How ICER Builds a Value Assessment, Beyond the Price

The short answer#

When the Institute for Clinical and Economic Review (ICER) rates a new drug, the price tag is the conclusion, not the starting point. Two separate engines drive the work. One grades how confident we can be that the drug actually improves health against the treatments patients already have. The other estimates the price at which that improvement would be a fair trade for everything else a health system could spend the same money on. The pair yields a value-based price range and a distinct warning light for whether a fair long-term price could still crowd budgets in the short term.

Key points#

• ICER grades clinical evidence first, using a lettered matrix, before any dollar figure enters the picture.
• A cost-effectiveness model then converts lifetime health gains into a price range, anchored to willingness-to-pay thresholds.
• A separate budget-impact check asks whether the health system can absorb the new spending soon, not just whether the price is fair per patient.
• Severity, caregiver effects, equity, and hope are weighed as context rather than folded into a single number.
• Every assessment is argued in public, with comment periods and an open committee vote.

Two questions, kept apart#

Most debates about drug prices collapse two questions into one: does the drug work, and is it worth the money? ICER's method insists on answering them in order and keeping them separate. Confidence about benefit is judged on the clinical evidence alone. Only afterward does the model ask what that benefit is worth in dollars. Blurring the two is exactly how a promising therapy gets oversold or a genuinely useful one gets dismissed on cost before anyone has agreed it helps.

Grading the evidence before pricing it#

The first engine is the Evidence Rating Matrix. As ICER describes it on its site and in the 2023 Value Assessment Framework, the matrix asks two things at once: how large is the net health benefit against a named comparator, and how certain are we of that judgment. The answer is a letter. An A ("Superior") signals high certainty of a substantial net benefit. A B ("Incremental") means high certainty of a smaller one. A C ("Comparable") means high certainty the drug roughly ties its comparator, and a D ("Negative") means high certainty it is worse.

When the data thin out, the certainty axis does the talking. Grades such as B+, C+, and C++ mark moderate rather than high confidence. P/I ("Promising but Inconclusive") flags a drug that looks helpful but leans on immature evidence, and I ("Insufficient") covers cases where the evidence cannot support any confident call. ICER's user guide traces the matrix back to an evidence-based-medicine grid built by a multi-stakeholder group. The point worth holding onto is that a drug can truly work and still earn a cautious grade, because the trials were short, small, or never tested against a real-world alternative.

Turning health into a price#

Only once the evidence carries a grade does the second engine start. It is a cost-effectiveness analysis that projects lifetime costs and lifetime health gains, then expresses the result as a cost per quality-adjusted life year (QALY). Alongside the QALY, ICER reports the equal value of life years gained (evLYG). That second metric answers a well-known criticism: because a QALY weights added years by quality of life, it can undervalue treatments for people living with chronic illness or disability by treating their extra years as worth less. The evLYG counts a year of life gained more evenly.

Those cost-per-health figures are measured against willingness-to-pay thresholds. The 2023 framework usually frames a health-benefit price benchmark as a band, with a lower bound near $100,000 per QALY gained and an upper bound near $150,000 per evLYG. Running the thresholds backward gives you the value-based price range, the set of prices at which the drug's cost would line up with the health it delivers. This is where the phrase "beyond the price" turns literal. The model never asks what a company decided to charge. It asks what a health system could reasonably justify paying for the health produced.

Can the system afford it right now#

A price can be fair per patient and still overwhelm a budget if enough patients start it at once. ICER treats that as its own question with its own tool, the potential budget-impact analysis. The framework sets an annual spending threshold pegged to growth in the national economy, a figure ICER updated in October 2025 from roughly $880 million to about $821 million in new spending per year. When projected uptake would push a drug past that line, the report raises a short-term affordability flag and warns of possible access barriers, even where the long-term price looks defensible. Long-term value and near-term affordability are different problems, and the method refuses to let one hide the other.

The things a model cannot hold#

Any model built on population averages will miss things, and the framework says so directly through what it calls contextual considerations and potential other benefits. These cover the severity of the illness, the value of hope for patients with few remaining options, the strain on caregivers, and equity concerns. ICER also runs special adaptations for ultra-rare diseases and for one-time or short-course therapies whose costs and benefits arrive at very different moments. Many of these factors are described in words rather than pressed into the cost-per-QALY figure, a deliberate signal that they still count even when they resist being counted.

Argued in the open#

None of this is settled privately. ICER's methods pages lay out a public sequence: a scoping phase, a draft evidence report open to comment, and a revised report presented at a public meeting. There, one of ICER's independent appraisal committees debates and votes on the central questions, and a policy roundtable of clinicians, patient representatives, and payers discusses how the findings might play out in practice. A final evidence report closes the loop. The votes and ratings are recommendations meant to inform decisions, not binding rules.

What the number can and cannot settle#

Read this way, an ICER assessment is a structured argument, not a final grade. It can line a drug up against its alternatives, put a size on the uncertainty, and translate health gains into a defensible price band. It cannot capture every value that matters to one particular patient, and it can only be as sound as the trial data and model assumptions available on the day it is written. The QALY remains contested, which is why the evLYG sits beside it. Knowing the moving parts, the evidence letter, the cost-per-health model, the affordability flag, and the public vote, is what lets you weigh the judgment instead of reacting to a single number.