When Guidelines Backfire: Why Evidence Struggles to Cross Borders Into Middle-Income Hospitals

Evidence-based medicine has transformed modern healthcare by replacing tradition and authority with rigorous clinical trials as the foundation of decision-making. Clinical guidelines have standardized care, reduced unwarranted variation between hospitals, and allowed proven interventions to be scaled across entire health systems. Yet a provocative new viewpoint published in the Journal of General Internal Medicine argues that this triumph has a blind spot: evidence does not travel intact. According to authors Vinícius Brenner Felice and Cassiano Teixeira, two intensive care physicians from Porto Alegre, Brazil, the effectiveness of any recommendation depends not only on the methodological rigor behind it but also on the health-system environment into which it is rolled out. In middle-income countries such as Brazil, they contend, the central question is often not simply what works, but what works here and now.

The authors draw on recent analyses showing that guideline implementation in low- and middle-income settings frequently collides with systemic barriers, including limited staffing, constrained infrastructure, fragmented care pathways, and difficulty keeping recommendations current or applying them consistently. In such contexts, the gap between efficacy and effectiveness widens dramatically. An intervention that performs beautifully under the controlled conditions of a well-funded trial may fail to deliver meaningful benefit when the surrounding ecosystem cannot support its safe implementation. This is not a marginal concern. Middle-income countries serve billions of people, and their hospitals increasingly face pressure to adopt the same cutting-edge standards promoted by international guideline bodies, even when the local scaffolding required to make those standards work is missing.

Mechanical ventilation provides one of the most striking illustrations of the problem. Debates about ventilator asynchronies, optimal driving pressure, fine-tuning of positive end-expiratory pressure, and advanced weaning strategies are scientifically relevant and clinically valuable in mature systems. But their relative importance shifts dramatically when baseline mortality among ventilated patients is extremely high. In a Brazilian university hospital cohort, mortality among mechanically ventilated patients reached 51 percent. In a nationwide analysis of the first 250,000 COVID-19 admissions in Brazil, mortality among invasively ventilated patients approached 80 percent. Under such circumstances, the authors argue, the most impactful intervention may not be refining ventilator settings at all, but rather stepping back to strengthen upstream and downstream processes: avoiding both delayed and unnecessary intubations, scaling up staff training, cutting down deep sedation, and rolling out structured liberation protocols.

The lesson here is counterintuitive but powerful. When outcomes are overwhelmingly poor, incremental physiological optimization may matter less than systemic redesign. A hospital struggling to keep ventilated patients alive at rates far worse than international benchmarks gains little from adopting the latest evidence on driving pressure thresholds if its fundamental processes of care remain broken. The viewpoint suggests that clinicians and policymakers in these settings face a triage problem of their own: not which patient to treat first, but which improvement to pursue first. And the answer, more often than not, lies in the unglamorous machinery of healthcare delivery rather than in the frontier of critical care science.

Surgical oncology raises similarly uncomfortable questions. It is entirely legitimate to debate the optimal perioperative or adjuvant therapy after esophagectomy, one of the most demanding operations in cancer medicine. Yet those discussions lose practical traction when perioperative mortality remains high because of limited capacity to rescue patients from complications. In the international OGAA cohort, patients undergoing esophagectomy in low- and middle-income countries experienced 90-day mortality of 9.4 percent, compared with 3.7 percent in high-income settings, despite similar rates of major complications. This discrepancy strongly suggests a failure-to-rescue phenomenon: patients suffer the same complications everywhere, but in some hospitals the systems needed to detect deterioration early and intervene decisively simply are not there.

The implications of that finding are stark. In hospitals where perioperative mortality approaches 30 percent, the authors write, it becomes ethically and organizationally necessary to ask whether centralization of care, volume consolidation, and investment in multidisciplinary perioperative pathways should take precedence over debates about the newest immunotherapy regimen. A patient operated on in a low-volume facility without robust rescue capacity faces a fundamentally different risk profile than a patient with the same tumor treated in a high-volume center, regardless of which drug regimen follows the surgery. The ethical dimension is hard to escape: offering the most advanced therapy within a fragile system may be less defensible than ensuring the patient survives the operation itself.

Stroke care offers another perspective on what the authors call contextual effectiveness. Establishing dedicated stroke teams and stroke units is widely recommended worldwide, yet their benefit in resource-limited environments often stems less from high-technology reperfusion therapies and more from getting the basics right. Evidence highlights the importance of early dysphagia screening, prevention of aspiration pneumonia, early mobilization, nutritional support, and coordinated discharge planning. Brazilian data indicate that dysphagia occurs in up to 76 percent of stroke patients and is associated with markedly increased pneumonia risk, and a recent meta-analysis confirmed that post-stroke dysphagia is linked to higher odds of pneumonia and mortality. The question, therefore, is not whether stroke units are worthwhile, but how they are structured. Units that focus narrowly on imaging and thrombolysis while overlooking fundamental supportive care may appear modern without actually being effective.

Sepsis management perhaps most clearly illustrates the limits of guideline transferability. The Surviving Sepsis Campaign has played a pivotal role in standardizing sepsis care across the globe, yet even its own recommendations acknowledge variability in resource availability across settings. The randomized trial by Andrews and colleagues in Zambia demonstrated that an early resuscitation protocol incorporating aggressive fluid administration, vasopressors, and transfusion actually increased in-hospital mortality compared with usual care. Similarly, the landmark FEAST trial showed that fluid boluses increased short-term mortality among African children with severe infection in environments lacking equivalent critical care support. These findings do not invalidate sepsis physiology or the global guidelines built upon it. Rather, they underscore a sobering principle: interventions proven beneficial in resource-rich contexts may turn harmful when scaled up without the infrastructure required to carry them through safely.

From these examples, the authors construct a more nuanced hierarchy of priorities. First-order priorities include adequate staffing, early recognition of deterioration, prevention of complications, timely access to antibiotics and oxygen, safe airway management, nutritional support, rehabilitation, and effective teamwork. Second-order priorities involve fine-tuning advanced therapies once basic systems are functioning reliably. Too often, they argue, health systems in middle-income countries attempt to import second-order debates into environments where first-order problems remain unresolved. This mismatch, they insist, is not merely technical but ethical. A more mature use of evidence-based medicine in such contexts may require filtering every recommendation through three lenses: external validity, implementation feasibility, and opportunity cost. If aspiration pneumonia after stroke remains common, investing in dysphagia screening and nursing training may yield greater benefit than adopting sophisticated imaging algorithms. If perioperative mortality remains high, strengthening rescue capacity may matter more than selecting the latest targeted therapy. If septic patients die because they cannot access timely airway support or monitoring, strict adherence to care bundles may not address the core failure.

Ultimately, the viewpoint concludes that evidence does work in countries like Brazil, but not as a plug-and-play solution. Its real-world effectiveness hinges on the resilience of the system that sustains it. Guidelines are most valuable when they help clinicians and policymakers sort out priorities, spell out prerequisites, and avoid decontextualized interventions. In fragile settings, the authors suggest, true sophistication may lie not in scaling up the newest recommendation but in stepping back to ask the right question at the right time. Often, that question is not which advanced therapy should come next, but rather why the system is still falling short on the basics. For a global medical community increasingly shaped by shared guidelines and universal benchmarks, that reframing may be the most important prescription of all.

Subject of Research: Contextual effectiveness and transferability of evidence-based clinical guidelines in middle-income health systems

Article Title: Does Evidence Travel Well? Rethinking Guidelines and Real-World Effectiveness in Middle-Income Health Systems

Article References: Felice, V. B., & Teixeira, C. (2026). Does Evidence Travel Well? Rethinking Guidelines and Real-World Effectiveness in Middle-Income Health Systems. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10867-9

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10867-9

Keywords: evidence-based medicine, clinical guidelines, middle-income countries, global health, mechanical ventilation, sepsis, stroke care, surgical oncology, failure to rescue, health systems, Brazil, implementation science