Cost‑Effectiveness Analysis in Public Health

Expert-defined terms from the Advanced Certificate in Economic Evaluation in Public Health Interventions course at LearnUNI. Free to read, free to share, paired with a professional course.

Download PDF Free · printable · SEO-indexed
Cost‑Effectiveness Analysis in Public Health

The ACER is the ratio of total costs to total health outcomes for a single inter… #

It provides a baseline measure of efficiency but does not compare alternatives directly.

Example #

In a vaccination program, the ACER might be $150 per disability‑adjusted life year (DALY) averted.

Practical application #

Useful for preliminary screening of interventions when resources are limited.

Challenges #

ACER can be misleading if multiple interventions are compared because it ignores the incremental nature of decision‑making and may overstate the value of a less efficient option.

A BIA estimates the financial consequences of adopting a new health intervention… #

It focuses on affordability rather than value for money.

Example #

A health department evaluates the 5‑year budget impact of introducing a new hepatitis B vaccine, projecting incremental expenses and potential savings from avoided treatments.

Practical application #

Helps policymakers decide whether an intervention can be funded given fiscal constraints.

Challenges #

Requires detailed cost data, assumptions about uptake rates, and may be sensitive to changes in population demographics or policy environment.

The CET represents the maximum amount a decision‑maker is prepared to pay for a… #

g., per QALY or DALY). It serves as a benchmark to judge whether an ICER is acceptable.

Example #

In many low‑ and middle‑income countries, the CET is set at one to three times the gross domestic product per capita.

Practical application #

Enables consistent ranking of interventions across disease areas.

Challenges #

Determining an appropriate CET is contentious; thresholds may not reflect true opportunity costs and can be politically driven.

CEA compares the relative costs and outcomes (often expressed in natural units l… #

The result is usually expressed as an ICER.

Example #

Comparing a school‑based nutrition program to a community‑wide health education campaign, CEA quantifies the additional cost per child gaining one healthy weight year.

Practical application #

Guides resource allocation decisions by identifying interventions that provide the greatest health benefit per dollar spent.

Challenges #

Requires robust data on both costs and effects; results can be sensitive to methodological choices such as discount rates and perspective.

CUA is a form of CEA that incorporates patient preferences for different health… #

It allows comparison across diverse health programs.

Example #

A CUA of a new antiretroviral regimen may report an ICER of $800 per QALY gained compared with standard therapy.

Practical application #

Facilitates cross‑program priority setting when health benefits differ in nature (e.g., mental health vs. infectious disease).

Challenges #

Utility elicitation can be methodologically complex; cultural differences may affect preference measurement.

The discount rate adjusts future costs and health outcomes to their present valu… #

Standard rates range from 3 % to 5 % per annum.

Example #

A cost incurred in year 5 is multiplied by (1 + discount rate)^‑5 to obtain its present value.

Practical application #

Ensures comparability of interventions with different time horizons, such as preventive vaccines versus acute treatments.

Challenges #

Choice of discount rate can substantially alter conclusions; different rates for costs and outcomes add complexity.

An intervention is strictly dominated if it is both more costly and less effecti… #

Extended (or weak) dominance occurs when an intervention’s ICER is higher than that of a more effective option, making it inefficient.

Example #

A screening test that costs $200 more and prevents 0.5 fewer cases than another test is strictly dominated.

Practical application #

Dominated options are excluded from further analysis, simplifying the decision set.

Challenges #

Accurate identification of dominance requires precise estimates; measurement error may misclassify interventions.

HALY is a generic metric that combines mortality and morbidity into a single fig… #

QALYs and DALYs are specific types of HALY.

Example #

A program that reduces malaria incidence may produce 10,000 HALYs over a decade.

Practical application #

Provides a common language for comparing disparate health programs.

Challenges #

Selection of the appropriate HALY type influences results; data on disability weights may be scarce.

The ICER is the ratio of the difference in costs to the difference in effectiven… #

It quantifies the additional cost required to gain one extra unit of health outcome.

Example #

An ICER of $1,200 per QALY indicates that each additional QALY achieved by a new drug costs $1,200 compared with the comparator.

Practical application #

Central decision‑making metric in CEA; interventions with ICERs below the CET are considered cost‑effective.

Challenges #

Sensitive to small differences in effectiveness; uncertainty around the ICER often requires probabilistic analysis.

INB translates cost‑effectiveness into monetary terms #

INB = (λ × ΔEffect) – ΔCost, where λ is the willingness‑to‑pay threshold. A positive INB indicates that an intervention is cost‑effective at the given λ.

Example #

With λ = $5,000 per DALY, ΔEffect = 0.8 DALYs, and ΔCost = $3,200, INB = $1,800, suggesting cost‑effectiveness.

Practical application #

Facilitates statistical testing and regression analysis in CEA.

Challenges #

Requires specification of λ, which may be uncertain; interpretation depends on the chosen threshold.

A Markov model is a mathematical framework that simulates the progression of a c… #

A Markov model is a mathematical framework that simulates the progression of a cohort through a series of health states over discrete time cycles, allowing for recurring events and long‑term outcomes.

Example #

A chronic disease model with states “healthy,” “diseased,” “complication,” and “death” can estimate lifetime costs and QALYs of a therapeutic regimen.

Practical application #

Captures disease dynamics where events can recur, such as infection cycles or treatment failures.

Challenges #

Requires assumptions about transition probabilities; model complexity can increase computational burden.

NMB is calculated as (λ × Effect) – Cost #

It expresses the value of an intervention in monetary terms, facilitating comparison across alternatives. The intervention with the highest NMB is deemed optimal.

Example #

For λ = $3,000 per QALY, an intervention achieving 0.5 QALYs at $1,200 yields an NMB of $300.

Practical application #

Simplifies decision analysis and allows for probabilistic sensitivity analysis using standard statistical software.

Challenges #

Dependent on the chosen λ; if λ is mis‑specified, NMB may mislead decision makers.

Opportunity cost represents the health benefits forgone when resources are alloc… #

In CEA, it is often expressed as the health gain that could have been achieved elsewhere with the same spending.

Example #

Funding a new screening program may prevent 1,000 DALYs but simultaneously displace a vaccination program that could have averted 1,500 DALYs; the latter loss is the opportunity cost.

Practical application #

Provides an economic rationale for setting realistic CETs based on actual health system constraints.

Challenges #

Quantifying opportunity costs requires comprehensive data on all competing interventions, which is rarely available.

PSA incorporates parameter uncertainty by assigning probability distributions to… #

g., ICERs).

Example #

Assigning a beta distribution to utility weights and a gamma distribution to cost parameters, then running 10,000 simulations to produce a cloud of ICER points.

Practical application #

Provides decision makers with the probability that an intervention is cost‑effective at various thresholds.

Challenges #

Requires computational resources and expertise in statistical modeling; results can be sensitive to the choice of distributions.

A QALY combines length of life with quality of life, where each year of perfect… #

A QALY combines length of life with quality of life, where each year of perfect health equals 1 QALY and less-than-perfect health is weighted by a utility value between 0 and 1.

Example #

A patient living 5 years with a utility of 0.8 accrues 4 QALYs (5 × 0.8).

Practical application #

Enables comparison of interventions across disease areas by standardizing health outcomes.

Challenges #

Utility measurement can be culturally sensitive; some argue QALYs undervalue certain health states or populations.

A reference case defines a set of methodological assumptions (e #

g., perspective, discount rate, time horizon) that should be consistently applied in CEAs to enhance comparability and credibility.

Example #

The WHO reference case recommends a societal perspective, 3 % discount rate, and lifetime horizon for chronic disease evaluations.

Practical application #

Researchers align their analyses with the reference case to facilitate policy uptake.

Challenges #

Strict adherence may limit flexibility needed for context‑specific analyses; updating reference cases can be slow.

Sensitivity analysis explores how changes in key parameters affect the results o… #

Sensitivity analysis explores how changes in key parameters affect the results of a CEA, identifying which variables drive uncertainty.

Example #

Varying vaccine efficacy from 70 % to 95 % and observing the impact on the ICER.

Practical application #

Helps prioritize data collection and informs robustness of conclusions.

Challenges #

One‑way sensitivity may oversimplify interactions; comprehensive multi‑parameter analysis can be resource‑intensive.

The societal perspective includes all costs and benefits regardless of who incur… #

g., transportation), and indirect costs such as productivity losses.

Example #

In evaluating a smoking cessation program, costs include counseling fees, participant travel, and lost workdays due to quitting attempts.

Practical application #

Provides the most comprehensive estimate of an intervention’s economic impact, supporting broader policy decisions.

Challenges #

Data on indirect costs may be scarce; assigning monetary values to intangible outcomes can be controversial.

The time horizon defines the period over which costs and effects are measured in… #

It should be long enough to capture all relevant differences between interventions.

Example #

A lifetime horizon is appropriate for chronic disease interventions, while a 5‑year horizon may suffice for short‑term treatments.

Practical application #

Determines the extent of future benefits included, influencing the ICER.

Challenges #

Longer horizons increase uncertainty; projecting costs and outcomes far into the future may require strong assumptions.

WTP represents the maximum amount a society or individual is prepared to spend f… #

g., per QALY). It is often used interchangeably with the cost‑effectiveness threshold.

Example #

A WTP of $50,000 per QALY is frequently cited in high‑income country analyses.

Practical application #

Guides interpretation of ICERs and informs reimbursement decisions.

Challenges #

Estimating WTP empirically is difficult; values may vary across populations and over time.

Yield refers to the amount of health benefit generated per unit of resource inve… #

Yield refers to the amount of health benefit generated per unit of resource invested, often expressed as cases averted, DALYs saved, or QALYs gained per dollar spent.

Example #

A malaria net distribution program may yield 0.02 DALYs averted per $1 invested.

Practical application #

Helps compare efficiency of interventions with different scales and modalities.

Challenges #

Accurate measurement requires reliable epidemiological data; yield can be context‑dependent.

A zero‑cost intervention is one that incurs no additional financial outlay for t… #

A zero‑cost intervention is one that incurs no additional financial outlay for the implementing agency, often because costs are borne by another party or are already embedded in existing services.

Example #

A health promotion campaign that uses existing community radio channels without extra fees.

Practical application #

Attractive in resource‑constrained settings where any additional spending is scrutinized.

Challenges #

Hidden costs (e.g., staff time) may be overlooked, leading to underestimation of true resource use.

Costs associated with negative side effects of an intervention, including medica… #

Incorporating adverse event costs ensures a more accurate net cost estimate.

Example #

Adding the cost of managing chemotherapy‑induced neutropenia to the total cost of a cancer regimen.

Practical application #

Helps decide whether the benefits of a high‑risk intervention outweigh its additional costs.

Challenges #

Data on frequency and severity of adverse events may be limited; valuation of patient discomfort is complex.

The BCR is the ratio of total benefits (expressed in monetary terms) to total co… #

A BCR > 1 indicates that benefits exceed costs.

Example #

A BCR of 1.8 means $1.80 of benefit for every $1 spent.

Practical application #

Provides a straightforward metric for policymakers assessing the overall economic return of a program.

Challenges #

Translating health outcomes into monetary values can be ethically contentious; BCR does not convey the distribution of benefits.

CBA compares the monetary value of all benefits of an intervention with its tota… #

Unlike CEA, CBA expresses health outcomes in monetary terms.

Example #

Valuing a reduction in disease burden using a VSL (value of statistical life) and subtracting program costs to calculate net benefit.

Practical application #

Allows direct comparison of health programs with non‑health projects (e.g., infrastructure).

Challenges #

Requires robust methods to assign monetary values to health outcomes; may undervalue intangible benefits.

NPV aggregates all discounted future costs and benefits into a single present‑va… #

NPV aggregates all discounted future costs and benefits into a single present‑value figure, indicating whether an intervention yields a net gain (> 0) or loss (< 0).

Example #

An intervention with discounted benefits of $5 million and discounted costs of $3 million has an NPV of $2 million.

Practical application #

Used in investment decisions and long‑term health program planning.

Challenges #

Sensitive to discount rate selection; future benefits that are highly uncertain can dominate NPV calculations.

External validity refers to the extent to which CEA results can be applied to se… #

External validity refers to the extent to which CEA results can be applied to settings or populations beyond the original study environment.

Example #

A CEA conducted in urban hospitals may have limited external validity for rural clinics due to differing cost structures.

Practical application #

Encourages analysts to report context‑specific parameters to aid adaptation.

Challenges #

Differences in healthcare systems, price levels, and disease epidemiology can hinder transferability.

Fixed costs are expenses that do not change with the volume of services delivere… #

In CEA, they are allocated across the expected number of units treated.

Example #

A laboratory’s $200,000 annual rent is a fixed cost that must be spread over all tests performed.

Practical application #

Accurate allocation of fixed costs prevents underestimation of per‑unit costs.

Challenges #

Determining appropriate allocation bases (e.g., number of patients, time) can be arbitrary.

HIA is a systematic process that predicts the health effects of a policy, progra… #

HIA is a systematic process that predicts the health effects of a policy, program, or project before it is implemented, often incorporating economic evaluation elements like CEA.

Example #

An HIA of a new public transport system may estimate reductions in traffic‑related injuries and associated cost‑effectiveness.

Practical application #

Informs policymakers about potential health consequences, supporting evidence‑based decisions.

Challenges #

Requires interdisciplinary expertise; predictions may be uncertain.

This metric expresses the additional cost required to prevent one additional cas… #

It is calculated as ΔCost ÷ ΔCases Averted.

Example #

An incremental cost of $10,000 to avert 100 extra cases yields $100 per case averted.

Practical application #

Useful when the primary outcome is disease incidence rather than QALYs or DALYs.

Challenges #

Does not capture severity differences between cases; may overlook broader societal benefits.

In probabilistic analysis, the joint probability distribution defines the simult… #

In probabilistic analysis, the joint probability distribution defines the simultaneous behavior of multiple correlated parameters, ensuring realistic simulation of combined uncertainty.

Example #

Correlating drug price and adherence rates to reflect that higher prices may reduce adherence.

Practical application #

Improves the credibility of PSA results by preserving relationships between variables.

Challenges #

Requires data on parameter correlations, which are often unavailable.

Long‑term follow‑up refers to the extended observation period after an intervent… #

Long‑term follow‑up refers to the extended observation period after an intervention to capture delayed effects, cost offsets, or sustainability of benefits.

Example #

Monitoring cardiovascular outcomes for ten years after a lifestyle intervention.

Practical application #

Provides evidence on durability of health gains, influencing the choice of time horizon.

Challenges #

Attrition, changing standards of care, and data collection costs can limit feasibility.

A Markov cohort model tracks a hypothetical group of individuals as they transit… #

A Markov cohort model tracks a hypothetical group of individuals as they transition between defined health states over successive cycles, assuming that all members in a state are identical.

Example #

A cohort of patients with chronic kidney disease moving between “stage 3,” “stage 4,” “dialysis,” and “death.”

Practical application #

Enables estimation of lifetime costs and outcomes for chronic conditions.

Challenges #

The “memoryless” property may oversimplify diseases where history influences future risk.

NHB expresses the health gain of an intervention after accounting for its costs,… #

It allows direct comparison of health outcomes without monetary conversion.

Example #

With λ = $20,000 per QALY, an intervention costing $4,000 and delivering 0.3 QALYs yields an NHB of 0.1 QALYs.

Practical application #

Facilitates decision making when monetary thresholds are controversial.

Challenges #

Requires a chosen λ; interpretation may be less intuitive for non‑economic audiences.

The opportunity cost of capital reflects the return that could be earned on the… #

The opportunity cost of capital reflects the return that could be earned on the funds if they were invested elsewhere, influencing the choice of discount rate for financial costs.

Example #

Using a 5 % discount rate to reflect the average return on government bonds.

Practical application #

Aligns health economic analyses with broader fiscal policy considerations.

Challenges #

Capital markets fluctuate; selecting a single rate may not capture sector‑specific financing conditions.

A partial economic evaluation focuses on a subset of cost components (e #

g., only direct medical costs) or outcomes, without conducting a full CEA.

Example #

Estimating only the program’s implementation costs without measuring health outcomes.

Practical application #

Useful for early‑stage assessments or when data on outcomes are unavailable.

Challenges #

May mislead decision makers if omitted costs or benefits are substantial.

Weighting involves assigning a utility value to each health state based on popul… #

Weighting involves assigning a utility value to each health state based on population preferences, typically derived from instruments like EQ‑5D or standard gamble methods.

Example #

A health state “moderate arthritis” receives a utility weight of 0.6.

Practical application #

Enables conversion of survival time into QALYs for CUA.

Challenges #

Preference heterogeneity across cultures; methodological differences can produce varying weights.

Resource allocation is the process of distributing limited health resources amon… #

CEA provides evidence to inform this process.

Example #

Allocating a fixed budget to fund a mix of vaccination, screening, and health education programs based on their cost‑effectiveness.

Practical application #

Supports transparent and evidence‑based decision making at national or institutional levels.

Challenges #

Political pressures, equity considerations, and stakeholder interests may conflict with pure efficiency criteria.

A sensitivity coefficient quantifies the degree to which changes in a specific p… #

A sensitivity coefficient quantifies the degree to which changes in a specific parameter affect the ICER or NMB, often expressed as the slope of the relationship.

Example #

A 10 % increase in drug price leading to a 5 % rise in the ICER yields a sensitivity coefficient of 0.5.

Practical application #

Identifies high‑impact parameters for targeted data collection or further research.

Challenges #

Linear approximations may not capture non‑linear effects; interactions among parameters can complicate interpretation.

In some models, the probability of moving between health states varies over time… #

In some models, the probability of moving between health states varies over time, reflecting disease progression or aging effects.

Example #

The probability of developing hypertension increases with each successive year in a cohort model.

Practical application #

Improves realism of long‑term simulations for chronic diseases.

Challenges #

Requires detailed longitudinal data; increases model complexity.

The utility scale ranges from 0 (equivalent to death) to 1 (perfect health), tho… #

The utility scale ranges from 0 (equivalent to death) to 1 (perfect health), though negative values are allowed for health states considered worse than death.

Example #

A utility of 0.85 indicates a health state 85 % as desirable as perfect health.

Practical application #

Standardizes the measurement of health‑related quality of life across studies.

Challenges #

Different elicitation methods (e.g., time trade‑off vs. visual analogue scale) may produce divergent utilities.

Variable costs change proportionally with the volume of services delivered, such… #

Variable costs change proportionally with the volume of services delivered, such as consumables, labor per patient, or medication.

Example #

The cost of syringes increases with each additional vaccination administered.

Practical application #

Critical for estimating incremental costs when scaling up an intervention.

Challenges #

Distinguishing between truly variable and semi‑fixed costs can be ambiguous.

Threshold adjustment updates the CET over time to reflect changes in macro‑econo… #

Threshold adjustment updates the CET over time to reflect changes in macro‑economic conditions, health system capacity, or societal preferences.

Example #

Raising the CET from $30,000 to $35,000 per QALY following a 5 % increase in GDP per capita.

Practical application #

Keeps cost‑effectiveness decisions aligned with current fiscal realities.

Challenges #

Frequent adjustments may create instability; political influences can distort the process.

July 2026 intake · open enrolment
from £90 GBP
Enrol