Our analysis reveals that maternal health-seeking behavior in the DRC is driven less by physical infrastructure or observable facility readiness, and more by how women perceive and experience care within the health system. Across two rounds of linked household and facility data, perceived quality of care consistently emerged as the strongest predictor of key outcomes—including initiation of ANC, receipt of PNC, and discussing potential complications. Women who reported respectful treatment, clear communication, and cleanliness and courtesy were significantly more likely to initiate care with the health system, independent of objective facility readiness. Overall, for ANC1, positive perceptions increased the odds of care seeking by +29-61% (odds ratio) and were significant for 85% of the perceptions that were evaluated. For PNC, positive perceptions increased the odds of care seeking by +63-316% (odds ratio) and were significant for 77% of the perceptions that were evaluated.

Using the same dataset, others have found that supply-side readiness has modest impact on ANC clinical quality. [18] However, this was limited to only looking at women who had already chosen to attend ANC, and our findings here suggest that structural readiness alone is insufficient to drive care-seeking without corresponding improvements in patient experience and perceived responsiveness. This highlights the individual-dependent nature of health-seeking behavior, where outcomes hinge on interpersonal trust and perceived responsiveness of care.

Similarly, perceptions of affordability were more predictive of care utilization than direct measures of financial hardship—a dynamic shaped by both individual-level expectations and population-level economic vulnerability. This relationship was further reflected in the specifics of coping behaviors: women in financially constrained households often sold assets to pay for care. The type of asset sold—such as food reserves versus agricultural equipment—was associated with distinct care trajectories, suggesting that the economic decisions available to a household (i.e., the kind of assets they own) affect their choices, and how they make trade-offs depends on the anticipated benefit of getting medical care. Taken together, the results indicate that individuals’ perceptions of healthcare services—more than structural readiness or explicit costs—are central determinants of maternal care-seeking behavior. Understanding how these perceptions are formed, reinforced, and transmitted is therefore critical—not only because they guide personal decisions, but because they shape collective trust and community norms around service use.

These perceptions are not secondary or peripheral aspects of health systems—they are fundamental drivers of care utilization. Women’s perceptions of quality of care are formed primarily through tangible interpersonal experiences with providers (e.g., being treated with respect, receiving clear explanations, feeling heard) and through direct assessments of facility conditions (e.g., cleanliness, drug availability, and confidence in the system’s ability to deliver effective care). In fragile health systems, these impressions carry disproportionate weight. If care is perceived as unsafe, disrespectful, or futile, women may forgo it—even when services are technically available and helpful. Conversely, when care is perceived as responsive, kind, and trustworthy, it can catalyze care-seeking and continued engagement.

This dynamic becomes even clearer when recognizing that no interaction with the health system exists in isolation: a negative experience during antenatal care can undermine trust in subsequent skilled deliveries or childhood immunizations, and one woman’s encounter with disrespect or stock-outs can ripple through her social network, discouraging others from seeking care.

Moreover, in places with pervasive financial insecurity such as rural DRC, these perceptions are further shaped by households’ ongoing evaluations of whether accessing care is justified, given their limited resources.

As expected, perceived unaffordability reduces care-seeking. More telling, however, are the specific coping strategies households employ—such as selling food reserves or livestock—which reveal both economic vulnerability and varying levels of commitment to pursuing care. Our findings reflect this nuanced behavioral calculus, echoing broader calls for a “quality revolution” that positions trust, person-centeredness, and safety not as optional extras, but as essential pillars of effective health systems [19], [20], [21] that can both improve the actual care delivered and increase coverage. Consistent with this vision, our results underline the importance of improving perceptions of quality, which requires both investments in the experience of care (e.g., improved cleanliness) as well as communicating improvements that have occurred (e.g., more convenient open hours). This will ensure that health systems are not merely accessed, but trusted, valued, and sustainably utilized.

Beyond these individual and perception-based determinants, our findings highlight how health financing structures also shape maternal care-seeking behavior. Specifically, we observe notable differences in women whose closest facility was either receiving PBF or DFF, suggesting that the chosen financing model influences not only the types of services delivered but also how these services are experienced by patients and perceived by communities.

In DFF facilities, women’s perceptions play a more significant role in their decisions to seek care. Without strict performance-based conditions, facility responsiveness to improving patient experience may vary considerably, thereby elevating the importance of patient perception as an indicator of quality. As a result, women’s subjective experiences and assessments become central factors guiding their decisions about whether to initiate or continue care. By contrast, PBF models standardize service provision around specific incentivized indicators, reducing variability in service delivery across facilities. This standardization can diminish the relative influence of patient perceptions in shaping care-seeking behavior. Thus, financing models not only shape how providers deliver services but also influence how patients perceive and respond to care.

Furthermore, we observed that DFF facilities often performed particularly well in service areas aligned with high-burden local priorities, such as malaria care. Compared to PBF, where centrally defined incentive structures guide investments, the greater flexibility in resource allocation under DFF may encourage facilities to respond more directly to local epidemiological conditions and community-specific health needs. This complements the more standardized improvements observed under PBF—particularly in vaccination, tuberculosis, prenatal and postnatal services—where performance indicators were more explicitly targeted [22].

These findings align with the recommendations of the 2022 Lancet Commission on Primary Health Care Financing, which emphasized that local autonomy, flexibility, and facility-level accountability are critical components of high-quality primary care systems [23]. According to the Commission, overly rigid financing frameworks may limit a facility’s ability to address localized health priorities, especially where the epidemiology and social barriers are variable and require adaptive responses. More broadly, this contrast between DFF and PBF shows how financing models shape not only how well the health system performs, but also how patients experience and respond to care being made available. Given that expressed demand is not equivalent to local health needs, this complexity needs to be further explored and reflects a systems-dependent dynamic, where the structure and design of the delivery platform play a key role in determining the success of an intervention.

We note that while these findings offer important insights into how perceptions and financing structures shape maternal care-seeking behavior and reinforce previous results, they should be interpreted with caution. The use of cross-sectional data at each time point limits causal inference and prevents us from observing within-person changes over time. Temporal ambiguity in key survey items—such as the sequencing of complications and provider interactions—further constrains our ability to determine directionality. Additionally, differences in geographic coverage between baseline and endline may introduce unrelated variation, limiting comparability across timepoints.

Our findings demand a redefinition of health system performance, placing patient experience firmly alongside structural readiness and clinical quality as a core objective—not merely as an indicator. Respectful, responsive, and trusted care should be explicitly tracked as critical components of high-quality service delivery. Metrics capturing trust, interpersonal quality, and perceptions of affordability must be integrated systematically into routine quality assessments, using independently collected data to ensure objectivity and prevent bias. Additionally, tracking the types of assets households sell and the proportion experiencing financial hardship could provide actionable insights into perceived affordability and economic vulnerabilities that influence care-seeking behaviors.

Financing models could incentivize these patient-centered metrics, but caution is required to avoid perverse incentives, such as providers pressuring patients into favorable reporting—a practice that can rapidly undermine the trust that we have found to be so important. Crucially, efforts to scale DFF should integrate incentives explicitly tied to facility responsiveness to patient feedback, reinforcing frontline autonomy and adaptability. Positioning DFF as a genuinely people-centered delivery platform contrasts favorably with more rigid PBF approaches, making health systems not only more responsive, but fundamentally more trusted by the communities they serve.

Contrary to traditional approaches to investing in primary healthcare, substantially greater emphasis must be placed on factors shaping patient perceptions—friendliness, cleanliness, and affordability—if we aim to fundamentally improve coverage. Perception data is not merely a satisfaction score; it is a proxy for trust, a powerful predictor of utilization, and the critical missing link in primary health care systems that aspire to be genuinely people centered.