The HEOR Workbench
A self-guided course and calculator set for health economics and outcomes research, from Boyce Data Science.
How to use this workbench
Health economics and outcomes research (HEOR) asks what difference a treatment, device, or program makes to patients in practice, and whether that difference justifies what a health system spends to deliver it. The learning panels cover the concepts in order, and the calculators let you run the same methods on your own numbers.
Who this workbench serves
Patient advocates reading a health technology assessment (HTA) appraisal for the first time. Research staff and students who keep hearing QALY and ICER in meetings without a place to look them up. Registry and real-world data teams whose data sets end up as inputs to an economic model. No economics background is assumed, only comfort with arithmetic.
How the panels fit together
Outcomes come first, then costs, then the cost-effectiveness panel that combines them; the later panels place the methods in the decision-making context where they are used. Each learning panel ends with a knowledge check, and answering it correctly marks the panel complete in the bar above. Nothing is stored on a server: use Save session to keep your work in a file on your own computer, and Load session to bring it back.
The learning path
The calculators
Use them with the worked examples in the text, or bring your own numbers: a therapy your community is watching, a program your organization runs, a scenario from a published appraisal.
- ICER calculator and cost-effectiveness plane. Enter costs and effects for two strategies, get the ratio, and see it plotted against a threshold you control.
- QALY builder. Assemble quality-adjusted life years from health states, utility weights, and durations, with and without discounting.
- Budget impact worksheet. Walk a population funnel from covered lives to net budget impact.
- HEOR study planner. A fill-in scaffold covering perspective, comparator, horizon, outcomes, and costs that you can print or download for your own project.
They are teaching instruments. The budget impact worksheet models a single year with flat uptake, the QALY builder discounts in whole-year steps, and neither propagates parameter uncertainty. Published analyses use multi-year uptake curves, probabilistic sensitivity analysis, and payer-specific unit costs. Use these panels to understand and interrogate an analysis, not to produce one for submission.
What is HEOR, and who uses it?
Health economics and outcomes research generates evidence on the value of health interventions: not only whether a treatment works in a trial, but what difference it makes to patients' lives and what a health system spends to deliver that difference.
Outcomes research and health economics
- Outcomes research studies the end results of care as experienced in practice: survival, function, symptoms, quality of life, hospitalizations, adherence. It often uses data from routine care rather than from controlled trials.
- Health economics studies how scarce healthcare resources are allocated and how they could be allocated better. Its core tool is economic evaluation, the systematic comparison of the costs and consequences of alternative options.
HEOR fuses the two: outcomes research supplies what happens to patients, health economics weighs that result against its cost and against the alternatives.
Why scarcity forces the question
Every health system, public or private, operates inside a finite budget. Money spent on one therapy is money not spent on nurses, screening programs, or a different therapy. Economists call the value of the best forgone alternative the opportunity cost. HEOR exists because "does it work?" is not enough to allocate a fixed budget; a decision-maker also needs to know compared to what, for whom, and at what cost.
Given what we would have to give up elsewhere, do the additional benefits of this intervention justify its additional costs, compared with what we do now?
Who uses HEOR evidence
| Audience | What they do with it |
|---|---|
| Payers and insurers | Coverage and formulary decisions, price negotiation, prior-authorization criteria. |
| HTA bodies | National or regional recommendations on whether a technology should be funded. Examples include NICE in England, Canada's Drug Agency, IQWiG and the G-BA in Germany, and the PBAC in Australia. |
| Manufacturers | Reimbursement submissions, and trial designs that capture the endpoints an economic model will need. |
| Clinicians and guideline groups | Weighing cost alongside efficacy when recommending care pathways. |
| Patients and advocates | Arguing for access, making sure the outcomes patients care about are counted, and challenging the assumptions inside a model. |
Sources that feed an economic evaluation
HEOR draws on randomized trials, but it leans heavily on other sources: observational studies, patient registries, insurance claims, electronic health records, patient surveys, and decision-analytic models that stitch these together. That breadth gives the work its real-world relevance, and it also creates the central methodological challenge, because non-randomized evidence needs careful design to be credible. The panel on real-world evidence returns to this.
HEOR is sometimes described as putting a price on life, or as a rationing tool. Its defenders make close to the opposite argument: that every system already makes these trade-offs, and that an explicit method makes them consistent and open to challenge, including challenge from patients about what should count as a benefit. Both readings show up in public debate, and it is useful to recognize which one an argument is starting from.
Knowledge check
Choose an answer for each question, then check your score. You can retry as often as you like.
Measuring outcomes: QALYs and PROs
Before anyone can ask whether a treatment justifies its cost, there has to be a defensible measure of what it delivers. This panel covers the ladder of outcome measures, from laboratory values up to the quality-adjusted life year.
The outcome ladder
| Level | Examples | Strength and limitation |
|---|---|---|
| Surrogate endpoints | Biomarker levels, tumor shrinkage, forced expiratory volume | Fast to measure; may not track how patients feel or how long they live. |
| Clinical events | Survival, hospitalization, disease progression | Unambiguous and meaningful; may take years to observe. |
| Patient-reported outcomes (PROs) | Symptom scales, function scores, quality-of-life instruments | Record the patient's own experience; require validated instruments and careful handling of missing responses. |
| Preference-weighted measures | QALYs built from utilities | Comparable across diseases, which is what makes them the common currency of economic evaluation. |
Utilities: putting a weight on a health state
A utility is a number expressing how good or bad a health state is, anchored so that 1 is full health and 0 is dead. States judged worse than death can score below 0. Utilities come from asking people to trade quantity of life against quality of life, using methods such as:
- Time trade-off (TTO). How many years in full health would you accept in exchange for ten years in this state?
- Standard gamble. What risk of death would you accept from a treatment that restores full health?
- Generic preference instruments. Questionnaires such as the EQ-5D or the SF-6D map a patient's answers to a utility through a country-specific value set. Many HTA bodies prefer these because the resulting numbers compare across diseases.
The QALY
The quality-adjusted life year multiplies time spent in a health state by the utility of that state:
Two years at utility 0.8 gives 1.6 QALYs; one year in full health gives 1.0. The power of the measure is that a therapy which mostly extends life and a therapy which mostly improves quality of life can be compared on one scale, which is what a system allocating a single budget across many diseases needs.
Without treatment, a patient lives three years at utility 0.6, giving 1.8 QALYs. With treatment, four years at utility 0.7, giving 2.8 QALYs. The treatment gains 1.0 QALY. Try variations in the QALY builder.
Criticisms of the QALY
- Generic instruments can be insensitive to the changes that dominate daily life in a specific condition, a recurring problem in rare, rapidly progressive, and pediatric diseases.
- Whose preferences count is contested. General-population value sets can weigh a health state very differently from the people living in it.
- On equity grounds, a QALY counts the same wherever it accrues, which sets aside severity, rarity, and end-of-life context. Several HTA systems respond with severity or rarity modifiers.
- Some jurisdictions restrict the measure outright. In the United States, federal law limits the use of QALY-based thresholds in certain coverage and payment programs, and no federal body sets an official cost-effectiveness threshold.
These criticisms have not displaced the QALY from HTA practice, though they shape how carefully analysts handle it, and they point to the places where patient input changes what a model produces.
Knowledge check
Choose an answer for each question, then check your score.
Measuring costs and perspective
The cost side of an evaluation looks like simple addition, but two choices quietly determine the answer: whose costs are counted, which is perspective, and how future costs are valued, which is discounting.
Categories of cost
| Category | Examples |
|---|---|
| Direct medical | Drug acquisition, administration, clinic visits, hospitalizations, monitoring laboratory tests, management of adverse events. |
| Direct non-medical | Travel to appointments, home modifications, paid caregiving, durable medical equipment. |
| Indirect, or productivity | Lost work time for patients and for unpaid family caregivers, early exit from the workforce, reduced productivity while working. |
Perspective: which costs are counted
The perspective of an analysis defines which of those categories enter the arithmetic:
- Payer perspective. Only the direct medical costs borne by the payer. Most reimbursement-facing analyses use this one.
- Healthcare sector perspective. All formal healthcare costs, whoever pays them.
- Societal perspective. Everything: medical costs, patient and family out-of-pocket costs, and productivity losses.
Perspective is not a technicality. A therapy that keeps a caregiver in the workforce can look expensive on a payer's ledger and cost-saving from society's. In conditions with heavy family caregiving burdens the choice of perspective can reverse the conclusion, which is why advocates reading an analysis should check which perspective it used and make the case for the one that reflects the disease's real footprint.
It is usually declared in one sentence of the methods section. Read that sentence before anything else in the cost analysis, because every number that follows depends on it.
Discounting: how future costs and effects are valued
Costs and benefits that occur in the future are discounted, reducing their present value at a fixed annual rate, on the reasoning that people and institutions prefer benefits sooner and costs later. Rates around 3% per year are common in United States analyses, and the NICE reference case in England applies 3.5% to both costs and health effects. Reference-case analyses discount health effects as well as money.
Discounting changes the answer most when costs are paid up front and benefits arrive over decades, which is the situation for gene therapies and other one-time treatments. A single large payment whose benefits accrue over thirty years looks very different at 0%, 3%, and 5%, so sensitivity analyses routinely vary the rate.
Other costing decisions to check
- Time horizon. How far into the future the analysis runs. Chronic disease and one-time curative treatments usually call for a lifetime horizon, since a short horizon can hide either long-term costs or long-term benefits.
- Costing year and currency. All costs should be inflated to a single stated year, and that year should be named.
- Charges compared with costs. United States hospital charges are list prices rather than resources consumed, so careful analyses convert them to costs or to standardized payment amounts.
Knowledge check
Choose an answer for each question, then check your score.
Cost-effectiveness and the ICER
Now the cost side and the outcome side come together. Cost-effectiveness analysis asks how much additional health a new strategy produces for the additional money it costs. The answer takes two forms: a single ratio, the incremental cost-effectiveness ratio (ICER), and a picture, the cost-effectiveness plane.
The ICER
When effects are measured in QALYs, the ICER reads as dollars per QALY gained. An intervention costing $40,000 more and gaining 0.5 QALYs has an ICER of $80,000 per QALY. The word incremental is doing the work in that sentence: the comparison is always against a specific alternative, usually current standard of care, never against doing nothing in the abstract.
The cost-effectiveness plane
Plot the increment as a point, with incremental effect on the horizontal axis and incremental cost on the vertical axis. The quadrant the point falls in tells you how to read the result. Drag the point below, or move the threshold slider, and the interpretation underneath updates.
| Quadrant | Meaning | Decision logic |
|---|---|---|
| Southeast | More effective, less costly | The new strategy dominates the comparator, so no ratio is needed. |
| Northeast | More effective, more costly | The case that needs a judgment: compare the ICER against a willingness-to-pay threshold. |
| Northwest | Less effective, more costly | The new strategy is dominated, so it is rejected. |
| Southwest | Less effective, less costly | Uncommon and uncomfortable: the question becomes whether the saving compensates for the health given up. |
Thresholds: what a system will pay for a QALY
A willingness-to-pay threshold turns an ICER into a yes or no signal. Reference points in common use:
- United States. No official threshold exists, and federal law restricts QALY-based thresholds in some coverage and payment programs. Published analyses commonly benchmark against a range of roughly $50,000 to $150,000 per QALY, and the Institute for Clinical and Economic Review (ICER, the organization, not the ratio) uses $100,000 to $150,000 per QALY in its value assessments.
- England. NICE applied a range of £20,000 to £30,000 per QALY for more than two decades. Following a government decision, the range rose to £25,000 to £35,000 per QALY for technology appraisals from April 2026. The highly specialised technologies route, used for very rare conditions, applies a higher threshold, and severity modifiers can raise the acceptable ratio further.
- Heuristics tied to national income. Rules of thumb keyed to gross domestic product per capita have been used for cross-country comparisons and are increasingly criticized as a basis for national decisions.
A threshold embodies a judgment about how much health the system believes it displaces elsewhere when it spends more here. Reasonable people disagree about the right number, and in England the recent change was made by government rather than by the appraisal body.
Uncertainty: no ICER is a single number
Every input, whether a utility, an event rate, or a price, is uncertain, so credible analyses stress-test the result:
- One-way sensitivity analysis varies one parameter at a time across a plausible range, often displayed as a tornado diagram that ranks parameters by how far each moves the ICER.
- Probabilistic sensitivity analysis (PSA) assigns distributions to all parameters and re-runs the model many thousands of times, producing a cloud of points on the plane and a cost-effectiveness acceptability curve, which shows the probability that the strategy is cost-effective at each threshold.
- Scenario analysis swaps structural assumptions, such as a different perspective, time horizon, or discount rate.
Ask what the comparator was, whose perspective the analysis took, over what horizon it ran, and how far the result moves in the sensitivity analyses. A headline ratio quoted without those answers cannot be interpreted.
Knowledge check
Choose an answer for each question, then check your score.
Budget impact and the analysis family
Cost-effectiveness answers whether something is good value. A payer's second question is different: can we afford it in this budget year? This panel maps the family of economic analyses and introduces the models that produce their numbers.
The classic analysis types
| Type | Effects measured in | Used when |
|---|---|---|
| Cost-minimization analysis (CMA) | Assumed equal | Outcomes are demonstrably equivalent, as with biosimilars, so only costs are compared. |
| Cost-effectiveness analysis (CEA) | Natural units | Cost per life-year, per case detected, or per hospitalization avoided; intuitive within a single disease. |
| Cost-utility analysis (CUA) | QALYs | The HTA workhorse, comparable across diseases. The literature often calls it CEA loosely. |
| Cost-benefit analysis (CBA) | Money | Health gains are converted to money, and a positive net benefit supports adoption. Less common in clinical HTA, more common in public health and policy. |
Budget impact analysis
A budget impact analysis (BIA) estimates the net change in a specific payer's spending over a short horizon, usually one to five years, if a new intervention is adopted. It is not a judgment about value: a therapy can be excellent value per QALY and still strain a budget, or poor value and inexpensive. The logic is a funnel:
- Start from the covered population.
- Narrow to the eligible patients, using prevalence, diagnosis, and indication criteria.
- Apply expected uptake, the share of eligible patients who move to the new therapy each year.
- Multiply by net cost per treated patient, which is the cost of the new therapy minus the costs it displaces, such as the old therapy and avoided hospitalizations.
The budget impact worksheet walks this funnel with your own numbers.
Cost-effectiveness asks whether each unit of health justifies its price. Budget impact asks what the total bill will be, for this payer, starting in the coming year. Reimbursement dossiers almost always need both.
Decision models
Trials rarely run long enough to observe lifetime costs and QALYs, so analysts build decision-analytic models that project beyond the observed data:
- Decision trees lay out branching paths with probabilities and payoffs, and suit short-horizon, one-off decisions such as whether to screen.
- Markov, or state-transition, models move patients among health states such as stable, progressed, and dead in repeated cycles, accumulating costs and QALYs in each state. They are the standard choice for chronic and progressive disease.
- Partitioned survival and microsimulation models are used where survival curves or individual-level variation drive the economics, as in oncology and rare disease.
Every model is a set of explicit assumptions, which is what makes it possible to challenge one. Patient communities often hold the knowledge that decides whether a model's structure is credible: natural history, caregiving burden, and what progression means in daily life.
Knowledge check
Choose an answer for each question, then check your score.
Real-world evidence and HTA
Trials show what a therapy can do under controlled conditions. Health systems pay for what it does in practice. Real-world evidence closes part of that gap, and in rare disease it is frequently the only evidence available.
Real-world data feeding HEOR
| Source | Typical use in HEOR |
|---|---|
| Patient registries | Natural history, progression rates, and survival, which become model transition probabilities; external comparator groups for single-arm trials. |
| Claims data | Resource use and costs, treatment patterns, adherence, and utilization at scale. |
| Electronic health records | Clinical detail that claims lack: laboratory values, functional scores, devices, and symptoms recorded in notes. |
| Patient-generated data | Surveys, wearables, and digital measures, which record function and symptoms between clinic visits. |
Standardizing these sources to common data models and shared terminologies is what lets them be reused across studies and jurisdictions. Many programs treat that infrastructure work as the thing that decides whether an economic model's inputs can be traced and reproduced.
Where real-world evidence enters the economic case
- Baseline risk and natural history: what happens without the new therapy, which is essential where a placebo group is unethical or infeasible.
- Long-term extrapolation: registries anchor survival projections beyond the end of trial follow-up.
- Real-world costs and resource use: hospitalizations, equipment, and caregiving intensity by disease stage.
- Utilities from real populations: PRO instruments administered in routine care or through a registry.
- Outcomes-based agreements: coverage tied to real-world performance, monitored through registries and claims.
The HTA landscape
| Body | Where | Character |
|---|---|---|
| NICE | England | Explicit cost-per-QALY framework, with a separate highly specialised technologies route for very rare conditions. |
| CDA-AMC | Canada | Canada's Drug Agency, formerly CADTH, issuing reimbursement recommendations to public plans. |
| IQWiG and the G-BA | Germany | Added-benefit assessment drives price negotiation, with less weight on the QALY. |
| PBAC | Australia | Statutory committee; listing requires demonstrated cost-effectiveness. |
| ICER | United States | Independent nonprofit with no statutory power, whose value assessments nonetheless inform payer negotiations. |
In the European Union, the Joint Clinical Assessment adds a shared clinical review layer while pricing and economic decisions stay national. It has applied to new cancer medicines and advanced therapy medicinal products since January 2025, extends to selected high-risk medical devices during 2026, reaches orphan medicines in January 2028, and covers all new medicinal products from 2030. The first joint assessment report, for an orphan medicine in pediatric low-grade glioma, was published in June 2026.
What makes rare disease harder
- Small populations produce wide confidence intervals and single-arm trials, which puts heavy weight on registry-based external controls and natural history studies.
- High per-patient prices push ratios well past conventional thresholds. Some systems respond with rarity or severity modifiers, special appraisal routes, or managed access agreements that continue collecting evidence.
- Generic utility instruments can miss what progression takes from patients and families, so condition-specific PROs and caregiver burden data are often added to strengthen the case.
- Caregiver and societal costs are frequently large and frequently excluded, which turns the choice of perspective into an advocacy question.
Every number in an economic model is an evidence claim: a utility someone measured, a cost someone tallied, a transition probability someone estimated. Communities that build rigorous registries, capture the outcomes patients care about, and understand how a model is put together are in a position to influence those inputs rather than only receiving the decision.
Knowledge check
Choose an answer for each question, then check your score.
ICER calculator and cost-effectiveness plane
Enter the total cost and total effectiveness of two strategies. The calculator computes the increments and the ratio, plots the result on the cost-effectiveness plane, and interprets it against a threshold you set. It pairs with the panel on cost-effectiveness and the ICER.
Step 1: define the comparison
Step 2: read the increments
Step 3: judge it against a threshold
Make the new strategy both cheaper and more effective, and watch it become dominant. Then set the incremental effect close to zero and watch the ratio explode, which is why tiny effect differences make ratios unstable. Then slide the threshold across your ratio: the decision flips while the evidence stays the same, because the threshold is a policy choice.
QALY builder
Model a patient pathway as a sequence of health states. Give each state a duration and a utility weight, and the builder totals the quality-adjusted life years with and without discounting. Build one pathway for usual care and one for treatment, then compare them. It pairs with the panel on measuring outcomes.
Pathway 1: without the intervention
| Health state | Years in state | Utility | QALYs |
|---|
Pathway 2: with the intervention
| Health state | Years in state | Utility | QALYs |
|---|
State durations come from natural history data in registries and cohort studies. Utilities come from EQ-5D-type instruments administered to patients, or from published utility catalogs. When you read a published model, each duration and each utility should trace to a citable source; where one traces only to an assumption, that is the place to ask questions.
Budget impact worksheet
Walk the funnel from covered lives to net budget impact for a single year. It pairs with the panel on budget impact and the analysis family.
Step 1: the population funnel
Step 2: cost per treated patient per year
Step 3: results
Multi-year uptake curves where adoption grows over time, market-share shifts among existing therapies, separate treatment of incident and prevalent patients, and results expressed per member per month. This worksheet is the single-year skeleton those refinements are built on.
HEOR study planner
A scaffold for scoping an economic evaluation or outcomes study, whether it is yours or one you are appraising. Fill it in, then print it or download your answers. Use Save session in the header to keep everything in a file you control.
Frame the decision
Analysis design choices
Evidence plan
They follow the reporting items reviewers look for, set out formally in the CHEERS 2022 checklist for economic evaluations, linked on the Sources panel. Filling this page in confidently is a reasonable test of whether you can read or scope an economic evaluation.
Glossary
Every acronym used in this workbench, spelled out, with a short definition. Type in the box to filter.
Sources and check dates
The reference points in this workbench are snapshots rather than fixed quantities. Each site below opens in a new tab. To reproduce a figure, open the source and read the current value; where a number here differs, the source is right and this page is out of date.
Thresholds, methods, and reporting standards
- NICE health technology evaluations: the manual (PMG36). The reference case for England: perspective, discount rate, utility instrument, and evidence requirements. Checked September 2026.
- NICE, changes to the cost-effectiveness thresholds. The move to £25,000 to £35,000 per QALY and when it took effect. Checked September 2026.
- CHEERS reporting guideline, EQUATOR Network. The checklist behind the study planner fields.
- ISPOR good practices reports. Task force guidance on budget impact analysis, modeling, and real-world evidence.
- Institute for Clinical and Economic Review. United States value assessment reports and the benchmarks used in them.
Health technology assessment bodies
- Canada's Drug Agency (CDA-AMC), formerly CADTH.
- IQWiG and the Federal Joint Committee (G-BA), Germany.
- Pharmaceutical Benefits Advisory Committee (PBAC), Australia.
- European Commission, health technology assessment. Joint Clinical Assessment scope and rollout dates. Checked September 2026.
Real-world data and evidence
- FDA real-world evidence program. Guidance on real-world data sources and study designs.
- OHDSI. The OMOP common data model and the open-source analytics built on it.
- Real-World Evidence Literature Map, Boyce Data Science. A corpus of real-world evidence papers by disease area, data source, and method.
The HEOR Workbench, Boyce Data Science. Nothing you type is sent anywhere: the calculators run in your browser, and Save session writes a file to your own computer.