Buildings Alone Don’t Heal: Ethiopian Study Finds Infrastructure Fails to Predict Real Care Quality

A well-stocked clinic with running water, reliable electricity, and a full pharmacy shelf sounds like a good clinic. But a sweeping national study from Ethiopia suggests that the link between what a health facility has and what actually happens inside its consultation rooms is far weaker than health planners have long assumed. Researchers at the Ethiopian Public Health Institute analyzed nationally representative data from the 2021-22 Ethiopia Service Provision Assessment, triangulating three distinct lenses on quality: physical audits of facilities, direct observation of clinical encounters, and exit interviews with patients. Their conclusion is uncomfortable for anyone who equates infrastructure with performance. Structural readiness, measured with World Health Organization tracer indicators, correlated only weakly with how faithfully providers followed evidence-based guidelines, and it played almost no role in determining whether patients themselves rated their care as excellent.

The study, published in BMC Health Services Research, examined three frontline services that anchor primary healthcare across low- and middle-income countries: antenatal care, family planning, and sick-child care. For each service, the research team constructed a structural readiness index ranging from zero to one, capturing the presence of essential equipment, commodities, laboratory capacity, and trained staff. The results showed moderate readiness across the board. Sick-child services averaged 0.51, antenatal care 0.54, and family planning 0.62. On paper, then, Ethiopian facilities were neither destitute nor exemplary; they possessed roughly half to two-thirds of the structural elements that international standards consider necessary for safe, effective service delivery.

What happened inside the consultation rooms told a much grimmer story. Trained observers watched clinicians work through real patient encounters and scored their adherence to national and international clinical guidelines. Providers completed, on average, only 22 percent of the recommended clinical steps for sick-child care and 24 percent for family planning. Antenatal care fared somewhat better at 47 percent, but even there, more than half of the evidence-based actions that should accompany every visit were simply skipped. These are not abstract metrics. A missed clinical step might mean failing to check a child’s vaccination status, not assessing danger signs in a pregnant woman, or omitting counseling on contraceptive side effects. Each omission represents a lost opportunity to prevent illness or death, and the aggregate picture is one of clinical practice falling dramatically short of what guidelines demand.

The most technically revealing finding concerns the relationship between the two quality domains. Using correlation analysis and restricted cubic splines, a flexible statistical method for detecting nonlinear associations, the researchers found that structural readiness predicted process quality only weakly. More striking was the dispersion: facilities with identical readiness scores varied enormously in how well their providers adhered to guidelines. Two clinics could boast the same equipment checklist and commodity availability, yet one might deliver 40 percent of recommended care steps while the other delivered barely 10. This scatter is the statistical signature of a system in which the binding constraint is not what facilities own but what providers do. Infrastructure, in other words, is a necessary but profoundly insufficient condition for good care.

The patient perspective added a third, equally dissonant layer. Between 27 and 39 percent of clients rated their overall service as excellent, a figure that might seem encouraging until it is juxtaposed with the observed clinical performance. Yet even among satisfied patients, trust was fragile: fewer than 40 percent expressed full confidence that the facility could deliver quality care if they became seriously ill the next day. That gap between satisfaction and confidence is analytically important. Patients may leave a visit pleased with how they were treated while still harboring deep doubts about the system’s capacity to handle a genuine emergency. Such conditional trust has consequences, potentially shaping decisions about where to seek care, whether to delay treatment, and whether to return at all.

So what actually drove those high patient ratings? The researchers turned to multivariable logistic regression, adjusting for a range of facility and client characteristics, and the answer was emphatically interpersonal. Shorter waiting times, respectful and clear communication from providers, and the availability of medicines were the dominant predictors of excellent quality ratings. Structural readiness and observed clinical process quality contributed little, if anything, to the odds of a patient awarding top marks. In a sense, patients were grading the encounter rather than the medicine: they noticed whether they were kept waiting, whether they were spoken to with dignity, and whether they walked away with the drugs they needed. The clinical content of the visit, largely invisible to them, barely registered.

This pattern exposes a fundamental measurement problem in global health. For decades, health systems in low- and middle-income countries have leaned on structural indicators, such as the WHO’s Service Availability and Readiness Assessment framework, as proxies for quality, largely because they are cheap and easy to collect through facility audits. The Ethiopian data demonstrate why that shortcut is dangerous. A facility can score well on every tracer item and still deliver care that ignores most of the clinical guideline. Conversely, a modestly equipped facility might house a clinician who practices meticulous, patient-centered medicine. If funders and ministries reward structural scores alone, they risk optimizing a metric that bears little relationship to health outcomes or to the experiences of the people the system exists to serve.

The findings arrive at a consequential moment. The Sustainable Development Goals have pushed governments to expand access aggressively, and Ethiopia has built one of the largest primary healthcare networks in sub-Saharan Africa. But the study’s authors frame the stakes plainly: expanding access without guaranteeing quality does not improve population health. Patients who reach a facility only to receive a fraction of recommended care have gained the illusion of coverage rather than its substance. The disconnect also carries a psychological dimension, because patients who sense respectful treatment may not realize that critical clinical steps were omitted, allowing substandard care to hide behind a pleasant bedside manner.

The policy implications cut in two directions at once. The authors are careful not to dismiss infrastructure: equipment, medicines, water, and electricity remain essential preconditions for any functional health system, and their absence makes good care impossible. But the data argue that infrastructure investment, pursued alone, has hit a ceiling. Quality improvement efforts must now directly target what happens in the clinical encounter, through strategies such as supportive supervision, clinical mentoring, guideline-based job aids, and feedback on observed practice. Simultaneously, the strong influence of waiting time, communication, and medicine availability on patient trust suggests that service organization and supply chains deserve attention not merely as logistics problems but as core determinants of whether people believe in their health system.

For a global health community increasingly focused on measuring quality rather than assuming it, the Ethiopian study offers a rare and valuable trifecta: nationally representative scope, three independent measurement approaches, and a consistent statistical story. Its message is likely to resonate far beyond Ethiopia’s borders, because the reliance on structural proxies is a global habit, not a local one. The clinics of Addis Ababa and the rural health posts of the southern regions have delivered a lesson that applies wherever health systems are judged by their inventories rather than their practice: a building can be ready for care, but readiness does not perform the care. People do, and it is their hands, their words, and their adherence to evidence that ultimately determine whether a visit to the clinic heals or merely happens.

Subject of Research: Relationships between health facility structural readiness, clinical process quality, and client-reported care experience in Ethiopia

Article Title: Does health facility infrastructure predict performance? Disconnects between service readiness, process quality, and client experience in Ethiopia

Article References: Tollera, G., Lemlem, A., Berhanu, K., Alemu, A., Girmay, A. M., & Tadele, A. (2026). Does health facility infrastructure predict performance? Disconnects between service readiness, process quality, and client experience in Ethiopia. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15578-9

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15578-9

Keywords: Ethiopia, health facility readiness, quality of care, process quality, client experience, service provision assessment, antenatal care, family planning, sick-child care, health systems, WHO tracer indicators, patient satisfaction