From the founder
Approval Is Not Access
When the FDA approves a medicine, the country reads it as a finish line. For the patient, it is the start of a second approval, one nobody announces.
The second approval lives in payer coverage policies: hundreds of documents, written plan by plan, that decide whether the approved medicine is covered, under which benefit, reviewed by whom, and on what evidence. Labels and coverage policies do not always move together. Labels widen. Indications expand. Evidence accumulates. The policy on file at a health plan can describe a drug as it was two label revisions ago, and the patient meets the difference at the pharmacy counter or in a denial letter.
We built KAIS to map that gap.
KAIS is built on a dated, cited library of payer rules. Every rule traces to the payer's own document, with the source and the retrieval date recorded, and the document's own effective date where it states one. One part of that library every count agrees on: 572 payer-to-reviewer routing rows, mapping plans to the utilization management reviewers who actually decide. The library as a whole is larger than any one team's question. For a given drug, what matters is the slice: which plans hold criteria for it, how recently those criteria were checked, and where they diverge from the current label.
When a label moves, we can say which policies predate it. A policy that predates a label change is not automatically wrong. It is a reason to inspect. Sometimes the criteria still match. Sometimes they impose a requirement the label never asked for. Sometimes they omit an indication the label now carries. Those are different findings, and KAIS reports them as different findings, with the label's own words next to the payer's.
One current example. Louisiana Blue's policy 00884, effective August 1, 2026, accepts either biopsy or noninvasive testing to establish the required fibrosis stage for Rezdiffra. One additional requirement concerns patients with a BMI above 27: they must have tried lifestyle interventions for at least three months, with a goal of weight loss, a criterion the policy itself flags as an additional company requirement. The current prescribing information, revised July 2026, directs that Rezdiffra be used in conjunction with diet and exercise. It sets no three-month prerequisite. Identifying that difference is not a judgment about whether it is justified. It is a fact an access team needs before a case is filed, and it is the shape of the answer KAIS returns: the requirement that changed, the drug and the plan it affects, the chart evidence that would satisfy it, and the label's own words beside it, each document dated.
Every door uses that library, and they do not mix. Inside a practice, KAIS works in the practice's own protected space: it organizes cited chart evidence and the clinician's documented rationale for review and signature. The reference API offered to manufacturers and hubs compares only public material: labels, coverage policies, and reviewer routing. It accepts no patient records, and it exposes none.
The deepest version of this problem is disease interception. Modern medicine increasingly approves therapies whose value is measured in progression delayed and risk reduced: progression to symptomatic type 1 diabetes delayed, a recurrence pushed back, a decline slowed. Coverage logic was built for sickness, priced over a contract year, while those benefits accrue over years and members change plans in between. The result is a system that rations the future with paperwork written for the present.
Each actor in that chain holds a piece of the record. Policy intelligence services track the documents. Manufacturers learn about the gap from their hub vendors' case queues. Hubs learn about it one patient at a time. Practices learn about it from denial letters. Those pieces often sit in separate systems. Joining them on one dated record is the work KAIS is built to do: the label change, the payer criterion, the reviewer who decides, and the chart evidence that answers it.
That comparison is the company we are building: an access evidence layer for life sciences, queried by API, checked against current labels and current policies, and honest about its own limits. If a source cannot be retrieved, the answer says so. If a criterion is not held, the answer says not held. Where applicability is unresolved or an assertion is unsupported, the answer identifies that explicitly instead of smoothing it over. A referee of facts does not get to guess.
The reference API described here answers platforms and hubs through scoped sandbox keys. What we are doing now is putting the map in front of the people who live with the gap: the brand teams whose labels moved, the hubs whose queues fill with the consequences, and the platforms that serve them.
If you run access for a brand or a hub, bring us one drug and the payer set you care about. We will answer one concrete question against the current label and the current policies, with sources you can check. If your team would rather test it directly, ask for a scoped sandbox key.
Neals Maxilin, Founder and CEO, Apex Flow, Inc.
References
- Blue Cross and Blue Shield of Louisiana, Medical Policy 00884, resmetirom (Rezdiffra), current effective date August 1, 2026. Policy document
- Rezdiffra (resmetirom) prescribing information, revised July 2026 (DailyMed).