Over the past two decades, the global development discourse has undergone a paradigm shift toward evidence-based policy-making (EBPM), fueled by the rapid expansion of digital health infrastructures and the “data revolution.” International donors, multilateral agencies, and national governments have channeled significant resources into Health Management Information Systems (HMIS) and monitoring frameworks under the assumption that high-fidelity data automatically leads to more rational and equitable health outcomes (World Health Organization [WHO], 2023). However, this technocratic optimism often overlooks a persistent and troubling paradox: in many Low- and Middle-Income Countries (LMICs), particularly across Sub-Saharan Africa, the proliferation of data has not consistently translated into improved decision-making. Instead, health systems frequently grapple with a “decoupling” effect, where sophisticated data collection processes coexist with policy decisions driven by political expediency, patronage, and donor-defined metrics rather than localized clinical or epidemiological needs (Parkhurst, 2017).
This article argues that the primary barrier to effective evidence utilization is not a technical scarcity of information, but rather the complex political economy of the health sector. By examining the divergent paths of five African nations—Rwanda, Kenya, Ghana, Ethiopia, and Botswana—this analysis explores how institutional incentives, power dynamics, and structural constraints determine whether evidence is strategically utilized, selectively ignored, or rhetorically co-opted. Moving beyond traditional “capacity building” critiques, this study employs a Political Economy Analysis (PEA) framework to investigate the “Reporting Trap” and the influence of external funding flows on national health priorities. Ultimately, the paper seeks to provide a roadmap for transforming data from a passive administrative output into a dynamic instrument of power, equity, and systemic resilience in LMIC governance.
1. Rwanda: The High-Modernist State and the Metrics of Performance
Rwanda’s health sector is often characterized by what political scientists term “high-modernism”—a strong belief in scientific and technical progress as the primary driver of state legitimacy. Since the early 2000s, the Rwandan government has centralized health data through a robust Health Management Information System (HMIS), which serves as the backbone for national strategic planning. By prioritizing health expenditure at approximately 7.3% of its GDP, the state has signaled that data-driven outcomes are central to its “Vision 2050” development agenda (World Bank, 2024).
The primary mechanism for evidence integration in Rwanda is Performance-Based Financing (PBF). This system creates a direct, symbiotic link between verified data and fiscal disbursement. Health facilities are rewarded based on quantitative indicators—such as the number of assisted births or vaccination rates—ensuring that data collection is not merely an administrative burden but a prerequisite for financial survival. Consequently, this has fostered a culture of meticulous record-keeping that is rare in the region (Binagwaho et al., 2014).
However, the high-stakes nature of the PBF model introduces significant risks regarding data integrity and “metric fixation.” When funding is tied exclusively to specific indicators, there is an inherent institutional incentive to prioritize “measurable” outputs over “meaningful” health outcomes that are harder to quantify. The World Health Organization (2023) notes that while Rwanda’s administrative data is extensive, vital statistics—specifically cause-of-death data—remain partially modeled. This suggests a gap between the “countable” metrics used for financing and the “complex” epidemiological data required for long-term health policy.
Furthermore, the Rwandan experience highlights the tension between central control and local autonomy. While the top-down mandate for data use ensures high compliance, it may stifle local-level innovation. Frontline providers often view data as a tool for upward accountability to the Ministry of Health rather than a tool for downward accountability to the patients they serve. This dynamic underscores the reality that even the most efficient data systems can become instruments of state surveillance rather than democratic empowerment if not balanced by qualitative feedback loops.
2. Kenya: The Geopolitics of Devolution and Data Fragmentation
The 2013 devolution of health services in Kenya represented one of the most radical governance shifts in African history. By transferring health functions to 47 semi-autonomous counties, the Kenyan state sought to bring decision-making closer to the point of service delivery. In theory, this should have empowered county health management teams (CHMTs) to use localized evidence to address specific regional disparities, such as the high maternal mortality rates in the arid and semi-arid lands (ASAL) regions (Tsofa et al., 2017).
In practice, the devolution of data has been hampered by the “politicization of statistics.” County governors, who are elected officials, often view health data through the lens of political optics rather than clinical necessity. In some instances, evidence is selectively used to justify the construction of high-visibility infrastructure—such as new hospitals—which offers “political capital,” even when routine data suggests that funds would be better spent on basic primary care or medical supplies. This creates a disconnect between what the data prescribes and what the political cycle demands.
Technically, the transition to the District Health Information Software (DHIS2) has standardized data collection across Kenya, but the capacity to analyze this data remains uneven. While urban counties like Nairobi or Mombasa may possess the econometric expertise to interpret complex trends, rural counties often face a “data-rich, information-poor” (DRIP) syndrome. Without a cadre of trained health informaticians at the county level, the evidence generated by facilities often sits in digital silos, unread and unutilized for annual work plans.
Moreover, the Kenyan case illustrates the challenge of inter-county equity. As counties manage their own budgets, data becomes a weapon in resource competition. Counties that report “better” outcomes may be penalized with lower allocations from the national equitable share, creating a “perverse incentive” to under-report success or over-report need. This underscores a critical insight: for evidence to inform policy effectively, the institutional framework must ensure that “good data” does not lead to “bad fiscal outcomes” for local administrators.
3. Ghana: Donor Hegemony and the Conflict of Indicators
Ghana’s health system is frequently cited for its pioneering National Health Insurance Scheme (NHIS) and its sophisticated digital health architecture. The country has successfully integrated various data streams—including clinical, financial, and demographic data—into a centralized framework. Yet, the Ghanaian experience reveals a profound structural tension: the conflict between national sovereignty over data and the “indicator hegemony” of international donors (Agyepong et al., 2017).
Despite its middle-income status, Ghana remains heavily reliant on “vertical” funding from organizations like the Global Fund and Gavi. These entities require specific, often narrow, data reporting to satisfy their own global accountability standards. This forces Ghana’s Ministry of Health to maintain dual reporting tracks: one for national integrated health priorities and another for donor-specific projects. This “data dualism” often results in the prioritisation of donor-funded indicators (e.g., malaria bed net distribution) over critical but under-funded areas like non-communicable diseases (NCDs) or mental health.
The presence of the NHIS adds another layer of complexity to the evidence landscape. Because the NHIS generates massive amounts of claims data, it provides a unique window into healthcare utilization and provider behavior. However, this “claims-based evidence” is often used primarily for cost-containment and fraud detection rather than for improving the quality of clinical care. This highlights a missed opportunity to transform administrative evidence into a tool for systemic clinical audit and improvement.
Finally, Ghana’s journey shows that even “perfect” data systems cannot overcome fiscal constraints. When the national economy faces inflationary pressures or debt distress, evidence-based recommendations for health expansion are frequently vetoed by the Ministry of Finance. In this context, data does not “power” health policy; instead, it provides a retrospective account of what was lost during periods of austerity. This suggests that the ultimate arbiter of evidence use is often the fiscal reality of the state, not the quality of the data itself.
4. Ethiopia: The Grassroots Engine and the Aggregation Bottleneck
Ethiopia’s Health Extension Program (HEP) is globally recognized for its success in decentralized service delivery. By deploying over 40,000 Health Extension Workers (HEWs)—primarily women recruited from their own communities—Ethiopia has created a “bottom-up” data engine. These workers collect granular, household-level information on everything from latrine construction to contraceptive use, providing a level of “sociological evidence” that is rarely captured in traditional clinical systems (Assefa et al., 2019).
The strength of the Ethiopian model lies in the “trust-evidence loop.” Because HEWs are members of the community, the data they collect is generally more accurate than that gathered by external surveyors. This localized evidence allows for “micro-planning,” where health interventions are tailored to the specific needs of a village or district. For example, data indicating a local spike in diarrheal diseases can lead to an immediate community-led sanitation campaign, demonstrating a rare instance of data-to-action at the frontline.
However, Ethiopia faces a significant “aggregation bottleneck.” As this rich, qualitative, and quantitative data moves from the village (kebele) to the district (woreda) and eventually to the federal level, it undergoes massive compression and abstraction. By the time the data reaches the Federal Ministry of Health, the local context that made it actionable is often stripped away, leaving only “flat” statistics. This loss of granularity limits the ability of national policymakers to understand the nuanced drivers of health inequities in diverse regions like Tigray or Oromia.
Furthermore, the recent push for digital transformation through the Electronic Community Health Information System (eCHIS) has introduced new challenges. While digital tools reduce the burden of manual reporting for HEWs, they also require reliable electricity and internet connectivity—luxuries that are not always available in rural Ethiopia. The Ethiopian case thus serves as a reminder that the “last mile” of data collection is the most critical, yet the most vulnerable to the “digital divide” and the bureaucratic tendency toward over-simplification.
5. Botswana: Infrastructure Success vs. Epidemiological Complexity
Botswana represents a unique case in Sub-Saharan Africa: a stable, upper-middle-income country with a highly developed physical health infrastructure. With nearly 95% of the population living within 5 km of a facility, Botswana has solved the “access” problem that plagues many of its neighbors (Ministry of Health, Botswana, 2023). However, the country’s struggle with one of the world’s highest HIV/AIDS prevalence rates (approximately 20.8%) underscores that “facility evidence” is not the same as “epidemiological control.”
In Botswana, the evidence-based approach has historically been dominated by a “biomedical model.” The state has successfully used data to roll out universal Anti-Retroviral Therapy (ART), making it one of the first African nations to reach the UNAIDS 95-95-95 targets. However, the persistence of high new infection rates among young women suggests that the “social evidence”—data regarding gender-based violence, economic migration, and behavioral norms—has been undervalued in the policymaking process (UNAIDS, 2023).
The Botswana case also highlights the “Middle-Income Trap” of health data. Because the country is relatively wealthy, it receives less technical assistance and oversight from international agencies than poorer nations. While this increases “data sovereignty,” it can also lead to institutional complacency. Internal data systems may not be subject to the same rigorous external audits as those in donor-dependent countries, potentially masking systemic inefficiencies or gaps in service quality that “infrastructure-only” metrics fail to capture.
Conclusion
Ultimately, Botswana proves that data-driven policy must be multi-sectoral. The health system’s reliance on clinical data alone has limited its ability to address the “social determinants of health” that drive the HIV epidemic. To move forward, the country must integrate its robust biomedical data with social science evidence to create policies that are not just “evidence-based” in a laboratory sense, but “context-aware” in a societal sense. This requires a shift from viewing health as a technical output to viewing it as a complex social outcome.
The evidence presented across the case studies of Rwanda, Kenya, Ghana, Ethiopia, and Botswana suggests that the primary bottleneck in international development is no longer the scarcity of data, but the pathology of its application. For decision-makers, the transition toward a truly evidence-informed health system requires a departure from purely technocratic solutions in favor of a political-economy approach.
1. Institutionalizing Evidence: From Reporting to Accountability
Decision-makers must shift the primary function of data from upward reporting (to donors and central ministries) to downward accountability (to citizens and frontline providers).
- Actionable Step: Integrate data reviews into mandatory legislative and budgetary cycles. Resource allocation should not merely be “informed” by data but should require a “data-audit” justification for significant shifts in funding.
- The Goal: To create a “cost of ignoring evidence,” where political decisions that contradict clear epidemiological or service-delivery data require public or institutional justification.
2. Resolving the “Indicator Hegemony”: Reclaiming National Data Sovereignty
As seen in the Ghanaian and Ethiopian contexts, donor-driven priorities often create “vertical” data silos that fragment the health system.
- Actionable Step: National governments should enforce a “One National Information System” policy, requiring all external partners to align their metrics with national health goals. This reduces “indicator fatigue” and ensures that data collection serves the horizontal strengthening of the entire system rather than the monitoring of isolated disease-specific projects.
- The Goal: To ensure that local health needs—rather than the reporting requirements of the Global North—dictate the evidence agenda.
3. Bridging the “Aggregation Gap” with Digital Literacy
While digital tools like DHIS2 and eCHIS provide the infrastructure for data flow, they do not guarantee data literacy.
- Actionable Step: Invest in “Mid-Level Data Leadership.” Training should move beyond IT staff and data clerks to include County Directors, District Health Officers, and Hospital Managers. These leaders must be equipped not just to collect data, but to interrogate it for clinical and operational improvement.
- The Goal: To transform data from a passive administrative output into a “management intelligence” tool that informs daily facility-level decisions.
4. Expanding the Evidence Lens: Integrating Social and Behavioral Data
The case of Botswana highlights that biomedical and infrastructural data are insufficient for tackling complex public health crises.
- Actionable Step: Incorporate qualitative “social evidence”—such as patient satisfaction surveys, community feedback loops, and ethnographic studies—into formal health assessments.
- The Goal: To move from a “high-modernist” focus on facility density and drug supply to a holistic understanding of how social determinants (gender, poverty, and geography) influence health-seeking behavior.
The “Knowledge-to-Action” gap is fundamentally a gap in incentives. In systems where data is used for punishment or purely for donor compliance, it will be manipulated or ignored. In systems where data is a resource for problem-solving, equity, and resource optimization, it becomes an instrument of power.
For the modern development practitioner, the challenge is not to build better databases, but to build better institutional cultures. This requires a commitment to transparency, a willingness to confront political interference in statistics, and a shift toward “epistemic justice,” where the data generated at the community level is given the same weight as the models generated in international capitals. Only through this transformation can data move from being a “passive output” to an “active driver” of systemic resilience and health equity across the African continent.
References
- Agyepong, I. A., et al. (2017). The path to universal health coverage in Nigeria, Ethiopia, and Ghana: Evidence-based or politically driven? The Lancet Global Health, 5(5), e480-e481. https://doi.org/10.1016/S2214-109X(17)30102-1
- Assefa, Y., et al. (2019). The Health Extension Program of Ethiopia: Data-driven successes and challenges. Journal of Global Health, 9(2), 020415. https://doi.org/10.7189/jogh.09.020415
- Binagwaho, A., et al. (2014). Rwanda: Monitoring and evaluation under a decentralized health system. Global Health Action, 7(1), 26785. https://doi.org/10.3402/gha.v7.26785
- Ministry of Health, Botswana. (2023). Annual health statistics report. Government Printers, Gaborone.
- Parkhurst, J. (2017). The politics of evidence: From evidence-based policy to the good governance of evidence. Routledge.
- Tsofa, B., et al. (2017). Devolution and health planning in Kenya: Institutional barriers to evidence use. Health Policy and Planning, 32(4), 528-537. https://doi.org/10.1093/heapol/czw164
- UNAIDS. (2023). Country factsheets: Botswana 2023. https://www.unaids.org/en/regionscountries/countries/botswana
- World Bank. (2024). Current health expenditure (% of GDP) – Rwanda. World Bank Open Data. https://data.worldbank.org
- World Health Organization. (2023). World health statistics 2023: Monitoring health for the SDGs. WHO Press.
Felix Rutayisire is a researcher and evaluation specialist with expertise in international development, global governance, and evidence-based policy analysis. His work focuses on monitoring and evaluation, institutional effectiveness, and the implications of global policy shifts for development practice in Africa and beyond.