Empirical population and public health ethics: A review and critical analysis to advance robust empirical-normative inquiry

Introduction

The field of bioethics has experienced several important changes during the past several decades. One such development has been both the inclusion and generation of empirical data within the bioethical task of normative inquiry. ¹ Frequently characterized by an ‘empirical turn’ (Frith, 2010; Ten Have and Lelie, 1998) – and today referred to as empirical ethics – the philosophical and methodological implications of this transformation have been vigorously debated. A simultaneous discussion has taken place in the advancement of another area of normative inquiry that is anchored within public health: public health ethics. Although frequently contested in its origins (Petrini, 2010), Bayer and Fairchild (2004) suggest that this field first emerged, in part, from a series of infectious disease outbreaks that exposed the insufficiency of some of the contemporary bioethical frameworks, specifically those that aimed to provide moral evaluations for various public health actions (e.g. quarantine practices during the severe acute respiratory syndrome (SARS) outbreaks of those who were asymptomatic in the absence of diagnostic tests). As public health ethics has continued to evolve, some theorists have also begun to take up many of the issues espoused by social scientists and demographers studying concepts, theories and data related to population health and the social and structural determinants of health, thereby leading some to refer to this emergent field as population and public health ethics (PPHE). PPHE differs from traditional orientations of bioethics, in that these approaches aim to (1) emphasize moral evaluations related to the population, alongside those of the individual; (2) acknowledge that many ‘upstream’ interventions and influences on health occur outside of the health care system and (3) emphasize the prevention of illness and disease (Canadian Institutes of Health Research, 2012).

Notably, as PPHE advanced and took up various approaches to engaging with empirical data within normative inquiry, it appears to have done so without engaging the methodological and philosophical discussions previously considered within the realm of empirical ethics. And, as the current review critically interrogates, for the most part, those engaging in PPHE have not considered the task of data generation as an important component of their work. As a result, it remains unclear how far the discussions within empirical ethics can inform or differentiate empirical approaches within PPHE. The timing is, therefore, ideal to begin a dialogue about how PPHE can (and should) consider the generation of empirical data as an important part of the task of normative inquiry.

The evolution of empirical ethics: mapping the terrain

In the late 1980s and early 1990s, a variety of authors began to put forth pleas to bring empirical methods and theories into bioethical normative inquiry. The rationale behind these calls involved a range of issues related to dissatisfaction with the practical utility of bioethics from clinicians, as well as from some ethicists whose frameworks required additional context (e.g. those practising feminist and casuistry ethics; Borry et al., 2004; Musschenga, 2005; Parker, 2009). For example, some argued that, within clinical practice, much of the ethical theory did not sufficiently align with the realities faced by patients and clinicians (Hoffmaster, 1992). As Baron (1989) described the ‘taken-for-granted’ and demarcated roles of ethicists and practitioners at the time, ‘bioethicists tend to leave the “facts” of clinical medicine to the doctors; their task then is to apply elegant and compelling arguments from first principles of ethics … to these undisputed and indisputable facts’ (p. 3). Concurrently, sociological critiques argued that bioethics over-idealized rational thought, while concomitantly excluding pertinent cultural issues that can influence clinical interactions (Frith, 2010; Hedgecoe, 2004; Musschenga, 2005; Parker, 2009). For example, scholars have argued that bioethics was overly deductive in argumentation, and that the prescriptive potential of the present-day bioethical frameworks was derived solely on ‘rational’ justifications of ethical theories abstracted from reality (Hedgecoe, 2004; Lopez, 2004). Similar to the dissatisfaction expressed by practitioners, these critiques argued that bioethics over-focussed on developing and justifying ethical theories, with little attention to the practical utility or contextual conditions within patients’ and practitioners’ everyday and actual experiences.

Despite the call for more evidence in bioethical inquiry, there has also been disciplinary resistance to the so-called ‘empirical turn’, with some arguing that analytical skills that require in-depth training and experience in philosophy, as well as sophisticated understandings of normative theory, are required (cf. Brassington, 2013). Moreover, as philosophers and social scientists sought to bridge the disciplinary ‘gaps’ between their respective fields, a set of philosophical arguments against the so-called ‘empirical turn’ arose, thereby challenging a variety of meta-ethical ² and epistemological concerns. First, some bioethicists feared that, in submitting to the ‘empirical turn’ in bioethics, ethics could no longer hold true to the philosophical method that, heretofore, required a degree of abstraction in the development of ethical theory. For example, Callahan (1996) argued that bioethics should be constructed to disregard social and cultural factors in order to avoid the ‘dangerous’ realm of moral relativism (e.g. in which each individual has his or her own ethical ‘truths’ and a universal ethical theory need not apply). If, for example, normative theory emphasizes experience and context (e.g. historical, social and cultural contexts) over the development of a universal normative theory, critics argued that bioethics will lose its prescriptive character, and therefore moral relativism is unavoidable (Van der Scheer and Widdershoven, 2004).

Second, some argued that bringing the empirical into normative inquiry would result in a purely descriptive endeavour, because it is logically invalid to infer a conclusion based on a series of premises that are contingent on modality or are fact based and value ridden (Borry et al., 2004). For example, Hume’s Law (Hume and Selby-Bigge, 1896) suggests that an ‘ought’ statement cannot be derived from a set of ‘is’ premises (the ‘is-ought’ fallacy). Moore’s (1903) Principia Ethica raises concerns about the naturalistic fallacy, suggesting that it is fallacious to qualify an action as moral or good based on natural properties. For example, concepts that can be empirically measured (at least to some extent), such as desirability or pleasantness, should not be considered proxies or equivalent to moral or good (concepts that are not generally considered empirically measurable within a scientific epistemology), though these concepts may both be ascribed to the same social acts. Others argued that the ‘empirical turn’ is a manifestation of medicine’s emphasis on evidence-based medicine (Borry et al., 2005), rather than a turn based on logical reasoning within normative inquiry (Culpepper and Gilbert, 1999; Kerridge et al., 1998). For example, some authors have argued that these issues raise a fact-value problem in which the generation and presentation of evidence are inherently normative and can only offer partial and biased glimpses into the so-called ‘facts’ (though the importance of evidence within these discussions is not disavowed altogether; Dunn and Ives, 2009; Goldenberg, 2005). As a result, some argue that bringing empirical methods and theories into the normative risks the undue privileging of “Grade 1” or “gold standards” of evidence (Kerridge et al., 1998) and lead to clinical ‘decision trees’ that exclude opportunities for normative inquiry (Jacoby, 2006).

Nonetheless, many continued to argue in favour of advancing empirical-normative inquiry. For example, in defence of empirical ethics and in response to the is-ought fallacy, Reiter-Theil (2004) points out Hans Albert’s (2000) ‘Bridge-Principles’, in which he argues that the ‘is’ and the ‘ought’ must be separate, but that in order to have a duty to perform a given act (an ‘ought’), one must be empirically capable of doing so (an ‘is’). Moreover, others argued that, if transparently and carefully carried out, explicating assumptions related to the empirical and the normative can allay many of the concerns pertaining to the naturalistic fallacy and the fact-value distinction (De Vries and Gordijn, 2009; Molewijk and Widdershoven, 2012). For example, Molewijk and Widdershoven argue that all empirical approaches to ethics should be explicit about how they handle – from a meta-ethical perspective – the interactions between the empirical data, normative reasoning and the moral conclusion (e.g. acknowledging one’s rationale and methods within their work), thereby attenuating many of the meta-ethical concerns.

Typologies and methodologies of empirical ethics

While a variety of definitions have been advanced, most share the understanding that empirical ethics actively integrates empirical research and normative analysis of data and evidence regarding a moral issue, followed by a normative conclusion (McMillan and Hope, 2008; Molewijk and Widdershoven, 2012). A variety of authors have proposed various typologies to describe and classify the different ways in which this takes place. For example, Borry et al. (2004) describe three different roles of empirical research in bioethical reflection and decision-making. First, empirical research can be used to describe the moral question by examining ‘reality-revealing’ questions (what, why, how, who, where and when) pertaining to an ethical issue, as well as to potentially reveal new issues that were not previously considered an ‘ethical issue’ (e.g. the times that meals are served to various patients). Second, empirical research is used to assess the moral question and object (e.g. assessing the degree of normativity attributed to the ‘factual’). And, third, empirical approaches can be used to evaluate the normative decision-making process in order to delineate any unexpected or unintended consequences that arose and that will hold moral relevance for subsequent ethical deliberation.

Also in 2004, Molewijk et al. (2004) describe a set of typologies in order to reveal the extent to which theorists relate to empirical data as either positioning moral theory or a given social practice as constituting the ‘final moral arbiter’. Molewijk et al. (2004) clarify their rationale for this classification system: ‘If the morality of a social practice and moral theory are not in agreement with one another, which should be adapted: social practice, moral theory, both or neither?’ (p. 56). When it is the moral theory that requires revisions, Molewijk et al. consider this an inductive (bottom-up) approach; when it is the social practices that require transformation, it is referred to as a deductive (top-down) approach. While a variety of inductive or deductive approaches were described, Molewijk et al. argue in favour of what they termed integrated empirical ethics, in which normative and empirical are considered mutually constitutive (integrative) and interdependent; in this approach, ethicists and descriptive scientists work together to integrate moral theory and empirical data in order to reach a normative conclusion. More recently, in 2009, Kon (2009a) developed a target article for a special issue in the American Journal of Bioethics that sought to classify and define how empirical research informs normative ethics. After having this target paper critically reviewed by others in the field in the special issue, Kon (2009b) noted that virtually no attention had been given to the is-ought fallacy and he declared that a conceivable dialogical consensus had been met: ‘it is settled: the “is” can (and should) inform the “ought”!’.

From a conceptual perspective, it is important to point out a distinction that some authors have made regarding whether or not empirical ethics ‘counts’ if it is simply drawing on existing evidence (e.g. scientific literature, data repositories) versus engaging in the generation of evidence through the use of various social scientific methods. Whereas some authors suggest that empirical evidence requires the generation of new, context-specific evidence (Molewijk and Widdershoven, 2012), others suggest that the inclusion of evidence (whether generated within the task of normative inquiry or not) is just as likely to be subjected to the strengths and weaknesses ascribed to empirical ethics. For those engaging in data generation, many have suggested that qualitative methods – particularly, ethnographic methods – provide the ‘thick description’ needed to inform bioethical deliberations that is often absent within quantitative or statistical descriptions of a problem (Kon, 2009a, 2009b; Hoffmaster, 1992).

Empirical approaches within PPHE: A snapshot of the field

It is worth reviewing the extent to which the generation of empirical data is considered or engaged within the task of normative inquiry within the expanding field of PPHE. For instance, there are a number of conceptual frameworks developed within PPHE approaches to normative inquiry – many of which seek to provide evidence-based decision-making tools for policy makers (Petrini, 2010). Among these, perhaps one of the most accepted frameworks (Walton and Mengwasser, 2012) is the Nuffield Council on Bioethics Stewardship Model (Nuffield Council on Bioethics, 2007) – a framework that seeks to provide policy makers with a tool to assess evidence in order to determine the moral permissibility of a public health policy and/or action. While the framework encourages rigorous approaches to evaluating evidence, concerns related to data generation and normative-empirical inquiry remain absent. Similarly, among those researchers and theorists who ‘brand’ themselves as being engaged in PPHE, their work tends to invariably draw on the already-existing evidence (with some important exceptions), rather than the active integration of empirical research and normative analysis regarding a specific moral issue. Powers and Faden’s (2006) influential book Social Justice, for example, eloquently argues that, in order to determine how complex social and economic environments influence health and well-being, normative theory and empirical data must be considered together, not in isolation. Nonetheless, further reading of their justification tacitly distils a set of demarcated roles between empirical researchers generating evidence and the ethicists engaging in normative inquiry about the evidence. It seems that for those working in PPHE, the task of generating and evaluating empirical data remains a linear, static process, generated by empirical researchers and passed to ethicists for moral evaluation. While an important and valuable endeavour, there remains a need to interrogate how empirical methods can (or ought to) be advanced within PPHE.

It is also worth exploring how the work of those outside of PPHE (e.g. those who may not ‘brand’ themselves as working within PPHE) might relate to the ways in which we may begin to conceptualize an empirical PPHE. For example, the emerging field of impact ethics aims to engage in bioethical discussions and considers science as an important (but subservient) tool for promoting human good, and, many of the published works in this area have engaged in debates relevant to public health. However, they, too, do not appear to emphasize or integrate empirical methods within their work (cf. Impact Ethics, 2013). There may also be much to learn from other areas such as narrative ethics, in which the individual or patient is explicitly acknowledged as also representing a ‘true’ ethicist. For example, narrative ethics acknowledges that most ethical acts and decisions are enacted by the patient, not the ethicist or clinician; indeed, this work recognizes that the ethical decision-making practices of individuals have value and ‘matter’ (cf. Charon and Montello, 2002).

Disciplines within the social sciences (e.g. sociology, anthropology) also have a strong tendency to both generate data related to population and public health, followed by a conclusion about what actions ought to be taken. Yet, there are also important differences with how one might conceptualize and advance empirical-normative inquiry within PPHE. Most importantly, learning lessons from the growing pains experienced by the wider field of empirical bioethics, the practical integration of empirical-normative processes in PPHE would shift away from traditional empirical and social scientific research approaches in which moral conclusions are considered largely ‘rational’ and ‘scientific’, rather than explicitly (and transparently) engaging and integrating evidence with normative theory. In doing so, for example, it becomes possible to avoid some of the aforementioned meta-ethical fallacies (e.g. fact-value distinctions) that tend to arise within scientific endeavours that seek to advance a degree of moral progress.

While the vast majority of those working in (or parallel to) the field of PPHE have not incorporated data generation within their work, there is an exemplary study by Wolff and De-Shalit (2007) that merits discussion. Wolff and De-Shalit (2007) conducted an empirical study theoretically grounded on a modified operational approach to Martha Nussbaum’s (2000) capabilities theory to identify groups that are among the least advantaged members of society. By turning to social science methods, the authors explore whether public policies that seek to ameliorate one form of disadvantage will inadvertently create another form of disadvantage or, instead, produce an overall good. While Wolff (2009) describes the aim of this research as ‘not to engage in normative consequences or analysis of the social determinants of health’, they do indicate that their use of empirical methods is to provide more realistic context in their political philosophy, thereby offering a potential ‘glimpse’ into how one might begin to conceive of an empirical PPHE.

Operationalizing an empirical PPHE

In order to advance direction for where empirical PPHE efforts should be addressed, a broad operational definition that positions empirical PPHE as actively engaging in the generation of data is the most helpful at this point – particularly when considering that many working in PPHE have already at least considered or ‘used’ evidence within their normative inquiry. Specifically, the following definition regarding empirical PPHE is proposed: the active integration of individual- (e.g. experiential) and population-level (e.g. data pertaining to the health–wealth gradient) empirical research and normative analysis of the generated data regarding a moral issue, followed by a normative conclusion. This research agenda requires interdisciplinary approaches and collaborations, with ethicists and empirical researchers working side by side, including models in which ethicists engage in and with empirical research, as well as models in which empirical researchers engage with normative theory. Indeed, this research agenda suggests a dynamic process in the generation, and constant reevaluation, of ethical imperatives in population and public health: the normative informs the empirical questions to be asked; the answers to those questions demand critical reflection on the validity of the ethical norms and values and so on.

What sort of data can be generated in empirical PPHE?

In order to further explore the extent to which the generation of evidence can contribute to empirical PPHE, a set of modified typologies put forth by De Vries and Gordijn (2009) and influenced by others are offered (Birnbacher, 1999; Molewijk et al., 2004; Sulmasy and Sugarman, 2001). The choice of typologies and divisions is for heuristic and analytic purposes and is not meant to be exhaustive. Within each typology, a set of illustrative examples of traditional approaches to empirical bioethics as described by De Vries and Gordijn (2009) are provided. Next, a set of conceptual and theoretical issues pertaining to population health that are unresolved and could potentially benefit from empirical PPHE approaches to normative inquiry are offered. Finally, within each typology, an example from public health approaches to HIV prevention and health promotion is provided. As will be highlighted, HIV represents an ideal issue to begin an exploration into issues pertaining to empirical PPHE approaches, given that there are a variety of issues that arise within and outside of the clinic, provide opportunities for prevention and concern both individual- and population-level outcomes and interests.

Advancing empirical PPHE

There is a value in explicitly examining and questioning whether and how evidence and normative analysis should be bridged in seeking a moral conclusion in order to acknowledge and attend to the various meta-ethical questions previously raised. For example, while it remains unclear whether or not the various meta-ethical questions require resolution (Molewijk and Widdershoven, 2012), there is value in explicating how the philosophical methods and normative theory employed consider or position issues such as moral relativism, abstraction and the various meta-ethical fallacies that empiricism is often charged with when entering the normative domain. Empirical PPHE requires engagement in practical philosophy; if it is to be action-guiding and claim ethical validity, we need a measure for such moral authority. Thus, this field should advance a set of careful and concerted reflexive practices to critically interrogate empirical data, as well as our various normative assumptions and biases (e.g. by drawing on various theoretical orientations from the sociology of bioethics (cf. De Vries, 2004)).

This importance is underlined when it is noted that various global actors are advocating for the generation of PPHE-relevant data, making the timing ideal to engage in meaningful discussions about how PPHE can and should generate data within normative inquiry. For example, within the Canadian context, the Canadian Institutes of Health Research’s Institute of Population and Public Health identifies the development of methods and theory within the realm of PPHE as a part of their strategic mandate, and the use of empirical evidence is leveraged as one potential pathway through which to achieve this aim. There will be much to learn by engaging in interdisciplinary approaches, including working with those from fields that have already been engaging in similar (but conceptually, philosophically and methodologically different) empirical-normative endeavours. Indeed, in advancing theoretical and conceptual orientations pertaining to an empirical PPHE, the intention is not to create a new or ‘siloed’ discipline or field, but rather to open up new interdisciplinary spaces and opportunities for those of us in this field to work together in order to advance a degree of rigour in normative and empirical inquiry in population and public health research.

The generation of evidence in PPHE will have a distinct set of implications for various potential biases, generalizability issues, as well as the normative assessments of population-based evidence. As illustrated in the conceptual and substantive examples, the ‘kinds’ of evidence evaluated within PPHE are distinct from empirical bioethics, in that it requires additional considerations of population-level empirical methods and theoretical frameworks. Thus, there is a value added in generating these issues from within the rapidly developing sphere of population health intervention research (PHIR). Within PHIR, for example, population-level interventions are defined as those that have the capacity to ‘shift the distribution of health risk by addressing the underlying social, economic and environmental conditions’ (Hawe and Potvin, 2009). In studying population-level interventions, researchers can employ methods that are experimental, quasi-experimental, observational and both qualitative and/or quantitative. Indeed, this leaves those engaging in PPHE with questions pertaining to how previously established “gold standards” of evidence should be applied within a normative-empirical analysis. For example, within this approach, studies that use non-experimental designs (e.g. natural experiments) may be likely to yield more meaningful evidence for an empirical PPHE than a randomized controlled trial (e.g. in which researchers frequently attempt to ‘control for context’). Finally, the complexity of interpreting population-based results reemphasizes the need for interdisciplinarity and critical approaches in empirical PPHE. As Dawson (2012) articulates, scientific challenges and nuances that flow from empirical population-based studies should not be underestimated and require careful and ongoing scientific and ethical consideration.

The conflation of ‘facts’ and ‘values’ in empirical work can be problematic; however, explicating how the normative conclusions and recommendations are derived from a normative analysis of empirical data can serve to provide transparent (and ethical) guidance as to how and/or why a conclusion is derived. In doing ‘ethical’ empirical-normative inquiry, the source origins of both the methodological and philosophical frameworks should be addressed to the fullest extent possible (Hurst, 2010), along with clear, transparent rationale with respect to how the empirical and normative interacted to arrive at a moral conclusion (Molewijk and Frith, 2009). To achieve this will require that those conducting empirical work move beyond exclusive considerations of research ethics to integrating an ongoing consideration of how their findings can be bridged with a (inherently normative) recommendation for action.

Conclusion

Advancing empirical PPHE approaches will be most helpful by actively engaging in the generation of new data. These empirical-normative approaches to PPHE should not be conceptualized or practised as linear, but rather as a constant feedback loop in which the normative informs the empirical questions to be asked and new evidence constantly directs the recommendations for morally robust public health practices. These critical approaches to empirical PPHE can provide opportunities to develop more philosophical and empirical rigour within this field. It may be, after all, a very difficult task to ‘do the right thing’ in our efforts to improve population health without engaging in a set of robust, transparent and ethical empirical and normative practices.