Integrating Morality and Science: Semi-Imperative Evidentialism Paradigm for an Ethical Medical Practice

Abstract

Evidence-based medicine (EBM) supplies the best available data, yet clinicians still face low-value care, surrogate-driven reversals, and pseudoscientific claims. We propose Semi-Imperative Evidentialism (SIE), a normative framework that links evidential warrant to proportionate professional duties while preserving patient autonomy. Using a targeted narrative review in philosophy of science, bioethics, and clinical epidemiology, we distilled six binary attributes to classify activities as Science, Pseudoscience, or Non-science. Scientific items enter a two-tier ladder—Tier 1 (established clinical evidence) or Tier 2 (emerging or preclinical evidence)—with status re-scored as randomized trials, living meta-analyses, and post-marketing safety signals accrue. SIE maps tiers to action: Tier 1 should be offered or strongly recommended, with reasons documented if declined; Tier 2 should be discussed with explicit consent, preferably within trials or registries; Pseudoscience should be refused or discontinued with corrective education; Non-science may be acknowledged as contextual support when safe and non-substitutive. Worked examples—antiarrhythmic suppression post–myocardial infarction (CAST) and “complementary cancer cures”—illustrate earlier and more transparent course-correction. SIE provides a fallibilist bridge from evidence to duty, constraining discretion without eroding autonomy; prospective audits and cluster trials should test its impact on prescribing and consent.

1. Introduction

Evidence-based medicine (EBM) is a fundamental aspect of modern medical practice and constitutes the central building block of contemporary medicine. Medical practitioners are instructed to provide patients with optimal therapy throughout their academic career. This methodology requires observation, application, comprehension, analysis, and decision-making grounded in robust scientific evidence [1,2]. Beyond serving as a technical framework, EBM raises ongoing epistemological debates about whether medicine should sacrifice its humanistic and intuitive dimensions to conform to the perceived authoritarianism of evidence hierarchies [3,4,5,6]. This controversy extends beyond academia and has a widespread impact on medical practice worldwide, as it frequently discusses how the purported authoritarianism of EBM intersects with the limitations of medical intuition. This discussion can also include the potential impact of EBM when ethical dilemmas arise [7,8,9,10].

Semi-Imperative Evidentialism (SIE) is introduced here as a normative framework that links levels of evidential support to corresponding physician duties, thereby uniting empirical rigor with ethical responsibility. The aim of this framework is to promote a balanced approach and to place scientific evidence at the heart of ethical medical decision-making, while at the same time acknowledging the intrinsic value of clinical intuition and patient-specific considerations. Philosophically, this paper contributes to the ongoing debate on the demarcation of science and its ethical implications in applied domains. SIE advances this discussion by proposing a bridge between epistemic warrant and professional normativity—a link insufficiently articulated in previous accounts of medical evidentialism.

2. Methods

We conducted a targeted narrative review in philosophy of science, bioethics, and clinical epidemiology to derive demarcation attributes and the tier-to-duty mapping. Sources included PhilPapers, JSTOR, and Google Scholar for philosophy; PubMed/MEDLINE for clinical epidemiology; and major guidelines/position statements (e.g., General Medical Council). We included conceptual or empirical works that (i) advanced criteria distinguishing science, pseudoscience, and non-science; (ii) discussed clinician duties; or (iii) analyzed clinical reversals/surrogate outcomes.

For this paper, SIE is presented as a practical ethical framework that links graded evidence to proportionate clinician duties while preserving patient autonomy. “Evidentialism” is used in its standard sense—agents should proportion their beliefs and actions to the strength of the available evidence. Our thesis professes that Vinay Prasad et al.’s claim that a novel ethic is indispensable for maintaining the patient’s interest at the center of healthcare is profoundly consistent with this. He borrows the legal phrase “semper necessitas probandi incumbit ei qui agit,” meaning “The necessity of proof always lies with he who lays charges.” Although initially conceived as a legal principle, this notion holds significant applicability in the realm of medical interventions, where those endorsing a specific treatment or diagnostic technique must bear the burden of proof [11].

We first demarcate any proposed activity into these epistemic domains based on six binary attributes (developed from the demarcation literature and clinical epidemiology): Science, Pseudoscience, or Non-science.

• Science: We treat an activity as scientific when it advances testable, potentially falsifiable claims about disease mechanisms or outcomes and subjects those claims to disciplined methods (controls, transparency, error-checking, and communal critical scrutiny). The relevant question is not whether a phenomenon is “quantifiable” in the abstract, but whether the methods invite empirical risk and possible correction. It relates to concepts or treatments that can be empirically verified and tested through rigorous scientific methodologies. In the field of medicine, a therapy is considered scientific if it has been subjected to rigorous peer-reviewed research and controlled, repeatable experiments that provide significant evidence of its effectiveness and safety. Disciplines such as psychology and behavioral medicine are classified as Science when they offer testable hypotheses and falsifiable predictions. Within Science, SIE distinguishes two evidence tiers used for ethical duties at the bedside:

  1.1

  Tier 1: Clinical science: Clinical studies, such as randomized clinical trials (RCTs) for therapies or meticulously executed case–control studies with minimal biases for diagnostic tests, exemplify this field of medical science. Clinical science is the fundamental basis of EBM. Tier 1 comprises high-certainty clinical evidence (e.g., randomized clinical trials for therapies, well-designed diagnostic accuracy studies); clinicians have a strong duty to offer or strongly recommend to eligible patients and to communicate benefits and harms.

  1.2

  Tier 2: Pre-Clinical Science: On the other hand, pre-clinical science explores the foundational understanding of biological processes, substances, and surrogate outcomes. This course focuses on the theoretical foundations and possible therapeutic implications of novel findings occurring at the level of cells, molecules, or genes. Although pre-clinical science contributes significantly to the development of medical knowledge and creates a foundation for future clinical applications, it cannot epistemologically determine the efficacy of clinical therapies. Its applicability is limited to scholarly journals and laboratory settings where it acts as a catalyst for further clinical studies. It is inappropriate and potentially deceptive to rely solely on pre-clinical science in daily clinical practice since it has not been subjected to exhaustive tests necessary to verify its safety and efficacy in human subjects. A duty to discuss the intervention with patients, clearly disclosing the uncertainty surrounding its efficacy and/or safety. If offered, it should ideally be within a research or trial setting, or with explicit acknowledgment of its investigational nature and with robust informed consent. The potential benefits must be carefully weighed against the uncertainties and potential risks.

• Pseudoscience: falsified information masquerading as scientific knowledge. Despite the lack of empirical support and methodological rigor of science, it often asserts its scientific legitimacy and authority. Pseudoscientific therapies may be measurable or observable, but they claim that they do not fall within the scope of scientific investigation, or that they manipulate scientific language and evidence in such a way as to mislead or confuse. This is particularly detrimental due to the risk of harm to patients, waste of resources and a decline in confidence in legitimate medical practices. Pseudoscience is considered to be ethically problematic from the point of view of the SIE, since it undermines the commitment to evidence-based practice and logical decision-making. A clear ethical duty to refuse to offer or endorse the intervention and to educate the patient about the lack of credible evidence, potential harms (including opportunity costs), and the distinction between evidence-based approaches and pseudoscience.
• Non-science: This category is distinct. It does not refer to interventions proposed as alternatives to scientific treatments for specific conditions, but rather to aspects of care that address the patient’s broader context, such as psychological, spiritual, social, economic, and cultural factors. These are not subject to the same tiering as interventions but are integral to ethical medical practice. A distinct ethical duty exists to address non-scientific aspects relevant to patient well-being. As the General Medical Council (GMC) stipulates, doctors must “adequately assess a patient’s condition(s), taking account of their history including relevant psychological, spiritual, social, economic, and cultural factors” [12]. Furthermore, in contexts like palliative care, the GMC highlights the doctor’s role in “providing psychological, social and spiritual support to patients; and supporting those close to the patient” [13]. SIE emphasizes that these non-scientific elements are crucial for holistic care and patient-centeredness. The ethical duty here is to acknowledge, explore, and appropriately integrate these factors into the overall care plan, ensuring they do not conflict with or falsely substitute for evidence-based interventions for medical conditions. This delineates the use of non-science in the broader context of patient care from the evaluation of specific non-scientific interventions proposed for medical efficacy (which would likely be classified as pseudoscience or lacking evidence if making medical claims without a scientific basis). Classifying a ritual, belief, or lifestyle practice as Non-science is not a pejorative verdict. It simply acknowledges that the activity’s primary purpose is existential support rather than measurable biomedical effect. SIE encourages clinicians to honor—and, when appropriate, actively facilitate—patients’ spiritual, cultural, or personal traditions so long as they do not claim to cure disease or displace evidence-based care. In this way the framework safeguards both scientific integrity and the patient’s right to pursue sources of meaning that enhance coping, resilience, and well-being.

SIE prioritizes patient-important clinical outcomes over surrogates; surrogates may inform hypotheses but should not determine treatment in the absence of supportive clinical endpoints from randomized trials [14]. Rather than invoking the categorical imperative, SIE is analogous to a Kantian hypothetical imperative: if clinicians aim to practice good medicine consistent with fiduciary duties, then they ought to proportion recommendations to the best available evidence and disclose residual uncertainty [15]. This grounds duties in both warrant and the ends of medicine (healing, nonmaleficence, respect for autonomy), not in data alone.

According to SIE framework, one must consider it unethical for physicians to utilize pre-clinical information in clinical scenarios. However, such conduct is entirely ethical when considering this information to generate clinical hypotheses as researchers or in collaboration with other scientists. This explicit distinction ensures that medical practice maintains a solid foundation based on clinically validated evidence, protects patients from unsubstantiated hypotheses and promotes responsible and ethical developments in medical science (Figure 1).

This knowledge serves as the foundation for the inclusion of the term “semi” in SIE. Although the necessity of basing medical practice on solid evidence may appear obvious, the term “semi” is crucial to comprehend the extent and constraints of this framework through the SIE. The medical community is constantly reminded that the status quo of the present may not hold true tomorrow. It is aware of the potential for biases to influence evidence and the dynamic nature of the medical knowledge landscape. This part of the SIE approach underlines the importance of ongoing re-evaluation and critical analysis of medical evidence in order to ensure that practices continue their applicability and are firmly rooted in the latest scientific knowledge. We retain “Semi-Imperative” to mark two features. First, the duties are semi-deontic—they are pro tanto obligations graded by evidential strength and defeasible by informed refusal (Tier 1) or by uncertainty that shifts the duty to shared decision-making (Tier 2). Second, medicine is fallibilist: claims remain revisable; thus, duties are conditional on contemporaneous warrant and are periodically re-scored. “Semi”, therefore, signals graded, conditional normativity, not half-measures.

The tier definitions and domain classification above constitute the methodological substrate. The mapping of tiers to specific professional duties (offer/recommend; discuss/consent within research; refuse/educate) is operationalized in the Results via the SIE Tier–Patient Discretion Matrix and worked examples.

3. Results

3.1. Operationalizing the Three Epistemic Domains

To keep the screening step reproducible, we distilled the literature on demarcation criteria into six binary attributes (Table 1). Popper’s falsifiability criterion provides the backbone of attribute (ii): a claim counts as scientific only if it can, in principle, be refuted by empirical evidence [16]. This principle grounds the insistence that any purported therapy must make testable predictions about outcomes. Lakatos extends this by distinguishing progressive from degenerating research programs: a scientific program remains viable only if it generates novel predictions that are later corroborated, rather than endlessly rescuing itself with ad hoc adjustments [17].

Thagard’s analysis of pseudoscience highlights that what distinguishes it is not the absence of data per se, but the persistence of anomalies without self-correction and the reliance on post hoc or conspiratorial reasoning [18]. This maps directly onto our Pseudoscience (PS) category: therapies that ignore disconfirming evidence or reinterpret failure as vindication fall under this heading. Pigliucci warns against simplistic bright-line tests, arguing that demarcation is best understood as a family resemblance problem where multiple overlapping criteria define science rather than a single essence [19]. Our six-attribute schema reflects this pluralism by integrating empirical, theoretical, and sociological markers into a structured but flexible matrix. Finally, Hansson stresses that pseudoscience is not merely bad or incomplete science but a form of intellectual misrepresentation: it appropriates the symbols of legitimacy while systematically evading critical standards [20]. This observation grounds the ethical duty within SIE: physicians must not merely decline to prescribe pseudoscientific therapies but actively counter their deceptive framing through education.

Within this framework, an intervention is coded as Science (S) when it (i) articulates an empirical claim about physiological change, (ii) is in principle falsifiable, (iii) does not conflict with well-corroborated physical or biological laws, and (iv) is offered as a treatment or prevention modality for a specific condition. An activity is marked Pseudoscience (PS) when it makes therapeutic claims while violating at least one of attributes (ii)–(iii) or persisting despite a body of contrary evidence. Non-science (NS) denotes practices that address existential, cultural or spiritual dimensions of health, make no causal biomedical claims, and are never presented as substitutes for medical therapy.

Borderline cases are adjudicated by examining which attribute(s) fail: for instance, acupuncture for chemotherapy-induced nausea (empirical claim, plausible mechanism under investigation) [17] is coded S-Tier 2, whereas homeopathic prophylaxis for malaria (claim violates dose–response principles and lacks falsifiable mechanism) falls under PS. Mindfulness meditation, when used to cultivate resilience without claiming direct tumor regression, remains in NS.

Table 1. Decision grid for classifying activities into Science (S), Pseudoscience (PS) or Non-science (NS).

Attribute (Yes/No)	Science (S)	Pseudoscience (PS)	Non-Science (NS)
1. Empirical claim of physiologic efficacy (disease-specific)	Yes	Yes	No
2. Falsifiable mechanism/testable prediction	Yes	No (or ad hoc excuses)	N/A
3. Compatible with well-established scientific laws	Yes	Often violates or ignores	N/A
4. Evidence base currently available	Tier 1/2 data (see Figure 1)	Low-quality, contradictory, or none	Not applicable
5. Intended as substitute for standard therapy	Possibly	Typically promoted as substitute	Never
6. Primary purpose (therapeutic vs. meaning-making)	Treat/prevent pathology	Treat/prevent pathology	Meaning, identity, community

Legend: The six-attribute decision grid operationalizes demarcation by drawing on major philosophical accounts of science. Science (S) reflects Popper’s criterion of falsifiability and Lakatos’s demand for progressive research programs, which together require that interventions generate testable predictions and yield cumulative, corroborated knowledge [21]. Pseudoscience (PS) embodies Thagard’s view of theories that stagnate by accumulating anomalies, defending themselves with ad hoc explanations, and refusing revision, as well as Hansson’s account of pseudoscience as the strategic misuse of scientific symbols to claim false legitimacy [22]. Non-science (NS) follows Pigliucci’s pluralist approach, recognizing that not all human practices seek empirical adjudication; cultural, spiritual, and existential activities are meaningful in their own domains but cannot be evaluated as biomedical claims [23].

Table 1. Decision grid for classifying activities into Science (S), Pseudoscience (PS) or Non-science (NS).

Attribute (Yes/No)	Science (S)	Pseudoscience (PS)	Non-Science (NS)
1. Empirical claim of physiologic efficacy (disease-specific)	Yes	Yes	No
2. Falsifiable mechanism/testable prediction	Yes	No (or ad hoc excuses)	N/A
3. Compatible with well-established scientific laws	Yes	Often violates or ignores	N/A
4. Evidence base currently available	Tier 1/2 data (see Figure 1)	Low-quality, contradictory, or none	Not applicable
5. Intended as substitute for standard therapy	Possibly	Typically promoted as substitute	Never
6. Primary purpose (therapeutic vs. meaning-making)	Treat/prevent pathology	Treat/prevent pathology	Meaning, identity, community

Legend: The six-attribute decision grid operationalizes demarcation by drawing on major philosophical accounts of science. Science (S) reflects Popper’s criterion of falsifiability and Lakatos’s demand for progressive research programs, which together require that interventions generate testable predictions and yield cumulative, corroborated knowledge [21]. Pseudoscience (PS) embodies Thagard’s view of theories that stagnate by accumulating anomalies, defending themselves with ad hoc explanations, and refusing revision, as well as Hansson’s account of pseudoscience as the strategic misuse of scientific symbols to claim false legitimacy [22]. Non-science (NS) follows Pigliucci’s pluralist approach, recognizing that not all human practices seek empirical adjudication; cultural, spiritual, and existential activities are meaningful in their own domains but cannot be evaluated as biomedical claims [23].

3.2. The SIE Tier and Patient Discretion Matrix

A core output of the SIE framework is a calibrated matrix that maps the tier of evidence supporting an intervention to the corresponding physician’s duty and the appropriate level of patient discretion. This matrix operationalizes the normative bridge between evidence and ethics by assigning clear, action-guiding verbs to each domain: physicians have a strong duty to offer Tier 1, a duty to discuss Tier 2, a clear duty to refuse pseudoscience, and a supportive duty toward non-science. The goal is to eliminate ambiguity while preserving patient autonomy (

Table 2. The SIE Tier and Patient Discretion Matrix.

Evidence Tier	Physician Duty	Patient Discretion	Example
Established Science (Tier 1)—high-certainty RCTs/definitive evidence	There is a strong ethical duty to offer or strongly recommend the intervention and to document reasons if not followed.	May refuse with documented acknowledgment	Statins post-myocardial infarction
Emerging Pre-clinical Science (Tier 2)—promising but uncertain evidence (early-phase trials, robust observational data, strong biologic plausibility)	Clinicians have a duty to discuss openly, emphasize uncertainty, and preferably deliver within research or registry settings; explicit consent is required.	Shared decision-making. Acceptance usually requires signed acknowledgment or trial consent.	Phase II CRISPR-based gene therapy; prostate-specific antigen (PSA) screening
Pseudoscience (PS)—biomedical claims that contradict established knowledge or lack any credible empirical support	Clinicians have a clear duty to refuse or discontinue the intervention, to explain the lack of evidence and potential harms (including opportunity cost), and to redirect to evidence-based alternatives.	May pursue independently and at own cost after documented counselling; physician should not facilitate.	High-dose vitamin C infusions for Stage IV cancer
Non-science (NS)—spiritual, cultural or social practices not offered as biomedical treatments	Clinicians have a supportive duty to acknowledge and, when compatible with safety, integrate these practices as contextual support; always clarify that they are not disease therapies.	Fully patient-led; encourage if meaningful to well-being and not harmful or substitutive.	Patient prayer, mindfulness meditation, attendance at religious services

Legend: The matrix translates graded evidential support into proportionate ethical obligations, preserving scientific integrity while respecting autonomy.

).

In SIE, every scientific claim remains on probation. Tier 1 and Tier 2 are not end-states but way-stations that can shift as new data arrive. Post-marketing surveillance, living meta-analyses, and large pragmatic trials may demote a once-standard therapy from Tier 1 to Tier 2—or even to pseudoscience—when safety signals or null effects emerge. Conversely, a well-conducted confirmatory RCT or high-quality real-world study can promote a promising Tier 2 intervention to Tier 1. This dynamic character reflects Lakatos’s insistence that only progressive research programs deserve to remain scientific and Hansson’s point that pseudoscience involves claims to legitimacy despite contrary evidence. The framework, therefore, mandates re-evaluation and rapid updates when pivotal findings are published, ensuring that ethical duties track the best contemporaneous knowledge.

3.3. Application of SIE to Clinical Reversals

The SIE framework, with its emphasis on dynamic evidence assessment and Bayesian updating, is particularly well-suited to identify and adapt to medical reversals—instances where interventions once considered standard of care are later found to be ineffective or harmful [24]. Two prominent examples illustrate how SIE could have facilitated an earlier and more ethically sound re-classification of evidence:

• The Cardiac Arrhythmia Suppression Trial (CAST)—Flecainide [25]:

  −

  Initial Situation: Flecainide and similar antiarrhythmic drugs were widely used post-myocardial infarction (MI) to suppress ventricular premature depolarizations (a surrogate outcome), based on the plausible pathophysiological theory that suppressing these arrhythmias would reduce mortality.

  −

  SIE Application: Initially, this practice might have been classified as Tier 2 (strong theoretical basis, some observational support for surrogate outcome). However, SIE requires continuous vigilance for higher-quality evidence on hard clinical endpoints (e.g., all-cause mortality) and explicitly discounts surrogate-only rationales.

  −

  Reversal: The CAST trial, a Tier 1 RCT, revealed that flecainide and encainide increased mortality despite effectively suppressing arrhythmias [25].

  −

  SIE’s Earlier Re-classification: Under SIE, the CAST findings trigger immediate downgrading from Tier 2 to “contraindicated for this indication,” with a professional duty to cease prescribing, document the rationale, and educate peers and patients about the reversal. The framework’s periodic re-calibration mechanism would ensure such pivotal evidence leads to immediate practice change.

3.4. Navigating Ethical Dilemmas with SIE: Special Access Programs and CAM

The SIE framework also provides a structured approach to navigating complex ethical dilemmas arising from the misinterpretation or misapplication of evidence, particularly concerning Special Access Programs (SAPs) and Complementary and Alternative Medicine (CAM).

First, SAPs provide patients with the authorization to use medical interventions, including devices and drugs that have not been formally approved for widespread distribution. These initiatives, such as compassionate use, early access, expanded access, patient programs and managed access, are present in many countries [26]. Typically, medical interventions available through SAPs lack formal approval because they have not yet been evaluated to the satisfaction of regulatory authorities, rendering them “experimental” or “investigational” to some extent. SAPs have been specifically designed to ensure accessibility to these interventions in several different situations. For patients, a major justification for SAPs is compassion. Beyond potential health gains, patients may also value the opportunity to pursue additional treatment options and avoid the potential regret of not having done so [27].

The COVID-19 pandemic supplied a cautionary Tier-2 case. Early in vitro data and small, methodologically weak studies suggested that ivermectin might inhibit SARS-CoV-2 replication. Some of the first studies were later retracted. Subsequently, larger, higher-quality RCT and systematic reviews—e.g., the TOGETHER Trial and ACTIV-6—did not demonstrate clinically meaningful benefit on hospitalization or time to recovery, and meta-analyses converged toward null effects when high-bias studies were excluded [28,29,30].

−

SIE Application: When early signals exist but confirmatory trials are negative or equivocal, SIE mandates re-scoring to “do not offer outside a trial,” paired with a duty to educate about the evidentiary trajectory and the risks of diverting patients from proven care.

Another dilemma is CAM, which encompasses a wide range of healthcare modalities that were developed independently, in part or in whole, from the evidence-based medicine framework. Alternative medicine is utilized in place of evidence-based medicine, while complementary medicine is utilized in conjunction with it [31]. Despite the increasing use of terms like integrative and functional medicine, which are often portrayed as “holistic” approaches, these practices share similarities with CAM and typically lack scientific evidence to support their efficacy. Herbal remedies, acupuncture, meditation, yoga, and tai chi, in addition to consultations with naturopathic, homeopathic, and chiropractic practitioners, are prevalent CAM practices [32].

Patients who choose alternative approaches instead of evidence-based treatments may experience adverse effects when using CAM. For example, research has demonstrated that individuals who exclusively relied on alternative medicine to treat cancer experienced considerably elevated mortality rates in comparison to those who underwent conventional therapy [33,34,35].

For a variety of reasons, proponents of CAM frequently assert that the scientific method, which forms the basis of EBM, does not apply to their field. The initial assertion put forth is that the impacts of CAM are too subtle to be measured using traditional scientific methods [36].

−

SIE Application: SIE’s Pseudoscience tier directly addresses many CAM practices. It imposes an ethical duty on clinicians to refuse to offer interventions lacking a credible scientific basis or those contradicted by robust evidence. Crucially, this duty is paired with an obligation to educate patients about the lack of evidence, potential harms of abandoning proven therapies, and the distinction between evidence-based approaches and pseudoscience. This empowers patients to make informed choices rather than decisions based on misinformation or unsubstantiated claims.

At the same time, SIE underscores that the Non-science/Pseudoscience distinction has broader implications beyond the bedside. Non-science practices, when non-substitutive, may enhance coping, adherence, and overall well-being [37,38]. By contrast, pseudoscientific claims that substitute for effective therapy undermine both prevention and treatment at scale [39]. For this reason, the framework recommends that clinicians document pseudoscience-related counselling in the chart and to report significant safety signals through institutional or regulatory channels when appropriate.

4. Discussion: SIE as a Scientific Moral Rule

The SIE framework presented in this paper offers a pragmatic approach to a persistent challenge in modern medicine: the effective integration of evidence-based principles with core ethical duties in everyday clinical decision-making. Unquestionably, the longstanding claim that the practice of medicine, according to the Hippocratic tradition, is inherently a moral undertaking is founded on the fiduciary relationship between physician and patient [40]. At the same time, medicine is also a social practice, situated within institutions and communities that shape both the dissemination of evidence and the exercise of professional duty.

Current systems often leave clinicians coping with how to ponder varying levels of evidence, communicate uncertainties, and respect patient autonomy, particularly when faced with novel therapies, compassionate use requests, or the pervasive influence of pseudoscience. SIE aims to bridge this gap by providing a structured, stepwise algorithm that is both epistemologically sound and ethically robust. In the field of medical ethics, it is essential that clinical guidelines based on a meticulous assessment of evidence are constantly incorporated into personal patient data and patient preferences. The imposition of rigid requirements ought to be restricted to extraordinary situations, as they have the potential to undermine the fundamental nature of the physician-patient relationship. Instead, their purpose should be to aid clinicians in reaching the most optimal and efficient decision possible [41]. Clinical guidelines, when derived from robust evidence, should be individualized to the patient’s clinical profile and preferences, rather than imposed rigidly. Mandatory requirements should be limited and clearly specified—for example: (i) public-health imperatives where Tier-1 interventions prevent serious transmissible harm; (ii) decision-incapacity with a legally authorized representative; or (iii) life-threatening conditions without Tier-1 options where access occurs under SAP or IRB-approved protocols with robust consent. Otherwise, guidelines should inform, not override, individualized shared decision-making.

One of the primary innovations of SIE is its explicit tiering of evidence, coupled with assigned ethical duties and a clear articulation of patient discretion at each tier. This transforms abstract concepts into a concrete bedside tool, as illustrated by the Evidence Tier/Patient Discretion Matrix. Unlike vague exhortations to “consider the evidence” or “respect autonomy,” SIE provides specific guidance. For example, the duty to offer and strongly recommend a Tier 1 intervention is distinct from the duty to offer only within a research setting with robust consent for a Tier 2 intervention.

SIE also directly confronts the challenges posed by SAP and CAM, as detailed in Section 3. For SAP, SIE’s tiering system can help delineate when compassionate use might be ethically justifiable versus when it represents a false hope. For CAM, SIE’s classification provides a clear basis for clinicians to refuse to offer pseudoscientific interventions and instead fulfill their duty to educate patients, especially given the documented harms associated with choosing unproven alternative treatments over evidence-based care for serious conditions. SIE emphasizes shared decision-making, transparency about evidence quality, and documentation of informed refusal when patients decline Tier 1 care. At the same time, the Non-science/Pseudoscience distinction has public health implications. Non-substitutive Non-science practices—such as rituals or spiritual support—may enhance coping and adherence, whereas substitutive pseudoscience undermines both prevention and treatment at scale. For this reason, SIE recommends clinicians document pseudoscience-related counselling in the chart and report major safety signals through institutional or regulatory channels when appropriate.

However, the implementation of SIE is not without challenges. It requires a cultural shift towards greater transparency about uncertainty and a commitment from clinicians to engage in more structured communication. Time constraints in busy clinical settings may be perceived as a barrier. Healthcare systems would need to support this through training and potentially through the development of decision-support tools, as suggested in the research agenda.

4.1. SIE and the Philosophy of Medicine

By integrating these demarcation-sensitive epistemological, ethical components and explicit professional duties, SIE contributes to the broader philosophy of medicine across three levels: bedside judgment, professional norms, and the social context of care. It offers a model that attempts to:

• Reconcile the “Art” and “Science” of Medicine: While firmly anchored in scientific evidence, SIE clarifies that individualized care remains indispensable: clinical recommendations are evidence-constrained but value-informed, integrating patient goals, comorbidity profiles, and context. In SIE, the “art” is not license to deviate from evidence but the disciplined application of Tier-linked duties—offer (Tier 1), conditional offer with explicit consent (Tier 2), and refusal with education (PS)—to the particulars of a patient’s situation.
• Provide a normative bridge from evidence to action: SIE is prescriptive, but it does not infer duties from data alone. It joins a fallibilist epistemology (all claims remain defeasible) to a bridge principle that mediates the is–ought step: when the relevant community of inquiry (clinicians and scientists using transparent, reproducible methods) establishes robust warrant for a clinical claim, physicians incur a pro tanto duty to offer it because this best serves medicine’s fiduciary ends—beneficence, nonmaleficence, and respect for autonomy. By robust warrant we mean high- or moderate-certainty evidence on patient-important outcomes (not surrogates alone), typically from replicated RCT or convergent methods with coherent mechanisms and an acceptable risk–benefit profile. The duty is pro tanto—overridable by contraindications or informed refusal. Accordingly: for Tier 1 (Established Science), clinicians should offer/strongly recommend and document reasons if declined; for Tier 2 (Emerging Science), uncertainty shifts the duty to shared decision-making under explicit consent, preferably within a trial or registry; for Pseudoscience (PS), clinicians should refuse or discontinue, correct misinformation, and redirect to evidence-based options; for Non-science (NS), clinicians may acknowledge and support practices as contextual care when safe and non-substitutive. All classifications are rescorable as new evidence emerges.
• Challenge reductionist views of EBM: SIE guards against algorithmic reductionism by requiring transparency about uncertainty, documentation of informed refusal when Tier-1 care is declined, and periodic re-verification of standards through living syntheses and post-marketing surveillance. Thus, the framework operationalizes “evidence-based” as an ongoing, auditable practice rather than a static checklist.

In short, SIE is neither algorithmic nor intuitionist; it is a fallibilist, role-moral practice that ties graded evidential warrant to graded professional duties, with autonomy preserved through tier-specific communication.

4.2. Limitations of the Proposal

While the SIE framework offers a structured approach to integrating evidence and ethics, several limitations should be acknowledged. First, SIE is a conceptual framework derived from a narrative review of demarcation theory, clinical epidemiology, and bioethics; it has not yet undergone prospective empirical validation. This article advances a philosophical-normative framework grounded in demarcation theory and clinical epidemiology; empirical validation remains to be done.

Second, the tiering process inevitably involves expert judgment. Borderline cases and rapidly evolving studies risk misclassification. To mitigate this, implementation should include (i) explicit operational definitions, (ii) dual independent ratings with adjudication, and (iii) reporting of inter-rater reliability (e.g., κ statistics), with periodic re-ratings triggered by predefined evidence thresholds (new phase 3 trials, living meta-analyses, post-marketing safety signals).

Third, time and workflow constraints may limit uptake in busy clinical environments. Effective dissemination will require training in evidence appraisal and communication, embedding of decision-support tools within the electronic health record (EHR), and audit-and-feedback cycles aligned with service priorities.

Fourth, SIE’s mapping from evidence tiers to professional duties may be interpreted as rigid if not paired with shared decision-making. The framework explicitly requires documentation of informed acceptance or refusal for Tier 1 care and structured consent for Tier 2 care, but real-world fidelity to these steps warrants measurement and improvement support.

Fifth, scope limitations. SIE is designed for bedside ethics and may not fully capture population-level trade-offs in public health or resource allocation. Future work should test how the Science/Non-science/Pseudoscience distinctions interact with prevention and health-promotion strategies at scale.

Sixth, terminology and philosophical underpinnings. Demarcation remains contested in philosophy of science; reasonable scholars differ on the weight given to falsifiability, research-program progress, and sociological markers [42]. SIE adopts a pluralist, fallibilist stance, but the possibility of conceptual disagreement is a limitation that empirical testing cannot eliminate.

Finally, unintended consequences. There is a risk of stigmatizing patients who use Non-science practices or, conversely, minimizing harms from Pseudoscience that substitutes for effective care [43]. Implementation should pair clinician education with respectful counseling scripts, chart documentation of pseudoscience-related counseling, and clear referral to evidence-based alternatives, with safety signals reported through institutional or regulatory channels when appropriate.

5. Conclusions and Future Directions

The SIE framework offers a structured and ethically informed approach to integrating evidence into clinical practice. It seeks to move beyond the often-siloed considerations of EBM and bioethics, providing a unified workflow that addresses the epistemological challenges of evidence assessment and the moral imperatives of patient care. The framework directly addresses critical issues such as the management of low-value care, the interpretation of clinical data, and the navigation of compassionate use and pseudoscience, offering concrete guidance where current approaches often fall short.

The goal is to promote a medical culture where the principles of SIE are embedded in routine practice, ensuring that patient care is evidence-based, ethically sound and truly patient-centered. Adhering to this dynamic approach necessitates constant adaptation to new evidence, maintaining ethical and scientific rigor in healthcare practices.