Indicators, Measurement and Evaluation Frameworks

Introduction and context

Health-care evaluation converts broad questions—“Is the service effective?”, “Is it equitable?”, and “Is it worth the investment?”—into systematically collected evidence. The central instrument of this conversion is the indicator: a defined, observable measure that provides information about a phenomenon, intervention, service, or system. Indicators do not constitute evaluation by themselves. They become useful when linked to a theory of change, a clear denominator, a credible data source, an appropriate comparison, and a decision rule.

In health-care management, indicators support several functions: accountability to patients and regulators, quality improvement, resource allocation, strategic planning, risk management, commissioning, and research. However, measurement can also distort behaviour. If a hospital rewards only shorter length of stay, it may discharge patients prematurely; if a clinic rewards only the number of consultations, it may privilege volume over continuity and appropriateness. Consequently, postgraduate health-care managers must understand not only how to calculate indicators, but also what they mean, what they omit, and how they may influence professional practice.

This chapter presents the indicator hierarchy of input, process, output, outcome, and impact; explains structure–process–outcome, RE-AIM, and balanced-scorecard frameworks; and examines reliability, validity, sensitivity, benchmarking, and target setting. The emphasis is on selecting measures that are technically sound, ethically defensible, and useful for managerial decisions.

Learning objectives

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

1. Distinguish input, process, output, outcome, and impact indicators.
2. Construct numerators, denominators, time periods, and inclusion criteria for health-care measures.
3. Apply the structure–process–outcome, RE-AIM, and balanced-scorecard frameworks.
4. Assess reliability, validity, sensitivity, specificity, responsiveness, and data quality.
5. Interpret rates, ratios, proportions, and risk-adjusted comparisons appropriately.
6. Use benchmarking and target setting without encouraging gaming or inequity.
7. Design an evaluation measurement plan for a health-care programme.
8. Recognise ethical, governance, equity, and technological issues in measurement.

Indicators as a measurement system

An indicator is a quantitative or qualitative signal used to describe, monitor, compare, or judge performance. A useful indicator has at least five components:

• Concept: what is being measured, such as continuity of care or medication safety.
• Operational definition: the precise rule for identifying events and cases.
• Numerator: the number of qualifying events or individuals.
• Denominator: the population or opportunity from which the numerator arises.
• Time and setting: the period, service, geography, and population to which the measure applies.

For a proportion, the general expression is:

The numerator must normally be a subset of the denominator. For example, “patients receiving a medication reconciliation within 24 hours” should have as its denominator all eligible admissions, not all outpatient visits. A rate differs from a proportion because it relates events to person-time or another exposure of time:

The unit might be infections per 1,000 patient-days or deaths per 100,000 population-years. A ratio compares two quantities that need not be subsets of one another, such as nurses per occupied bed.

The indicator hierarchy

Indicators are often arranged along a causal or programme logic chain. Inputs enable activities; activities produce outputs; outputs contribute to outcomes; sustained outcomes may contribute to impacts. The relationships are not automatic, because context, implementation quality, patient characteristics, competing interventions, and external shocks intervene between levels.

Input indicators

Input indicators describe the resources invested in a programme. They include funding released, staffing levels, training hours, equipment availability, medicines, infrastructure, and information-system functionality. Inputs are necessary but do not demonstrate quality or benefit. A fully funded programme may still have poor processes, low uptake, or inequitable outcomes.

Managers should distinguish nominal resource availability from usable capacity. For example, the number of installed dialysis machines is less informative than the number operational, staffed, maintained, and available to eligible patients. Input measures should therefore include quality and availability conditions where relevant.

Process indicators

Process indicators describe what providers or systems do. Examples include the percentage of eligible patients screened for sepsis, the time from triage to assessment, or the proportion of prescriptions reviewed by a pharmacist. Process indicators are particularly useful when the causal link between a recommended practice and an outcome is strong.

A process measure should specify eligibility and acceptable exceptions. A “100% compliance” target may be inappropriate when clinical contraindications exist. The denominator should exclude cases for which the intervention is not indicated, while exclusions should be transparent and auditable. Otherwise, apparent improvement may result from changing the denominator rather than improving care.

Output indicators

Outputs are immediate products or services delivered. Examples include completed immunisations, discharged patients, counselling sessions, laboratory reports issued, or staff trained and assessed as competent. Outputs are closer to operational productivity than outcomes. A high number of counselling sessions does not establish that patients changed behaviour or improved health.

Output measures are valuable for capacity and workload planning, but they should be paired with indicators of appropriateness, experience, equity, and outcomes. Volume without quality can create incentives for superficial completion.

Outcome indicators

Outcomes represent changes in health status, behaviour, experience, knowledge, function, or service use that plausibly result from care. Examples include improved glycaemic control, reduced pain, better patient-reported experience, fewer avoidable readmissions, or increased treatment adherence.

Outcomes are often influenced by factors beyond the provider’s control. Interpretation therefore requires case-mix adjustment, comparison with a baseline or control group, and attention to follow-up completeness. Patient-reported outcome measures can capture benefits not visible in administrative data, but they require validated instruments and careful handling of missing responses.

Impact indicators

Impact indicators concern sustained, broad, and often population-level effects, such as reduced maternal mortality, increased healthy life expectancy, reduced health inequalities, or lower disability-adjusted burden. Impacts usually take longer to emerge and are affected by socioeconomic conditions, policy, environment, and multiple interventions. They are therefore important for strategy but often unsuitable as short-term performance measures for a single clinical unit.

From indicator definition to measurement specification

A measurement specification is the technical contract that makes an indicator reproducible. It should state:

1. Indicator name and purpose.
2. Population, setting, and eligibility criteria.
3. Numerator and denominator.
4. Unit of measurement and scaling factor.
5. Data source and extraction method.
6. Reporting frequency and responsible owner.
7. Exclusions, risk-adjustment variables, and missing-data rules.
8. Desired direction of performance.
9. Known limitations and interpretation cautions.

For example:

Timely antibiotic administration for suspected sepsis: numerator = eligible adult emergency admissions receiving the first indicated antibiotic within one hour of documented recognition; denominator = all eligible adult emergency admissions with suspected sepsis during the reporting period; unit = percentage; exclusions = documented allergy or clinical decision not to administer antibiotics; source = electronic prescribing and clinical records; reporting = monthly.

A measure should be tested through a data-definition review and an inter-rater or parallel extraction exercise. If two trained abstractors interpret the same records differently, the specification requires refinement.

Major evaluation frameworks

Structure–process–outcome framework

Donabedian’s structure–process–outcome model remains a foundational framework for quality evaluation.

• Structure refers to the relatively stable attributes of the setting: facilities, staffing, governance, technology, training, and organisational capacity.
• Process refers to what is done for and with patients: assessment, diagnosis, communication, treatment, coordination, and follow-up.
• Outcome refers to the effects on patient health, experience, function, safety, and equity.

The framework encourages a causal interpretation: appropriate structure increases the likelihood of good processes, and good processes increase the likelihood of desirable outcomes. Yet the relationship is probabilistic rather than deterministic. A well-equipped hospital may deliver poor care; a resource-constrained service may achieve excellent results through effective teamwork.

A balanced evaluation should therefore combine all three domains. Structure indicators are often easier to collect and change rapidly, but they can become compliance exercises. Outcome indicators are more meaningful to patients but may be slow, noisy, and confounded. Process indicators often offer the most immediate improvement leverage when evidence-based care pathways are established.

RE-AIM framework

RE-AIM is designed particularly for evaluating implementation and public-health interventions. Its five dimensions are:

• Reach: the proportion and representativeness of the intended population that participates.
• Effectiveness: beneficial and adverse outcomes, including effects across subgroups.
• Adoption: the proportion and representativeness of settings or staff that take up the intervention.
• Implementation: fidelity, adaptations, costs, and consistency of delivery.
• Maintenance: continuation of the intervention and its effects over time.

RE-AIM prevents an evaluation from equating efficacy in a selected sample with successful real-world implementation. An intervention may be highly effective among participants but have limited public-health value if few eligible people are reached. Conversely, broad reach with weak effectiveness may produce little benefit.

RE-AIM is useful when evaluating, for example, a digital hypertension programme. Reach might include enrolment among eligible patients and representation of older adults; effectiveness might include blood-pressure change and adverse effects; adoption might include the proportion of clinics using the programme; implementation might assess fidelity and staff workload; maintenance might assess continued use after grant funding ends.

Balanced scorecard

The balanced scorecard translates strategy into a portfolio of measures rather than relying on a single financial or activity metric. In health care, four commonly adapted perspectives are:

1. Patient and community: experience, trust, access, equity, and outcomes.
2. Internal processes: safety, timeliness, coordination, and clinical reliability.
3. Learning and growth: workforce capability, culture, innovation, and staff engagement.
4. Financial and stewardship: sustainability, productivity, cost, and responsible use of resources.

The scorecard should express a strategy map: learning and growth support better processes; better processes improve patient outcomes and experience; these outcomes support sustainable use of resources. The model is not a universal requirement to have exactly four perspectives. A public hospital may require a fifth perspective for equity, or a humanitarian organisation may prioritise community accountability.

A scorecard becomes weak when it contains too many disconnected measures. Each measure should have an owner, a reporting cycle, a desired direction, and an explicit decision attached to it. Managers should avoid averaging unlike indicators into a single score unless the weighting, missing-data treatment, and value judgements are transparent.

Measurement quality

Reliability

Reliability is the consistency of a measurement under stable conditions. A measure can be reliable without being valid: a faulty thermometer may give the same incorrect reading repeatedly.

Important forms include:

• Test–retest reliability: stability when the same construct is measured again.
• Inter-rater reliability: agreement among observers or abstractors.
• Internal consistency: coherence among items intended to measure one construct.
• Data-system reliability: completeness, timeliness, and consistent extraction across periods.

For categorical ratings, Cohen’s kappa is often used to assess agreement beyond chance between two raters. Its interpretation depends on prevalence and marginal distributions; a low kappa does not always mean poor raw agreement. For continuous measures, the intraclass correlation coefficient may be appropriate when assessing agreement or consistency among raters. Reliability statistics should not substitute for clear definitions and training.

Validity

Validity concerns whether a measure supports the intended interpretation and use. It is not a permanent property of an instrument independent of context.

• Content validity: coverage of relevant aspects of the construct.
• Construct validity: consistency with theoretical expectations, including convergent and discriminant evidence.
• Criterion validity: relationship with a credible reference or criterion when one exists.
• Face validity: whether the measure appears reasonable to stakeholders; useful but weak as evidence.
• Internal validity: credibility of causal conclusions within an evaluation.
• External validity: relevance or transportability to other populations and settings.

A measure of “patient-centred care” based only on waiting time has limited content validity. A validated patient-reported experience instrument may be stronger, but translation, literacy, mode of administration, and cultural context still require assessment.

Sensitivity, specificity, and responsiveness

When an indicator or test identifies a condition, sensitivity and specificity are central. Sensitivity is the probability that the test is positive when the condition is present; specificity is the probability that the test is negative when the condition is absent.

These formulas apply to a defined reference standard and a specified threshold. They do not directly provide the probability that a positive result is correct; positive and negative predictive values depend on prevalence.

For performance monitoring, responsiveness is the ability to detect meaningful change over time. A measure may be reliable but insufficiently responsive if it has a ceiling effect, floor effect, or overly broad categories. The minimally important difference should be considered alongside statistical significance.

Data quality and missingness

Data quality includes accuracy, completeness, consistency, timeliness, uniqueness, and validity of format. Missing data are not automatically random. Patients with poor access may be less likely to complete surveys; busy clinicians may omit documentation during emergencies. Reporting only complete cases can bias results.

A measurement plan should describe the amount and pattern of missingness, investigate its causes, and report denominators transparently. Imputation may be appropriate in research under defensible assumptions, but it should not conceal poor operational data capture. A sudden improvement after an electronic-record change may reflect a coding alteration rather than a real clinical improvement.

Comparing performance: benchmarking and adjustment

Benchmarking is the systematic comparison of performance with a reference point. Possible benchmarks include a historical baseline, peer organisations, a national standard, an internal best-performing unit, or a clinically established threshold. Comparisons are meaningful only when definitions, populations, time periods, and data quality are sufficiently comparable.

Crude comparisons can be misleading because hospitals serve different populations. Risk adjustment may account for variables such as age, comorbidity, disease severity, emergency status, or socioeconomic context. However, adjustment should not remove disparities that organisations have a responsibility to address. It should also avoid adjusting for mediators that are consequences of poor care, such as complications caused by the service itself.

For rare outcomes, rates fluctuate substantially by chance. Confidence intervals, control charts, or multi-year aggregation can help distinguish random variation from meaningful change. A funnel plot can identify unusually high or low results while accounting for different sample sizes, although outliers require investigation rather than automatic sanction.

Benchmarking should be developmental rather than purely punitive. Managers should ask: Is the difference real? Is it clinically important? Is it inequitable? What process explains it? Can the practice be transferred? What unintended consequences might follow from pursuing the benchmark?

Target setting and performance management

Targets translate aspirations into expected performance. A defensible target is specific, measurable, time-bound, relevant, and supported by a credible rationale. Common approaches include:

• Historical improvement: a defined improvement from a stable baseline.
• External standard: a regulatory, professional, or evidence-based threshold.
• Best achievable performance: a benchmark from comparable high-performing services.
• Threshold target: a minimum level required for safety or compliance.
• Stretch target: an ambitious level used for innovation, not punishment.

Targets should consider baseline performance, uncertainty, case mix, capacity, equity, and the time required for change. A target of 95% may be suitable for a highly reliable safety process but unrealistic for a complex outcome with incomplete follow-up. Where harm rises with performance failure, a minimum safety threshold may be more appropriate than ranking.

Target setting should be accompanied by balancing measures. If a hospital targets shorter emergency-department stays, it might also monitor reattendance, admission appropriateness, patient experience, and mortality. Targets should be periodically reviewed because fixed goals can become obsolete or incentivise threshold behaviour.

Worked example: constructing and interpreting an indicator

A clinic wants to monitor follow-up for patients newly diagnosed with hypertension. In one quarter, 240 eligible patients were identified. Of these, 198 had a documented follow-up visit within 30 days. The indicator is:

The numerator is 198 patients and the denominator is 240 eligible patients. If 20 patients were excluded because they moved out of the service area before follow-up and the exclusion rule was defined in advance, the revised denominator would be 220 and the percentage would be 90.0%. This illustrates why exclusions must be clinically justified and consistently applied: changing the denominator can materially change the apparent result.

The manager should not conclude that care improved merely because the percentage rose. The analysis should examine whether follow-up was timely, whether blood pressure improved, whether patients from deprived areas were less likely to return, and whether the documentation system captured visits consistently. A balanced set might include follow-up completion, median time to review, blood-pressure control at three months, patient-reported understanding, and equity stratification.

Applied case: evaluating a virtual diabetes-management service

This is an illustrative case.

Background

A regional health system introduces a virtual service for adults with type 2 diabetes. Patients receive remote consultations, medication review, home glucose education, and escalation to in-person care when clinically necessary. The stated goals are to improve glycaemic control, reduce avoidable travel, and preserve continuity during periods of high demand.

Problem

After six months, the programme reports a high number of virtual consultations and a fall in face-to-face appointments. Executives must determine whether the service is effective, equitable, safe, and sustainable. A volume-only evaluation would be inadequate.

Analysis and measurement design

The team develops a logic model. Inputs include trained clinicians, secure software, devices, and protected consultation time. Processes include eligibility assessment, medication reconciliation, escalation screening, and documented shared decision-making. Outputs include completed consultations and education sessions. Outcomes include change in [[MATH:HbA_{1c}], hypoglycaemia events, patient experience, and unplanned hospital use. Longer-term impacts include diabetes complications and health inequalities.

The team applies the three frameworks together. Structure–process–outcome is used to examine whether the service has safe infrastructure, reliable clinical processes, and beneficial patient results. RE-AIM is used to assess reach among older adults, adoption by clinics, implementation fidelity, and maintenance after initial funding. A balanced scorecard tracks patient experience and equity, internal safety processes, workforce learning, and financial sustainability.

A measurement specification defines eligible patients, the 65 ext{-}day follow-up window, data sources, and missing-data rules. Results are stratified by age, language, deprivation, disability, and digital access. A comparison with similar patients receiving usual care is preferred to a simple before–after comparison because secular trends and changes in case mix may otherwise be mistaken for programme effects.

Findings and response

The service reaches 62% of eligible patients, but reach is substantially lower among people without broadband access. Glycaemic outcomes improve modestly among participants, while medication-reconciliation completion is high. However, several patients report difficulty recognising when a virtual consultation should be replaced by urgent in-person assessment. Staff value flexibility but report increased documentation time.

The management response is not to expand the programme solely because consultation volume is high. The service adds telephone and community access routes, strengthens escalation protocols, audits adverse events, and monitors outcomes by subgroup. It also introduces a balancing measure for delayed escalation and a workforce measure for documentation burden.

Outcome and learning points

The evaluation demonstrates that implementation success is multidimensional. The programme is effective for many enrolled patients but not yet equitable in reach or fully mature in safety processes. The case illustrates five principles: outputs are not outcomes; digital access is an equity determinant; frameworks are complementary; subgroup analysis is essential; and balancing measures protect against narrow target pursuit.

Ethics, governance, and regulation

Health-care indicators involve personal and sometimes sensitive information. Governance should address lawful data use, purpose limitation, data minimisation, access control, retention, audit trails, and secure linkage. Patients should not be disadvantaged because their data are difficult to collect or because they decline an unnecessary secondary use.

Indicators can create ethical tension when transparency conflicts with privacy, or when public reporting risks stigmatising clinicians serving high-need populations. Small-cell suppression may be necessary to prevent re-identification. Patient and public involvement improves construct validity and legitimacy, especially for measures of experience, dignity, and equity.

Measurement systems should distinguish improvement from surveillance. If staff believe every deviation will trigger punishment, documentation may become defensive and adverse events may be concealed. A just culture combines accountability for reckless conduct with learning from system weaknesses and honest human error.

Technology, AI, and contemporary developments

Electronic health records, patient portals, remote monitoring, and interoperability can improve timeliness and granularity of measurement. Automated dashboards can identify missed follow-up, generate run charts, and stratify performance. Artificial intelligence may support case finding, natural-language extraction, risk prediction, and anomaly detection.

These technologies introduce risks. Algorithmic indicators may reproduce biased documentation, perform poorly in underrepresented groups, or be difficult to explain. Data drift can reduce validity when clinical practice or coding changes. Automation bias may cause managers to accept a dashboard without checking the underlying data. Every AI-enabled measure requires human oversight, subgroup validation, version control, cybersecurity, and a process for contesting erroneous classifications.

Contemporary evaluation increasingly combines routine data with patient-reported outcomes, experience measures, qualitative evidence, implementation science, and real-world evidence. The strongest systems do not treat quantitative indicators as superior to qualitative insight; they use each to answer different questions.

Best-practice checklist

Before adopting an indicator, ask:

1. What decision will this indicator inform?
2. Is the construct important to patients, professionals, or the public?
3. Are numerator, denominator, eligibility, exclusions, and time period explicit?
4. Is the data source complete, timely, secure, and reproducible?
5. Is the measure reliable, valid, and responsive enough for its purpose?
6. Could case mix or access differences distort comparison?
7. What balancing measures will detect unintended consequences?
8. Will results be stratified to reveal inequity?
9. Who owns the measure and what action follows an adverse result?
10. How will the measure be reviewed, retired, or revised?

Chapter Summary

Indicators are structured signals that support health-care evaluation, but their value depends on precise definitions, credible data, appropriate comparisons, and thoughtful interpretation. Input, process, output, outcome, and impact indicators describe different points in a programme logic chain and should be used together rather than treated as interchangeable.

The structure–process–outcome framework links organisational conditions, care delivery, and results. RE-AIM examines reach, effectiveness, adoption, implementation, and maintenance in real-world programmes. The balanced scorecard aligns patient, process, workforce, and stewardship perspectives with strategy. Measurement quality requires attention to reliability, validity, sensitivity, specificity, responsiveness, missingness, and data governance.

Benchmarking and target setting must account for uncertainty, case mix, equity, and unintended consequences. A mature evaluation system combines quantitative indicators with qualitative evidence, patient perspectives, balancing measures, and ethical oversight. Technology and AI can improve measurement, but only when transparency, validation, security, and human judgement remain central.

Key Terms

Indicator: A defined measure used to describe or assess a phenomenon or performance.

Input indicator: A measure of resources available to a programme or service.

Process indicator: A measure of activities performed during care or implementation.

Output indicator: A measure of immediate products or services delivered.

Outcome indicator: A measure of short- or medium-term change resulting from care or intervention.

Impact indicator: A measure of sustained, broad population-level effect.

Numerator: The count of qualifying events or cases in an indicator.

Denominator: The eligible population or opportunity from which the numerator arises.

Rate: Events related to a specified amount of person-time or exposure.

Reliability: Consistency of a measurement under stable conditions.

Validity: The extent to which evidence supports the intended interpretation and use of a measure.

Sensitivity: The probability that a test is positive when the condition is present.

Specificity: The probability that a test is negative when the condition is absent.

Responsiveness: The ability of a measure to detect meaningful change over time.

Benchmarking: Systematic comparison with a defined reference point.

Risk adjustment: Statistical adjustment for differences in patient or population characteristics affecting outcomes.

Balancing measure: A measure used to identify unintended effects of pursuing another target.

RE-AIM: A framework covering reach, effectiveness, adoption, implementation, and maintenance.

Balanced scorecard: A strategy and performance-management system using multiple organisational perspectives.

Review Questions

Short Answer Questions

1. What distinguishes an outcome indicator from an output indicator?
2. Why must the numerator and denominator of an indicator be explicitly defined?
3. Name the five dimensions of the RE-AIM framework.
4. What is the difference between reliability and validity?
5. Why are balancing measures important in target-based performance management?

Long Answer Questions

1. Explain the structure–process–outcome framework and discuss its strengths and limitations in evaluating a hospital service.
2. Compare RE-AIM with the balanced scorecard. In what circumstances would a health-care manager use each framework?
3. Discuss how case mix, missing data, and data quality can affect benchmarking of clinical outcomes.
4. Describe a systematic method for developing and approving a new health-care indicator.
5. Examine the ethical and governance issues involved in using electronic health records and AI-generated indicators.

Analytical / Application Questions

1. A hospital reports that 94% of eligible patients received a discharge summary, but excludes patients whose records contain incomplete demographic information. Analyse the possible validity and equity problems and propose a revised measurement approach.
2. Design an indicator set for a community hypertension programme using at least one input, process, output, outcome, and impact indicator. For each, specify a numerator, denominator or unit, data source, and potential limitation.
3. A virtual-care programme increases completed consultations but also increases emergency readmissions among patients with limited digital access. Using RE-AIM and a balanced scorecard, evaluate the programme and recommend management actions.

Multiple Choice Questions with Answers

Question 1

Which indicator level measures the resources available to deliver a health programme?

A. Input

B. Process

C. Outcome

D. Impact

Correct Answer: A. Input

Question 2

A numerator of 198 eligible patients receiving follow-up is divided by a denominator of 240 eligible patients. What is the percentage?

A. 72.5%

B. 82.5%

C. 90.0%

D. 121.2%

Correct Answer: B. 82.5%

Question 3

Which RE-AIM dimension concerns the proportion and representativeness of settings or staff that take up an intervention?

A. Reach

B. Effectiveness

C. Adoption

D. Maintenance

Correct Answer: C. Adoption

Question 4

Which statement best describes validity?

A. Consistency of results under stable conditions

B. The extent to which evidence supports the intended interpretation of a measure

C. The number of observations collected per month

D. The speed with which data are reported

Correct Answer: B. The extent to which evidence supports the intended interpretation of a measure

Question 5

A test identifies 90 true positives and 10 false negatives. What is its sensitivity?

A. 10%

B. 50%

C. 90%

D. 100%

Correct Answer: C. 90%

Question 6

Which is an example of a balancing measure for a target to reduce emergency-department length of stay?

A. Number of staff meetings

B. Emergency-department floor area

C. Patient reattendance within 72 hours

D. Number of computers installed

Correct Answer: C. Patient reattendance within 72 hours

Question 7

Why can crude comparisons of hospital mortality be misleading?

A. Mortality is never an outcome indicator

B. Hospitals may serve populations with different levels of illness and risk

C. All mortality data are qualitative

D. Risk adjustment always eliminates inequity

Correct Answer: B. Hospitals may serve populations with different levels of illness and risk

Question 8

Which balanced-scorecard perspective most directly includes workforce capability and organisational learning?

A. Learning and growth

B. Patient and community

C. Financial stewardship

D. Internal processes

Correct Answer: A. Learning and growth

Question 9

What is the main purpose of a denominator in a performance indicator?

A. To identify the data-entry software used

B. To define the eligible population or opportunity from which the numerator arises

C. To guarantee that the measure is valid

D. To remove all variation from results

Correct Answer: B. To define the eligible population or opportunity from which the numerator arises

Question 10

A programme reaches many patients but produces little improvement and is poorly maintained after funding ends. Which conclusion is most appropriate?

A. It demonstrates strong performance on every RE-AIM dimension

B. Reach alone proves effectiveness

C. Its implementation success is mixed and requires multidimensional assessment

D. Maintenance is irrelevant once reach is high

Correct Answer: C. Its implementation success is mixed and requires multidimensional assessment