The In-Between Is Where Care Actually Happens

For older adults, most of what determines outcomes never reaches a clinical encounter, yet we continue to design as though it does.

The unit we fund, staff, and measure, the clinical encounter, is not the unit where outcomes are actually made. Encounters are bounded, billable, and brief, while outcomes are continuous, building in the days after a discharge, in the weeks between visits, and in the quieter moments when something begins to shift, and no one is quite sure what it means. We have built an extraordinary apparatus around the encounter, but nothing comparable exists for the time between them. For older adults, whose care is inherently longitudinal, multi-condition, and shared across settings and people, that gap is no longer a nuance. It has become the central design problem.

For decades, we have optimized within silos.

Health systems focused on clinical delivery, policy shaped incentives, technology improved discrete workflows, community-based organizations carried what did not fit elsewhere, and families absorbed whatever remained, often without recognition or support.

Each contribution mattered, but none, on its own, was ever going to be sufficient. The challenges we now describe as structural, including uneven access, rising cost, workforce strain, and the daily fragility of care for older adults, are not the result of any single failure. They are the predictable output of a system that was never designed to operate as a whole.

A growing body of evidence reinforces the idea that a significant portion of health outcomes is shaped by factors outside clinical care, including social, behavioral, and environmental conditions. That reality only sharpens the disconnect between how we design the system and how outcomes are actually produced.

The more difficult question is whether we continue refining those silos or accept that the next phase of progress requires something more integrated, more coordinated, and more grounded in how people actually live with care.

From an operational perspective, the in-between is not abstract. It shows up when a family member notices a subtle change but does not know how to escalate it, when a care plan moves between settings without sufficient context, or when a non-clinical barrier quietly undermines a well-intentioned clinical decision. Responsibility is shared across multiple people, while accountability is often diffuse or unclear. These are not edge cases; they are the texture of longitudinal care, and they sit outside the boundaries of any single organization.

That reality is what makes collaboration, in its usual form, insufficient. The question is not whether organizations partner, but how they operate together. Collaboration as an operating model means information, responsibility, and judgment moving across boundaries with precision and purpose. It requires recognizing that no single entity has a complete view of the person, and that the earliest signals are usually seen by those closest to the day-to-day experience.

This is not just an infrastructure challenge. It is a shift in mindset. Organizations must see themselves as participants in an ecosystem rather than as isolated service providers. Leaders have to be candid about what their systems can and cannot see, and there has to be discipline around which information actually changes what happens next.

The instinct to solve fragmentation by sharing more data across more stakeholders is understandable. In practice, it often produces the opposite effect. Without prioritization and ownership, more information does not lead to better decisions; it leads to diffusion. The opportunity is not to share everything, but to surface what matters in time to act.

This also requires engaging across sectors that have spent decades operating under different incentives, languages, and assumptions, and organizing those contributions around the individual's lived experience rather than institutional boundaries. There are signs this shift is underway.

Payment models are pulling toward outcomes and continuity. Technology is increasingly capable, though not yet consistently trusted, to connect what has long been disconnected. Community-based organizations are gaining recognition for the role they have always played. There is growing acknowledgment that families and informal caregivers are not peripheral to care but are often the most consistent presence within it.

Even so, movement is not the same as progress. Progress requires intentional design and a willingness to ask different questions.

For those responsible for care delivery and outcomes, that means looking more closely at where the current model creates blind spots. Where, today, do we have visibility into the time between encounters, and where do we not? Who is most likely to see the earliest signal, and how easily can that signal be acted on? When information moves across settings, does it change what happens next, or does it simply accumulate? These are not abstract considerations; they are design choices. They also begin to reframe the role of the broader care network not only as participants in care delivery but also as contributors to insight, continuity, and early action.

The risk, as always, is that we recognize the need for change without fully committing to it, continuing to invest in isolated improvements while leaving the underlying fragmentation intact. The opportunity is to build something that reflects how older adults and their families actually live, with continuity where there is currently handoff, with information that moves with the person rather than stopping at organizational boundaries, and with practical support for the people, paid and unpaid, who hold it all together.

This is not a small shift, but it is necessary. The work is not to build a better encounter, but to build a care system that holds together in the spaces between them, because that is where older adults actually live and where their outcomes are ultimately determined.

This is where I have landed, and it is also where the next set of conversations needs to deepen.

‍