Case Study 12.2 — The Coverage Decision: How a Payer Decides Whether to Buy a Chronic Therapy for Millions

DataField.Dev

Case Study 12.2 — The Coverage Decision: How a Payer Decides Whether to Buy a Chronic Therapy for Millions

What this case is. A structural walk-through of the decision a health plan, employer, or national health system actually faces when an effective chronic therapy arrives for a very large population. It is written from the payer's chair, deliberately, because that chair is almost never occupied honestly in public argument — payers are usually cast either as villains rationing care for profit or as prudent stewards defending the common purse, and neither cartoon survives contact with the arithmetic.

The premise you must accept to read this fairly: budgets are finite. Not as a rhetorical device, not as a pretext. A plan that spends more on one thing spends less on another or charges more in premiums, and premiums are paid by people, many of whom have less money than the people this decision is about. There is no version of this problem where the money comes from nowhere.

The premise the payer must accept in return: the drug works. This is not a case where a payer can retreat into evidentiary doubt. STEP 1 reported roughly −15% of baseline body weight at 68 weeks against roughly −2.4% on placebo. SELECT reported a 20% relative reduction in major adverse cardiovascular events in adults with established cardiovascular disease and overweight or obesity without diabetes — an absolute move of roughly 8% to roughly 6.5%, about 1.5 percentage points, a number needed to treat of roughly 65 to 70 over about three years. Declining to cover this is declining to fund a benefit that exists.

Part 1 — Why this decision is different from every other coverage decision

Payers approve expensive drugs constantly. Many cost far more per patient per year than a GLP-1 agonist. What makes this decision structurally novel is not the price. It is the denominator.

THE SHAPE OF A NORMAL COVERAGE DECISION vs THIS ONE

  ORPHAN / SPECIALTY DRUG
    very high price per patient  ×  very small eligible population  =  manageable total
    ▓▓▓▓▓▓▓▓▓▓                       ▪                                 ▓▓▓

  ORDINARY CHRONIC DRUG (generic statin, antihypertensive)
    very low price per patient   ×  very large eligible population  =  manageable total
    ▪                                ▓▓▓▓▓▓▓▓▓▓                        ▓▓▓

  GLP-1 AGONIST FOR OBESITY
    high price per patient       ×  very large eligible population  =  ??????????
    ▓▓▓▓▓▓▓                          ▓▓▓▓▓▓▓▓▓▓                        ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓

  Every existing tool in a formulary manager's kit was built for one of the first two
  shapes. The third is new, and the tools do not fit it.

Multiply a plausible per-member cost by the share of a covered population meeting eligibility criteria and the result frequently exceeds the plan's entire pharmacy budget — not its budget for a category, its budget. That number is the reason the debate exists. It is not a bluff.

Part 2 — The four arguments a medical director will make, and what is right about each

"Premiums are set a year ahead." A plan cannot absorb an unbudgeted expansion mid-year. A benefit that requires a premium increase gets one, and the increase is paid by every member — including members who will never take the drug, many of whom are lower-income than the average member who will. This is a real distributional question that the phrase "insurers should just cover it" does not answer; it relocates.

"It is chronic therapy." Benefit reverses on discontinuation, because the drug overrides an intact regulatory system rather than replacing a missing hormone. The plan is not buying a course of treatment. It is buying an indefinite obligation, indexed to a growing eligible population, for a drug still under patent.

"We will not be here to collect." In a fragmented insurance market, members change plans every few years. A payer funding prevention today whose payoff arrives in fifteen years is, with high probability, funding it on a competitor's behalf. Every payer knows this and discounts long-horizon prevention accordingly. This is the wrong-pockets problem, and it is a defect of fragmentation rather than of character — which is exactly why single-payer systems, whose members do not leave, make different prevention decisions with the same evidence.

"We have been here before." Fenfluramine combinations were withdrawn after an association with valvular heart disease. Sibutramine was withdrawn after a cardiovascular signal. Rimonabant was withdrawn in Europe and never approved in the US over psychiatric effects. A payer who covered three weight drugs that looked good on weight and bad on outcomes learned a rule: wait for hard endpoints. SELECT is the first time that rule has been satisfied in this category.

Part 3 — The three arguments against, and what is right about each

The evidence is now what payers said they wanted. For years the category was declined for want of outcomes data. Outcomes data arrived. A payer that shifts to a budget argument after receiving the evidence it demanded is entitled to do so — budgets are real — but should say plainly that the objection has changed, rather than continuing to imply the evidence is thin.

Category exclusions are not clinical decisions. A blanket exclusion of "agents used for weight loss" does not evaluate anything. It is the residue of a period in which obesity was coded as cosmetic, and it survives in plan documents long after the people who wrote it have stopped believing it. Whatever a plan decides, deciding by category is deciding without looking.

Utilization management is also attrition. Prior authorization, step therapy, documentation requirements, quantity limits, and required program enrollment all have defensible rationales — and all reliably lose a fraction of the eligible population, non-randomly. The fraction lost is concentrated among people with less time, less continuity of care, less familiarity with appeals, worse English, and nobody to make phone calls for them. A plan can be doing responsible stewardship and rationing by inconvenience simultaneously, and the test of which is happening is whether anybody measures how many eligible people the gate turns away.

Part 4 — Where the line gets drawn, and who is standing on each side

Faced with a total it cannot fund, a payer does what payers do: it narrows the population. SELECT made that possible in a way it had not been before, because it did not only supply a hard endpoint — it supplied a defined subgroup. A plan can now cover people with established cardiovascular disease and overweight or obesity, cite an outcomes trial, and decline everyone else on clinical rather than categorical grounds.

That is genuine progress and it is also a line, and lines have two sides.

On the covered side. People who meet the criterion get a therapy with demonstrated hard-endpoint benefit. Their absolute risk was high, so their absolute benefit is largest. This is, by most standards, good allocation.

On the excluded side. People with obesity and no established cardiovascular disease — the great majority of people who want this drug. Their absolute risk is lower, so even if the relative benefit held, their absolute benefit would be smaller and their number needed to treat larger. The payer's logic is coherent. But the excluded population includes people for whom the drug would relieve real impairment now: joint pain, sleep apnea, mobility, the metabolic trajectory that produces the qualifying cardiovascular disease a decade later. Excluding them is not obviously wrong. It is also not costless, and the cost is borne by people who will never be counted.

And the exclusion has a shape. Access requires coverage, or cash, or the time and literacy to appeal. Obesity's prevalence is patterned by income, occupation, education, and geography. Put those together and the drug is most available where the burden is lightest — the inverse care law, described in 1971, operating exactly as described.

Part 5 — What happens to the people who lose coverage

This is the part that is easiest to skip and hardest to see.

An employer drops the weight-management category at renewal — a decision made once, in a benefits meeting, affecting a whole workforce at once. Or a plan tightens criteria. Or a member changes jobs and the new formulary is different.

A person stops. Appetite returns, because it was never gone, only suppressed. Weight is regained over months to a couple of years. The cardiovascular risk reduction erodes with it. Years later, some fraction of those people have an event the therapy would have delayed or prevented.

Nobody records the cause. The event is separated from the coverage decision by years, mediated by a dozen other factors, spread thinly across an enormous population. Chapter 11's framework explains why this is invisible where insulin rationing was not: interruption here is not acutely lethal, so there is no attributable death, no coroner's finding, no name, no headline, and therefore no legislation. The harm is potentially far larger than insulin rationing and almost entirely illegible.

It is worth stating clearly that this is a prediction from structure, not a measured finding. The studies that would measure it — linked payer-and-outcomes cohorts following people who discontinue after coverage loss, natural experiments around formulary changes, discontinuation registries — are ordinary pharmacoepidemiology, and nobody is obliged to run them.

Part 6 — The honest summary

The payer faces a genuine dilemma with no costless resolution:

The drug works, and the benefit is real.
The total cost at population scale exceeds what the budget contains.
Narrowing by evidence is the most defensible way to ration, and it still leaves people on the wrong side of a line who would have benefited.
Money not spent here is spent elsewhere, or returned as lower premiums, and those uses have beneficiaries too — usually invisible ones.
The harms of the decision are diffuse, delayed, and unmeasured, which makes them politically weightless regardless of their size.

Anyone who tells you this is simple is selling either outrage or complacency.

Discussion questions

1. Write the medical director's memo recommending against covering the category, at its strongest, in 300 words. Then write the rebuttal, also at its strongest. Which one did you find easier, and what does that tell you about your priors?

2. The wrong-pockets problem means a fragmented insurance market systematically underfunds long-horizon prevention. Name two mechanisms that could correct it without requiring a single payer, and give a real objection to each.

3. SELECT let payers replace a category exclusion with an evidence-based subgroup. Is that unambiguously an improvement? Argue that it is. Then argue that a narrow evidence-based line can do more harm than a broad categorical one, because it is harder to contest.

4. Design a coverage policy for this category, given a fixed budget that cannot cover everyone eligible. State your criterion, and then state explicitly who is excluded by it and what you expect to happen to them. The exercise is the second half.

5. Part 5 describes a harm that is real, potentially very large, and invisible. Suppose you are persuaded it is occurring. What follows — for a regulator, for an employer choosing a plan, for a researcher, and for a patient deciding whether to start? Give a different answer for each, and note which of the four can act without new data.

6. A person is on the excluded side of the line: obesity, no established cardiovascular disease, no coverage, cannot pay cash. Using only §12.11's toolkit, what can they actually do? Be honest about how short the list is, and say which item on it is a policy question rather than a personal one.