Health economics in nursing research: what you need to know to include economic evaluation methodology in your research

Abstract

Due to limited resources and constant, ever-changing healthcare challenges, health economics is essential to support healthcare decisions while improving health outcomes. Economic evaluation methodology facilitates informed decision-making related to the efficient allocation of resources while positively impacting clinical practice. In this paper, we provide an overview of economic evaluation methods and a real-world example applying one method of economic evaluation (cost-utility analysis) in nursing research.

Cost-benefit analysis, Cost-consequence analysis, Cost-effectiveness analysis, Cost-minimization analysis, Cost-utility analysis, Economic evaluation, Economics, Healthcare quality, Nursing research, Quality-adjusted life years

Learning objectives

• To understand the different types of health economic evaluation and when they should be applied.

• Understand the basic steps in undertaking a cost-utility analysis and the type of economic evaluation preferred by many decision-making bodies internationally.

• To interpret the results of a cost-utility analysis.

Introduction

Health Economics is a field of research applying economic theories and methods to understand and explain how society makes decisions regarding their health behaviours, healthcare use, and health resource allocation.¹ Economic evaluation methodology is a tool to guide decision-making around the efficient allocation and equitable distribution of healthcare resources and manage the demand for healthcare services, health promotion, and prevention.^2,3

Economic evaluation is defined as ‘the comparative analysis of alternative courses of action in terms of both their costs and consequences’.⁴ The costs and consequences are synonymous with the inputs and outputs or outcomes of a given intervention. The inputs or resources required to deliver the programmes/intervention are valued by assigning a cost based on the market price. Similarly, the outputs are the resources saved or value created because of the programme/intervention and are valued using different techniques depending on the type of economic evaluation.

Why economic evaluation is essential for research

Policymaking bodies such as the UK National Institute for Health and Care Excellence require the submission of economic evidence to support funding allocation decisions for pharmaceuticals and medical technologies.^5,6 And increasingly, economic evidence is required at all levels of healthcare policy to support decision-making around interventions in healthcare. There is a paucity of nursing research applying economic evaluation methodology,⁷ but this trend is indeed changing positively with an increase in evaluations in nurse-led models of care. Taylor et al.⁸ assessed the cost-effectiveness of REACH-HF and home-based cardiac rehabilitation compared with the usual medical care for heart failure in the UK, while Driscoll et al.⁹ assessed the cost-effectiveness of a heart failure nurse practitioner inpatient service from a health system perspective in Australia.

However, a significant portion of clinical and nursing research lacks a comprehensive view of economic implications with the economic analysis often performed post hoc.⁷

Overview of economic evaluation methodology

There are different types of economic evaluation depending on how the outputs are expressed, namely, cost-minimization analysis, cost-effectiveness analysis (CEA), cost-utility analysis (CUA), cost-consequence analysis, and cost-benefit analysis.⁴ The choice of evaluation depends on the questions asked and the specific policy to be informed by the results. These different types of economic evaluations are summarized in Table 1.

Table 1

Types of economic evaluation

EE	Costs/input			Consequences/outputs/outcomes			Result
	Identify	Measure	Value	Identify	Measure	Value	Comparison	Presentation
CMA	Resources consumed in delivery of the programme/intervention	Quantify resources	Assign monetary value ($)	Outcomes generated	Quantify outcome	Value using appropriate techniques	Equal effectiveness between programmes being compared	Cost comparison
CCA	Resources consumed in delivery of the programme/intervention	Quantify resources	Assign monetary value ($)	Outcomes generated	Quantify outcomes	Value using appropriate techniques	Total costs Total outcomes of each alternative	Total costs Total consequences
CBA	Resources consumed in delivery of the programme/intervention	Quantify resources	Assign monetary value ($)	Outcomes generated	Quantify benefits	Monetary value benefits	Compare total costs/inputs and total benefit	Net monetary benefit
CEA	Resources consumed in delivery of the programme/intervention	Quantify resources	Assign monetary value ($)	Effectiveness of each alternative	Quantify effectiveness of each alternative	Natural units of effectiveness	Compare the difference in total costs and total effectiveness	ICER = $\frac{\mathrm{Difference}\mathrm{in}\mathrm{costs}}{\mathrm{Difference}\mathrm{in}\mathrm{effectiveness}}$
CUA	Resources consumed in delivery of the programme/intervention	Quantify resources	Assign monetary value ($)	Health-related quality of life Survival or quantity of life	Preference-based PROMs Survival analysis	QALYs or DALYs	Compare the difference in total costs and total utility	ICER = $\frac{\mathrm{Difference}\mathrm{in}\mathrm{costs}}{\mathrm{Difference}\mathrm{in}\mathrm{QALYs}}$

CBA, cost-benefit analysis; CCA, cost-consequence analysis; CEA, cost-effectiveness analysis; CMA, cost-minimization analysis; CUA, cost-utility analysis; DALYs, disability-adjusted life years; ICER, incremental cost-effectiveness ratio; PROMs, patient-reported outcome measures; QALYs, quality-adjusted life years.

In the healthcare setting, CUA and CEA are recommended by regulatory bodies worldwide.^10,11 However, CUA has the attraction of providing a comprehensive and standardized framework for comparing costs and outcomes across disparate programmes or interventions and patient populations.⁴

Cost-effectiveness analysis

Outcomes in CEA are expressed in natural units or physical units of effect that are usually clinically or bio-medically focused, such as the number of infections averted, units of blood pressure reduced, or, more commonly, the number of life years gained.¹² The result in CEA is presented as the cost per unit of outcome or outcome per unit of cost. As such, CEA is used when comparing interventions with common or similar outcomes, e.g. within a particular healthcare speciality or programme, to maximize technical efficiency.^12,13 Technical efficiency is required when operating within a fixed budget, and there is need to choose between a set of alternatives, so the decision-maker has to maximize the outcome level within the budget constraint.

Cost-utility analysis

Cost-utility analysis is sometimes referred to as a form of CEA. However, unlike CEA where the outcome is programme specific, the unit of outcome in CUA is generic therefore allowing for the comparison of health consequences across disparate conditions and programmes, such as a programme preventing stroke, surgery for hernias, and end-of-life care.² This type of analysis is specifically recommended for programmes where health-related quality of life (HRQoL) is an important outcome, where morbidity and mortality are important or where a combination of outputs is accrued, and a single unit of outcome is required.

When undertaking a CUA, the outputs reflect society’s utility or value for the given outcome.¹⁴ The outcome is assessed as the quality of life or HRQoL resulting from each alternative while the value that society places on this health outcome is obtained using different methods such as standard gamble, time trade-off, or contingent valuation.¹⁵ However, there are patient-reported outcome measures (PROMs) assessing HRQoL that are specifically designed to quantify HRQoL and the utility that society places on that quality of life. These are referred to as multi-attribute utility instruments or preference-based PROMs. The score obtained by these PROMs is referred to as the utility score. Utility scores take on values ranging from 0 (equivalent to death) to 1 (equivalent to full health); negative values indicate outcomes worse than death.^16,17 In CUA, the utility score is combined with the quantity of life accrued by the intervention into one generic metric referred to as the quality-adjusted life year (QALY). A QALY of 1 is equivalent to 1 year spent in full health. Several preference-based PROMs are used in clinical practice, such as the most widely used EuroQol instruments, EQ-5D-3L and EQ-5D-5L.¹⁸Table 2 provides examples of preference-based PROMs and details on how their utility values are obtained and from which populations/society.

Table 2

Generic preference-based patient-reported outcome measures

Instrument	Dimensions	Utility elicitation method	Scoring algorithms
EQ-5D-3L EQ-5D-5L	Mobility, self-care, usual activities, pain/discomfort, and anxiety/depression	Time-Trade-Off Visual Analogue Scale Contingent valuation	Adult general population samples from several countries including Australia
HUI2	Sensation, mobility, emotion, cognition, self-care, pain, and fertility	Visual Analogue Scale Standard Gamble	Canada (age groups 5–37, 12–16, 8–16), the USA (18–89), Australia (15+), the UK (general population), and Uruguay (8–17 age group)
HUI3	Vision, hearing, speech, ambulation, dexterity, emotion, cognition, and pain	Visual Analogue Scale
SF-6D	Physical functioning, role limitation, social functioning, pain, mental health and vitality	Standard Gamble	Adult general population samples from the UK, Japan, Hong Kong, Australia, and Brazil
AQoL-4D AQoL-6D AQoL-7D AQoL-8D	Independent living, happiness, mental health, coping, relationships, self-worth, pain, and senses	Time-Trade-Off	Australia (adult general population sample).
QWB	Mobility, physical activity, and social activity	Visual Analogue Scale	The USA (adult general population sample)
15D	Health mobility, vision, hearing, breathing, sleeping, eating, speech, elimination, usual activities, mental function, discomfort and symptoms, depression, distress, vitality, and sexual activity.	Visual Analogue Scale	Finland (adult general population sample)

15D, 15 dimensions; AQoL, assessment of quality of life; EQ-5D, EuroQol 5 dimensions; HUI, health utility index; QWB, quality of well-being scale; SF-6D, short form 6 dimensions.

Other generic units of outcome measurement applied in CUA include the more commonly used disability-adjusted life year (DALY). A DALY is a measure used to quantify a population’s overall disease and injury burden. Disability-adjusted life years consider both the years of life lost due to premature mortality and the years lived with disability, adjusted for the severity of the disability.¹⁹ Results of CUA are expressed as the cost per QALY or costs per healthy year.

Step-by-step approach: how to conduct a cost-utility analysis

As previously highlighted, the methods to identify, measure, and value the inputs for CEA and CUA are identical; however, the assessment of outputs or outcomes differs. Often, programmes or interventions have multiple outcomes; however, these are aggregated into a single generic outcome in CUA, the QALY.

Central illustration summarizes the steps involved in undertaking a CUA.

To demonstrate the steps in Central illustration, we will consider a hypothetical example of a cost-utility study alongside a randomized control trial to assess the cost-utility of telehealth services compared with usual care of face-to-face services for outpatient care of hypertension in adults.

Central illustration

Step 1: defining the question

In defining the question, the alternatives or programmes being compared, also referred to as comparators, must be stated as should the study’s perspective and time horizon. The other elements to be explicitly stated are the population in which the services are provided and the output of interest, as this informs the type of evaluation to be undertaken. The comparators should always include usual care, the most used alternative, or the existing approach, which is compared against the new intervention(s) or treatment(s). Comparators in the example above are the new intervention, telehealth services vs. face-to-face services (usual care).

The study perspective is determined by the decision-maker and, therefore, dictates whose costs and consequences are assessed. The healthcare payer perspective includes only costs to the healthcare payer such as the cost of treatment and other services provided in the delivery of the care.²⁰ Conversely, the healthcare sector perspective ‘accounts for all monetary costs of healthcare, regardless of who bears the cost’²⁰ and therefore includes patient out-of-pocket costs as direct medical costs such as payments for tests, procedures, and medications that are not covered by the healthcare system. The societal perspective, as the name suggests, includes all costs to society.²⁰ This perspective includes costs to other sectors beyond healthcare, such as education and social services, and costs to patients and their families as out-of-pocket medical and non-medical expenditures. Examples of non-medical expenditures include parking fees and, most significantly, the cost of lost productivity due to reduced capacity to work or early retirement for patients and/or carers such as following a stroke. Last is the limited societal perspective, which includes costs to the healthcare sector and to patients and their families; however, it stops short of the societal perspective by excluding the impact on other sectors outside healthcare.²⁰

The societal perspective is critical in the analysis of home care models of care where the carer costs are pertinent.²¹ Such programmes may be cost-effective from a health system perspective but not a societal perspective once informal carer costs are considered.

The time horizon is the duration over which costs and outcomes are assessed and must be appropriately determined to consider all the benefits accrued from an intervention.⁴ If the analysis is alongside a clinical trial, the time horizon is equivalent to the duration of the trial but can be extrapolated to an extended period.

Any perspective could be taken in this case; however, because the management of hypertension involves significant out-of-hospital care and informal carer support, a societal or limited societal perspective is preferable. Because the study is being conducted alongside a randomized control trial, the time horizon will coincide with the duration of the trial.

Research question: what is the cost-utility of telehealth services compared with face-to-face services in managing adult hypertension from a limited societal perspective in Australia?

Step 2: identification, measurement, and valuation of costs

This involves identifying the resources required to deliver each alternative, measuring how much and attaching cost values. Examples of resource inputs include capital costs, labour, consumables, and patient costs such as travel and lost productivity. Measurement is implemented using (i) a bottom-up approach or micro-costing, where individual items are quantified, or (ii) a top-down approach or macro-costing, where aggregate quantities are attributed to each alternative. Valuation involves attaching market prices/costs to each item. For example, in Australia, out-of-hospital care is valued based on the medical benefit schedule (MBS), pharmaceuticals are based on the pharmaceutical benefit scheme, while hospitalizations are valued based on diagnosis-related group (DRG) costs. Non-market items such as volunteer time are valued based on market prices of personnel undertaking similar roles as the volunteers.

Because this is a prospective study, a bottom-up or micro-costing approach with real-time recording of cost data will be used. To quantify out-of-pocket costs, patients will fill out a cost diary including costs incurred weekly, time off work to travel and attend face-to-face appointments, gap payment for clinician or allied health visits or prescriptions, over-the-counter medications, transport and parking costs, and informal care costs. The number of visits and matching MBS or DRG codes will be applied to quantify health system costs (see Table 3).

Step 3: identification, measurement, and valuation of outcomes

This is determined by the type of analysis being conducted.

Table 3

Quantification of costs and outcomes over 12 months

	Telehealth			Usual care
Cost item	Unit cost	Quantity	Total	Unit cost	Quantity	Total
Health system costs
Hospital in-patient care	$1600	8 days	$12 800	$1600	4 days	$6400
Out-patient care	$150	2 visits	$300	$150	2 visits	$300
GP visits	$35	0	0	$35	4 visits	$140
CT scan	$150	2 scans	$300	$150	2 scans	$300
Medication	$30	12 months	$360	$30	12 months	$360
Patient out-of-pocket cost
Hospital parkinga	$5	6 tickets	$30	$5	10 tickets	$50
Time off workb	$65	70 h	$4550	$65	42 h	$2730
Gap fees for CT scan	$100	2 scans	$200	$100	2 scans	$200
Gap fees for GP visit	$25	0	0	$25	4 visits	$100
Total costs			$18 540			$10 580

Outcomes
	EQ-5D-5L score	Quantity of life	QALY	EQ-5D-5L score	Quantity of life	QALY
Baseline	0.68	0.5	0.105	0.68	0.5	0.04
Follow-up	0.89			0.76
Change in utility score	0.21			0.08

^aTwo 1 h tickets for each outpatient visit, 1-h ticket for the CT scan and GP visit

^b8 h for each day in hospital, 2 h for each outpatient visit, an hour for a CT scan and GP visit.

This is a CUA alongside a clinical trial. The trial will establish the effectiveness of telehealth services compared with face-to-face services in controlling the clinical outcome, in this case, blood pressure. Because it is a CUA, we will prospectively assess HRQoL using the EQ-5D-5L at baseline and follow-up. To facilitate the calculation of QALYs, patient survival over the study period or stated time horizon will be established using survival analysis calculations. To obtain QALYs, the utility score is multiplied by survival. This is summarized in Table 3.

Step 4: quantifying costs and outcomes

Following Steps 1–3, the costs and outcomes of each alternative are quantified and presented.

We will now compute the total costs and total outcomes accrued for each alternative for the entire study period. Because the societal perspective was chosen, both hospital and patient out-of-pocket costs incurred with telehealth and usual care will be quantified. Quality-adjusted life years are estimated by combining the duration of survival with the utility score estimated from the EQ-5D-5L responses. The change in QALYs between baseline and follow-up is then calculated for each alternative. See Table 3 for a hypothetical quantification.

Step 5: calculating the incremental cost-effectiveness ratio

The incremental cost-effectiveness ratio (ICER) is the main result of CEA and CUA and is a summary statistic reflecting the economic value of one alternative over another. It is defined as the ratio of the difference in costs of both alternatives to the difference in their outcomes.⁴

The ICER is a ratio of the difference in costs (Step 3) and the difference in the QALY change between alternatives calculated in Step 4 above.

ICER = 7960/0.65 = $12 246.15 per QALY

Step 6: making the cost-effectiveness decision

Following the calculation of the ICER, an inference to the cost-effectiveness of one alternative compared with the other is provided, but the funding/policy decision is based on the cost-effectiveness threshold. This is a predetermined cost per unit of outcome above which an intervention is considered cost-effective. This threshold varies from one country or healthcare system to another. Although there is no specific threshold value in Australia, the arbitrary benchmark is interventions below $50 000 per QALY gained are likely to be funded²² while NICE has a threshold of between £20 000 and £30 000 per QALY.²³ As such, an intervention could be cost-effective in one country’s healthcare system but not another. For example, the use of quadruple therapy in patients with heart failure with reduced ejection fraction is cost-effective compared with triple therapy and double therapy in the USA at an ICER of $81 000 and $51 081, respectively, which is below the US threshold of $150 000.²⁴ However, this would not be cost-effective in Australia where the threshold is $50 000 per QALY.

The ICER obtained in Step 5 is $12 246.15 per QALY. Firstly, this is in Quadrant A of the cost-effectiveness plane (CEP), meaning telehealth is more effective but also more costly than usual face-to-face care for hypertension. Secondly, this ICER is below the Australian government threshold of $50 000 per QALY and is, therefore, considered a cost-effective intervention in the Australian setting.

Uncertainty in cost-effectiveness results

Like all statistical calculations, there may be uncertainty in the estimation of inputs and outputs or in the methodology and, therefore, the need to account for uncertainty in the ICER. This uses CEPs and cost-effectiveness acceptability curves (CEACs). The CEPs incorporate bootstrapped pairs of ICERs, visually representing the uncertainty inherent in both cost and outcome estimates; see Figure 1. In Quadrants B and C, the decision is straightforward; usual care is preferred to because telehealth is less effective regardless of cost. In Quadrant D, telehealth is said to have dominated usual care because it costs less yet is more effective. In Quadrant A, the cost-effectiveness decision is not straightforward; the decision-maker must determine if they can afford the intervention based on the cost-effectiveness threshold. The likelihood of an intervention being cost-effective across a range of cost-effectiveness thresholds can be presented in a CEAC where bootstrapped pairs of costs and outcomes are presented, depicting the uncertainty in estimates of the ICER.

Figure 1

Cost-effectiveness plane for decision-making. CE, cost-effectiveness; ICER, incremental cost-effectiveness ratio.

Uncertainty is also shown using one-way or two-way sensitivity analysis that is where one or two input parameters are varied over a plausible range to test if the ICER or cost-effectiveness decision is consistent with the variations.²⁵ This analysis shows which variables the ICER is sensitive to or the drivers of the ICER. In probabilistic sensitivity analysis, probability distributions are assigned to cost and outcome parameters. This ensures that, rather than assuming fixed values, the assigned distributions capture the inherent uncertainty surrounding their true values. Gamma distributions are used for cost inputs and beta distributions for outcomes.²⁵

One-way sensitivity analysis: In the above example, the greatest cost item is hospital stay; this variable (both number of days and unit cost) could be varied across the median value or interquartile range and the ICER recalculated accordingly.

Two-way sensitivity analysis: Both hospital stay and the change in HRQoL may be varied at the same time to test the impact on the ICER.

Probabilistic sensitivity analysis: In this case, distributions will be created for all the input variables and a CEP and CEAC generated showing the probability of telehealth being cost-effective compared with usual care across different thresholds.

Software

Economic analysis can be undertaken through software such as Microsoft Excel, TreeAge Pro exist,²⁶ Stata, R, Python, and MATLAB.^27,28

Reporting and dissemination

Standard reporting of economic analysis must follow recommended guidelines or checklists for research. The most commonly used checklist is the Consolidated Health Economic Evaluation Reporting Standards developed by the International Society for Pharmacoeconomics and Outcomes Research.²⁹ For dissemination to decision-makers on the other hand, different decision-making bodies provide specific guidelines on how economic evaluation results should be presented. These guidelines are publicly available.^11,30

Challenges and limitations of conducting economic evaluation in the clinical setting

Most economic evaluations are conducted alongside clinical trials, usually with generous budgets for data collection and patient follow-up. Several factors must be considered for economic evaluations to become a main-stay in quality improvement and routine clinical care including (i) understanding the value of this methodology in routine health decision-making, (ii) availability of resource use data that reflects the real cost of interventions or models of care beyond activity-based funding, (iii) availability of PROMs data as part of routine healthcare, (iv) decommercialization or subsidizing of subscription-based PROMs for use in routine care databases, and (v) availability of health economic expertise to clinicians/health services.

Conclusion

Demonstrating that a new therapy is effective is no longer sufficient to guarantee the best healthcare practice. Nurses and allied health professionals must have a perspective combining healthcare expertise with economic proficiency to address the complex nature of modern healthcare problems. Health economic evaluation must be considered fundamental to nurses and allied health professional research if we are to find solutions to tackle healthcare challenges in the modern world.

Funding

The authors received no financial support for this article's research, authorship, and/or publication.

Definition of key terms

Term	Definition
Cost-effectiveness plane	‘Represents differences in costs and health outcomes on a graph, with effectiveness on the x axis and cost on the y axis’.31
Economic evaluation	‘The comparative analysis of alternative courses of action in terms of both their costs and outcomes’.32
Healthcare payer perspective	‘This perspective includes only those monetary costs (e.g. treatment costs and other health service resource use associated with disease management) incurred by a (typically third party) healthcare payer (e.g. Medicare/Medicaid, British national health service, a health maintenance organization, etc)’.20
Healthcare sector perspective	‘This perspective is similar to the healthcare payer perspective but accounts for all monetary costs of healthcare, regardless of who bears the cost. A key distinction between the healthcare sector and healthcare payer perspectives is that the healthcare sector perspective includes patients’ out-of-pocket cost’.20
Incremental cost-effectiveness ratio	‘Difference in the change in mean costs in the population of interest divided by the difference in the change in mean outcomes in the population of interest’.31
Limited societal perspective	‘This perspective accounts for cost components beyond those captured by the healthcare sector perspective, including patient time, patient transportation, unpaid caregiver time, and productivity loss. It excludes spill over impacts affecting sectors other than healthcare, such as education’.20
Quality-adjusted life years	‘A measure of the state of health of a person or group in which the benefits, in terms of length of life, are adjusted to reflect the quality of life’.31
Sensitivity analyses	‘A means of exploring uncertainty in the results of economic evaluations’.31
Societal perspective	‘A societal perspective is broader than the limited societal perspective. It represents the overall public interest by including all resources that could be used for other purposes’.20
Time horizon	This is the time over which costs and outcomes are measured.

Author notes

Conflict of interest: No conflict of interest to declare.