Assessing the quality of care in low- and middle-income countries: a rapid review of the landscape of assessment tools

Key findings This rapid review on the ‘tools used to assess each component of quality of care—according to Donabedian’s model—in LMICs, and their reported validity and applicability in routine health system settings’, presented a very complex landscape. What tools have been used to assess the quality of care in LMICs, and how can these tools be categorised according to Donabedian’s model (input, process, outcome)? A wide range of tools has been employed to assess the quality of care in LMICs, reflecting varying degrees of complexity, heterogeneity and alignment with Donabedian’s model of quality (structure/input, process and outcome). These tools can be broadly categorised into institutional tools—such as SARA, DHS tools and the PSNCQQ—and investigator-developed tools, which are often tailored to specific research contexts. Under the input component, institutional tools like SARA and DHS provide scalable, standardised assessments of infrastructure, equipment, workforce and essential medicines. Investigator-developed tools were also widely used, often drawing inspiration from the institutional instruments. In the process component, most studies employed investigator-developed tools to assess clinical practices, provider–patient interactions and adherence to protocols. These tools are context-specific and flexible but vary considerably in their methodological rigour. Regarding the outcome component, patient satisfaction was the most frequently assessed aspect, commonly measured through adapted investigator-developed tools or institutional instruments such as the PSNCQQ. While these tools are often validated and offer valuable insights into users’ perceptions, they primarily capture subjective experiences. Objective health outcomes were rarely assessed, likely due to contextual constraints such as limited resources, fragmented data systems and insufficient capacity for longitudinal follow-up. A key finding of our review is the marked heterogeneity of the tools employed across studies. Several factors might explain this diversity. Indeed, the health systems of LMICs are themselves heterogeneous, varying widely in governance structures, financing mechanisms, service-delivery models and existing data infrastructures.98 These systemic differences are likely to shape distinct measurement needs that encourage the development of context-specific tools. For instance, while institutional tools like SARA and DHS offer standardised and comparable metrics, they are often considered insufficiently granular to capture the nuances of daily clinical practice or the specificities of local disease burdens. This limitation might drive researchers to design investigator-developed tools that more closely align with national or subnational priorities—whether maternal health, infectious diseases or non-communicable conditions—thereby contributing to variability in tool scope and content. Moreover, heterogeneity might also reflect the historically fragmented architecture of donor-driven and vertical programmes in LMICs. For example, over time, global health initiatives in fields such as HIV, malaria, tuberculosis and Reproductive, Maternal, Newborn, and Child Health have each promoted their own measurement frameworks, resulting in multiple parallel tools that continue to coexist rather than converge. This layering effect could produce a landscape in which health facilities and researchers could choose among—or combine—multiple, non-harmonised approaches, further reinforcing diversity in measurement practices. On the other hand, operational and methodological constraints—such as limited staffing, insufficient training in data collection, weak routine information systems and unreliable record-keeping—shape what is practically measurable in real-life settings. These constraints often necessitate simplifying, adapting or selectively applying tools to fit local capacity, resulting in additional variation in the depth and rigour of assessments. Finally, the absence of a global consensus on minimum quality-of-care standards across all levels of the health system means that LMIC stakeholders retain substantial discretion in determining which indicators to prioritise. As a result, the tools in use reflect both the flexibility afforded by this lack of standardisation and the need to respond pragmatically to resource and capacity constraints. These contextual, institutional and methodological factors might explain why quality of care assessment in LMICs relies on a heterogeneous array of tools that differ in scope, depth, content focus and alignment with Donabedian’s framework. What information was reported regarding the validity of the tools used in the included studies? Across the included studies, comprehensive information on validity of the quality assessment tools was sparse and inconsistently reported. Most tools—particularly investigator-developed—underwent only face or content validity checks, often based on expert opinion or adaptation from prior tools, without formal documentation of the validation process. For tools measuring subjective constructs, such as patient or provider experience and patient satisfaction, psychometric testing (eg, factor analysis, internal consistency measures like Cronbach’s alpha) was rarely conducted, and inter-rater reliability assessments were seldom reported. A few exceptions included standardised tools such as the PSNCQQ, which had documented construct validity and reliability coefficients, but these were applied in limited settings and without cross-cultural revalidation. Institutional tools, while methodologically standardised, also lacked detailed reporting of validation procedures in the reviewed papers. Their reliability is typically assumed based on institutional origin, yet empirical evidence of context-specific adaptation or testing was largely absent. What information was reported regarding the applicability of the tools used in the included studies? (Did included studies provide evidence on the cost-effectiveness of the tools? Did included studies’ distribution and content demonstrate the relevance of the tools to varying levels of technical capacity and the overall health systems?) The applicability of tools used to assess quality of care in LMICs was rarely examined in a structured manner across the included studies. Cost-effectiveness was not reported in any study, and very few tools demonstrated clear alignment with national systems’ operational and financial constraints. Notably, most assessments were one-off, research-driven initiatives, with only four led by national health authorities, underscoring the reality that these tools were largely designed for academic purposes rather than for sustained, system-integrated quality monitoring. While the implementation of institutional tools offers broad validity and scalability, it seems often resource-intensive and typically relies on external funding. Conversely, investigator-developed tools, though often contextually tailored, frequently lacked standardisation and were rarely embedded into routine practice. They also tend to oversimplify quality constructs or overlook systemic factors that influence care. Actually, the absence of cost-effectiveness reporting found in our review does not necessarily imply that such evidence is nonexistent in the broader quality-of-care literature. Rather, it highlights a gap in LMIC research, where economic evaluation of quality-assessment tools is rarely undertaken. This gap likely stems from several intersecting factors.98 First, many quality assessments in LMICs are conducted within donor-driven or vertical programmes that emphasise data collection for programmatic monitoring but seldom require or sponsor economic evaluations of the measurement tools themselves. Second, Ministries of Health often operate under resource constraints, prompting prioritisation of service delivery and workforce needs over cost analyses of assessment tools. Third, methodological challenges—such as fragmented information systems, limited routine data and the difficulty of attributing costs across multiple partners—make economic evaluation technically demanding. These contextual limitations reduce the likelihood that implementers or researchers will systematically measure the cost, cost-efficiency or return on investment of quality of care assessment tools. Insights from high-income countries (HICs) illustrate that even in resource-rich settings, the cost-effectiveness of quality of care assessment is not consistently examined. Economic evaluations in HICs tend to focus on system-level quality improvement strategies—such as hospital accreditation, clinical audits, pay-for-performance mechanisms or electronic quality-monitoring platforms—rather than on the cost-effectiveness of individual assessment tools.99–101 Although these studies demonstrate that robust quality of care assessment mechanisms can be costly,100 their findings are not readily transferable to LMIC contexts due to substantial differences in workforce capacity, financing, digital infrastructure and service-delivery models. Thus, even if our search strategy had explicitly incorporated economic evaluation terms (eg, cost analysis, implementation cost, economic evaluation of quality tools), the yield of cost-related evidence for LMICs would likely still have been minimal. The absence of LMIC-focused economic evaluations therefore appears to reflect a structural void in the global literature rather than a limitation of our review strategy. Finally, the limited reporting on the operational applicability of the tools—such as resource requirements, training needs, integration with routine systems or sustainability—further illustrates the lack of systematic attention to feasibility in real-world health system environments. These findings indicate that the economic, logistical and institutional conditions under which quality of care tools are implemented remain an underexplored dimension of the field. Addressing this gap will be essential for informing future investments in quality of care assessment, particularly in settings where resources are scarce and measurement capacity must be carefully aligned with health-system priorities. Comparison with existing reviews on quality assessment tools Our findings deepen and extend the contributions of previous reviews by Fracolli et al,102 Brizuela et al103 and Quach et al,104 which focused respectively on primary care, maternal and newborn health and child health facility assessments. These earlier reviews documented the diversity of available quality-of-care assessment tools—including SPA, SARA, WHO Hospital Care Tools, Primary Care Assessment Tool (PCAT)—and provided detailed descriptions of their content and methodological characteristics. However, their analytical scope remained confined to specific clinical sectors. As a result, they did not fully capture how such tools are deployed across the broader, more heterogeneous health-system landscape in LMICs. By examining a wider spectrum of clinical areas and levels of care, our study was able to identify cross-cutting patterns of tool selection and use that remain invisible in more narrowly focused analyses. A central contribution of our review is the demonstration of a substantial gap between validated tools described in previous literature and the tools actually used in empirical studies conducted in LMICs. Several standardised tools highlighted in earlier reviews—such as the WHO Primary Care Evaluation Tool, the ADHD Questionnaire for Primary Care Providers, the General Practice Assessment Questionnaire, PACOTAPS and the PCAT—were entirely absent from the studies included in our sample. Although these tools have been validated for assessing various dimensions of quality of care, their operational applicability in LMIC contexts remains insufficiently examined. Their absence in practice likely reflects constraints related to time, training requirements, data infrastructure and financial resources, which collectively reduce their feasibility for routine use in resource-limited health systems. Furthermore, many of these tools primarily emphasise input and process components of care, mirroring the partial coverage observed in several of the tools identified in our review. In relation to previous literature, our study both confirms and extends several recurrent observations. Consistent with Brizuela et al,103 we found a predominance of input and process components coverage over outcome component. Similar to Quach et al,104 our review highlighted that even the most comprehensive tools—such as WHO Hospital Care Assessment tools—remain insufficient to capture all relevant components of care quality and are infrequently employed in LMIC settings. Echoing the conclusions of Fracolli et al,102 our findings show that validated tools for quality of care assessment remain largely absent in empirical studies in LMICs, despite their potential usefulness. Taken together, these comparisons underscore persistent gaps in the harmonisation, comprehensiveness and practical uptake of quality of care assessment tools across diverse clinical domains and health-system levels.