Fragmentation Is Not a Yes-or-No Question

Ask any health system leader whether their system is “fragmented,” and you will usually get a defensive answer. Fragmentation sounds like an indictment — as if some systems have it and well-run ones do not. That framing gets in the way of fixing anything.

The most useful starting point is this: every healthcare system is fragmented to some degree. Not because any of them are poorly designed, but because of how healthcare is actually delivered. A single patient’s care can touch multiple providers, multiple financing mechanisms, multiple levels of care, multiple governance bodies, and a wide range of professions and specialties, often within the same year, sometimes within the same week. Coordinating all of that perfectly, every time, is not something any health system in the world has achieved. Even systems regularly held up as models of integration, such as the UK’s NHS, Kaiser Permanente in the United States, or Denmark’s national health service, still have areas where coordination breaks down.

So, the question “is this healthcare system fragmented?” is the wrong one to ask, because the answer is almost always yes. The more productive questions are: how fragmented is it, where specifically is the fragmentation occurring, and what is it actually costing patients and the system?

A working definition

Fragmentation occurs when healthcare services, financing, governance, information systems, or patient pathways are insufficiently coordinated, resulting in inefficiency, duplication, gaps in care, inequity, or poor outcomes. At its core, fragmentation is a failure of coordination between the components of a health system, not a failure of any single component on its own. A hospital can be excellent. A primary care clinic can be excellent. An insurance scheme can be well designed. And the system connecting them can still fail the patient who has to move between all three.

A continuum, not a category

Because fragmentation is a matter of degree, it helps to think of it on a spectrum rather than as a label. At one end, a system is highly fragmented: multiple disconnected actors, multiple funding sources, and multiple information systems, with the patient left to bridge the gaps. Moving along the spectrum, systems become fragmented but with limited coordination mechanisms, then partially integrated with some referral pathways in place, then coordinated with shared pathways and governance, and at the far end, integrated: a patient-centered system with aligned governance, financing, services, and information. No health system we are aware of sits fully at that far end across every part of care. Most systems are a mixture: strong coordination in some areas, real gaps in others.

That mixture is exactly why a single label like “fragmented” or “integrated” tells you very little. It is more useful, and more actionable, to describe a system as having a fragmentation profile: for example, high financial fragmentation, high governance fragmentation, moderate service fragmentation, and high information fragmentation, rather than one flat verdict. A profile like that tells you where to start.

Why this framing matters

Once fragmentation is understood as a continuum that exists at multiple levels simultaneously, the task facing health system leaders changes. It is no longer “eliminate fragmentation,” which is not a realistic goal for any system anywhere. It becomes “know precisely where your system sits, on which dimensions, and reduce fragmentation where it is doing the most damage to patients and to cost.” That is a measurement problem as much as it is a policy problem, and it is where the rest of this series goes next: first mapping the different levels and types of fragmentation, then what they cost patients in practice, then what has actually worked to reduce them.

At B&I, this is the question VAO was built to help answer: not whether a health system is fragmented, but exactly where, how severely, and what closing those gaps is worth. If your organization is working through that question, we would welcome the conversation.