Economic Evaluation and Resource Allocation

Introduction

Health-care systems operate under scarcity. Needs and expectations are extensive, while budgets, clinical staff, facilities, medicines, technologies and managerial capacity are limited. Economic evaluation provides a systematic way to compare alternative uses of these resources. It does not ask only whether an intervention works; it asks whether the additional health gained is worth the additional resources forgone elsewhere.

Economic evaluation is therefore an important component of evidence-informed and value-based health-care management. It supports decisions about adoption of medicines, screening programmes, service redesign, workforce models, capital investment and population-health interventions. However, economic evidence is not a substitute for ethical judgement, clinical expertise or public deliberation. Rather, it makes the opportunity costs and trade-offs of decisions more explicit.

Learning objectives

By the end of this chapter, the reader should be able to:

1. Distinguish cost analysis, cost-effectiveness, cost-utility, cost-benefit and budget-impact analysis.
2. Identify, measure and value relevant costs and outcomes.
3. Interpret QALYs, ICERs, discounting and sensitivity analysis.
4. Appraise economic models and their assumptions.
5. Apply economic evidence to organisational and policy resource allocation.
6. Recognise ethical, distributional, governance and implementation limitations.

1. The foundations of economic evaluation

Scarcity, choice and opportunity cost

Scarcity means that using a resource for one purpose prevents its use for another. The opportunity cost of an intervention is the value of the best alternative use of the resources committed to it. For example, funding an expensive oncology medicine may displace primary-care appointments, nursing posts or preventive services. A favourable clinical result does not eliminate this trade-off.

Economic evaluation compares at least two alternative courses of action. The comparator may be current practice, no programme, a less intensive service or another treatment. A study that measures only the cost of one programme is a cost analysis, not a full economic evaluation, unless it also compares outcomes and alternatives.

Perspective

Perspective determines which costs and consequences are included. Common perspectives are:

A perspective must be stated before analysis. Mixing perspectives can produce misleading conclusions. For example, excluding unpaid caregiver time while claiming a societal perspective understates the cost of a long-term disability programme.

Types of cost

Relevant costs may include:

• Direct medical costs: medicines, consultations, diagnostics, procedures, hospitalisation and rehabilitation.
• Direct non-medical costs: transport, accommodation and paid social care.
• Indirect costs: productivity losses, absenteeism and reduced unpaid work.
• Intangible costs: pain, anxiety and loss of dignity. These may be difficult to value but are not necessarily irrelevant.
• Capital costs: buildings, equipment and information systems.
• Opportunity costs: the value of benefits forgone when resources are diverted.

Fixed costs do not vary with short-run activity, whereas variable costs change with volume. A decision may concern marginal cost—the additional cost of one more unit of service—rather than average cost. Average cost is total cost divided by total output and can be inappropriate for incremental decisions.

2. Main forms of economic evaluation

Cost analysis and cost description

Cost analysis estimates the resources consumed by a programme or disease. Cost-of-illness studies may describe the economic burden of diabetes, cancer or mental illness, but they do not establish whether one intervention is better value than another. They are useful for planning and advocacy, but should not be interpreted as evidence of cost-effectiveness.

Cost-minimisation analysis

Cost-minimisation compares costs when alternatives are assumed to have equivalent outcomes. This assumption must be demonstrated, not merely asserted. If effectiveness differs, cost-minimisation is inappropriate and may favour a cheaper but inferior option.

Cost-effectiveness analysis

Cost-effectiveness analysis (CEA) compares costs with outcomes measured in natural units, such as strokes prevented, life-years gained, infections avoided or blood-pressure points reduced. It is particularly useful when outcomes are clinically specific and a single outcome captures the main objective.

The incremental cost-effectiveness ratio is:

Here, C₁ and E₁ are the cost and effect of the new intervention, and C₀ and E₀ are those of the comparator. ΔC is incremental cost, measured in currency, and ΔE is incremental effect, measured in natural health units. The ICER has units such as dollars per life-year gained. It is meaningful only when the alternatives, perspective, time horizon and outcome definition are clearly specified.

Cost-utility analysis

Cost-utility analysis (CUA) commonly uses quality-adjusted life-years (QALYs), combining length and quality of life. A QALY is usually calculated as:

where uₜ is the health-related quality-of-life utility during period t, bounded conventionally between 0 (equivalent to death) and 1 (full health), and Δt is the duration of that period in years. Health states worse than death may have negative utilities in some instruments.

For a person experiencing utility 0.8 for two years and 0.5 for one year, QALYs equal (0.8 × 2) + (0.5 × 1) = 2.1 QALYs. QALYs allow comparison across disease areas, but they involve value judgements about health states, methods of measurement and whose preferences count.

CUA is often preferred for broad resource allocation because it captures both survival and quality of life. Nevertheless, QALYs may inadequately represent equity, caregiver effects, mental health, or the special value society places on particular groups. Distributional consequences should therefore be reported alongside the QALY result.

Cost-benefit analysis

Cost-benefit analysis (CBA) expresses both costs and benefits in monetary terms. A basic net-benefit measure is:

A positive net benefit indicates that monetised benefits exceed costs, subject to the assumptions of the analysis. The benefit–cost ratio is:

Monetising health benefits can support comparison with non-health investments, but valuation may rely on willingness-to-pay, human-capital methods or stated-preference studies. These methods raise concerns about income-related inequality: people with greater ability to pay may appear to place a higher monetary value on health.

Budget-impact analysis

Budget-impact analysis (BIA) estimates the financial consequences of adopting an intervention for a specified population, payer and time horizon. Unlike CEA, BIA asks whether the intervention is affordable within a real budget. It should include eligible population size, uptake, treatment displacement, implementation costs, market shares and timing of expenditure.

A simplified annual budget impact is:

A programme can be cost-effective but unaffordable in the short term, or affordable but not cost-effective. Decision makers need both value information and cash-flow information.

3. Measuring costs and outcomes

Identification, measurement and valuation

Costing has three distinct stages. First, identify which resources are used. Second, measure quantities, such as staff hours, bed-days or doses. Third, value those quantities using unit prices or opportunity costs. The general cost identity is:

The subscript i denotes each resource category. Quantities and prices must refer to the same period, currency and perspective. Double counting must be avoided; for example, a hospital payment should not be counted as both a provider cost and a separate treatment cost in the same perspective.

Costs may be estimated from primary data, administrative databases, published tariffs, micro-costing or gross-costing. Micro-costing measures resources in detail and is useful when technologies differ substantially. Gross-costing applies an average cost to service units and is less demanding but less precise.

Time horizon and discounting

The time horizon should be long enough to capture all important differences in costs and outcomes. Chronic interventions may require a lifetime horizon. A short horizon can make preventive interventions appear unattractive because costs occur immediately while benefits occur later.

Discounting converts future costs and outcomes into present values. A general expression is:

where PV is present value, Xₜ is the cost or outcome occurring at time t, r is the annual discount rate expressed as a decimal, and t is the number of years from the present. The expression assumes a constant rate and discrete annual periods. For example, a cost of 121 currency units occurring in two years at a 10% rate has a present value of 121 ÷ 1.1² = 100.

Discount rates are policy-sensitive. Costs and health effects may be discounted at the same or different rates depending on national guidance. Analysts must follow the relevant jurisdictional standard, report the rate transparently and test alternative rates. Discounting should not be used to conceal uncertainty.

4. Decision rules and interpretation

Dominance

An intervention is strictly dominated if it costs more and produces fewer benefits than another option. It should be removed before ICER calculation. Extended dominance occurs when a combination of other options produces a more favourable incremental cost-effectiveness relationship than the option under consideration.

Consider three strategies ordered by increasing effectiveness:

The ICER of B versus A is 40 ÷ 0.5 = 80 currency units per effect unit. The ICER of C versus B is 80 ÷ 1.0 = 80. If a fourth strategy had an ICER lower than 80 compared with B, B might be subject to extended dominance, depending on the full frontier analysis. Incremental comparisons must be made after ordering options by effect and removing dominance.

Thresholds and net monetary benefit

An ICER is interpreted against a decision-maker’s willingness-to-pay threshold, λ, representing the maximum acceptable cost per unit of health outcome. An intervention is considered cost-effective when its ICER is below λ, subject to uncertainty and other criteria.

Net monetary benefit (NMB) expresses the same decision in monetary terms:

If NMB is positive, the intervention generates more monetised health value than its incremental cost at threshold λ. NMB is particularly useful in probabilistic analysis and when comparing multiple strategies. λ is not a universal scientific constant; it reflects social preferences, budget constraints and policy judgement.

Worked example: CEA and CUA

A hospital is considering a remote monitoring service. Current care costs 1,000 per patient and produces 0.70 QALYs. Remote monitoring costs 1,400 and produces 0.76 QALYs.

Incremental cost = 1,400 − 1,000 = 400.

Incremental effect = 0.76 − 0.70 = 0.06 QALYs.

At a threshold of 10,000 per QALY, the intervention appears cost-effective. At a threshold of 5,000, it does not. This conclusion assumes that the estimates are unbiased, the alternatives are correctly specified, the time horizon captures relevant effects and the threshold is appropriate. It does not mean that adoption is automatic; capacity, equity, implementation feasibility and budget impact remain relevant.

5. Uncertainty and sensitivity analysis

Economic evaluations contain parameter uncertainty, structural uncertainty and heterogeneity. Parameter uncertainty concerns values such as costs, probabilities and utilities. Structural uncertainty concerns model form, assumptions about disease progression or included pathways. Heterogeneity concerns genuine differences among patient groups.

Deterministic sensitivity analysis

One-way sensitivity analysis varies one parameter at a time. It identifies influential assumptions but may ignore correlations. Two-way analysis varies two parameters. Scenario analysis examines coherent alternatives, such as short versus lifetime horizons or different treatment-uptake rates. A tornado diagram may rank parameters by their effect on the outcome.

Probabilistic sensitivity analysis

Probabilistic sensitivity analysis assigns probability distributions to uncertain parameters and repeatedly samples from them. It estimates the probability that each strategy is cost-effective at different thresholds. Common choices include beta distributions for probabilities and utilities, gamma or log-normal distributions for positive costs, and normal distributions when appropriate. Distribution choice must reflect the data and parameter bounds, not convenience.

Cost-effectiveness acceptability curves show the probability that each strategy is cost-effective across values of λ. They do not show that an intervention is clinically effective with a particular probability; they show decision uncertainty conditional on the model and distributions.

Value of information

When uncertainty may lead to costly wrong decisions, further research can have economic value. Expected value of perfect information estimates the value of eliminating all parameter uncertainty; expected value of partial perfect information examines selected parameters. These methods can prioritise research, although estimates depend on the decision model and population affected.

6. Models and critical appraisal

Decision trees are suitable for short-term pathways with discrete events. Markov models represent movement among health states over repeated cycles and are useful for chronic disease. Individual-level simulation can represent patient histories, competing risks and complex interactions. Discrete-event simulation is useful when waiting times, queues and resource constraints are central.

A model should be appraised for:

1. Structure: Are health states, pathways and comparators clinically plausible?
2. Data: Are estimates relevant, current and transferable to the target population?
3. Internal validity: Does the model reproduce its input data and expected relationships?
4. External validity: Are predictions reasonable in the intended setting?
5. Transparency: Are assumptions, code, sources and calibration methods documented?
6. Uncertainty: Are parameter, structural and heterogeneity uncertainties addressed?
7. Validation: Have clinical experts, statistical checks and independent data been used?

Models should not create false precision. A result reported as 12,347 currency units per QALY may appear exact even when uncertainty is substantial. Managers should examine ranges, drivers and decision consequences rather than rely on a single point estimate.

7. Applied case: introducing a digital heart-failure monitoring service

The following is an illustrative case; figures are hypothetical and are not a recommendation for actual procurement.

Background and problem

A regional health system experiences frequent readmissions among adults with heart failure. It is considering a digital monitoring service that transmits weight and symptom data to a specialist team. The service may reduce admissions but requires software, training and clinical review time.

Analysis

The payer perspective is selected, with a five-year horizon. Current care costs 8,000 per patient over the period and yields 2.40 QALYs. Monitoring costs 8,900 and yields 2.47 QALYs. Costs and outcomes are discounted according to the applicable national guidance. The incremental cost is 900 and incremental effect is 0.07 QALYs:

At a threshold of 20,000, the base case is potentially cost-effective. However, the first-year budget impact is substantial because the full implementation cost occurs before all avoided admissions are realised. If 2,000 patients are eligible and uptake is 40%, 800 patients receive the service. At an incremental cost of 900 per patient, the simplified five-year incremental programme cost is 720,000, before considering fixed implementation costs and any savings from avoided admissions already included in the per-patient estimate.

Sensitivity analysis identifies admission reduction, staff workload and utility improvement as the main drivers. A scenario with low uptake produces a smaller total budget impact but may reduce learning and fixed-cost efficiency. A scenario with high referral volume creates a staffing bottleneck.

Approach and solution

The health system establishes a staged adoption plan: a limited pilot, predefined clinical and economic outcomes, interoperability testing, staff training and a review after twelve months. The evaluation includes readmissions, mortality, patient-reported quality of life, equity of access, staff time, alert burden and unplanned contacts. Patients without smartphones are offered an equivalent non-digital pathway to avoid avoidable exclusion.

Outcome and learning points

The base case supports conditional adoption, not unconditional scale-up. The decision combines cost-effectiveness, affordability, implementation capacity and equity. The case illustrates that a favourable ICER can coexist with a difficult short-term budget impact. It also demonstrates why digital interventions require evaluation of workflow, data governance and access, not merely software licence cost.

8. Resource allocation in practice

Economic evaluation can inform several organisational decisions:

• Portfolio prioritisation: compare programmes using consistent outcomes and thresholds.
• Disinvestment: identify low-value or obsolete services, while considering contractual and ethical consequences.
• Procurement: include acquisition, maintenance, training, consumables, interoperability and disposal costs.
• Workforce planning: evaluate whether an intervention shifts costs to scarce staff or reduces workload.
• Prevention: use an appropriate long-term horizon and account for delayed benefits.
• Equity-sensitive planning: examine results by age, deprivation, geography, disability, ethnicity and disease severity.

A technically sound evaluation may still fail in practice if it ignores implementation. Decision makers should distinguish efficacy from effectiveness, account for real-world adherence and assess whether the organisation can deliver the intervention at the assumed quality.

9. Ethical, governance and contemporary issues

Resource allocation raises distributive questions: Should priority go to those with the greatest need, the greatest capacity to benefit, the youngest, the most disadvantaged or the least costly to treat? Maximising QALYs may conflict with equity. Approaches such as equity weights, distributional cost-effectiveness analysis and deliberative public engagement can make these trade-offs explicit, but they do not remove the need for judgement.

Analysts should disclose funding, conflicts of interest, data exclusions, model assumptions and deviations from protocols. Patient confidentiality must be protected when using linked administrative or digital data. Procurement decisions should not be based on vendor-sponsored models without independent validation.

Artificial intelligence may support risk prediction, demand forecasting, automated costing and identification of high-cost pathways. However, predictive accuracy does not establish cost-effectiveness. AI models require assessment of calibration, bias, explainability, data drift, cybersecurity, human oversight and the costs of false positives and false negatives. Generative AI may assist documentation but should not replace expert verification of economic assumptions.

Contemporary evaluation increasingly addresses climate effects, resilience, social care, informal caregiving and distributional outcomes. These extensions are valuable when relevant, but analysts should avoid adding poorly measured outcomes simply to make a study appear comprehensive.

Best-practice checklist

A high-quality economic evaluation should:

1. State the decision problem, population, alternatives and perspective.
2. Use a clinically credible comparator.
3. Identify all relevant costs and outcomes.
4. Select a time horizon that captures important consequences.
5. Report quantities, unit costs, currency year and discount rates.
6. Use an appropriate model and justify its structure.
7. Address dominance and calculate incremental results correctly.
8. Present uncertainty transparently, including scenario and probabilistic analyses where suitable.
9. Examine equity, implementation, affordability and ethical implications.
10. Disclose funding, conflicts, data sources and limitations.

Chapter Summary

Economic evaluation compares alternative uses of scarce health-care resources. Its central concept is opportunity cost: resources allocated to one intervention cannot simultaneously produce benefits elsewhere. Perspective determines which costs and outcomes are included, while the time horizon determines whether delayed effects are captured.

Cost-effectiveness analysis uses natural clinical units; cost-utility analysis commonly uses QALYs; cost-benefit analysis monetises costs and benefits; and budget-impact analysis estimates affordability within a specified population and budget period. Incremental analysis, dominance assessment and ICER interpretation are essential. QALYs combine quality and length of life, while discounting expresses future values in present terms.

Sensitivity analysis reveals how conclusions change under uncertainty. Decision trees, Markov models and simulation models can support complex evaluations, but their credibility depends on transparent assumptions, relevant data and validation. Finally, economic evidence must be integrated with equity, ethics, feasibility, workforce capacity, governance and public values. Value-based allocation is therefore a structured judgement process, not mechanical adherence to a single ratio.

Key Terms

Opportunity cost: The value of the best alternative use of resources committed to a decision.

Perspective: The viewpoint determining which costs and outcomes are included.

Cost-effectiveness analysis: Comparison of incremental costs with outcomes measured in natural health units.

Cost-utility analysis: Economic evaluation using preference-based measures such as QALYs.

Cost-benefit analysis: Evaluation expressing both costs and benefits in monetary terms.

Budget-impact analysis: Estimation of the financial effect of adopting an intervention over a defined period.

QALY: A year of life adjusted for health-related quality of life.

Utility: A preference-based value assigned to a health state.

ICER: Incremental cost divided by incremental effect.

Net monetary benefit: Monetised incremental health benefit minus incremental cost.

Willingness-to-pay threshold: Maximum acceptable cost per unit of health outcome.

Discounting: Conversion of future costs or outcomes into present values.

Dominance: A situation in which an option is more costly and less effective than another.

Sensitivity analysis: Examination of how results change when assumptions or parameters vary.

Probabilistic sensitivity analysis: Repeated analysis using probability distributions for uncertain parameters.

Decision tree: A model representing discrete pathways and events.

Markov model: A model representing transitions among health states over cycles.

Micro-costing: Detailed measurement and valuation of individual resource inputs.

Distributional cost-effectiveness analysis: Evaluation incorporating the distribution of costs and health benefits across groups.

Review Questions

Short Answer Questions

1. What is opportunity cost, and why is it central to health-care resource allocation?
2. Distinguish cost-effectiveness analysis from budget-impact analysis.
3. Define a QALY and state how it combines length and quality of life.
4. What does an ICER measure?
5. Why is perspective important in an economic evaluation?

Long Answer Questions

1. Explain the principal forms of economic evaluation and identify an appropriate use for each.
2. Discuss the assumptions, strengths and limitations of QALYs in priority setting.
3. Explain how discounting affects the evaluation of preventive and chronic-care interventions.
4. Describe how a Markov model can be appraised for credibility and usefulness.
5. Discuss why a cost-effective intervention may nevertheless be unacceptable for immediate adoption.

Analytical / Application Questions

1. An intervention costs 2,500 per patient and produces 1.20 QALYs, while current care costs 2,000 and produces 1.15 QALYs. Calculate the ICER and interpret it against thresholds of 8,000 and 15,000 per QALY.
2. Design a budget-impact analysis for a digital diabetes-management programme. Specify the perspective, population, uptake assumptions, time horizon, cost categories and outcomes that should be reported.
3. A programme improves average QALYs but primarily benefits affluent urban patients. Explain how an equity-sensitive evaluation could examine this result and what additional evidence a decision maker should request before scale-up.

Multiple Choice Questions with Answers

Question 1

Which concept represents the value of the best alternative use of resources committed to an intervention?

A. Average cost

B. Opportunity cost

C. Fixed cost

D. Sunk cost

Correct Answer: B. Opportunity cost

Question 2

An intervention costs 1,400 compared with 1,000 for current care and produces 0.76 rather than 0.70 QALYs. What is its ICER?

A. 400 currency units per QALY

B. 667 currency units per QALY

C. 6,667 currency units per QALY

D. 24,000 currency units per QALY

Correct Answer: C. 6,667 currency units per QALY

Question 3

What is the principal outcome measure commonly used in cost-utility analysis?

A. Hospital bed-days

B. QALYs

C. Market share

D. Staff turnover

Correct Answer: B. QALYs

Question 4

What is the main purpose of budget-impact analysis?

A. To estimate affordability within a specified budget and population

B. To prove that all clinical outcomes are equivalent

C. To monetise every health benefit

D. To eliminate uncertainty from a model

Correct Answer: A. To estimate affordability within a specified budget and population

Question 5

In the formula PV = Xₜ ÷ (1 + r)ᵗ, what does r represent?

A. The outcome utility

B. The annual discount rate

C. The sample size

D. The incremental effect

Correct Answer: B. The annual discount rate

Question 6

An intervention that costs more and produces fewer health benefits than another option is described as:

A. Cost-saving

B. Strictly dominated

C. Extendedly beneficial

D. Budget neutral

Correct Answer: B. Strictly dominated

Question 7

Which analysis varies uncertain parameters repeatedly using assigned probability distributions?

A. One-way sensitivity analysis

B. Break-even analysis

C. Probabilistic sensitivity analysis

D. Micro-costing

Correct Answer: C. Probabilistic sensitivity analysis

Question 8

What does net monetary benefit calculate?

A. Total cost divided by total patients

B. Monetised incremental health benefit minus incremental cost

C. Future cost multiplied by the discount rate

D. Average utility multiplied by market share

Correct Answer: B. Monetised incremental health benefit minus incremental cost

Question 9

Which perspective generally includes costs and consequences falling on patients, providers, payers and society?

A. Hospital perspective

B. Payer perspective

C. Patient perspective

D. Societal perspective

Correct Answer: D. Societal perspective

Question 10

Why can a cost-effective intervention still be difficult to adopt immediately?

A. It necessarily has worse clinical outcomes

B. Its ICER is always negative

C. It may have an unaffordable short-term budget impact or exceed implementation capacity

D. Economic evaluation excludes all implementation issues

Correct Answer: C. It may have an unaffordable short-term budget impact or exceed implementation capacity