Beyond Primary, Secondary, and Tertiary:
A Practice-Derived Operational Architecture for Healthcare System Planning, Delivery, and Financing

Author: Dr. Syed Sabahat Azim, MBBS, Ex-IAS - Founder & CEO, Zoya Technologies (Zoyel); Founder & Chairman, Glocal Healthcare Systems Pvt Ltd

Abstract

Healthcare systems worldwide are planned, financed, and regulated around a classification framework - the primary, secondary, and tertiary tier system - that was conceptualized in 1920 and formalized at Alma-Ata in 1978. Despite its ubiquity, this framework systematically fails to answer the questions most material to health system planners: where should a facility be located? What infrastructure does it require? Which services can be delivered digitally? How should financing be structured?

The framework proposed in this paper emerged not from theoretical exercise but from fifteen years of operational healthcare system design and implementation across eastern and northern India, encompassing more than 1,100 hospital beds across West Bengal, Uttar Pradesh, Bihar, and Odisha. Repeated operational observation revealed that two variables - time sensitivity (acuity) and infrastructure dependency (intervention requirement) - are far more predictive of planning requirements than institutional hierarchy or medical specialty. A 2×2 classification matrix built on these two axes yields four operationally coherent quadrants, each with distinct implications for facility location, workforce, digital delivery architecture, financing design, and cluster development. The framework's larger thesis is that healthcare infrastructure should be organised according to operational systems logic rather than institutional tradition or specialty hierarchy.

1. Introduction: The Failure of the Tier Model

Every health system requires a taxonomy. Without one, planners cannot decide where to build, insurers cannot design rational benefit packages, and policymakers cannot allocate capital. The primary-secondary-tertiary framework has provided that taxonomy since the Dawson Report of 1920 and its formal adoption into World Health Organization doctrine through the Declaration of Alma-Ata in 1978. It is the framework taught in every school of public health, embedded in every national health policy, and assumed in every multilateral health financing discussion.

The persistence of the framework reflects institutional continuity and the absence of a widely adopted operational alternative rather than demonstrated superiority for contemporary planning problems. This paper offers one.

The primary-secondary-tertiary classification exists in two parallel and mutually incompatible forms. According to the World Health Organization, the levels of care constitute a comprehensive approach to health systems: primary care is the first point of contact, secondary involves specialized acute care, and tertiary offers highly specialized, complex, or rehabilitative care. In market usage, the tiers describe institutional complexity and cost: general practice and first-contact care (primary), standard hospital procedures including surgery, medicine, and routine hospitalization (secondary), and complex specialist or sub-specialist care (tertiary), with a contested quaternary tier for extreme complexity. These two frameworks are not reconcilable - a community health worker providing preventive nutrition education is primary under WHO logic but has no natural place in the market hierarchy; a high-volume specialist cataract surgeon is tertiary under market logic but largely rehabilitative under WHO logic.

More importantly, neither version answers the questions that matter most to those who must actually design and build health systems. Neither tells a planner where to site a facility, what that facility needs to contain, whether its services can be delivered digitally, or how the financial incentives around it should be structured. The consequence of this analytical gap is visible across health systems worldwide: hospitals designed for complex care overwhelmed by outpatient demand that could be managed in lower-cost settings; insurance schemes creating friction precisely where the system should be frictionless; digital health investment made without principled guidance on which services can safely migrate to remote delivery; specialist care either over-distributed (reducing volume and outcomes) or under-distributed (reducing access), with no clear principle to govern the balance.

This paper argues that a 2×2 matrix built on two operationally meaningful axes resolves these problems. The framework is not a theoretical proposal. It is a codification of operational insight derived from fifteen years of healthcare system design and implementation in resource-constrained settings across India.

2. Origin and Method: A Practice-Derived Framework

The framework proposed in this paper emerged from approximately fifteen years of healthcare system implementation experience through Glocal Healthcare Systems Pvt. Ltd., spanning multiple states in eastern and northern India including West Bengal, Uttar Pradesh, Bihar, and Odisha, with the network ultimately encompassing more than 1,100 hospital beds.

The earliest operational experiments began in Sonamukhi, in Bankura district of West Bengal - a semi-rural region with severe access deficits and limited public health infrastructure. The initial vision attempted to establish a fully comprehensive hospital model: a facility that could serve the population across the full clinical spectrum. While clinically ambitious, this model rapidly revealed fundamental structural tensions. Comprehensive care under one roof required infrastructure intensity, workforce depth, and utilization density that could not be sustained in a semi-rural catchment. The economics of a full-service facility assumed patient volumes that were incompatible with geographic dispersion. The workforce requirements assumed specialist concentration that was incompatible with the talent market. And the clinical model assumed that all categories of care had similar location requirements - which operational reality quickly disproved.

Patients with acute emergencies could not travel. Patients needing highly specialised elective procedures were already travelling - long distances, to concentrated centres in Chennai, Hyderabad, and Kolkata - despite the existence of nominally nearer facilities. The geographic behavior of patients was not irrational; it was a precise signal about the different location requirements of different types of care.

Over time, as the network expanded across districts and states, these observations crystallized. Lack of emergency and stabilization infrastructure at district level produced immediate mortality - preventable deaths attributable entirely to distance. Highly specialized procedural centres, by contrast, could not be made financially or operationally viable at district level regardless of policy intent; they required the volume, talent, supply chain, and ecosystem that only concentration could produce. The conventional primary-secondary-tertiary framework provided no conceptual tool to distinguish between these two realities. It treated them both as questions of institutional tier rather than questions of operational logic.

The 2×2 framework proposed in this paper emerged from repeated operational iteration: what are the two variables that most reliably predict the infrastructure, workforce, geography, and financing requirements of a given service? The answer, tested across hundreds of clinical categories across multiple states over more than a decade, was consistent: time sensitivity (is this care acutely time-critical, or can it be planned?) and infrastructure dependency (does this care require physical intervention in a specialized facility, or can it be delivered through information and conversation?). These two variables - acuity and intervention dependency - define the four quadrants of the framework.

3. The Proposed Framework

3.1 The Two Axes

Axis 1 - Acuity: Acute versus Non-Acute. The first axis separates conditions involving a serious and immediate risk of death, disability, or irreversible damage from conditions that are non-acute, stable, or elective. This is not a measure of severity in the conventional sense - a non-acute condition can be severe. It is a measure of time sensitivity: does delay between presentation and care materially worsen the outcome? Acute conditions are those for which the answer is yes. The practical implication is location: acute care must be geographically proximate to the population it serves because the patient cannot travel. Non-Acute care can, in principle, be accessed at greater distances.

Axis 2 - Delivery Modality: Interventional versus Non-Interventional. The second axis separates care that requires a physical intervention in a specialised facility - surgery, procedural medicine, device implantation, physical therapy requiring specialist equipment - from care that does not. Non-interventional care encompasses diagnosis, monitoring, counselling, medication management, rehabilitation guidance, preventive services, and longitudinal follow-up. The critical insight is that non-interventional care can, in principle, be delivered through any medium capable of transmitting clinical information: digital platforms, telephone, community health workers, and kiosk-based diagnostics. The practical implication is delivery architecture: non-interventional care is inherently scalable through digital and community channels; interventional care is not.

3.2 The Four Quadrants

3.3 Specialties Span Multiple Quadrants

The framework classifies episodes and modes of care delivery rather than entire medical specialties. This distinction is essential and should not be misread as a weakness. On the contrary, the fact that specialties span multiple quadrants is evidence that planning should be organised around operational care characteristics rather than specialty labels.

Oncology illustrates this most clearly. Diagnostic consultation and survivorship follow-up fall in Q3 (Non-Acute Non-Interventional). Radiotherapy and oncologic surgery fall in Q4 (Non-Acute Interventional). Emergency neutropenic sepsis may fall in Q1 or Q2 depending on severity and the intervention required. Chemotherapy delivery spans multiple quadrants depending on the toxicity profile, urgency, monitoring requirements, and infrastructure dependency of the specific regimen. An oncology planning framework derived from this classification would therefore specify different location, infrastructure, and financing requirements for each component of oncology care - a far more useful output than the single label of 'tertiary' that the existing framework provides.

Dialysis presents a further instructive case. While renal failure is a chronic disease state, interruption of dialysis rapidly produces acute physiological deterioration. Dialysis therefore behaves operationally closer to acute support infrastructure than to elective specialty care. It requires equipment-intensive but comparatively modular infrastructure that can be deployed at district or urgent-care level without requiring the full ecosystem of a specialist cluster. This operational observation generates a specific planning implication - dialysis coverage should follow a proximity logic closer to Q2 than Q4 - that the tier classification cannot generate.

4. Evidence Base by Quadrant

4.1 Q1 & Q2: Acute Care - The Time-Mortality Evidence

The evidence base for time-sensitive acute care is among the most robust in clinical medicine. MacKenzie and colleagues' landmark evaluation of trauma centre care in the United States demonstrated a 25 percent reduction in mortality associated with trauma centre designation - an effect attributable primarily to geographic access and institutional readiness rather than individual clinical skill. The door-to-balloon time literature for acute myocardial infarction - establishing the 90-minute benchmark for percutaneous coronary intervention - provides the clearest quantification of the time-mortality relationship in acute interventional care. For stroke, the IST-3 trial and subsequent meta-analyses established that each 15-minute delay in thrombolysis is associated with measurable outcome deterioration.

These findings have a common structural implication: the value of acute care infrastructure lies substantially in its geographic deployment, not merely in its technical quality. A world-class cardiac catheterization laboratory located three hours from the majority of its catchment population cannot deliver the outcomes that a competent unit located within sixty minutes can. The planning implication is direct: Q1 and Q2 facilities must be located according to population access standards rather than market viability.

4.2 Q3: Non-Acute Non-Interventional - The Case for Digital Delivery

The evidence for digital and community-based delivery of non-acute non-interventional care has grown substantially over the past two decades. The systematic review evidence on technology-enabled hypertension management shows consistent reductions in blood pressure across intervention types, with the Omboni meta-analysis demonstrating a 3.5 mmHg reduction in systolic blood pressure associated with smartphone app interventions. The diabetes self-management technology review by Greenwood and colleagues found consistent improvements in HbA1c across intervention types. The TIM-HF2 trial demonstrated that structured telemonitoring reduced all-cause mortality and improved quality of life in a protocol-managed cohort, though effect sizes in unselected populations have been more modest.

These results should not be interpreted as evidence against digital Q3 delivery. Many early telemedicine programmes represent immature implementations of transitional technology architectures - fragmented, poorly integrated with clinical workflows, and lacking the longitudinal patient data, algorithmic triage support, and escalation pathways that a well-designed digital chronic care system would incorporate. The relevant comparison is not between a poorly designed digital programme and well-resourced in-person care, but between a well-designed digital system - incorporating structured data, validated workflows, physiological monitoring, and AI-supported guardrails - and the fragmented, inconsistent, often inaccessible in-person chronic care that most populations in low- and middle-income countries actually receive. On that comparison, the case for digital Q3 delivery is considerably stronger.

The evidence for community health worker interventions in chronic disease management is particularly robust in low- and middle-income settings, where Lewin and colleagues' Cochrane review found consistent improvements in chronic disease outcomes attributable to community-based non-physician delivery.

4.3 Q4: Non-Acute Interventional - Volume, Outcomes, and the Cluster Effect

The case for allowing Non-Acute Interventional care to concentrate geographically rests on one of the most consistently replicated findings in health services research. Luft and colleagues, publishing in the New England Journal of Medicine in 1979, first demonstrated that hospitals performing higher volumes of complex procedures had significantly better outcomes. Birkmeyer and colleagues quantified this relationship across major procedures: mortality for pancreatic resection was 3.8 percent at the highest-volume hospitals versus 16.3 percent at the lowest-volume hospitals; for esophagectomy, 3.4 versus 12.2 percent.

The Indian experience offers some of the most compelling illustrations of what this dynamic produces at health system level. Aravind Eye Care, founded in Madurai in 1976, now performs over 400,000 surgeries annually, achieving complication rates comparable to NHS benchmarks at a fraction of the cost. The efficiency is not despite the volume; it is because of it. Task-shifting, specialized theatre design, supply chain optimization, and team training are viable only above volume thresholds that a geographically dispersed model could never reach.

Porter's work on economic clusters and Krugman's economic geography framework provide the theoretical underpinning: spatial concentration of related economic activity generates externalities - shared labour pools, specialized suppliers, knowledge spillovers - that improve productivity and quality beyond what any dispersed model can achieve. The healthcare cluster is not an exception to this principle; it is one of its clearest manifestations.

5. Implications for Health System Architecture

5.1 Location and Infrastructure Planning

The framework generates differentiated location principles for each quadrant. Acute facilities (Q1 and Q2) must be located according to population access standards, not market viability or administrative convenience. The appropriate measure is travel time, not geographic distance: a fifty-kilometre highway journey and a ten-kilometre mountain road journey may represent radically different access realities. The framework proposes a ninety-percentile one-hour travel time standard as the planning target, with three hours as the minimum acceptable threshold in resource-constrained contexts. These standards should be explicit obligations in national planning documents.

Non-Acute Non-Interventional care (Q3) should be distributed as widely as possible, including through digital and community channels that reduce the effective distance between patient and provider to zero. The planning priority for Q3 is connectivity, device access, community health worker training, and digital platform design - not physical facility construction. Non-Acute Interventional facilities (Q4) should be permitted and, through appropriate referral and transport policies, encouraged to concentrate where ecosystems are viable. The policy obligation is to fund patient pathways to these clusters - not to prevent cluster formation in the name of distribution equity.

5.2 Workforce by Quadrant

Each quadrant requires a distinct workforce model. Q2 (Acute Interventional) demands full multi-disciplinary specialist teams with continuous availability, anesthetic depth, and intensive care capacity: this is the most resource-intensive workforce model and the principal constraint on geographic distribution of acute interventional care. Q1 (Acute Non-Interventional) can be staffed substantially by trained nurses, paramedics, and community health workers supported by structured protocols and digital decision support - a workforce model accessible at district level.

Q3 (Non-Acute Non-Interventional) is precisely the domain where task-shifting generates its greatest returns. Community health workers, trained nurses, pharmacist-led chronic disease management programmes, and AI-supported digital platforms can deliver the longitudinal monitoring, medication management, and behavioral support that constitute the majority of Q3 care - at a fraction of the cost of physician-led in-person delivery.

Q4 (Non-Acute Interventional) requires the deepest specialist concentration but also enables the most efficient specialist utilization: high-volume specialist clusters allow surgeons and proceduralists to operate at the top of their training, supported by specialised teams and supply chains. The cluster model is not merely a market outcome; it is the operationally superior workforce utilization model.

5.3 Digital Infrastructure and Artificial Intelligence

The non-interventional axis of the framework provides a principled basis for digital health investment that the tier classification cannot generate. Any service that is genuinely non-interventional is, in principle, a candidate for digital or community delivery. This is a broader category than most health systems have recognized: it encompasses not only routine chronic disease monitoring but also mental health counselling, physiotherapy guidance, medication management, certain forms of diagnostic interpretation, nutritional support, and a substantial proportion of first-contact triage.

Digital infrastructure should not be treated as an optional technological overlay on a physical delivery system. In the non-interventional quadrants - Q1 and Q3 - digital systems are increasingly the primary delivery architecture. Artificial intelligence is likely to disproportionately transform the non-interventional quadrants. In Q3, AI-supported monitoring, pattern recognition, and proactive escalation can extend the effective reach of clinical supervision across large patient populations. In Q1, AI-assisted triage can improve the consistency and accuracy of clinical prioritization.

5.4 Insurance and Financing Design

The framework has significant implications for insurance benefit design. The dominant dysfunction of health insurance worldwide - creating friction precisely where the system should be frictionless, and permitting overuse where gatekeeping would be most valuable - can be traced to the application of uniform incentive structures across qualitatively different categories of care.

A framework-informed insurance design would apply co-payments and utilization management primarily to Q3 care, where the risk of low-value utilization is highest and the cost of deterrence is lowest. Acute pathways (Q1 and Q2) should be largely free at the point of use - the evidence consistently shows that co-payments on emergency presentations deter necessary as well as unnecessary care, with adverse outcome effects that outweigh administrative savings. Acute interventional infrastructure cannot be financed purely on utilization economics because the value of the system lies partly in continuous readiness capacity. Trauma theatres, ICU beds, blood banks, and emergency specialist teams must exist before they are needed; Q2 therefore behaves economically more like strategic infrastructure than conventional market service delivery.

Pre-authorization adds value in Q3 and Q4, where care is planned and clinical alternatives exist. It adds no value and considerable harm in Q1 and Q2, where care is urgent and alternatives are dangerous. Networks should be designed to ensure that all acute care is in-network by geography and that Q4 specialist clusters are appropriately incentivized for high-volume referral.

6. Limitations and Critiques

6.1 The Boundary Problem

The most immediate conceptual challenge is the definition of the acute-non-acute boundary. Many presentations are neither clearly time-critical nor clearly elective. The patient with escalating heart failure who has not yet decompensated, the child with worsening asthma, the elderly patient with early septic foot - these sub-acute presentations occupy a clinical middle ground the binary axis does not easily accommodate. The framework's position is that this is a challenge of clinical implementation rather than a conceptual flaw: the planning implications of the two categories are sufficiently distinct to justify the binary distinction, and the clinical grey zone - like all clinical grey zones - requires clinical judgment rather than algorithmic classification.

6.2 Mental Health as a Validation of the Non-Interventional Axis

Rather than presenting an exception to the framework, mental health services - much like oncology - serve as a prime validation of its core thesis: operational characteristics, rather than medical specialty, should dictate delivery architecture. The traditional tier system has long struggled to classify mental health effectively, often isolating it from broader health system planning. Under the proposed 2×2 framework, mental health maps seamlessly onto the operational quadrants.

The vast majority of mental health care is non-acute and non-interventional, placing it firmly in Q3. Here, digital and personal-device-based delivery is not merely a fallback but often the superior modality, effectively bypassing the severe barriers of societal stigma and privacy concerns that deter in-person attendance in many societies.

7. Conclusion

The primary-secondary-tertiary framework has served its historical purpose. It gave health systems a shared vocabulary at a moment when that vocabulary was urgently needed. But it has calcified into a set of assumptions that now impedes more than it enables - obscuring the practical distinctions that matter for planning, financing, and delivering care in a world defined by digital technology, specialized clinical clusters, and time-sensitive emergency medicine.

The 2×2 framework proposed in this paper is not a theoretical exercise. It emerged from the operational experience of building health systems in settings where the consequences of misclassification are immediate and visible: preventable deaths attributable to distance from emergency care; specialist services that cannot sustain themselves when dispersed to meet bureaucratic distribution mandates; chronic disease loads that overwhelm hospital outpatient departments that were never designed for them. The framework is a codification of what operational experience in those settings consistently revealed.

The framework's central thesis is not merely that this is a better classification system. The larger argument is that healthcare infrastructure should be organized according to operational systems logic rather than institutional tradition or specialty hierarchy. That reorganization - grounded in acuity and intervention dependency - generates coherent, differentiated, and actionable guidance on where to build, how to staff, which services to digitize, and how to finance the result. That is the standard by which any health system planning framework should be judged, and it is the standard by which this one deserves to be tested.