Designing an economic evaluation study is a complex but crucial process for informing healthcare and public health policy. I will guide you step-by-step, using a running example to illustrate the concepts.

Running Example: We want to evaluate a new tele-rehabilitation program for post-stroke patients compared to traditional in-person rehabilitation.

1) Choose Evaluation Type

The choice of evaluation type depends on the nature of the intervention, the measurable outcomes, and the decision-making context.

• Cost-Effectiveness Analysis (CEA):


• Outcome: Natural health units (e.g., life-years gained, cases averted, disease-free days, reduction in blood pressure).


• Ratio: Cost per unit of effect (e.g., $/life-year gained, $/stroke averted).


• Pros: Outcomes are intuitive and easy to understand for clinicians.


• Cons: Cannot directly compare interventions for different diseases or with different types of outcomes (e.g., a cancer drug vs. a mental health intervention).


• When to use: When the outcome is clinically meaningful and comparable within a disease area, but not necessarily across different health conditions.


• Example Application: Cost per additional patient achieving functional independence (e.g., modified Rankin Scale score ≤2).


• Cost-Utility Analysis (CUA):


• Outcome: Quality-Adjusted Life Years (QALYs) or Disability-Adjusted Life Years (DALYs). These combine quantity (life-years) and quality (utility/disability weights) of life into a single metric.


• Ratio: Cost per QALY gained or Cost per DALY averted.


• Pros: Allows comparison across different diseases and interventions because the outcome (QALY/DALY) is generic. This is often preferred by policy makers for resource allocation decisions.


• Cons: Utility values can be difficult to measure, controversial, and may not fully capture all aspects of patient experience.


• When to use: When quality of life is a significant outcome, or when you need to compare interventions across different health areas. This is generally the most recommended type for informing broad health policy.


• Example Application: Cost per QALY gained. (This is generally a strong choice for chronic conditions and rehabilitation).


• Cost-Benefit Analysis (CBA):


• Outcome: All outcomes (health and non-health) are converted into monetary values.


• Ratio: Net monetary benefit (Total Benefits - Total Costs) or Benefit-Cost Ratio (Total Benefits / Total Costs).


• Pros: Provides a clear monetary value of the intervention's worth, allowing direct comparison with any other investment (e.g., education, infrastructure).


• Cons: Monetizing health benefits (e.g., value of a life, value of pain reduction) is ethically complex and methodologically challenging (e.g., willingness-to-pay, human capital approach). Often highly controversial.


• When to use: When a societal perspective is paramount and the willingness/ability to monetize health outcomes is accepted, or when a significant portion of benefits are non-health related and easily monetized.


• Example Application: Extremely difficult for a rehabilitation program, as monetizing improved function or reduced caregiver burden is complex and contentious.


• Cost-Minimization Analysis (CMA):


• Outcome: Assumes that the outcomes of the comparator interventions are identical.


• Ratio: Simply compares total costs. The cheaper option is preferred.


• Pros: Simplest type of economic evaluation.


• Cons: Requires strong evidence that outcomes are truly equivalent, which is rare in practice. If outcomes are not identical, it's misleading.


• When to use: When two or more interventions achieve the exact same health outcomes with the same safety profile (e.g., two generic drugs with identical efficacy).


• Example Application: Not appropriate for our example, as a new tele-rehabilitation program is unlikely to have identical outcomes to traditional in-person care.

Decision for Running Example: Given that quality of life and functional independence are key outcomes for post-stroke patients, and we'd ideally want to compare this intervention to others across the healthcare system, a Cost-Utility Analysis (CUA) measuring cost per QALY gained is the most appropriate and robust choice. CEA (cost per functionally independent patient) would also be a strong second option if broader comparisons weren't the primary goal.

2) Define Study Question

A well-defined study question specifies the intervention, comparator, population, perspective, and outcomes.

• Template: "Is [Intervention A] cost-effective/cost-utility compared to [Intervention B] for [Target Population] in [Setting] over [Time Horizon] from a [Perspective]?"


• Running Example Study Question:

3) Perspective

The perspective determines which costs and benefits are included in the analysis.

• Healthcare System (Payer) Perspective:


• Costs Included: Direct medical costs borne by the healthcare system (e.g., provider visits, hospitalizations, medications, diagnostic tests, intervention costs).


• Benefits: Health outcomes for the patient.


• When to use: Most common in countries with publicly funded healthcare systems, or for decisions within health insurance companies.


• Running Example: My initial question used this. It would include the costs of the tele-rehab program itself (equipment, platform, staff time), costs of traditional rehab, and any subsequent medical costs (e.g., for complications, readmissions).


• Societal Perspective:


• Costs Included: All costs, regardless of who bears them. This includes direct medical costs, direct non-medical costs (e.g., patient travel, childcare, informal caregiver time), and indirect costs (e.g., productivity losses due to illness or premature death, both for patients and caregivers).


• Benefits: Health outcomes and all other societal benefits (e.g., increased productivity, improved quality of life for caregivers).


• When to use: Often considered the most comprehensive and ideal perspective, especially for public health interventions, but more challenging to measure.


• Running Example (Societal Perspective): This would add patient travel time/costs, caregiver time, and productivity gains/losses from patients returning to work or caregivers having more time. This is often preferred for a public health or broader policy decision as it captures the full impact.


• Patient Perspective:


• Costs Included: Out-of-pocket expenses, travel costs, time costs, lost wages, informal care burden.


• When to use: Useful for understanding the financial burden on patients, but not typically used for broader policy decisions on its own.


• Other Perspectives: Employer perspective, government perspective (broader than just healthcare).

• Revised Study Question: "Is a new tele-rehabilitation program cost-effective (specifically, cost-utility) compared to traditional in-person rehabilitation for adult post-stroke patients in a community setting over a 5-year time horizon from a societal perspective?"

4) Comparator(s)

The choice of comparator(s) is critical, as it defines the "opportunity cost" – what is being given up to adopt the new intervention.

• Running Example:


• Primary Comparator: Traditional in-person rehabilitation. This is the current standard of care and the most logical alternative to evaluate against.


• Potential Secondary Comparators (if relevant):


• No rehabilitation (less likely for stroke, but could be "usual care" which might include minimal rehab).


• Alternative tele-rehabilitation platforms (if multiple exist).


• A more intensive or less intensive traditional program.

5) Time Horizon

The time horizon must be long enough to capture all relevant costs and effects of the intervention and disease progression.

• Considerations:


• Duration of intervention effect: How long does the benefit last?


• Latency of effects: Do the benefits appear immediately or much later?


• Disease progression: How does the disease evolve over time, and how does the intervention alter that trajectory?


• Lifespan: For chronic conditions, a lifetime horizon might be appropriate, often requiring modeling.


• Running Example:


• Post-stroke rehabilitation can have long-lasting effects on functional independence, and the consequences of poor rehabilitation (e.g., increased dependency, institutionalization, recurrent stroke) can span many years.


• A 5-year time horizon (as in our question) is a reasonable minimum for short- to medium-term outcomes. However, a lifetime horizon would be ideal to capture all QALYs gained and future costs averted, likely requiring a modeling approach (e.g., Markov model) to extrapolate beyond the study's observed data.


• If using a 5-year horizon, we would need to justify why this is sufficient or acknowledge the limitations.

6) Outcome Measures

These are the specific metrics used to quantify the effects of the intervention.

• Running Example (CUA - Cost per QALY gained):


• Life Years (LYs): Measured by survival data. This might come from clinical trials, registries, or population life tables.


• Quality of Life (Utility) Scores:


• Measured using validated generic preference-based instruments like:


• EQ-5D (EuroQol 5-Dimension): Very common, simple questionnaire, generates health state descriptions that can be converted to utility values using country-specific tariffs.


• SF-6D (Short Form 6-Dimension): Derived from the SF-36/SF-12, also generates utility values.


• HUI (Health Utilities Index): Another option.


• Measurement Timing: Collect utility scores at baseline and at multiple follow-up points throughout the study (e.g., 3 months, 6 months, 1 year, annually).


• Calculation: QALYs are calculated by multiplying the time spent in a particular health state by the utility value of that state, then summing these products over the time horizon.


• Other Potential Outcome Measures (for CEA if we chose that):


• Functional Independence: e.g., modified Rankin Scale (mRS) score ≤ 2 or ≤ 3.


• Activities of Daily Living (ADL) / Instrumental ADL (IADL) scores: e.g., Barthel Index, FIM score.


• Reduction in stroke recurrence.


• Reduced caregiver burden: Could be measured using a scale like the Zarit Burden Interview.

7) Discounting

Discounting accounts for society's time preference for both money and health. A unit of cost or benefit today is valued more highly than the same unit in the future.

• Why:


• Time preference for consumption: People generally prefer to receive benefits sooner and incur costs later.


• Opportunity cost of capital: Money available today can be invested to yield returns.


• When: Any study with a time horizon longer than one year.


• Rates: Standard practice is to apply a consistent discount rate to both costs and effects. Common rates range from 3% to 5% annually. The specific rate often depends on national guidelines (e.g., NICE in the UK uses 3.5%, CADTH in Canada uses 1.5% for costs and 1.5% for effects, US uses 3%).


• Running Example: Given our 5-year (or potential lifetime) time horizon, discounting is essential. We would likely use a 3% annual discount rate for both costs and QALYs, but also test different rates in sensitivity analysis.

8) Data Collection Strategy

This depends on whether you're conducting a trial-based evaluation or a model-based evaluation.

• Trial-Based Economic Evaluation (Primary Data Collection):


• Approach: Integrate economic data collection into a prospective clinical trial (e.g., an RCT).


• Pros: High internal validity, directly links costs and outcomes observed in the same cohort.


• Cons: Can be expensive and time-consuming, limited by trial duration, may not capture all real-world costs/effects, generalizability issues.


• Running Example:


• Effect Data: Collect EQ-5D scores, survival data, and clinical outcomes (e.g., mRS) directly from trial participants at specified intervals.


• Cost Data:


• Intervention Costs: Micro-costing (detailed resource consumption for staff time, equipment use, platform subscription, internet costs, patient support materials, training).


• Healthcare Resource Use: Patient-reported questionnaires (e.g., frequency of GP visits, specialist visits, hospitalizations, medication use), linking to administrative databases (hospital records, claims data).


• Non-Medical Costs: Patient/caregiver questionnaires (travel costs, time off work/informal care time).


• Model-Based Economic Evaluation (Secondary Data Synthesis):


• Approach: Synthesize data from various sources (published literature, clinical trials, observational studies, registries, expert opinion) into a mathematical model (e.g., decision tree, Markov model).


• Pros: Can extrapolate beyond trial duration, incorporate long-term effects, use real-world data, compare multiple interventions, fill data gaps, more flexible.


• Cons: Highly dependent on the quality of input data, assumptions can impact results significantly, transparency is crucial.


• Running Example:


• Even if we have a trial, a model (e.g., Markov model with states like "independent," "dependent," "institutionalized," "death") would be necessary to project QALYs and costs over a lifetime horizon and beyond the trial's observation period.


• Input Data:


• Effectiveness: Efficacy of tele-rehab vs. traditional rehab from clinical trials.


• Transition Probabilities: Rates of moving between health states (e.g., from independent to dependent) for post-stroke patients from literature/registries.


• Utility Values: From published literature for various stroke health states.


• Costs: Unit costs of healthcare resources, costs of institutionalization, costs of caregiver support from administrative data, published studies.

9) Cost Categories & Estimation

A comprehensive economic evaluation must capture all relevant costs from the chosen perspective.

• Running Example (Societal Perspective):


• Intervention Costs:


• Tele-rehabilitation Program:


• Equipment (tablets, sensors, monitoring devices, internet access for patients).


• Software/Platform subscription and maintenance.


• Staff salaries (rehabilitation therapists, IT support, program coordinators).


• Training for staff and patients.


• Program development and administrative overhead.


• Traditional In-person Rehabilitation:


• Staff salaries (rehabilitation therapists).


• Facility costs (rent, utilities).


• Equipment (gym equipment, assistive devices).


• Healthcare Resource Use (Disease-Related):


• Hospitalizations: Stroke-related readmissions, complications.


• Outpatient visits: GP, neurologist, other specialists.


• Medications: Antiplatelets, statins, blood pressure meds, antidepressants.


• Diagnostic tests: Imaging, lab tests.


• Emergency department visits.


• Institutionalization: Long-term care facilities.


• Direct Non-Medical Costs:


• Patient travel costs: To and from clinics (for in-person rehab or other medical appointments).


• Caregiver costs: Informal care provided by family/friends (time, out-of-pocket expenses).


• Home modifications, assistive devices.


• Indirect Costs:


• Productivity losses/gains: Patient's lost wages due to illness or inability to work, or gains from returning to work.


• Caregiver productivity losses: Lost wages due to providing care.


• Estimation Methods:


• Micro-costing: Detailed bottom-up approach (e.g., timing how long a therapist spends with a patient, counting exact supplies).


• Gross costing: Top-down approach, using average costs from administrative databases (e.g., average cost of a hospital day, average cost of a GP visit).


• Unit costs: From national guidelines or published literature (e.g., cost per hour of therapist time).


• Surveys: For patient out-of-pocket and time costs.

10) Analysis Plan (e.g., ICER, Sensitivity Analysis)

This outlines how the collected data will be analyzed to answer the study question and address uncertainty.

• Primary Analysis:


• Calculate Total Costs: Sum all relevant costs for each intervention group over the time horizon.


• Calculate Total Effects (QALYs): Sum QALYs for each intervention group over the time horizon.


• Calculate Incremental Costs (ΔC): Cost (Tele-rehab) - Cost (Traditional Rehab).


• Calculate Incremental Effects (ΔE): QALYs (Tele-rehab) - QALYs (Traditional Rehab).


• Calculate Incremental Cost-Effectiveness Ratio (ICER):

• Interpretation of ICER:


• Place the ICER on the cost-effectiveness plane.


• Compare the ICER to a pre-defined willingness-to-pay (WTP) threshold. This is the maximum amount society is willing to pay for one additional unit of effect (e.g., $50,000 - $150,000 per QALY gained in the US; 1-3 times GDP per capita in many countries).


• Decision Rule: If ICER < WTP threshold, the intervention is considered cost-effective.


• Uncertainty Analysis: This is crucial because economic evaluations rely on estimates and assumptions.


• Deterministic Sensitivity Analysis:


• One-way sensitivity analysis: Vary one parameter at a time (e.g., discount rate, key cost item, effectiveness estimate) over a plausible range and observe the impact on the ICER.


• Multi-way sensitivity analysis: Vary multiple parameters simultaneously.


• Tornado diagram: Graphically presents the parameters that have the greatest impact on the ICER.


• Probabilistic Sensitivity Analysis (PSA):


• Assign probability distributions to all uncertain input parameters (e.g., normal, beta, gamma distributions).


• Perform Monte Carlo simulations (run the model thousands of times, drawing random values for each parameter from its distribution).


• Output:


• Cost-effectiveness scatter plot: Shows the distribution of ICERs in the cost-effectiveness plane.


• Cost-effectiveness acceptability curve (CEAC): Shows the probability that the intervention is cost-effective at different WTP thresholds. This is the gold standard for presenting uncertainty.


• Subgroup Analysis: Investigate if the cost-effectiveness differs for specific patient groups (e.g., by age, stroke severity, socioeconomic status).

11) Ethical Considerations

Beyond general research ethics, economic evaluations have specific ethical dimensions.

• Informed Consent: Ensure participants understand the economic data being collected (e.g., questions about income, resource use) in addition to clinical data.


• Confidentiality and Data Security: Especially important for sensitive financial and health information.


• Equity and Access:


• Does the intervention disproportionately benefit certain groups?


• Are the costs and benefits distributed fairly across different populations?


• If the intervention is found cost-effective, how will equitable access be ensured?


• Resource Allocation: The study's findings directly inform resource allocation, which has ethical implications regarding who receives care and who might be denied. Be transparent about methodology and assumptions.


• Researcher Bias: Acknowledge potential conflicts of interest (e.g., funding sources) and maintain objectivity.


• Opportunity Cost: Explicitly acknowledge that recommending one intervention means not funding another.

12) Dissemination Plan

Communicating findings effectively to relevant stakeholders is crucial for impact.

• Target Audience:


• Policy makers/Decision-makers: Ministry of Health, health technology assessment (HTA) bodies, hospital administrators, insurance companies.


• Clinicians: Doctors, therapists, nurses.


• Patients and Patient Advocacy Groups.


• Researchers/Academics.


• General Public.


• Channels:


• Peer-reviewed publications: High-impact health economics, clinical, and public health journals.


• Health Technology Assessment (HTA) Reports: Formal submissions to national or regional HTA bodies (e.g., NICE, CADTH, ICER).


• Policy Briefs/Executive Summaries: Concise, non-technical summaries tailored for busy decision-makers, highlighting key findings and policy implications.


• Presentations: At scientific conferences, professional society meetings, policy forums.


• Lay Summaries: Easy-to-understand versions for patients and the public.


• Media Engagement: If findings are highly significant, consider engaging with media outlets.


• Stakeholder Workshops: Engage directly with decision-makers and patient groups to discuss findings and implications.


• Key Messages: Clearly articulate the ICER, the probability of cost-effectiveness, and the implications for patient care and resource allocation. Highlight limitations and areas for future research.

This comprehensive guide should provide a solid framework for designing your economic evaluation study. Remember that economic evaluation is an iterative process, and engaging a multidisciplinary team (health economists, clinicians, statisticians, patients) is highly recommended.