not be.

Proposed indicators for measuring ANC service integration

We then proposed indicators for measuring ANC service integration for each of the three ANC service integration models using the criteria of being a ‘good’ indicator as guiding principles along with the review of the literature on existing ANC service integration measures. Several criteria were prioritised during the indicator development process namely, that the indicator incorporate current guidelines and recommendations for best practices for the content of care and be measurable with extant data in LMICs. Indicators for which these two criteria could not be met were excluded from further consideration. A number of different data sources can potentially be used to measure ANC service integration in LMICs including household survey data, health facility assessment data and routine health information system (RHIS) data. Each of these data sources are unique in the way in which data is collected and all have potential limitations as to how useful they may be in measuring ANC service integration. We reviewed common tools for each data source and assessed the feasibility of using each data source to measure ANC service integration in LMICs. We used both household survey data (DHS) and health facility assessment data (SPA and SARA) for this analysis as both of these data sources were accessible for a multitude of LMICs.49–51 While RHIS data is readily available for country use, it can be difficult to access more broadly and was not available at the time of this study to our research team. As such, for the purposes of this analysis, we have excluded its use.

Operationalising the ANC service integration indicators

Continuity of care

Figure 1A and online supplemental table 7 contain information on the proportion of women receiving each of the four key interventions (ANC1, ANC4, SBA and PNC) across the continuum of care by country. For all countries, the most common service received was at least one ANC visit which was near universal and varied from 95% in Kenya and Malawi to 98% in Senegal. The proportion of women receiving four or more ANC visits varied greatly from 33% in Burkina Faso to 88% in Sierra Leone. In all countries except Sierra Leone, the proportion of women who received four or more ANC visits showed a steep decline and was nearly half the proportion for one ANC visit. The proportion of women receiving skilled birth attendance ranged from 63% in Sierra Leone to 91% in Malawi. In four out of five countries, skilled birth attendance was higher than four or more ANC visits while in three out of five countries skilled birth attendance was lower than PNC for the woman. The proportion of women who received PNC within 24 hours also varied greatly, from 28% in Kenya to 85% in Burkina Faso and Senegal. In two countries PNC was the second most commonly received intervention while in two countries it was the least commonly received intervention.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Continuum of care estimates, by country. ANC, antenatal care; DHS, Demographic and Health Survey; PNC, postnatal care; SBA, skilled attendant at birth.

A depiction of the different combinations of maternal health services within the continuum of care that women received is presented in table 2. Very few women received no services at all (less than 4% in all countries) while the proportion of women receiving all services varied among countries from 15.9% in Kenya to 53.5% in Sierra Leone. In all five countries, receiving all four services was in the top three most common combinations (shown in green in table 2). However, the most common combination was different across countries. The most common combination of services in Senegal and Sierra Leone was all four services (46.2% and 53.5%, respectively). For Burkina Faso and Kenya, the most common combination of services was receiving three out of four services. For Burkina Faso, this included all services except four or more ANC visits (43.3%) while in Kenya this included all services except PNC for women (25.6%). Finally, in Malawi the most common combination of services was receiving only two services, ANC and SBA (25.2%). Combinations excluding ANC1 were uncommon in all countries (less than 1%).

View this table:

• View inline
• View popup

Table 2

Proportion of women who received each continuum of care outcome combination, by country

Figure 1B and online supplemental table 8 contain information on the proportion of women who dropped-out of the continuum of care from one service to the next from ANC1 through PNC. This provides a visual cascade of services within the continuum of care to indicate the proportions of women who moved from one service to the next and the proportion of women who dropped-out along the continuum. In four out of five countries (all but Sierra Leone) the largest dropout occurred between ANC1 and ANC4 (43%–66% decrease). In two countries (Kenya and Malawi), there was also substantial dropout in the later phase of the continuum with a more than 50% decrease from SBA to PNC. In Sierra Leone, the largest dropout occurred from ANC4 to SBA, a 34% decrease.

Coordinated care

Figure 2A and online supplemental table 9 contain information on client-level provision of coordinated care for the three countries that had direct observation data. Client-level provision of coordinated care assessed the proportion of women who received an ANC consultation and received one assessment and management intervention, one nutrition intervention, one infectious disease specific intervention and one counselling intervention at the first ANC visit. The proportion of women receiving coordinated care ranged from 12.3% in Senegal to 33.9% in Kenya. On average, women received more than two-thirds of the required interventions (68.1% in Malawi, 69.6% in Senegal and 75.0% in Kenya). There was substantial variability in which services were most commonly delivered, with no one service being consistently delivered across countries.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Coordinated care estimates, by country. ANC, antenatal care; IFA, iron and folic acid.

Figure 2B and online supplemental table 9 contain information on facility-level co-location of services for all five countries. Facility-level co-location of services assessed the proportion of facilities that offered both ANC and malaria services. In all countries, more than 60% of the facilities had co-location of services with the proportion of facilities with co-location of services ranging from 64.0% in Malawi to 92.8% in Sierra Leone.

Figure 2C and online supplemental table 9 contain information on facility-level co-location of services and service readiness for all five countries. Facility-level co-location of services and service readiness assessed the proportion of facilities that offered both ANC and malaria services and also had the trained staff and guidelines, equipment, diagnostics and medicines required to deliver ANC and malaria services. Almost no facilities (less than 2%) met the criteria for co-location of services and service readiness across all countries. On average facilities had 54.8% (Malawi) to 71.5% (Burkina Faso) of the items required for co-location of services and service readiness. Some patterns across domains were seen. For example, diagnostic capacity was the least available domain in all countries, ranging from 19.8% in Sierra Leone to 54.7% in Senegal. In addition, haemoglobin testing capacity was the least available item in four out of five countries. However, there was substantial variability across the remaining domains in terms of performance. For example, the availability of equipment was a strength in Burkina Faso (95.7%), but a weakness in Malawi (44.2%).

Comparison of indicators of ANC integration

A comparison of ANC integration estimates is presented in figure 3. Looking across ANC service integration estimates, there are large differences between estimates for ANC service integration depending on the model used, and in some countries, the ANC integration indicator definition within a model. No one integrated ANC indicator was consistently the highest estimate for ANC service integration. However, continuity of care was consistently the lowest estimate for ANC service integration. In some countries, coordinated care estimates were all similar (Malawi, Senegal). However, in other countries this was not the case (Burkina Faso, Sierra Leone).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Comparison of ANC service integration estimates, by country. ANC, antenatal care; DHS, Demographic and Health Survey; SARA, Service Availability and Readiness Assessment; SPA, Service Provision Assessment.