Every Function Green, the Aggregate Red

A specialty product with acceptable coverage, underperforming its forecast, and a working explanation that was wrong.

Every function’s data looked healthy on its own. Integrated across enrollment, coverage, affordability and dispensing, approvals were not converting to activations because no function owned the transitions between them. The fix was architectural: a channel redesigned around initiation, restructured hub handoffs, a rebuilt affordability bridge and explicit ownership of each transition.

Decision context

A specialty product with acceptable coverage was underperforming its forecast. The working explanation inside the company was a demand problem. That is almost always the first hypothesis, because demand is what Commercial is measured on and it is the thing that can be addressed with spend.

What the data actually showed

The picture came from stitching together sources that normally live apart: enrollment records from the hub, dispensing and fill data from the specialty pharmacy network, payer coverage status, and affordability program utilization.

Separately, each dataset looked healthy. Coverage had been won. Enrollments were being processed. Fills were occurring. The problem only became visible when they were put in the same view, because the number of patients entering at the top was materially larger than the number reaching a second fill, and the loss was concentrated at specific transitions rather than spread evenly.

Root cause versus symptom

The symptom was abandonment. The root cause was that no single function could see a patient across the whole path, so nobody was accountable for the transitions. A patient who was approved but never activated was, from Market Access’s point of view, a success. From the pharmacy’s point of view, that patient had never arrived and therefore did not exist.

The distinguishing test I use is simple. If every function reports green and the aggregate is red, the problem is at the seams, not inside the boxes.

What changed

• The channel architecture was redesigned around initiation rather than around fulfillment convenience, which touched the specialty pharmacy footprint and how patients were routed within it.
• The hub services were reworked, and the handoffs moved.
• The affordability bridge was redesigned so a covered patient was not still an unstarted patient.
• The reporting cadence changed, so the same patient-level view reached Finance, Commercial and Market Access at the same time rather than in three separate versions.
• Governance changed, which mattered more than any single design decision. Someone was named as accountable for each transition.

Duration and indicators

This kind of work takes a quarter to diagnose honestly and several quarters to see move, because persistence is a lagging indicator by definition. The indicators worth watching are time from enrollment to first fill, the proportion of approved patients who never activate, and refill continuation at the second and third fill. What typically does not improve on the same timeline is forecast accuracy, because the forecast was built on the old assumptions and has to be rebuilt rather than corrected.

Attribution

I am not going to publish the movement in those indicators. They are client-reported, I do not have permission to publish them, and I have no documented calculation package that an outside party could reproduce. Even if I did, I could not honestly separate my contribution from concurrent market growth, clinical evidence, pricing decisions and policy movement.

What I claim is the diagnosis and the architecture. The organization could see something it could not see before, and it assigned ownership to transitions that previously had none.

If every function reports green and the aggregate is red, the problem is at the seams, not inside the boxes.

Former employers and products referenced are matters of professional history. The views expressed are my own and are not made on behalf of any current or former employer, and no commercial outcome for any named product is attributed to me personally. All trademarks belong to their respective owners.