Integration of the HIV M&E System

Health-related data is collected every day throughout the different levels of the health system so that it can be used by actors within the system and beyond. Health professionals, managers and policymakers all have some role in generating, collecting and utilising this data to support decision-making, efficiency, effectiveness, accountability, priority setting and allocating resources. However, research in resource-limited settings suggests that the integration and quality of health-related data is inconsistent, and this results in a limited use of this data. This challenge has been further exacerbated in the last few decades in many low-to-middle-income countries (LMICs) where their health services are often vertically organised through disease-specific programmes (Koivu et al., 2017).

Defining the Challenge

The last few decades have seen many LMICs establish dedicated monitoring and evaluation (M&E) systems for their vertical disease-specific or Disease control programmes (DCPs). These M&E systems are frequently managed as distinct parallel systems to generate reliable, timely and accessible health information that will be used to manage DCPs (Kawonga, Blaauw, & Fonn, 2016). South Africa is one the countries that has established a vertical HIV M&E system to monitor its national HIV/AIDS programme, and through its decentralised District Health System (DHS), it is well placed to utilise information produced by this HIV M&E system. However, research indicates that the integration of the HIV M&E system with the District Health Information System (DHIS) is not being achieved (Kawonga, Fonn, & Blaauw, 2013). The focus of this paper is to unpack the non-integration of the HIV M&E systems with the DHIS.

Integration

Integration is commonly understood as the integration of service delivery functions in health systems (Kawonga, Blaauw, & Fonn, 2012). DCPs have been created to tackle one or more related diseases in a vertical approach, while general health services (GHSs) address a wider range of diseases in a horizontal approach (Koivu et al., 2017). The administrative (M&E) integration of DCPs is envisioned by bringing DCP middle managers under the authority of GHS middle managers. This would transfer the administrative authority of DCP activities to GHS middle managers, while DCP managers provide specialist support (Kawonga, Blaauw, & Fonn, 2016). More simply, this brings the functions of DCPs under the authority of GHS managers, while DCP managers provide support on how to produce optimal results through these functions. For example, the integration of M&E functions (data collection, collation, analysis, and use) would see horizontal GHS managers being responsible for the collection and collation of disease-specific data and consult with vertical DCP managers for technical assistance “on how to use these data for management” (Kawonga, Fonn, & Blaauw, 2013).

Integration is also envisioned through consolidating two or more disease-specific services at a single delivery site, integrating disease-specific services into primary care, ensuring continuity of care over time or across different levels of care, or collaborating across various government sectors (Kawonga, Blaauw, & Fonn, 2012). In more general terms, integration involves the merging of DCP and GHS policies, management, administrative procedures, implementing activities, governance, service delivery, and financing (Kawonga, Fonn, & Blaauw, 2013).

South Africa’s health sector reform has also prioritised the decentralisation of primary health care policy and services since the end of the apartheid era. This has relied on the “devolution of political and administrative authority from the national level to nine semi-autonomous provincial” governments, and the establishment of a DHS by shifting health management responsibility from provincial to district health clusters (deconcentration) (Koivu et al., 2017).

Organisational Structure

While there has been some progress in integrating HIV services across a range of general health services, there has so far been minimal achievement of integration of HIV-related data between HIV programmes and the DHIS (Kawonga, Blaauw, & Fonn, 2012). Vertical HIV/AIDS-specific programmes and services have been developed and implemented since 1994. These have been designed at a national level by an HIV/AIDS directorate and controlled by provincial and district level HIV managers through a dedicated HIV M&E system to monitor the HIV programme (Kawonga, Blaauw, & Fonn, 2012; Kawonga, Fonn, & Blaauw, 2013). There exist several sub-programmes within the HIV programme (e.g. HIV counselling and testing (HCT), prevention of mother-to-child HIV transmission (PMTCT), and ART) that each have an accompanying set of M&E forms and reporting mechanisms to the National Treasury. These HIV programmes are mostly funded through an HIV conditional grant that is allocated to each province by the National Treasury (Kawonga, Fonn, & Blaauw, 2013). In addition, almost 25% of the government’s HIV/AIDS budget is through foreign aid such as the Global Fund for AIDS, TB and Malaria (GFATM) and the Presidential Emergency Plan for AIDS Relief (PEPFAR) (Kawonga, Blaauw, & Fonn, 2012). Funding from these Global Health Initiatives (GHIs) has been criticised for fragmenting the coordination and planning of disease-specific programmes in recipient countries and this is believed to have had an impact in the alignment or integration of these programmes within the DHS (Kawonga, Blaauw, & Fonn, 2012).

This verticalisation and fragmentation contradicts the health sector’s attempts to integrate services, information and management under the control of horizontal GHS managers at the district level (Kawonga, Fonn, & Blaauw, 2013).

District Health Management Teams (DHMTs) are responsible for this management role and the DHIS was established to facilitate this (Kawonga, Blaauw, & Fonn, 2012). The DHIS is a fundamental component of the National Health Information System that collects public sector facility-level data that is used to produce a set of district health service indicators. Ideally, a disease-specific M&E system – such as the HIV M&E system – should be a sub-component of a system-wide information system, similar to the DHIS. This would enable DHMTs to integrate HIV data into overall district health system management (Kawonga, Blaauw, & Fonn, 2012). However, the research indicates that there is a dire lack of integration of these information systems.

Top-Down Approaches

As mentioned earlier, the HIV programme and M&E framework are both designed by technical experts at the national level. Experts at this level define which HIV data should be collected and reported on (this changes as services are modified or added) and design the data recording and collation forms. Research identified several limitations of this HIV M&E system that affected its efficiency and usability which hampered the availability of HIV data. In their case study, Kawonga, Blaauw & Fonn et al., (2012) argued that the HIV M&E system is designed in a top-down and uncoordinated fashion. This M&E system is characterised by a massive dataset, duplication of data collection, incomplete data capturing and the dismissal (non-use) of nationally mandated forms. In this case study, some of the data that was collected was not integrated and analysed. Data collection forms which were subsequently introduced were not checked for whether they duplicated existing forms or worsened administrative burdens. Additionally, health workers’ perceptions of these newly introduced forms were not considered before their implementation.

Kawonga, Blaauw, & Fonn (2016) case study revealed that there is no accountability for the top-down approach of the HIV programme activities. While it is known that decision-making regarding strategies, interventions and policies should occur in a consultative process with actors at the provincial and district levels, this process has rarely reached the district level. National stakeholders placed the blame on provincial-level actors for this lack of cascading consultations because their roles did not permit them to intervene at the district-level (“can’t talk to the districts directly”). The top-down culture was also reflected in national and provincial respondents' understanding of their engagement with district-level actors as informing and training them on new and revised HIV M&E forms. Instead of genuine engagement that involved feedback and debating the communication with higher-level actors was substantially dictatorial. The large sentiment among district actors was that being informed of changes was insufficient involvement for actors who were implementers.

Furthermore, Kawonga, Blaauw, & Fonn (2016) argue that districts should be guided and coordinated by the provincial management using a performance-based approach – which establishes standard outputs. Yet the study found that in practice there is a tendency to apply a rules-based approach – which establishes standard processes. This led to a greater focus on district managers abiding to set HIV M&E procedures (how to verify data and data submission deadlines) rather than ensuring high-quality outputs (completeness of data and optimal use of data). District managers felt confined by these rules. By being unable to modify the format or content of M&E collection tools, they were prevented from the opportunity to make them more usable and thus improve the quality and use of output data (Kawonga, Blaauw, & Fonn, 2016).

The autocratic management style of provincial and national actors which disempower lower-level managers’ agency resembles machine bureaucracy. This pressures district managers and health workers to adopt decisions that do not address their local needs. Additionally, they are unable to solve problems and are excluded from policy-making processes. It is also been documented that the country’s health system increasingly invests in provincial HIV programmes (personnel numbers and skills) while ignoring district capacity development (Kawonga, Blaauw, & Fonn, 2016).

Silos Approach

A significant weakness of the HIV M&E system is that its vertical design contradicts integration and centralisation because it inherently divides data into several vertical silos through excessive and parallel reporting practices. This has seen in the proliferation of multiple data collection tools reports with different formats and deadlines which all contribute to an increased burden on the stakeholders collecting and using this data. This approach has been shown to cause gaps in the data (missing data) and substandard data quality. With limited data sharing across DCPs and GHSs, planning and reporting of collaborative activities and the effective use of this data becomes significantly difficult.

Although M&E activities for the production of HIV prevention data are integrated with the DHIS, those for the production of HIV treatment data are not, which creates silos within a silo – within the vertical HIV programme, parallel M&E processes for prevention and treatment data. This silo approach could potentially foster division in the management of HIV sub-programmes and this would contradict the goals of an integrated HIV/AIDS response. The silo approach would also limit the availability and utility of ART treatment data at the district level because this data bypasses the DHIS. This results in the country’s national policy for integrated district management to be undermined (Kawonga, Blaauw, & Fonn, 2012).

HIV managers operating outside the domains of the DHS meant that they used HIV data in even smaller sub-programme specific silos, which usually excluded horizontal managers (Kawonga, Fonn, & Blaauw, 2013). This form of verticalization is even more extreme than the usual models of verticalization in the literature (Kawonga, Fonn, & Blaauw, 2013).

Uncoordinated Organisational Roles

Any differentiated organisation must clearly assign different roles explicitly. The case study by Kawonga, Blaauw, and Fonn (2016) observed that the DHMIS mandated overlapping HIV data collection tasks for both district and programme managers. The ART M&E standard operating procedures (only for ART data) specified that HIV programme managers should periodically inspect the quality of ART data, yet it does not specify an ART data role for district managers (Kawonga, Blaauw, & Fonn, 2016). In practice, both groups of managers visited health facilities to perform HIV data quality inspections. However, this was done on separate days and district managers also looked at all health data in addition to HIV-specific data (Kawonga, Blaauw, & Fonn, 2016). This duplication and overlapping of roles created inefficiency in integration.

Overall, the HIV programme managers led HIV monitoring and organisational practices encouraged this (Kawonga, Blaauw, & Fonn, 2016). Programme managers significantly used HIV M&E data more than district managers as they were required to review local HIV service performances and propose improvements during quarterly district meetings as well as present data on HCT uptake meetings at provincial and national levels. A further exclusion of district managers was seen through their development of annual district health plans which detailed activities and targets for all services in their respective districts. These plans omitted many HIV activities and were instead developed by provincial HIV programme managers. National respondents in the study were aware of this but justified these HIV-specific plans because of the requirements of the HIV-specific conditional grants. Yet, they also implied that the lack of capacity by district managers prevented their involvement in this process.

Conclusion

The essay explores the integration challenges present in the district health information systems, in particular the HIV M&E system, in South Africa. The vertical design of HIV M&E systems, along with their inadequate integration into broader the DHIS, has resulted in fragmented data collection, limited data sharing, and ultimately, compromised decision-making processes. The essay went into detail regarding the inefficient organisation of the health information system, its top-down approach and the creation of silos. It also highlights the need for clear delineation of roles, improved coordination mechanisms, and a shift towards outcome-focused rather than process-driven approaches. Addressing these issues is crucial not only for optimizing the management of HIV/AIDS programs but also for enhancing overall health service delivery and outcomes in LMICs.

References

Kawonga, M., Blaauw, D., & Fonn, S. (2012). Aligning vertical interventions to health systems: a case study of the HIV monitoring and evaluation system in South Africa. Health research policy and systems, 10(1), 1-13.

Kawonga, M., Blaauw, D., & Fonn, S. (2016). The influence of health system organizational structure and culture on integration of health services: the example of HIV service monitoring in South Africa. Health Policy and Planning, 31(9), 1270-1280. https://doi.org/10.1093/heapol/czw061

Kawonga, M., Fonn, S., & Blaauw, D. (2013). Administrative integration of vertical HIV monitoring and evaluation into health systems: a case study from South Africa. Glob Health Action, 6, 19252. https://doi.org/10.3402/gha.v6i0.19252

Koivu, A., Hederman, L., Grimson, J., & Korpela, M. (2017). Vertical interventions and parallel structures: A case study of the HIV and tuberculosis health information systems in South Africa. Journal of Health Informatics in Developing Countries, 11(2).