National P&T committee decision engine — clinical evidence, ICER/QALY cost-effectiveness, and budget impact converging on a formulary listing verdict
Before any medicine can be considered for a national formulary, a hospital P&T list, or the WHO Essential Medicines List, the manufacturer must submit a structured dossier: clinical evidence, a health-economic model, and a budget impact projection. Formats differ — NICE's single technology appraisal template, AMCP's Format for Formulary Submissions, WHO's EML application — but the underlying logic is identical: no substantive review begins until the paperwork is complete.
A complete submission bundles several distinct workstreams into one package:
• Clinical module — pivotal trial reports, systematic literature review, indirect treatment comparison versus the relevant comparator(s) • Economic module — a cost-effectiveness or cost-utility model (typically a Markov cohort or partitioned-survival model) producing an incremental cost-effectiveness ratio • Budget module — a population-level budget impact analysis projecting net cost to the payer over 3–5 years • Administrative module — regulatory approval status, proposed price, patient access scheme details, plain-language summaries
Missing or internally inconsistent modules are the single most common reason a dossier is returned before review even starts.
Formulary decision-making happens at multiple levels simultaneously. National health technology assessment (HTA) bodies — NICE (England & Wales), HAS/Transparency Committee (France), G-BA/IQWiG (Germany), CADTH (Canada), PBAC (Australia) — issue reimbursement recommendations that shape national coverage. The WHO Expert Committee on the Selection and Use of Essential Medicines maintains a global reference list intended to guide procurement in low- and middle-income countries, independent of any single country's pricing. Beneath these sit thousands of local Pharmacy & Therapeutics (P&T) committees at hospitals and insurance plans, who decide formulary placement for their own population using national HTA output as one input among several.
The intake gate mirrors a regulatory refuse-to-file check: reviewers verify all required sections are present, internally consistent, and use an acceptable economic model structure before allocating scarce clinical and economic reviewer time. A dossier failing this check is returned to the sponsor for resubmission rather than rejected outright — the clock resets, but the door is not permanently closed.
The WHO Essential Medicines List explicitly prioritizes public-health relevance and affordability over commercial novelty — a medicine can be clinically excellent yet never appear on the EML if lower-cost alternatives address the same need adequately.
Many submissions arrive alongside a proposed managed entry agreement — a confidential discount, a price-volume cap, or a coverage-with-evidence-development scheme that ties continued listing to real-world data collection. These arrangements let committees provisionally list a promising but uncertain therapy while formally deferring the full cost-effectiveness judgment until more mature data — including post-marketing outcomes — becomes available.
Before a single dollar is discussed, clinical reviewers must establish whether the drug actually works better than what patients already have access to. Formulary committees increasingly apply the GRADE framework — the same system used in clinical practice guidelines — to rate the certainty of evidence, not just its existence, and to distinguish a genuinely superior therapy from one that merely looks impressive against a weak comparator.
GRADE rates the certainty of a body of evidence from High to Very Low based on five domains: risk of bias in trial design and conduct, inconsistency across studies, indirectness of the population or comparator studied, imprecision of the effect estimate, and publication bias. A single well-designed randomized trial with a large effect can earn "High" certainty; several small, heterogeneous, open-label trials rarely will — regardless of how favorable the headline result looks.
Regulatory approval frequently requires only a placebo-controlled trial, but formulary decisions require comparison against the therapy patients would otherwise receive. When no head-to-head trial exists, reviewers construct network meta-analyses or matching-adjusted indirect comparisons — statistical bridges between separate trial populations. These methods carry real uncertainty: unmeasured differences between trial populations ("effect modifiers") can distort the estimated relative benefit in either direction.
Many pivotal trials measure a surrogate endpoint — tumor response rate, LDL reduction, HbA1c change — rather than a hard outcome like survival or major cardiovascular events, because surrogates mature faster and require smaller trials. Committees must judge how reliably the surrogate predicts the outcome patients actually care about; a validated surrogate (blood pressure for stroke) earns more trust than a novel or disease-specific one with a thin validation history.
The same surrogate-endpoint tension that drives FDA accelerated approval controversies recurs at the formulary stage: a drug can be legally approved on a surrogate yet still be denied listing if the committee judges the evidence too immature to support a durable clinical benefit claim.
Safety review runs in parallel to efficacy review: adverse event rates, discontinuation rates, and any post-marketing signals are weighed against the magnitude of expected benefit. A therapy with a modest efficacy edge but a concerning rare adverse event profile may score worse overall than a therapy with smaller relative benefit but a clean, well-characterized safety record — particularly for chronic, non-life-threatening indications where the harm-benefit calculus is less forgiving.
Once clinical benefit is established, health economists translate it into a single comparable number: the incremental cost-effectiveness ratio, or ICER — the additional cost required to gain one additional quality-adjusted life year (QALY) relative to the current standard of care. This number, plotted against a willingness-to-pay threshold, is the fulcrum on which most formulary cost-effectiveness decisions turn.
A QALY combines length of life with quality of life on a single 0-to-1 utility scale, where 1.0 represents a year in perfect health and 0.0 represents death. Utility weights are typically derived from standardized instruments — the EQ-5D questionnaire, time trade-off, or standard gamble methods — applied to representative patient or general-population samples. A treatment that extends life by two years at a utility of 0.5 produces the same 1.0 QALY gain as one that extends life by one year at full health.
Economists translate trial data into a lifetime model — most commonly a Markov cohort model (patients transition between defined health states each cycle) or a partitioned-survival model (patients are apportioned across progression-free, progressed, and dead states over time). The model extrapolates beyond the trial's observed follow-up, layers in costs (drug acquisition, administration, monitoring, adverse event management, downstream care) and utility weights for each health state, and produces a lifetime cost and lifetime QALY estimate for the new drug versus the comparator.
ICER = (Cost_new − Cost_comparator) / (QALY_new − QALY_comparator)
Plotting incremental cost against incremental QALY produces the cost-effectiveness plane with four quadrants: a drug that costs less and works better is "dominant" (automatic accept); one that costs more and works worse is "dominated" (automatic reject); the two trade-off quadrants — costs more but works better, or costs less but works worse — are where the ICER and the willingness-to-pay threshold actually do the deciding work.
No threshold is scientifically "correct" — it is a policy choice reflecting the opportunity cost of displacing other health spending. Critics note a rigid threshold can systematically disadvantage treatments for rare diseases (small trials, high per-patient cost) and for conditions with no existing effective therapy. In response, several bodies apply severity or rarity modifiers: NICE's 2022 methods update allows a higher effective threshold for diseases with substantial absolute QALY loss, explicitly weighting fairness alongside pure cost-per-QALY efficiency.
A drug can be highly cost-effective (low ICER) and still be unaffordable at population scale, or expensive per patient (high ICER) yet trivially affordable if very few patients are eligible — cost-effectiveness and affordability are evaluated as two separate, sequential questions.
A favorable ICER answers "is this good value per patient?" — it does not answer "can the health system actually afford this at scale?" Budget impact analysis answers the second question: it projects the total net change in payer spending over a multi-year horizon, combining eligible population size, realistic uptake, and any spending offset from therapies the new drug displaces.
Following ISPOR good-practice guidance, the model starts from an epidemiology-based estimate of the eligible population (disease prevalence × diagnosis rate × treatment-eligible fraction), applies a projected market share/uptake trajectory, and multiplies by net per-patient cost each year. The output is a year-by-year net budget impact table — the figure payers actually use for annual pharmaceutical spending planning, distinct from the lifetime, discounted figures used in the cost-effectiveness model.
Real adoption rarely jumps to full eligible-population coverage on day one. Models typically apply an S-shaped diffusion curve — slow initial uptake among early-adopting prescribers, an acceleration phase, then a plateau — mirroring observed launch dynamics for comparable products. Later years may also build in anticipated price erosion from generic or biosimilar entry of either the new drug class or its comparator, which can materially soften long-run budget impact.
The clearest illustration is the direct-acting antiviral revolution in hepatitis C: individual regimens were unambiguously cost-effective (curing a chronic, costly disease for a bounded course of treatment), yet the sheer size of the eligible population created budget impact so large that several national payers implemented explicit rationing — restricting initial access to patients with more advanced liver disease — purely on affordability grounds, despite favorable ICERs.
When a favorable ICER collides with an unaffordable budget impact, committees frequently do not reject outright — they negotiate. Price-volume agreements cap total payer spend regardless of realized uptake; confidential rebates lower the effective net price below the list price used in public modeling; outcomes-based agreements tie payment to real-world effectiveness. These arrangements let a drug that fails a strict up-front budget test still reach patients under contained financial risk to the payer.
Budget impact caps are typically set as a fixed annual envelope per therapeutic area or per payer — meaning a positive committee decision for one drug can mechanically reduce the budget headroom, and therefore the approval odds, of the next drug reviewed in the same category that year.
All prior analysis converges on a single meeting: a multidisciplinary Pharmacy & Therapeutics committee weighs the clinical evidence grade, the ICER against threshold, and the budget impact against cap, then votes — not just yes or no, but on tier placement and any utilization restrictions attached to a "yes."
A well-governed P&T committee deliberately mixes specialties: prescribing physicians (often from the relevant therapeutic area and from primary care), clinical pharmacists, health economists, a biostatistician, nursing representation, and — increasingly, at bodies such as NICE and the WHO EML committee — patient or lay representatives who bring the lived-experience perspective that clinical and economic data alone cannot capture. Members serve fixed terms, and every participant discloses financial relationships with pharmaceutical manufacturers before each meeting.
Most committees require a quorum (commonly two-thirds of voting members) and pass listing decisions by simple or supermajority vote after open discussion of the clinical, economic, and budget dossiers. Any member with a disclosed financial conflict related to the product under review is recused from that specific vote — a rule enforced independent of how the recusal might swing the outcome, to preserve the integrity of the process rather than any particular result.
A "yes" vote is rarely unconditional. Common restrictions include prior authorization (a clinician must justify use against defined criteria before dispensing), step therapy (patients must first fail a cheaper alternative), quantity limits, and specialist-only prescribing. These tools let a committee extend conditional access to a drug whose evidence or economics do not yet support unrestricted first-line use, without denying access outright to patients who genuinely need it.
Manufacturers typically retain a formal right of appeal, usually limited to procedural grounds or new evidence rather than simple disagreement with the committee's judgment. Listed drugs are not decided once and forgotten: real-world utilization, outcomes, and safety data feed periodic re-review, and a drug listed under a coverage-with-evidence-development agreement can be delisted or have its tier changed if the promised confirmatory data fails to materialize or turns out unfavorable.
The WHO Expert Committee publishes its full meeting reports, application reviews, and reasoning publicly — a transparency standard increasingly adopted by national HTA bodies, on the premise that decisions rationing access to medicines carry a governance obligation to be publicly defensible.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Tier 1 — Preferred | |||
| Tier 2 — Preferred with PA | |||
| Tier 3 — Non-Preferred | |||
| Not Listed |