Efficiency

Efficiency in Health Care

Efficiency is broader than simply spending less or completing more activity. In health economics, it asks whether scarce resources are being used in ways that avoid unnecessary waste and produce the greatest feasible value. In health care it takes technical, productive, allocative, scale and dynamic forms, each measured with different methods, and it can conflict with equity, so efficiency claims need careful interpretation.

What an efficiency assessment needs to specify

An efficiency assessment compares resource use with what is produced and with what could feasibly have been produced instead. The conclusion depends on how the analyst defines the decision-making unit, its objectives and the alternatives available to it. A statement that one provider, intervention or health system is more efficient than another is difficult to interpret unless these elements are explicit.

An assessment should specify the following elements.

• The decision-making unit identifies the organisation, service, programme, intervention or health system whose efficiency is being assessed.
• The inputs identify the resources used to produce healthcare or health.
• The outputs and outcomes identify what the decision-making unit produces and what ultimately matters to patients or populations.
• The objective identifies what the decision-maker is trying to maximise or achieve.
• The feasible alternatives identify the realistic ways in which the resources could have been used.
• The time horizon identifies when resource use and resulting consequences are counted.
• The population identifies whose resource use, services and outcomes are included.
• Quality and case-mix adjustments address important differences between the services or populations being compared.
• The constraints identify limits such as budgets, workforce capacity, geography, technology and ethical obligations.

Efficiency is therefore always assessed relative to a defined objective, resource constraint and set of feasible alternatives. It is not an intrinsic property that can be established from expenditure or activity alone.

How inputs become outputs and outcomes

Efficiency begins with a production relationship in which inputs are used to generate outputs and, ultimately, outcomes. These categories should be kept distinct because a service can produce a high volume of activity without producing correspondingly valuable health outcomes. Selecting the wrong result measure can make avoidable or ineffective activity appear efficient.

Inputs are the resources used to deliver care or improve health. Typical inputs are listed below.

• Expenditure used to purchase healthcare resources.
• Staff time contributed by clinical, administrative and support personnel.
• Medicines, devices and other consumable resources.
• Beds, buildings, diagnostic equipment and other physical capacity.
• Digital systems, information and managerial capacity.
• Patient and caregiver time when the analytical perspective includes those resources.

Outputs are services or activities produced from those inputs. Typical outputs are listed below.

• Consultations, procedures and diagnostic tests.
• Admissions, discharges or completed episodes of care.
• Vaccinations, prescriptions or treatment courses.
• People screened, treated or reached by a programme.

Outcomes are consequences that matter to patients, populations or decision-makers. Typical outcomes are listed below.

• Survival, symptoms and functional ability.
• Quality-adjusted life years or other measures of health gain.
• Complications, readmissions or adverse events avoided.
• Patient experience, access and continuity of care.
• Equity and financial protection when these form part of the decision objective.

Producing more outputs does not necessarily mean producing better outcomes. Additional consultations, tests or procedures may have little value when they are unnecessary, ineffective, unsafe or poorly targeted.

Why a simple productivity ratio is not enough

A basic productivity ratio describes how much measured output is produced for each unit of input. It can be useful as a starting point, especially when the inputs and outputs are comparable and clearly defined. It does not, by itself, show that resource use is efficient.

A simple ratio can be written as follows.

$$\text{Output-to-input ratio} = \frac{\text{Measured output}} {\text{Measured input}}$$

For example, consultations per staff hour or procedures per £1 million describe how intensively measured resources are converted into services. A higher ratio could reflect better organisation, but it could also reflect lower quality, easier patients, omitted inputs or unnecessary activity.

A simple output-to-input ratio may fail to account for several things.

• Differences in service quality or patient outcomes.
• Differences in disease severity, complexity or case mix.
• Inappropriate or avoidable activity.
• Multiple inputs and multiple outputs.
• Resources omitted from the chosen perspective.
• Uncertainty and random variation.
• Whether a different allocation would produce greater value.

Productivity and efficiency are therefore related but distinct. Productivity measures an observed relationship between inputs and outputs, whereas efficiency compares observed performance with a feasible or desirable alternative.

Technical efficiency and unnecessary resource use

Technical efficiency concerns whether a decision-making unit obtains the maximum feasible output from a given set of inputs or uses the minimum feasible inputs required to produce a given output. It focuses on the production process rather than on whether the selected services are the most valuable ones to provide. A technically efficient provider operates on the estimated production frontier for the inputs, outputs and assumptions used in the analysis.

An output-oriented technical-efficiency measure can be represented as follows.

$$TE_{\text{output}} = \frac{\text{Observed output}} {\text{Maximum feasible output from the same inputs}}$$

An input-oriented technical-efficiency measure can be represented as follows.

$$TE_{\text{input}} = \frac{\text{Minimum feasible input for the observed output}} {\text{Observed input}}$$

Under these representations, a score of 1 indicates that the decision-making unit lies on the estimated frontier. A score below 1 indicates a gap between observed performance and the estimated best feasible performance.

The orientation should match the decision context. An input-oriented assessment asks how far resource use could be reduced while maintaining output, whereas an output-oriented assessment asks how far output could be increased without adding inputs.

Technical efficiency does not establish that the correct services are being delivered. A provider may produce an unsuitable or low-value mix of services without detectable production waste and therefore appear technically efficient.

Allocative efficiency and the choice among competing uses

Allocative efficiency concerns whether resources are distributed among activities in a way that best advances the decision-maker’s stated objectives. It requires comparison across competing uses because assigning resources to one activity prevents those resources from being used elsewhere. In healthcare, the objective may involve population health, welfare, access, equity or a specified combination of outcomes.

The objective matters because it changes what counts as an efficient allocation. In welfare economics, an allocation is Pareto efficient when no feasible change can make one person better off without making someone else worse off; this criterion says nothing about whether the resulting distribution is fair. Much health economic evaluation instead takes an extra-welfarist objective, such as maximising health gain from a fixed budget, so an allocative efficiency claim should state which objective it uses.

A simplified allocation problem maximises total value

$$\max_{x_1,\ldots,x_n} \sum_{i=1}^{n} V_i(x_i)$$

subject to the resource constraint

$$\sum_{i=1}^{n} C_i(x_i) \leq B$$

The terms are defined as follows.

• $x_i$ is the level of activity allocated to programme $i$.
• $V_i(x_i)$ is the value produced by that activity.
• $C_i(x_i)$ is the resource requirement associated with that activity.
• $B$ is the available budget or resource constraint.

A programme can be technically efficient but allocatively inefficient. This occurs when the programme avoids production waste but the resources assigned to it could produce greater value through another feasible use.

In a simplified economic model, allocative efficiency may be associated with the point at which marginal benefit equals marginal cost.

$$MB = MC$$

This condition is not a universal operational rule for healthcare. Fixed budgets, capacity limits, non-market prices, external effects, uncertain outcomes, ethical obligations and multiple objectives can all affect the appropriate allocation.

Productive efficiency and the lowest feasible cost

Productive efficiency concerns producing a specified output at the lowest feasible cost, given available production methods and input prices. It combines the avoidance of unnecessary input use with the selection of an appropriate, least-cost combination of inputs. The concept is therefore more demanding than a simple comparison of output volumes.

A provider may be technically efficient but not productively efficient. For example, it may obtain the maximum feasible output from its existing workforce while using a more expensive skill mix than another clinically appropriate production model.

Productive efficiency should also be distinguished from productivity. Productivity records the amount produced relative to inputs, while productive efficiency asks whether the output could have been produced at a lower cost through a feasible alternative production process.

Terminology differs between literatures, so the definitions behind a claim should be checked. A widely cited set of health economics definitions frames productive efficiency in terms of health outcomes, as the maximisation of health outcome for a given cost, or the minimisation of cost for a given outcome, which allows interventions with a common outcome measure to be compared. In the production-frontier literature, the least-cost idea is often called cost efficiency, and the term allocative efficiency can refer to choosing the least-cost mix of inputs at given prices rather than to allocating resources across programmes.

Scale efficiency and an appropriate operating size

Scale efficiency concerns whether a service or organisation operates at a size that makes effective use of its production capacity. A provider can be technically efficient at its current scale but still operate at a size that is too small or too large relative to the estimated production technology. Scale efficiency therefore separates the performance of the production process from the consequences of organisational size.

The relationship between changes in inputs and outputs can take several forms.

• Increasing returns to scale occur when output rises by a greater proportion than inputs.
• Constant returns to scale occur when output rises in the same proportion as inputs.
• Decreasing returns to scale occur when output rises by a smaller proportion than inputs.

A scale decision should not be based on unit cost alone. Centralising services might reduce measured production costs while increasing travel burdens, weakening geographical access or reducing system resilience.

Dynamic efficiency and resource use over time

Dynamic efficiency concerns how resource use, innovation and outcomes develop over time. It recognises that current expenditure may build future capacity and that short-term savings may create longer-term losses. The relevant comparison therefore extends beyond the immediate production period.

Dynamic efficiency may involve activities such as those below.

• Investment in new technologies and infrastructure.
• Workforce education and capability development.
• Organisational learning and process improvement.
• Prevention that reduces future disease or treatment needs.
• Research, innovation and the adoption of effective practices.
• Withdrawal or replacement of obsolete or low-value practices.

An allocation that appears efficient over a short period may be inefficient over a longer horizon if it weakens future capacity or prevents valuable innovation. Conversely, an intervention is not dynamically efficient merely because it is new; its expected benefits, risks and opportunity costs still require assessment.

How opportunity cost connects efficiency to scarcity

Efficiency analysis requires attention to opportunity cost because healthcare resources have alternative uses. The benefit produced by an activity is not sufficient evidence that the allocation is efficient. The analyst must consider what could have been achieved through the next-best feasible use of the same resources.

When a health system operates under a fixed budget, additional spending on one intervention may displace services elsewhere. The health or other value forgone through that displacement is part of the opportunity cost of the decision.

This connection explains why efficiency is central to health economics. Scarcity creates the need to choose, and efficiency concerns whether those choices make the best feasible use of the resources available.

How efficiency differs from cost saving

Cost saving means reducing expenditure, whereas improving efficiency means producing the same or greater value with fewer resources or producing greater value with the same resources. The two can occur together, but they are not interchangeable. A saving that disproportionately reduces health, access or quality may make resource use less efficient.

A cost reduction may fail to improve efficiency when it does any of the following.

• Reduces valuable health outcomes.
• Lowers quality or patient safety.
• Creates harmful waiting times or treatment delays.
• Transfers costs to patients, caregivers or another public service.
• Worsens access for high-need populations.
• Increases future expenditure.
• Removes a service whose value exceeds its opportunity cost.

An efficiency improvement does not always reduce total spending. Resources released from one activity may be redeployed to generate additional health or other valued outcomes elsewhere.

How efficiency differs from cost-effectiveness

Cost-effectiveness compares the additional costs and outcomes of alternative interventions. Efficiency is broader because it can concern production processes, input combinations, organisational scale, resource allocation across programmes and changes in performance over time. Cost-effectiveness evidence may inform efficiency, but it does not measure every form of efficiency.

A cost-effective intervention can contribute to allocative efficiency when it represents a better use of resources than the activities it displaces. However, a favourable cost-effectiveness result does not show that the intervention is being delivered technically or productively efficiently.

The reverse is also possible. A provider may deliver services with little observable production waste while delivering interventions that are not cost-effective relative to available alternatives.

Why quality and case mix matter

Efficiency comparisons can be distorted when measured activity is treated as though it were equivalent across providers and populations. Healthcare organisations may differ substantially in the needs, risks and complexity of the people they serve. They may also differ in outcomes and quality even when their recorded activity levels are similar.

An efficiency assessment may need to consider differences in the factors below.

• Mortality, complications and readmissions.
• Patient-reported outcomes and experience.
• Safety, continuity and appropriateness of care.
• Age, comorbidity and disease severity.
• Emergency status and treatment complexity.
• Socioeconomic conditions and referral patterns.
• Factors affecting performance that are outside the provider’s control.

Inadequate adjustment can make a provider serving more complex patients appear inefficient. It can also encourage avoidance of high-risk patients or reward higher activity even when that activity does not improve outcomes.

Adjustment should improve comparability without concealing meaningful differences in access, treatment or performance. The variables, assumptions and limitations of the adjustment should be reported transparently.

Efficiency does not replace equity

Efficiency and equity are related but distinct decision objectives. Efficiency concerns the relationship between resources and valued results, whereas equity concerns how health, access, costs and opportunities are distributed. Maximising an aggregate outcome does not necessarily produce a distribution that society regards as fair.

An allocation that maximises total health may fail to do the following.

• Prioritise people with the greatest severity of illness.
• Reduce avoidable health inequalities.
• Protect disadvantaged or underserved groups.
• Provide equal access for equal need.
• Distribute financial burdens fairly.

Equity considerations may legitimately alter a decision that would otherwise maximise aggregate outcomes. This is often described as an equity-efficiency trade-off, in which some aggregate health gain is given up to achieve a fairer distribution. These value judgements should be made explicit rather than being hidden within an unexplained efficiency claim.

Methods used to assess efficiency

No single empirical method measures every form of efficiency. The appropriate method depends on the decision question, available data, production assumptions and whether the analysis focuses on providers, programmes, interventions or whole systems. Different methods may produce different results because they construct the comparator or frontier differently.

Common approaches are listed below.

• Ratio and indicator analysis compares selected outputs or outcomes with selected inputs.
• Cost-effectiveness analysis compares the incremental costs and outcomes of alternative interventions.
• Data envelopment analysis uses linear programming to construct a frontier from observed decision-making units.
• Stochastic frontier analysis estimates a parametric production or cost frontier while separating inefficiency from statistical noise under specified assumptions.
• Productivity-index methods examine changes in productivity between periods.
• Regression and multilevel methods examine associations between resource use, outputs, outcomes and contextual factors.
• Benchmarking compares performance with selected peers, standards or best-observed practice.
• Programme budgeting and marginal analysis examines how reallocating resources between programmes could improve value.

Every method answers a bounded question. A technically sophisticated model does not compensate for unsuitable inputs, outputs, comparators or objectives.

Interpreting an estimated efficiency frontier

A production frontier represents the maximum estimated output attainable from given inputs, or the minimum estimated inputs required for a given output, under the model used. It is an analytical benchmark rather than proof of an absolute physical maximum. Its position depends on the data, sample, variables and assumptions included in the analysis.

Frontier results should be interpreted cautiously, for the reasons below.

• An observed frontier reflects the best performance in the available data, which may still contain waste.
• Measurement error can make a provider appear unusually efficient or inefficient.
• Omitted quality or outcome differences can distort the frontier.
• Results can change when inputs, outputs or comparison units change.
• Data envelopment analysis is sensitive to extreme observations and treats deviations from the frontier as inefficiency.
• Stochastic frontier analysis depends on the selected functional form and assumptions about statistical noise and inefficiency.
• Efficiency scores may not be comparable across studies using different models or datasets.

A frontier score is therefore evidence about relative performance within a defined analysis. It should not be presented as a context-free ranking or a complete judgement about provider quality.

Illustrative example of two hospitals with the same activity

A simple example shows why higher measured productivity cannot automatically be treated as greater efficiency. The figures are illustrative. Two hospitals each report 10,000 outpatient consultations during one year. Hospital A uses £2 million in measured operating resources, while Hospital B uses £1.6 million.

Their simple output-to-input ratios are as follows, first for Hospital A

$$\frac{10{,}000}{£2{,}000{,}000} = 0.005 \text{ consultations per pound}$$

and then for Hospital B.

$$\frac{10{,}000}{£1{,}600{,}000} = 0.00625 \text{ consultations per pound}$$

Hospital B records 25% more consultations per pound than Hospital A, as the following calculation shows.

$$\frac{0.00625}{0.005} - 1 = 0.25$$

That calculation establishes a difference in measured productivity, not a complete difference in efficiency. Before concluding that Hospital B is more efficient, the analyst would need to examine several further questions.

• Whether the hospitals treat patients with similar needs and complexity.
• Whether consultation quality and health outcomes are comparable.
• Whether the recorded resource measures include the same cost categories.
• Whether either hospital delivers unnecessary consultations.
• Whether important activities or outcomes are missing from the calculation.
• Whether the difference persists after accounting for uncertainty and contextual constraints.

If Hospital B treats less complex patients or achieves poorer outcomes, the simple ratio may overstate its relative performance. If quality, case mix and resource measurement are comparable, the ratio provides stronger evidence of a genuine efficiency difference.

Incentives can change measured and actual efficiency

Payment systems, targets and performance measures affect how organisations use resources and record activity. An incentive designed to increase efficiency may improve genuine performance, but it may also encourage behaviour that improves the measured indicator without improving patient outcomes. Efficiency measures should therefore be assessed for their behavioural consequences as well as their technical construction.

Possible unintended responses include the following.

• Selecting patients who are easier or less costly to treat.
• Increasing recorded activity without increasing valuable outcomes.
• Shifting costs to another organisation or sector.
• Reducing unmeasured aspects of quality.
• Delaying necessary expenditure beyond the measurement period.
• Coding cases differently to improve apparent performance.

A sound efficiency framework uses balanced measures, transparent definitions and safeguards against predictable distortions. It also examines whether improvements in one part of the system create costs or losses elsewhere.

Uncertainty in efficiency analysis

Efficiency estimates are uncertain because data, assumptions and model choices are imperfect. Treating a point estimate or rank as exact can create false confidence and encourage inappropriate comparisons. Uncertainty should be examined whenever it could change the interpretation or decision.

Relevant sources of uncertainty are listed below.

• Measurement error in inputs, outputs, outcomes or prices.
• Sampling variation and small numbers.
• Alternative model specifications.
• Missing variables and incomplete case-mix adjustment.
• Assumptions about the production or cost frontier.
• Changes in performance over time.
• External shocks or constraints affecting only some organisations.

Analysts can explore uncertainty through sensitivity analysis, alternative specifications, confidence intervals, bootstrapping or repeated analysis over time, depending on the method. Reports should show whether the main conclusion remains stable under plausible alternatives.

How to report an efficiency claim

A credible efficiency claim should state exactly what was compared and what the analysis can support. Reporting only an efficiency score or ranking hides the choices that produced the result. Transparent reporting allows readers to judge whether the comparison answers the intended decision question.

An efficiency report should identify the items below.

• The decision-making units and population included.
• The analytical perspective and objective.
• The inputs, outputs and outcomes measured.
• The relevant time period.
• The comparator or method used to construct the frontier.
• Adjustments for quality, case mix and contextual constraints.
• The treatment of prices, uncertainty and missing data.
• The distinction between productivity, technical efficiency, productive efficiency and allocative efficiency.
• Important equity, access or system-wide consequences.
• The limits on transferring results to other settings.

A reported improvement should also explain what changed. A higher score may result from better organisation, improved technology, altered patient selection, reduced quality, changed measurement or movement by other members of the comparison group.

Common misreadings of efficiency claims

Efficiency is often discussed as though it had one simple meaning, but different efficiency concepts answer different questions. Keeping those questions separate prevents apparent improvements from being interpreted too broadly. It also makes clear when an analysis supports only a limited claim.

Common misunderstandings include the following.

• Lower expenditure is not automatically greater efficiency because lower spending may also reduce valuable outcomes.
• Greater activity is not automatically greater efficiency because some activity may be unnecessary or harmful.
• Higher productivity is not automatically greater efficiency because quality, case mix and alternative uses may be omitted.
• Technical efficiency does not establish allocative efficiency because resources may still be assigned to the wrong activities.
• Cost-effectiveness does not establish technical efficiency because a worthwhile intervention may still be delivered wastefully.
• A frontier score of 1 does not prove perfect performance because the estimated frontier depends on the available data and model.
• Greater efficiency does not necessarily mean a smaller total budget because released resources may be redeployed.
• An aggregate efficiency gain does not establish that the distribution of benefits and burdens is equitable.

Forms of efficiency compared

The major forms of efficiency should be reported by name rather than being collapsed into a single label. Each form identifies a different source of possible improvement and requires different evidence.

These distinctions are complementary rather than competing. A complete assessment may need to consider several forms of efficiency together, while remaining clear about which conclusions the available evidence can and cannot support.

Sources

• Palmer S, Torgerson DJ. Economics notes: definitions of efficiency. BMJ. 1999;318(7191):1136.
• Jacobs R, Smith PC, Street A. Measuring Efficiency in Health Care: Analytic Techniques and Health Policy. Cambridge University Press. 2006.
• Charnes A, Cooper WW, Rhodes E. Measuring the efficiency of decision making units. European Journal of Operational Research. 1978;2(6):429-444.
• Aigner D, Lovell CAK, Schmidt P. Formulation and estimation of stochastic frontier production function models. Journal of Econometrics. 1977;6(1):21-37.
• Brouwer WBF, Culyer AJ, van Exel NJA, Rutten FFH. Welfarism vs. extra-welfarism. Journal of Health Economics. 2008;27(2):325-338.
• Wagstaff A. QALYs and the equity-efficiency trade-off. Journal of Health Economics. 1991;10(1):21-41.
• Mitton C, Donaldson C. Health care priority setting: principles, practice and challenges. Cost Effectiveness and Resource Allocation. 2004;2:3.
• Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th edition. Oxford University Press. 2015.