Walk into almost any practice and ask what it knows about its own panel. The answer is usually: quite a lot. The record holds the problem list, the last result, the overdue screening, the referral that was placed and never closed. None of that is missing. It is sitting there, correctly, in a system somebody paid a great deal of money for.
Then ask what happens next, and the answer gets vaguer. Somebody runs a report, when there is time. Somebody works a list, when there is somebody. The gap between what the record knows and what the patient experiences is not a knowledge gap. It is an operations gap, and it is where most healthcare technology quietly stops.
Insight is not the scarce resource
A decade of investment has gone into analytics: dashboards, risk scores, population views, predictive models. The output of all of it is a list. Lists are useful right up to the moment somebody has to work one.
The constraint in a busy clinic is not knowing which patients are high risk. It is that acting on the list competes with the schedule, and the schedule wins. A report that identifies two hundred overdue patients creates two hundred units of work for a team that did not have capacity before the report was run.
The middle is where care is won or lost
Between the insight and the outcome sits a long, unglamorous chain: confirm eligibility, place the order, reach the patient, find a slot, remind them, confirm the visit happened, get the result back, record it against the measure. Every link in that chain is a place where a patient falls out.
This is the missing middle. It is not clinical work and it is not analytics work, so it tends to belong to nobody in particular. In most practices it is absorbed by staff goodwill, which is a finite and badly monitored resource.
What working the middle actually requires
- Detection that runs continuously rather than when someone remembers to run a report
- Action that is automatic by default, with the clinician in the loop only where judgement is needed
- Follow-through that keeps watching until the loop closes, not until the order is sent
- Evidence written back into the record as it happens, so proving the outcome is not a separate project
- All of it inside the system the practice already uses, because a twelfth login is not a solution
Why this matters more under value-based contracts
When payment follows outcomes, the operations gap stops being an irritation and becomes a financial one. A program you were eligible for and did not execute looks identical, on a balance sheet, to a program you never had. The difference between the two is entirely in the middle.
The practices that do well under these contracts are rarely the ones with better insight. They are the ones that made the boring part reliable.