Drug cost is why Medicare recommendations go wrong

Almost every Medicare recommendation that goes wrong goes wrong the same way. The plan was cheaper on premium, the client enrolled, and in February she paid more at the pharmacy counter than she expected. Nothing was misrepresented. The comparison simply answered a question the client did not need answered. What she needed was her annual total for her eleven prescriptions, at the pharmacy she uses, under the plan's own rules. That number is knowable before she enrolls, and it is not the premium.

The only figure worth comparing is the annual total

A drug plan is a function, not a price. Feed it a medication list, a set of dosages and fill frequencies, a pharmacy, and a start date, and it returns a cost for the plan year. Two plans fed the same list can return totals that differ by more than a year of premium in either direction, which is why premium-first comparisons are not conservative or simplified. They are wrong in an unpredictable direction.

Where the drug sits: the formulary

Each plan publishes its own formulary, and tier placement is a plan decision rather than a property of the drug. A drug on a preferred generic tier under one plan can sit on a non-preferred tier under another, at materially different cost sharing, in the same county, in the same plan year. Most formularies run something like preferred generic, generic, preferred brand, non-preferred drug, and specialty, and some add a select or select care tier below the rest.

Two consequences follow, and both are easy to miss with a plan finder open in one tab and a client on the phone.

• A generic is not automatically cheap. Generic status tells you nothing about tier. A generic placed on a non-preferred tier can cost the client more than a brand placed on a preferred brand tier.
• Not on the formulary is a different problem from expensive. A drug that is not covered at all does not produce a high cost in a comparison. It produces a gap, and unless the comparison flags it, the gap shows up when the client tries to fill the prescription. Plans are required to provide a transition supply in defined circumstances for enrollees new to a plan, which buys time to pursue an exception but does not solve the case.

The three gates in front of a covered drug

Tier placement tells you the cost. Utilization management tells you whether the client will get the drug at all without work.

Each of these is appealable, and exceptions and tiering exceptions exist for a reason. But an appeal is a process with a timeline, and specialty tier placement is generally outside the reach of a tiering exception. The practical rule is that a restriction you find before enrollment is a conversation, and a restriction the client finds at the counter is a complaint.

The same drug, priced differently in March and September

Part D cost sharing moves through phases across the plan year. A plan may apply a deductible, and it may apply that deductible to only some tiers. After it is satisfied, the client pays the plan's cost sharing during initial coverage. Once out-of-pocket spending crosses the annual threshold, catastrophic coverage applies. Those thresholds and the annual out-of-pocket cap are set for each plan year, so the arithmetic changes on January 1 even for a client whose prescriptions have not changed at all.

This is why a monthly estimate is not a comparison. A client with two specialty medications may cross into catastrophic coverage early in the year, which makes her monthly cost front-loaded and her annual total far lower than twelve times her January bill. A client with several mid-tier maintenance drugs may never cross it, and her cost is close to flat. Feed both of them a monthly figure and you have told one of them something useless and the other something false. The comparison has to price the full plan year, month by month, in order, because the order is what determines when each phase begins.

Pharmacy choice is a pricing decision

Cost sharing is quoted against a pharmacy, and plans distinguish between preferred network, standard network, and out of network, usually with mail order priced separately again. The same drug, the same tier, and the same plan can produce noticeably different cost sharing at the pharmacy on the corner versus the one across town.

Three things follow for the fact-find. Ask which pharmacy the client actually uses, by name and location, rather than which chain she likes. Check whether that specific location is preferred under each plan you are comparing, because network status is per location. And price mail order honestly: it is frequently cheaper for maintenance medications and frequently unacceptable to a client who wants to talk to her pharmacist, which is a preference worth respecting rather than optimizing away.

Specialty medications add a further constraint, because plans often route them through a designated specialty pharmacy. If the client's current arrangement is not that pharmacy, the change belongs in the conversation before the application, not after.

Therapeutic equivalents, and who gets to agree

Sometimes the honest answer is that the client's medication list, not the plan, is the expensive variable. A different member of the same drug class, a generic where the client is on a brand, or a different device or formulation can sit several tiers lower and change the annual total substantially.

Flagging that is useful. Deciding it is not the agent's call. The prescriber decides, and the only defensible version of this conversation is the one where the agent says what the plan documents show and the client takes it to her doctor. A recommendation that quietly assumes a switch the prescriber has not agreed to is a recommendation built on a fact that is not true yet, and the client will discover that at the counter.

Why premium-first comparisons mislead

Consider two plans in the same county for a client taking two brand-name maintenance drugs and four generics, filling at the independent pharmacy near her house.

• Plan A has no premium, applies a deductible to the brand and non-preferred tiers, places both of her brand drugs on a non-preferred tier, one of them behind step therapy, and treats her pharmacy as standard rather than preferred.
• Plan B charges a premium, applies no deductible to the lower tiers, places one of her brand drugs on a preferred brand tier with no restriction, and counts her pharmacy as preferred.

Ranked on premium, Plan A wins by the full amount of Plan B's premium and the conversation is over in ten seconds. Priced across the plan year for her actual list at her actual pharmacy, the ranking can reverse, and the step therapy requirement on a drug she is already stabilized on is a problem that no amount of premium savings fixes. Nothing about this example is unusual. It is the ordinary shape of a Medicare comparison, and it is why premium is the wrong sort key.

The same logic decides the structural question of a Medicare Advantage plan with drug coverage versus a Medicare Supplement paired with a standalone Part D plan. That comparison involves provider access, out-of-pocket exposure, and underwriting, but the drug side of it cannot be settled on premium either, because the Part D plan is subject to every rule above.

What to model, in order

• The real list: every medication with dosage, strength, and fill frequency, not drug names alone
• Tier placement per plan, plus anything not on the formulary at all
• Prior authorization, step therapy, and quantity limits checked against the prescribed dose
• The full plan year across every coverage phase, in month order
• Cost at the client's named pharmacy, with preferred, standard, and mail order priced separately
• Low Income Subsidy status, which changes the arithmetic rather than adjusting it
• Therapeutic equivalents flagged where they move the annual total materially, for the prescriber to decide

Done by hand, that is twenty minutes per plan and nobody does it for more than two plans. Done by the platform, it is the default output rather than the deep dive, and the ranked list the client sees is ordered by what the year costs her rather than by what the first month costs her. That is the whole point of modeling drug cost instead of estimating it: not precision for its own sake, but a recommendation that still looks correct in February.

Solved Enroll is in private beta. Beta agencies are onboarded in cohorts ahead of a 2027 public rollout, and the live carrier list grows with each one. Join the waitlist for the next cohort.

The Medicare page covers eligibility, provider matching, and drug cost modeling as they run in the platform, and the AI Plan Recommender explains how the resulting totals feed a ranked, explainable recommendation.