Healthcare Ethical Decision-Making Factors and Frameworks

Original Editor - Stacy Schiurring based on the course by Andrea Sturm
Top Contributors - Stacy Schiurring and Jess Bell

Introduction

Ethical decision-making in healthcare is irreducibly complex and is shaped simultaneously by individual, relational, organisational, situational, and societal factors. This Physiopedia page provides a structured overview of the Five-Dimensions Framework of Factors in Ethical Decision-Making, the key frameworks that support ethical reasoning, and the implications of using AI as a reasoning tool in clinical contexts.

The Five-Dimensions Framework of Factors in Ethical Decision-Making

The Five-Dimensions Framework was developed through a programme of research grounded in the real-world experiences of practising physiotherapists across all World Physiotherapy regions. An international online survey of 559 physiotherapists from 72 countries identified over 40 distinct factors influencing ethical decision-making, grouped into five dimensions. A subsequent qualitative analysis formalised these into the framework, and piloting by Sturm and Fryer (2024) confirmed its applicability across diverse clinical contexts. The framework is non-hierarchical: all five dimensions interact simultaneously and hold equal weight.

The Five-Dimensions Framework of Factors in Ethical Decision-Making

Dimension I: Individual Factors. Individual factors are those intrinsic to the practitioner; they are grouped by how amenable they are to change. Dynamic factors (i.e., those capable of change through professional development) include attitude, awareness, communication, education (both formal and informal), experience (encompassing maturity, personal experience, professional experience, and role reversal), knowledge, moral intent(ion), responsibility (towards the patient's best interest and wider interests), skills, and values (both personal and professional).

A second group of factors were considered relatively consistent for the individual: emotions and feelings, empathy and mentalising ability, intuition and gut feelings, individuality, ego-strength (the capacity to uphold professional values under pressure, even at personal cost), locus of control (the extent to which a practitioner believes their decisions are self-directed rather than governed by external forces, such as systems or hierarchy), and field-dependence (the tendency to be influenced by the surrounding social context when forming judgements, rather than relying on independent analysis). Age and religion were understood as stable factors.

The five most frequently reported individual factors were experience, values, responsibility, education, and moral intention.

Dimension II: Relational Factors. The relational theme, notably the only theme newly constructed from the data rather than adopted from other disciplines, reflected a significant gap in prior ethics literature. Six factors were identified:

1. Characteristics of the patient/client: including commitment, cooperation, cultural background, dependency, education level, family status and structure, psychological profile, socio-economic status, and values
2. Epistemologies: the different forms of knowledge held by the practitioner and the patient/client, respectively
3. The patient/client–physiotherapist relationship: encompassing communication, respect, power dynamics, and connection with family or kin
4. Power asymmetries: between healthcare professionals, within the therapeutic relationship, and in the workplace
5. Significant others: including colleagues, management, mentors and role models, and other health professionals
6. Supportive network: the presence or absence of colleagues, ethics committees, professional bodies, and the broader team

The most frequently reported relational factors were significant others (particularly colleagues), characteristics of the patient/client (particularly their preferences, wants and needs), and the patient/client–physiotherapist relationship (particularly the dimension of respect for the patient/client).

Dimension III: Organisational Factors. Six organisational factors reflect the workplace conditions and structures that shape ethical decision-making:

1. Organisational group norms and culture: the philosophy and values set by workplace administrators
2. Organisational structure: including the influence of dominant voices within teams or hierarchies
3. Consequences: the potential impact of ethical decisions on the practitioner's career
4. External pressures: pressure to retain employment or remain employable, which can actively hinder moral agency
5. Workplace multidisciplinarity: the availability of an interdisciplinary team for discussion and referral
6. Code of ethics: both workplace-specific codes and those of national professional associations

The most frequently reported organisational factors were codes of ethics, external pressures, and organisational structure, alongside organisational group norms and culture.

Research indicates that workplace conditions can make it almost impossible to comply with a professional code of ethics, even when a practitioner holds strong ethical values. Ethical challenges are not simply a matter of individual character; they are shaped by the systems in which professionals work.

Dimension IV: Situational Factors. Six situational factors reflect the specific features of the context in which a decision must be made:

1. Characteristics of the setting: the environment in which the clinical encounter takes place
2. Characteristics of the issue: the type of ethical issue and the degree of ambiguity or certainty surrounding the relevant facts
3. Consequences of the decision for all parties involved: practitioners may prefer pragmatic, context-sensitive outcomes over a rigid framework application
4. Finances: spanning institutional finances, the patient/client's and family's financial status, the practitioner's own situation, and systemic corruption
5. Options available: the range of choices the practitioner perceives to be accessible
6. Immediate job context: workload demands and limited time for reflection

The most frequently reported situational factors were consequences, finances, and characteristics of the setting, alongside the immediate job context.

Dimension V: Societal Factors. Six macro-level factors operate beyond the individual and the immediate workplace:

1. Environment: the broader physical and social context in which practice occurs
2. Healthcare system(s): national and international health system structures, conditions, and funding models
3. National and cultural context: the most frequently reported societal factor, encompassing cultural norms, values, and expectations
4. Politics: governmental and policy-level influences, not always perceived as supportive of ethical practice
5. Religion: at a societal level, distinct from its role as an individual factor
6. Government influence: perceived barriers to ethical practice arising from governmental decisions

The authors also note that one participant specifically highlighted the responsibilities of educational institutions and professional bodies in responding to the ethical challenges arising within these broader societal realms.

The Ethical Reasoning Bridge

The Ethical Reasoning Bridge. The term "casuistry" used in the image above refers to "the process of answering practical questions via interpretation of rules or cases that illustrate such rules, especially in ethics."

Developed by Edwards, Delany, Townsend, and Swisher (2011), the Ethical Reasoning Bridge describes ethical practice as a dynamic, reflective movement between two poles of reasoning rather than the linear application of rules.

The normative, deductive pole (e.g., principles, codes, and universal standards) asks what ought to happen. The relational, inductive pole (e.g., the actual patient, their story, cultural context, relationships, and the power dynamics of the clinical encounter) asks what is actually happening.

The skilled ethical practitioner moves between these poles, testing normative positions against relational reality and vice versa. This movement, known as reflective equilibrium, is where moral agency develops. Edwards and colleagues argue that moral agency is both the process and the outcome; it is built through the practice of crossing, not acquired in advance.

The framework has direct implications for AI use. Standard AI can generate text resembling one pole or the other, but has no mechanism for the dialectical movement between them. A practitioner who adopts an AI output without engaging in this reflective process effectively stops crossing the bridge and foregoes the moral agency that crossing develops.

Please read the Ethical Reasoning Bridge Physiopedia Page for more in-depth information on this topic.

Human Reasoning and AI: Key Concepts for Clinical Practice

Dual-Process Theory and the Explore–Exploit Continuum

Kahneman's dual-process theory distinguishes System 1 thinking (i.e., fast, automatic, intuitive) from System 2 thinking (i.e., slow, effortful, deliberative). In healthcare ethics, System 1 is the moral sensitivity that signals something feels wrong before you can articulate why; System 2 is the careful bridge-crossing process.

A significant limitation of this binary model is its inadequate account of novelty, yet novelty is ubiquitous in ethical practice. Khalil and Brüne (2025) propose the explore–exploit continuum as a more accurate model; at any moment, a practitioner can exploit (applying familiar frameworks to recognised problems) or explore (suspending assumptions when genuinely new situations arise).

It is critical to distinguish uncertainty from novelty. Uncertainty means the type of problem is recognisable and existing frameworks apply, so careful exploitation is appropriate. Novelty means the situation has not been genuinely encountered before, so existing frameworks may not apply and genuine exploration is required — not simply more careful retrieval of familiar responses.

Tri-System Theory and Cognitive Surrender

Shaw and Nave (2026) extend the dual-process model with the Tri-System Theory, adding a third system for AI as an active reasoning partner.

Table 3. Components and Characteristics of the Tri-System Theory.

System 1 - Human	System 2 - Human	System 3 - AI
Fast · Intuitive Pattern recognition Emotions Gut feelings Moral sensitivity (i.e., knowing something feels wrong before you can say why)	Slow · Deliberative Careful reflection Ethical "bridge-crossing" Analysis The effortful work of ethical reasoning across the five dimensions	Artificial · External AI cognition: automated, data-driven, confidence-generating Simulates reasoning without performing it What happens to Systems 1 and 2 depends entirely on how the person engage with it

"Standard large language models, the technology behind ChatGPT and similar tools, do not reason; they predict. Given a sequence of words, they generate the statistically most likely continuation based on patterns learnt from enormous text corpora. The model has no understanding of the content in the sense that you and I are using that word." - Dr. Andrea Sturm, PT

A further structural issue is WEIRD bias. AI systems trained predominantly on Western, Educated, Industrialised, Rich, and Democratic text tend to centre individual autonomy and the four-principle bioethics model, while underrepresenting relational, organisational, and societal dimensions. This is not a temporary limitation; it is a property of how most systems are built.

Cognitive surrender is the uncritical adoption of AI outputs without verification, effectively outsourcing judgement entirely. It differs from cognitive offloading (using a tool strategically while remaining in control) and automation bias (a tendency to favour automated outputs). In cognitive surrender, System 2 does not participate, but rather the AI's answer simply becomes the practitioner's answer.

Tri-System Theory poses a critical question: when System 3 is present, what happens to Systems 1 and 2?

Shaw and Nave tested this empirically by running three controlled experiments with 1,372 participants. Each participant completed an adapted Cognitive Reflection Test. This validated reasoning task is specifically designed to require deliberate System 2 thinking to override an instinctive System 1 response. Crucially, the AI assistant available to participants was secretly manipulated: in some conditions, it gave correct answers, in others, it gave incorrect ones. Participants consulted the AI on more than half of all trials, even when not instructed to, which demonstrates how instinctively we reach for System 3. When the AI was correct, accuracy improved by 25 percentage points above the no-AI baseline. When the AI was wrong, accuracy dropped 15 percentage points below it. This means that the participants using faulty AI performed worse than those using no AI at all. Shaw and Nave describe this gap as the signature of cognitive surrender.

Applying This Knowledge in Clinical Practice

Before consulting AI: Form your own ethical understanding of the situation first. Identify relevant factors across all five dimensions and determine whether the situation is genuinely novel (requiring exploration) or a familiar type of problem where careful framework application is appropriate. Engage both poles of the Ethical Reasoning Bridge and activate System 2 before consulting System 3. Without a sufficient ethical knowledge base, there is no basis on which to recognise or override AI when it is wrong.

While consulting AI: Evaluate the output critically. Consider what cultural perspective it reflects and whether it engages relational, organisational, and societal dimensions or defaults to individual autonomy and four-principle bioethics. Ensure the prompt includes the specific clinical, cultural, and systemic context of the situation. Ask whether the output is genuinely responding to the novelty of the situation or retrieving a familiar response from familiar territory.

After consulting AI: Reflect on whether the output has changed your view, and if so, why. Any shift should be justifiable with reference to ethical theories, the Five-Dimensions Framework, or a framework such as the Ethical Reasoning Bridge. A useful final test is asking if you could explain this reasoning clearly and confidently to the patient, their family, and a senior colleague.

Resources

• Khalil R, Brüne M. Adaptive Decision‐Making “Fast” and “Slow”: A Model of Creative Thinking. European Journal of Neuroscience. 2025 Mar;61(5):e70024.
• Shaw SD, Nave G. Thinking-Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender. Available at SSRN 6097646. 2026.
• Sturm A, Ager AL, Roth R. Western ideals and global realities–physiotherapists’ views on factors that play a role in ethical decision-making: an international qualitative analysis. European Journal of Physiotherapy. 2024 Jan 2;26(1):12-24.