National Health Policy and Political Economy

Introduction

National health policy is not simply a technical plan for improving health. It is a public settlement about whose health needs receive priority, who pays, who provides services, which risks are collectively shared, and how scarce resources are distributed. Although epidemiology, economics and management provide essential evidence, policy choices are also shaped by institutions, interests, ideas, political coalitions and unequal power.

For health-care managers, understanding this political dimension is essential. A technically sound reform may fail because it threatens powerful interests, lacks administrative capacity, conflicts with institutional rules or is not perceived as legitimate. Conversely, a politically attractive policy may produce limited health gains if it is poorly designed or inadequately implemented.

This chapter examines the policy process from agenda setting through evaluation. It combines stakeholder analysis, power mapping, institutional theory and political-economy perspectives. The emphasis is not on treating policy as a linear sequence, but on understanding it as an iterative and contested process in which evidence, values and power interact.

Learning objectives

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

1. Explain the stages and non-linearity of national health policy processes.
2. Apply stakeholder analysis and power mapping to a health-policy problem.
3. Use institutional theory to explain policy continuity, resistance and change.
4. Analyse health reforms through political-economy frameworks.
5. Distinguish policy formulation, adoption, implementation and evaluation.
6. Assess equity, legitimacy, accountability and feasibility in policy decisions.
7. Develop a politically informed implementation and evaluation strategy.

1. Health policy as a political and institutional process

A health policy is an authoritative decision, set of decisions or deliberate non-decision by public authorities concerning health goals, financing, regulation, service delivery or population protection. Policies may be expressed in legislation, regulations, budgets, strategies, benefit packages, clinical standards or administrative instructions.

A useful distinction is between:

• Policy intent: what decision-makers officially seek to achieve.
• Policy design: the instruments, financing arrangements and rules selected.
• Policy implementation: how organizations and professionals translate rules into action.
• Policy effects: the changes in access, quality, expenditure, health outcomes and distribution.
• Policy legitimacy: the extent to which affected groups regard the policy as appropriate, fair and lawful.

A policy can succeed in one dimension and fail in another. For example, a financing reform may increase enrollment but leave service quality unchanged. A hospital autonomy policy may improve managerial flexibility while widening geographic inequalities. Evaluation therefore requires multiple criteria rather than a single measure of success.

Health policies are especially political because they allocate life chances and financial burdens. They determine who receives subsidized care, which occupations are regulated, whether private providers may operate, how medicines are priced and whether taxation or insurance contributions finance services. These decisions create winners, losers and groups whose interests are invisible unless distribution is explicitly examined.

2. The policy cycle—and why it is not linear

Textbooks often present five stages: agenda setting, formulation, adoption, implementation and evaluation. This sequence is analytically useful, but real policy processes are recursive. Evaluation can reopen agenda setting; implementation can alter the policy itself; and adoption may occur before design is fully settled.

Figure: The policy cycle is iterative, with power, institutions and ideas influencing every stage.

2.1 Agenda setting

Agenda setting concerns how a condition becomes recognized as a public problem requiring government action. Health conditions do not automatically reach the policy agenda. They compete for attention with employment, security, education, inflation and other concerns.

Three elements commonly interact:

1. Problem definition: data and narratives establish what the problem is. Maternal mortality may be framed as a clinical-quality failure, a gender-equity issue, a transport problem or a consequence of poverty.
2. Policy alternatives: feasible solutions must be available. A problem is more likely to gain attention when credible instruments already exist.
3. Political opportunity: elections, crises, leadership changes, court decisions, media attention or international commitments can open a policy window.

Kingdon’s multiple-streams model describes these elements as the problem, policy and politics streams. Policy entrepreneurs couple the streams by promoting a particular solution when an opportunity arises. The model explains why technically important problems may remain neglected for years and why sudden reform can occur after a crisis.

Agenda setting is also shaped by framing. A policy framed as universal solidarity may attract different allies from one framed as individual responsibility. Managers should therefore ask not only, “How large is the problem?” but also, “Who defines it, using which indicators, and whose experiences are excluded?”

2.2 Policy formulation

Formulation converts a recognized problem into possible courses of action. It involves defining objectives, target populations, benefits, financing, delivery arrangements, regulatory instruments, timelines and accountability mechanisms.

Common policy instruments include:

• legislation and regulation;
• taxation and subsidies;
• public procurement;
• insurance mandates and provider payment;
• information, standards and professional guidelines;
• direct public provision;
• contracting and public–private partnerships.

Formulation should compare options against explicit criteria such as effectiveness, equity, efficiency, fiscal sustainability, administrative feasibility, political acceptability and legal compatibility. Trade-offs must be made visible. A policy that maximizes average health may worsen distribution; a highly equitable benefit package may require taxes that are politically difficult to sustain.

Evidence informs formulation but does not determine it. Randomized trials may establish effectiveness, while implementation research, costing studies, qualitative inquiry and distributional analysis assess whether an intervention can work in a particular institutional setting.

2.3 Adoption

Adoption is the formal authorization of a policy through a legislature, cabinet, ministry, regulatory agency, court or budget process. Adoption may involve negotiation, amendment and compromise. A policy document is not necessarily a funded policy: legal authority, budgetary authorization and organizational responsibility must align.

A policy’s adoption prospects depend on the distribution of benefits and costs, the salience of the issue, coalition strength, veto points and the credibility of promised gains. The “veto-player” perspective emphasizes that actors whose agreement is required can block or dilute change. Federal systems, bicameral legislatures, constitutional courts and professional regulators may create multiple veto points.

2.4 Implementation

Implementation is the transformation of policy rules into organizational routines, professional practices and population experiences. It includes financing flows, procurement, staffing, information systems, training, supervision, referral arrangements and enforcement.

Top-down approaches emphasize clear objectives, resources, authority and compliance. Bottom-up approaches highlight street-level discretion: front-line workers interpret rules under conditions of workload, uncertainty and limited resources. Lipsky’s concept of street-level bureaucracy is particularly relevant to health systems, where nurses, clinicians, social workers and facility managers allocate time and attention in real time.

Implementation is therefore a process of translation. The same national policy may produce different results across regions because of administrative capacity, local politics, infrastructure, professional norms and informal practices.

2.5 Evaluation

Evaluation asks whether a policy was implemented as intended, achieved its objectives, represented good value and distributed benefits fairly. Four forms are especially useful:

• Process evaluation: Was the intervention delivered, to whom and with what fidelity?
• Outcome evaluation: Did indicators such as coverage, quality or utilization change?
• Impact evaluation: What changes can reasonably be attributed to the policy rather than to other causes?
• Economic evaluation: Were additional health gains worth the additional resources used?

Evaluation should include unintended effects. For example, a copayment may reduce unnecessary use but also discourage essential care among low-income households. A performance target may improve recorded immunization while encouraging inaccurate reporting or neglect of unmeasured services.

3. Stakeholder analysis and power mapping

A stakeholder is an individual, group or institution that can affect a policy, is affected by it, or has a legitimate interest in it. Stakeholder analysis prevents policy teams from treating “the public” as a single actor.

Important stakeholder categories include:

• elected leaders and political parties;
• ministries of health, finance and local government;
• public and private providers;
• health professionals and unions;
• patients, caregivers and civil-society organizations;
• pharmaceutical and technology firms;
• insurers and purchasers;
• donors and international agencies;
• courts, legislatures and regulatory bodies;
• journalists, researchers and opinion leaders.

A practical stakeholder assessment records each actor’s interests, resources, position, legitimacy, relationships and likely response. Position is not fixed: an association may support universal coverage in principle but oppose a proposed payment mechanism.

Power mapping

Power is the capacity to shape decisions, implementation or interpretation. It may be:

• Visible: formal votes, legal authority and budget control.
• Hidden: control over agenda access, evidence production or consultation rules.
• Invisible: shaping what is considered normal, affordable or politically possible.

A simple power–interest matrix is useful, but it should not be treated as a mechanical classification. A low-income patient group may have little formal power yet possess moral legitimacy, mobilization capacity and the ability to attract media attention. A ministry may have high formal authority but weak implementation control.

A stronger map also examines relationships. For example, a finance ministry may influence a cabinet committee; professional councils may influence clinical practice; a patient coalition may influence media framing; and donors may influence procurement standards. Network power can matter more than formal hierarchy.

Stakeholder analysis procedure

1. Define the policy question and affected population.
2. List direct, indirect and potentially excluded stakeholders.
3. Assess interests, resources, legitimacy and likely positions.
4. Map alliances, conflicts and dependencies.
5. Identify risks of capture, exclusion or tokenistic consultation.
6. Design engagement, negotiation and accountability mechanisms.
7. Reassess the map as policy conditions change.

Ethically, engagement is not merely a strategy for securing compliance. Participation should be meaningful, accessible and transparent about which decisions are genuinely open to influence.

4. Institutional theory and health policy

Institutions are the formal rules, organizations, norms and routines that structure action. Institutional theory helps explain why policies persist even when alternatives appear more efficient, and why similar reforms produce different outcomes in different countries.

Historical institutionalism

Historical institutionalism emphasizes path dependence. Earlier decisions create infrastructures, vested interests and expectations that make some future choices easier than others. A tax-funded health service, employment-based insurance system and fragmented donor-funded system each create different reform pathways.

Critical junctures—such as war, economic crisis, regime change or epidemic emergencies—can disrupt existing paths. Yet change is often layered rather than revolutionary: new programs are added to old arrangements, producing complexity and institutional ambiguity.

Sociological institutionalism

Organizations may adopt policies because they appear legitimate and modern, not solely because they are technically optimal. International accreditation, strategic plans and performance dashboards may be adopted ceremonially while everyday practices change little. This phenomenon, sometimes called decoupling, makes implementation evaluation essential.

Rational-choice institutionalism

This perspective examines how rules shape incentives. Ministries, providers, patients and politicians are assumed to pursue interests within institutional constraints. Payment systems illustrate the point: fee-for-service can encourage activity, while fixed budgets may encourage cost containment but risk under-provision. No incentive works independently of professional norms, information and enforcement.

Institutional analysis questions

Managers should ask:

• What formal rules authorize or constrain action?
• Which organization is responsible, and does it possess the required capacity?
• What informal norms influence behavior?
• Who benefits from existing arrangements?
• What path dependencies make change difficult?
• Where are the gaps between authority, responsibility and resources?

Institutional diagnosis often reveals that implementation problems are not simply failures of motivation. They may reflect incompatible reporting requirements, fragmented accountability, weak procurement systems or contradictory incentives.

5. Political economy of health policy

Political economy examines how power, resources and institutions shape economic decisions and their distribution. In health, it asks who pays, who gains, who bears risk, who controls production and whose preferences are represented.

Distributional conflict

Health reforms redistribute money, risk, status and organizational autonomy. Expanding public purchasing may benefit patients but reduce the bargaining power of private providers. Introducing generic medicine procurement may reduce pharmaceutical revenues while improving affordability. Decentralization may empower local governments but widen disparities where fiscal capacity differs.

A political-economy analysis should identify:

• material interests and expected gains or losses;
• ownership and market structure;
• fiscal constraints and taxation politics;
• labor relations and professional autonomy;
• gender, class, ethnicity, disability and geographic inequalities;
• external influence from donors, lenders and corporations;
• institutional veto points and accountability channels.

Ideas and narratives

Interests alone do not explain policy. Ideas define what is considered fair, efficient or feasible. Universalism, social insurance, market competition, charity and human rights imply different policy designs. Narratives can also depoliticize choices: describing a distributional decision as a purely technical efficiency reform may obscure who bears its costs.

Public choice and collective action

Public-choice approaches draw attention to rent seeking, concentrated benefits and diffuse costs. A small, organized industry may lobby effectively against a reform whose benefits are spread across millions of citizens. Collective-action problems can also weaken patient groups, especially when illness is stigmatized or affected populations are geographically dispersed.

Governance beyond government

Contemporary health policy involves networks rather than ministries alone. Governments regulate and finance, but insurers, firms, professional bodies, communities, donors and digital platforms also govern behavior. Network governance requires coordination, transparency and mechanisms to manage conflicts of interest.

6. From technical proposal to politically feasible policy

A high-quality policy option should pass four tests:

1. Substantive merit: Is there credible evidence that it can improve health or system performance?
2. Equity: Who benefits, who pays and who may be excluded?
3. Institutional feasibility: Can responsible organizations implement and regulate it?
4. Political legitimacy: Can it obtain and sustain lawful, publicly defensible support?

These tests do not mean abandoning ambitious goals. They encourage sequencing and coalition building. A reform may begin with a defined population, demonstrate results, establish administrative capacity and then expand. However, sequencing can also entrench inequality if early benefits are reserved for politically influential groups. Equity safeguards should therefore be explicit.

A policy brief for decision-makers should state the problem, affected groups, options, evidence, distributional consequences, financing, implementation risks, legal requirements, stakeholder positions and evaluation plan. It should distinguish evidence from assumptions and disclose uncertainty.

7. Implementation management and accountability

Implementation planning should link authority, resources and responsibility. A useful implementation design specifies:

• policy owner and accountable authority;
• operational objectives and milestones;
• financing and procurement arrangements;
• workforce requirements and role definitions;
• data systems and reporting burden;
• training, supervision and feedback;
• complaint, appeal and redress mechanisms;
• risk register and mitigation actions;
• monitoring indicators and review points.

Accountability has several dimensions. Vertical accountability concerns reporting to ministers or legislatures. Horizontal accountability involves regulators, auditors, courts and professional bodies. Social accountability enables communities and service users to monitor, question and influence providers.

Performance management must be designed carefully. Indicators should be valid, reliable, interpretable and linked to decisions. Excessive targets can distort priorities. Managers should combine quantitative measures with qualitative inquiry, independent audits and patient experience data.

8. Evaluation, learning and policy adaptation

Evaluation is strongest when designed before implementation begins. A theory of change links inputs to activities, outputs, outcomes and longer-term effects. It should identify assumptions and external factors.

For a policy intervention, a basic change measure may be expressed as:

Here, Y₀ is the baseline value, Y₁ is the value after implementation, and ∆Y is the observed change. This describes change but does not by itself establish causation. Attribution requires a credible counterfactual, such as a comparison group, interrupted time series or controlled implementation design.

Distribution should also be examined. An average improvement can conceal deterioration among disadvantaged groups. Results should therefore be disaggregated by income, sex, age, disability, ethnicity, geography and other ethically relevant characteristics, subject to privacy and statistical reliability.

Evaluation findings should feed back into policy. Adaptation is not necessarily failure; it may represent responsible learning in a complex system. Yet adaptation should be documented so that changes are transparent and unintended consequences can be assessed.

9. Contemporary developments: digital policy and artificial intelligence

Digital health changes the policy environment by creating new data infrastructures, private platforms and forms of surveillance. Electronic records, telemedicine and algorithmic decision support may improve access and coordination, but they can also reproduce existing inequalities where connectivity, literacy or language access is uneven.

Artificial intelligence can support forecasting, triage, fraud detection and resource allocation. Its use requires governance of data quality, privacy, cybersecurity, explainability, bias, professional accountability and procurement. An algorithm should not be treated as neutral merely because it is computational. Historical data may encode discriminatory access patterns, and automated decisions can make responsibility difficult to locate.

Policy-makers should require impact assessment, human oversight, auditability, mechanisms for appeal and clear allocation of liability. Technology should serve publicly defined health objectives rather than silently determining them.

10. Illustrative case: introducing a national essential-medicines purchasing reform

Background

Assume, for illustration, that the fictional country of Lydora experiences frequent stock-outs of essential medicines. Public facilities purchase through decentralized contracts, prices vary widely and patients often buy medicines privately.

Problem

The ministry proposes pooled national procurement using a centralized electronic tender, a preferred-generic list and framework contracts. The reform could lower prices and improve availability, but it threatens suppliers that benefit from fragmented purchasing and may create risks if the central system fails.

Stakeholder and power analysis

The ministry of health has formal authority but depends on the finance ministry for budget release and on regional warehouses for distribution. The finance ministry supports savings but is concerned about transition costs. Pharmaceutical distributors oppose centralization, while generic manufacturers support predictable volume contracts. Clinicians favor reliable supply but fear loss of therapeutic choice. Patients and civil-society groups support affordability but demand transparency. Procurement officials possess technical expertise and can either enable or delay implementation.

The power map therefore shows that formal authority is insufficient. A coalition must include finance officials, procurement professionals, regional managers, clinicians and patient representatives. Anti-corruption agencies and auditors are important accountability actors.

Approach

The ministry conducts a market assessment, publishes selection criteria, separates product-quality decisions from commercial negotiations and pilots pooled purchasing in three regions. It establishes a transparent tender portal, conflict-of-interest declarations, supplier performance clauses and an emergency procurement pathway. Regional facilities retain limited authority for urgent or specialized products. Training and inventory-management support accompany the procurement change.

The evaluation plan includes stock-out frequency, delivery timeliness, medicine prices, product quality, facility satisfaction and patient financial burden. Results are examined by region and facility type rather than reported only as a national average.

Outcome

In the illustrative scenario, the pilot reduces price variation and improves availability in participating facilities, but two regions experience delivery delays because warehouse capacity was underestimated. The ministry adjusts the implementation design by investing in storage, improving demand forecasting and retaining contingency suppliers.

Learning points

1. Technical economies of scale do not guarantee implementation success.
2. Procurement reform redistributes rents and therefore requires political analysis.
3. Piloting can generate evidence, but pilot sites may not represent the whole system.
4. Transparency and conflict-of-interest controls are central to legitimacy.
5. Flexibility and contingency arrangements protect resilience.

11. Challenges and limitations

Political analysis has several limitations. Stakeholder positions may be hidden or change rapidly. Power is difficult to measure, and analysts may reproduce the same exclusions they seek to identify. Models such as the policy cycle simplify complex processes and can encourage false certainty. International comparisons may overlook historical and institutional differences.

Evidence can also be selectively used. Economic evaluation may undervalue unpaid caregiving; administrative data may exclude people outside formal services; and consultation may privilege articulate, well-resourced groups. Ethical analysis must therefore consider voice, rights, procedural fairness and the consequences of non-decisions.

Policy-makers should avoid both technocracy and cynicism. Politics does not make evidence irrelevant, and evidence does not eliminate value conflict. The professional task is to make assumptions, trade-offs and power relations more visible while pursuing feasible improvements in population health and equity.

Best-practice checklist

Before recommending a national health policy, ask:

• Is the problem clearly defined and supported by multiple forms of evidence?
• Have affected and excluded groups been identified?
• Who gains, who loses and who pays?
• Which institutions can authorize, implement and regulate the policy?
• Are authority, resources and accountability aligned?
• What political windows and veto points exist?
• Are equity and human-rights safeguards explicit?
• Is implementation adaptable but transparent?
• Are monitoring, evaluation and independent accountability built in?
• Could digital tools create privacy, bias or access risks?

Chapter Summary

National health policy is a contested process of collective choice, not a purely technical exercise. Agenda setting depends on problem framing, policy alternatives and political opportunity. Formulation compares instruments and trade-offs; adoption requires authority and coalition support; implementation translates rules through organizations and front-line discretion; and evaluation assesses delivery, outcomes, attribution, value and equity.

Stakeholder analysis identifies interests, legitimacy, resources and relationships. Power mapping adds formal, hidden and invisible forms of influence. Institutional theory explains path dependence, norms, incentives, decoupling and the importance of aligning authority with capacity. Political-economy analysis examines distributional conflict, markets, labor, fiscal constraints, ideas and accountability.

Effective health managers combine evidence with political and institutional diagnosis. They engage stakeholders ethically, anticipate resistance, protect marginalized groups, design realistic implementation arrangements and use evaluation for learning. Contemporary digital and AI-enabled reforms require additional attention to privacy, bias, transparency and responsibility.

Key Terms

Agenda setting: The process through which conditions gain recognition as public problems requiring action.

Policy formulation: Development and comparison of policy options, instruments and implementation arrangements.

Policy adoption: Formal authorization of a policy by a competent political or regulatory authority.

Implementation: Translation of policy decisions into organizational routines and services.

Policy evaluation: Systematic assessment of implementation, outcomes, impacts, value and equity.

Stakeholder: An actor affected by, able to influence or legitimately interested in a policy.

Power mapping: Analysis of actors’ formal authority, resources, relationships and agenda influence.

Policy window: A period when problem recognition, available solutions and political conditions align.

Policy entrepreneur: An actor who promotes particular ideas and couples them to political opportunities.

Veto player: An actor whose agreement is necessary for policy change.

Path dependence: The tendency of earlier institutional choices to constrain later options.

Critical juncture: A period of major disruption in which institutional trajectories may change.

Street-level bureaucracy: Front-line public service work involving discretion in applying policy.

Political economy: Study of how power, resources and institutions shape economic and policy decisions.

Distributional conflict: Competition over who receives benefits and bears costs.

Legitimacy: Perceived appropriateness, fairness and lawful authority of a policy.

Social accountability: Public or community mechanisms for monitoring and influencing institutions.

Decoupling: Adoption of formal structures without corresponding change in everyday practice.

Theory of change: A logical account linking activities to outputs, outcomes and longer-term effects.

Review Questions

Short Answer Questions

1. Why is the health-policy process described as iterative rather than strictly linear?
2. Distinguish policy intent from policy implementation.
3. What are visible, hidden and invisible forms of power?
4. Explain path dependence in national health systems.
5. Why does an observed change in an outcome not automatically establish causation?

Long Answer Questions

1. Explain the five analytical stages of the health-policy process and discuss how they interact.
2. Compare stakeholder analysis with political-economy analysis in the design of a universal health coverage reform.
3. Discuss how historical, sociological and rational-choice institutional theories explain policy continuity and change.
4. Examine the ethical requirements for meaningful participation in national health-policy formulation.
5. Develop an evaluation framework for a national essential-medicines policy, including process, outcome, equity and accountability dimensions.

Analytical / Application Questions

1. A government proposes introducing user fees in public hospitals to reduce unnecessary utilization. Construct a stakeholder and power map, identify likely distributional effects and recommend safeguards.
2. A national digital triage algorithm performs well on average but poorly for rural linguistic minorities. Analyse the problem using institutional and political-economy perspectives, then propose a governance response.
3. A decentralization reform has improved local innovation but widened regional inequalities. Design a politically feasible policy adaptation that preserves local autonomy while strengthening equity and accountability.

Multiple Choice Questions with Answers

Question 1

Which combination best describes the three streams in Kingdon’s multiple-streams model?

A. Problem, policy and politics

B. Finance, workforce and technology

C. Legislation, litigation and procurement

D. Inputs, outputs and outcomes

Correct Answer: A. Problem, policy and politics

Question 2

Which activity is primarily associated with policy formulation?

A. Measuring post-policy mortality changes

B. Comparing regulatory, financing and service-delivery options

C. Counting votes after parliamentary adoption

D. Auditing historical expenditure only

Correct Answer: B. Comparing regulatory, financing and service-delivery options

Question 3

A stakeholder with high power and high interest should generally be:

A. Ignored until implementation ends

B. Engaged continuously and involved in negotiation

C. Given only publicly available information

D. Excluded to prevent political influence

Correct Answer: B. Engaged continuously and involved in negotiation

Question 4

Which example best illustrates invisible power?

A. A minister signing a regulation

B. A legislature voting on a bill

C. A dominant narrative defining market solutions as the only realistic option

D. A procurement officer issuing a tender

Correct Answer: C. A dominant narrative defining market solutions as the only realistic option

Question 5

Path dependence means that:

A. All policies inevitably converge over time

B. Earlier institutional choices constrain later reform options

C. Policy outcomes are determined only by current evidence

D. Implementation is independent of organizational history

Correct Answer: B. Earlier institutional choices constrain later reform options

Question 6

Which statement best reflects street-level bureaucracy?

A. Front-line workers apply policy without discretion

B. Front-line workers interpret and adapt rules under practical constraints

C. Only elected officials influence implementation

D. Professional norms have no effect on service delivery

Correct Answer: B. Front-line workers interpret and adapt rules under practical constraints

Question 7

A policy evaluation reports baseline coverage of 60% and post-implementation coverage of 72%. Using ∆Y = Y₁ − Y₀, what is the observed change?

A. 6 percentage points

B. 12 percentage points

C. 20 percentage points

D. 132 percentage points

Correct Answer: B. 12 percentage points

Question 8

Which finding is most clearly an implementation outcome rather than an impact outcome?

A. Reduced maternal mortality attributable to the reform

B. Improved population life expectancy

C. The proportion of facilities receiving the required training

D. Reduced household financial hardship over five years

Correct Answer: C. The proportion of facilities receiving the required training

Question 9

Why can a national policy produce different results across regions?

A. Regional results are always random

B. Institutional capacity, local politics and informal practices vary

C. National policies cannot influence local organizations

D. Evidence has identical meaning in every setting

Correct Answer: B. Institutional capacity, local politics and informal practices vary

Question 10

Which governance measure is most appropriate when introducing an AI-based clinical decision tool?

A. Remove all human oversight

B. Prevent users from appealing automated decisions

C. Require auditability, bias assessment and clear responsibility

D. Assume historical data are neutral

Correct Answer: C. Require auditability, bias assessment and clear responsibility