The production of critical theories in Health Systems Research and Education An epistemological approach to emancipating public research and education from private interests

This paper aims at offering alternative methodological perspectives in health systems research, to produce critical, theoretical knowledge in domains such as health policy and management of health care, organization of disease control, political economy of health and medical practice.We first examined the reasons to believe that worldwide economic agents have driven publicly funded schools of public health to adopt their preferred policies and to orient their priority research topics. We then studied whether this hidden leadership has also contributed to shape research methodologies, which we contrasted with the epistemological consequences of a quest for intellectual independence, that is, the researcher’s quest to critically understand the state of health systems and generalize results of related action-research. To do so, we applied concepts of what could be named the ‘French School of Critical Sociology’ to qualitative research methodologies in descriptive health systems research. To do so, we applied concepts of what could be named the ‘French School of Critical Sociology’ to qualitative research methodologies in descriptive health systems research.


I. Introduction
The human stakes of international health policies are immense: conservative estimates put the current avoidable death toll at about 20 million per year (if demographic coefficients from industrialised countries are applied to LMIC); and 150 million worldwide are yearly driven below the poverty line because of catastrophic health expenditure 3 , 4 .Poor or good policy also has huge economic and fiscal externalities: in 2007, the total expenditure on health as a proportion of GDP was 8.8 % in the European Region against 15.7 % in the USA .Given the size of this market-to-be, policy decisions are issues of fierce legal, financial and regulatory tactics, political influence and distortion of science 5 .
Evidence indicates that international policies were aligned on commercial interests and consequently overlooked universal access to individual care as an objective and as an evaluation criterion 6 .In LMIC, governments and international health agencies did not publish much data related to access to and quality of individual care.The The production of critical theories in Health Systems Research and Education 3 WHO, for instance, requires member states to monitor the density of community health workers7 but not hospital admission rates.
In theory, policies should be subject to scientific evidence on health systems but little is available and it often stands idle.Academics endowed with a social commitment have the responsibility to produce knowledge relevant to social actors striving for equitable health care systems.Accordingly, this paper aims at steering debates on health systems research (HSR) methodologies and content of health systems education ('-E') in domains such as health policy and management, organization of disease control programs, political economy of health and medical practice.
In the present paper, we examine the possibility that economic agents have driven publicly funded schools of public health to adopt their preferred policies and to neglect other research topics such as access to individual care in LMIC, overall health systems efficiency, psycho-social care, democratization of health sectors, effectiveness of universal health insurance coverage, international aid transaction costs, evaluation of private health care financing, non-managed care approaches to health care rationalization, etc.We then explore the epistemological consequences of a quest for intellectual independence from private interests in health systems research aimed at critically understanding the state of health systems (past and present) and at generalizing results of related action-research.Notice that we do not examine the methodological requirements of producing a body of professional, prescriptive (and descriptive) knowledge constructed to help introduce change in health care delivery and management.
We shall see that the objective of publicly funded HSR-E becoming independent from private interests does not only include the establishment of priority topics of research but also the methodologies to produce this knowledge while examining alternative options.Public health scientific knowledge largely builds upon bio-medical sciences in the domain of disease control and epidemiology.By contrast, the organization of individual health care services and health systems relies much more on human sciences.Clearly, schools of thought exist in these disciplines.To explore the methodological requirements of the qualitative research component of descriptive public health8 , we deliberately concentrated on one such school, which could be named the French School of Critical Sociology.We chose to apply some of its key concepts to HSR because of its coherence across time (intergenerational dialogue of key authors, over 100 years9 ), its relevance to contemporary public health and health care organization issues and consistency with the values of a socially, not commercially, motivated approach to public health and HSR.

II.
Why suspecting a capture of public research by private interests?
Public health and / or individual health care?
It has been shown that competition for funds and for international recognition pushes professionals towards vertical initiatives and that research practices contribute to the dominance of vertical strategies and limit the success of evidence based policymaking for strengthening health systems 10 .Given that contemporary vertical initiatives are part of a large scale move to privatize discretionary, individual health care in LMIC 6, it is important to examine the political economy of (public health) academic orientations, including historical factors, underpinning these individual researchers' preferences.
The historical role of public health in society, together with funding sources, has fundamentally influenced its modern operations.Public health can be traced back to the cholera epidemic deciphered by G. Snow in London in 1854. 11 For decades, public health mainly addressed infectious disease control.Before World War 2, European government-led policies gave to the poor "public health", large scale interventions designed to control some diseases and implemented by public services (Snow 1936) -while the upper classes were the only ones to access curative, individual care (delivered in the private for profit sector).The same pattern emerged in LMIC.Originally, disease control programs were created to protect the workforce (e.g.workers building the Matadi-Leopoldville railroad as a measure against malaria and those of the Panama Canal as a measure against yellow fever).Later, the so-called vertical programs expanded the range of beneficiaries to include all lower social classes and targeted more diseases (for instance trypanosomiasis and epidemic meningitis), some maternal and child conditions and included family planning in 'neo-Malthusian' programs 12 .From an epistemological perspective tradition thus contributed to shape the borders of contemporary public health and to nest disease control in its hub while access to individual health care was neglected by the public health research community.
However, such neglect has no scientific basis anymore: it is now believed that access to individual care is the key determinant of health status improvement.By 1994, the scientific community was convinced that the aggregate effect of medical care on life expectancy (LE) during the 20 th century was roughly five years, with a further potential of two years 13 .Around 1996, conditions amenable to medical care (among others TB, appendicitis, asthma, etc.) and those that related to national health policies (liver cirrhosis, accidental death, etc.) were deemed to account for 11 and 19 % of total mortality 14 , respectively.Inadvertently demonstrating the consequences of the collapse of health systems , the rapid fall in ex-USSR life expectancy (LE) in the 1990s -before that social pathologies could unleash their disastrous effects -confirmed the importance of health care.In 2002, a study revisiting the 80's in East Germany showed that 50-60 % of the improvement in LE during that time (estimated by decomposing LE by age and cause, in particular unnecessary, untimely deaths) was attributable to declining mortality from conditions responsive to medical care.The authors concluded that the potential effect on mortality of reducing differences in effectiveness or quality of medical care was greater than had been assumed.Among women in Poland, up to 80 % of premature deaths were related to conditions amenable to medical care 15 .Nowadays, some authors attribute up to 30 years LE gains to preventive and curative health services in the USA during the 20 th century 16 .In industrialized countries, with improvements in public health (both environmental interventions and prevention) largely completed, medical care remains undisputedly the major determinant of life expectancy 17 .In short, there is a wealth of evidence to support placing access to versatile care (as provided by family medicine and general hospitals) at the core of (inter)national health policies and education.

Un-evidence based health policies and public health education
That access to care (in family medicine and general hospital settings) could save up to 20 million lives yearly, and that this could be achieved spending just 8 % of GDP (as in Europe) instead of 16 % (as in the USA) are dramatic facts simply overlooked both by: (1) international policy makers and, (2) health system researchers in most schools of public health worldwide.
1.In 1993, the World Bank 18 dismissed access to care in LMIC as an objective for (inter)national policies and until recently, these followed suite.Although vertical programs have specific indications 19 , LMIC governments generalized them and financed access of the poor to family medicine and hospital care inadequately or not at all -despite the fact that their alleged cost was an invalid argument against them 20 At the end of the nineties, UN agencies started promoting community insurances, as a substitute for ailing public financing of private providers with an alleged public regulator 212223 After having promoted universal access to a limited number of health programs (defined by the MDGs), the United Nations made a U-turn and launched an all-15 Nolte, E., Scholz, R., Shkolnikov, V., & McKee, M. 2002 out campaign for free care at the point of delivery 24 -as they began promoting public private mixes in health care financing -alongside health care commoditization.
2. In academia, research topics followed these international policy convolutions.For instance: -The number of papers with the words "health", "insurance", "coverage" and "Africa" identified on Google scholar as key words multiplied by almost 10 between 1993 and 2010.Contrasting with this feverish activity, compulsory insurance systems never contributed much to overall health sector financing (always below 20 % in French speaking Africa 25 ).To assess the interest of authors in health care accessibility issues, we measured the same with two additional words: "access" and "care".Between 1993 and 2010, their proportion increased from 31 % to 44 % only-while insurance coverage rate remained between 3 and 6 % of the French speaking Africa population (25Séry&Letourmy 2006, p. 204).
-The number of papers with the words "community", "health", "insurance" and "Africa" multiplied by -In fact, public health researchers devoted many papers to describing mutual aid associations but 50 % of the 2010 papers identified with "community", "health", "insurance" and "Africa" did not include the word "evaluation".
This lack of scientific interest in the effect of insurance on access to care contrasts with disturbing facts found in LMIC conditions: -In Colombia, in spite of increased insurance coverage, access to care generally decreased, with exceptions in the subsidized system.-A similar paradox has been observed in Peru where the population who did not consult increased from 50.5 to 56 % between 2007 and 2008.At the same time, its Social Insurance coverage increased (from 26.6 to 44.7 in the population) 29 .
-In Ghana, user fees increased from 9 to 11 % of total health expenditure between 2007 and 2009, in spite of insurance coverage increasing from 0 % in 2003 to 60 % at the end of 2009 30 .
These examples and many others point to the existence of significant non-financial barriers to access to individual health care (at the side of financial ones) and ineffective protection provided by health insurances in LMIC -two facts conspicuously overlooked by the scientific literature and international organizations alike.
Notice that the repetitive, general failures of health insurance schemes in LMIC (see a few data above) are related to their focusing public financing on the poor.At first sight equitable, this principle leads to reducing rich / poor solidarity mechanisms.Targeting public funds to the poor leads middle classes purchasing health insurance on the market, rejecting taxes meant to finance health care and chronically under-financing of public health care for the poor.This focusing-on-lower-decile(s) principle is thus radically opposed to the grain of 'traditional' (since 1945) West-European health care systems.
In conclusion, criteria for assessing the products of health systems researchers paid with public funds should encompass the relevance of knowledge to policy makers and social actors involved in the quest for more equitable access to care, which assumes the capacity to criticize public decisions and to master the scientific tools and concepts needed to explore the determinants of health systems failures (and successes).

III. Reasons for extreme academic prudence
So why has the academic community been so little concerned with access to care, even while it was studying organizations meant to improve it?Basically, we contend, because universal access to individual care was not a priority theme of the main donors to schools of public health: the private sector (and in Latin America, the World Bank).Corporations have the will to orient public research to the best of their interest since they have a legally binding fiduciary dutyto maximize shareholder profits.Since 1953 US law does not prohibit companies anymore from donating money to causes from which they do not directly make profit.Zambrano, A., Ramírez, M., Yepes, F.J., Guerra, J.A., & Rivera, D. 2007. ¿Qué  Corporations also have huge resources to commit to this endeavor.While there are more than 100,000 charity foundations in the USA, one of them, the Bill and Melinda Gates Foundation has an operating budget exceeding WHO funds 31 .Therefore, and because of the reduced government funding of public universities, an increasing share of public research has become privately funded, often exceeding public financing of the same by more than 50% 32 .Accordingly, public universities developed several mechanisms leading their scientists to abide by private interests: 1. Career and promotion, job stability and access to institutional resources became conditioned on the importance of grants attracted by researchers -that is on the extent to which their research accords with the charity foundations' agendas.
2. Careers began depending on publishing, which is measured by bibliometric indicators 33 .High impact biomedical journals are often owned by shareholders investing in health care delivery, pharmaceuticals and health financing.This raises an issue of conflict of interest since these journals increasingly mix health policy papers with biomedical ones.They orient publications and thus research by selecting issues, authors, reviewers and methodologies -the latter being also influenced by those who pay for the studies 34 Professional associations also own scientific journals, which similarly generates a conflict of interest: if self-serving interests are recognized to dispel the myth of self-regulation 35363738 , these interests are likely to bias scientific policies as well.As a result, conflicts of interest in public health research financed by charity foundations controlled by corporate manufacturers of food, tobacco and pharmaceutical goods have been shown to be paramount 31 and with similar conflicts in Health Systems Research and Education (HSR-E).In public universities, private foundations often funded research into marketable devices (e.g.drugs, vaccines, and diagnostics) 41 and participation in 'global health initiatives' -public-private partnerships intended to control diseases and to open outlets for products.Their interests span from jobs to capital return and from trade to consultancy.Besides funding this research, it appears that these private foundations also oriented topic selection in public research, disseminating pro-market messages in public universities, publishing papers, making presentations at scientific meetings, hosting conferences and workshops and paying scientists as spokespersons 31 and by specializing research teams.There are thus reasons to suspect that private interests have managed to hijack publicly led health systems research, resulting in its neglect of medical and health care related knowledge (with the exception of managed care).This has not been done only by orienting priority research topics but also research methodologies.

IV. Epistemological requirements of an independent HSR-E
We now explore the methodological requirements for producing critical theories in HSR-E and contrast them with dominant practices.In public health, human sciences and qualitative research are well known for their ability to orient strategic management and contribute to managerial decision making.They also have a key role in assessing success conditions and determinants of health policies.To this end, Michel Foucault and Pierre Bourdieu proposed concepts particularly relevant to descriptive and interpretative HSR-E, which we now examine.

Foucault and the selection of research topics in the social sciences
Foucault studied the interface of human sciences with biology, economics and philology.Echoing Locke 42 , he explored their intellectual borders, which limit the domain of a science -the space within which the laws and language forms of a science are applicable and operationally effective.To qualify this border, Foucault used the concept of "representation possibility" -episteme, or the historical a priori that grounds knowledge and its discourses and thus represents the condition of its possibility within a particular, historical context.
The theoretical knowledge of HSR-E being partly descriptive (and partly prescriptive) and man being both its object and subject, virtually the entire domain of human sciences can find applications to it and contribute to marking the limits of its operational effectiveness.To the extent that research is publicly funded and HSR-E is an applied science, "relevance" to public interests should be a key criterion for defining its priorities.This requires making explicit the links of a particular theory to decision making and action (to improve the human condition).
The following example illustrates this problem of border definition with an issue key to quality of care: the survival of Hippocratic medicine.How can we convince doctors not to exploit the structural information asymmetry that characterizes their relationships with patients in LMIC settings (and makes health financing markets inefficient by definition 43 )?While health care becomes increasingly commoditized, professional cultures become increasingly materialistic and regulation is not very effective to ease their effects in this context.Since motivation is at least partly an unconsciousness issue, the description of doctors' ethos remains a borderline case study for the theoretical human sciences 44 .
Foucault defined human sciences' paradigms as historically evolving concepts.Man's functions and their limits (norms) are debated at the interface of human sciences and biology (public health addresses death and disease but historically, it has neglected bio-psychosocial suffering).From an economic standpoint, people have needs and desires (e.g.those of patients, professionals, civil servants and investors) -interests and profits -that result in irreducible (social) conflict situations and require (political) rules to tame their effects.From a philological viewpoint, there are systems of signs.Contemporary health policy and administrative speeches, for instance, are by and large scientistic in their expression, making them unquestionable by lay people and limiting the effectiveness of democratic mechanisms.Functions/norms, conflicts/rules and systems/signs together make up the study field of human sciences.Psychology mainly addresses the former, sociology explores conflicts and rules while their interpretation (signifying systems) is an issue of philology (language science) e.g. in literature and myths studies.The role of the concept of sense in philology is to show how a language can be the object of a representation.The (social, political, health) system study, instead stresses how a structure (say, a health system) precedes any sense (say, of patient-centered care, governance, transparency, accountability, etc.).Languages can be represented to the extent that both sign/sense and system are used.

Inter-or multi-disciplinarity? (Edgar Morin)
Economics and sociology, psychology, and philology are permanently intermingled at methodological (a theme further developed by E. Morin) and epistemological levels.In epidemiology, for instance, the study of the combined character of natural (disease) and cultural factors characterizes epidemiological research even when it explores the primary determinants of a disease.
The domain of human sciences applied to HSR-E is not defined by what cannot be explored by biomedical sciences.Rather, to the extent that applied sciences are geared towards decision making and action, HSR-E requests ad hoc combinations of theoretical, practical and empirical knowledge, yielded by blends of managerial, biomedical and human sciences methods.According to Edgar Morin, scientific assessment and representation should be interrather than multi-disciplinary in essence -meaning that instead of mere parallel use of disciplines to address a problem or a question (which is multi-disciplinarity), each discipline should influence the methods used by the other disciplines -which characterizes inter-disciplinarity.As Morin puts it, "The issue is to question the principle of 43 The World Bank.1997.Health, Nutrition, & Population Sector Strategy.The World Bank Group available from: http://go.worldbank.org/1BO4F9LL30. 44Professionals' motivation can be studied by action-research, which is an issue of professional, partly prescriptive knowledge, while the present paper addresses scientific knowledge only.
disciplines which divide up a complex object, made of numerous interrelations, interactions, interferences, complementarities, oppositions between constitutive entities, a principle whereby everybody is prisoner of a particular discipline. 45"

Topics and issues at the interface of HSR and human sciences
Foucault observed that the history of human sciences has been one of changes in the interaction of the three biological, economical and philological binomials.During the 20 th century, he said, human sciences reduced their preoccupation for functions, conflicts and meanings and increased their concern for norms, rules and systems.
In this incipient 21 st century, HSR-E lacks concern for and interpretation of conflicts, human functions and meaning while norms, rules and political structures (systems) are (over-?)emphasized.Biological norms are increasingly used to define rules for taming economic conflicts appearing in the health sector.The MDGs, for instance, conspicuously overlook access to family medicine and hospital care.People's demand for individual health care has been replaced by bureaucratically defined "health needs" ( 6Unger, De Paepe, Sen, & Soors 2010, p. 160)universal access to disease control programs.These 'needs' are a norm in public health speeches and, not by chance, a rule fixed by international trade agreements 46 .
As opposed to what happened in the 70's and 80's, health care services are no longer studied as the product of social, political and economic conflicts.Possibly because public research is increasingly financed by private interests, health care is increasingly studied as an issue of rules justified on alleged health systems functionality 47 .In schools of public health, signifiers and non-signifiers are no longer viewed as characteristics of human behavior.With the closer alignment of bioscience and policy, non-signifiers are progressively overlooked.QALYs and DALYs indicators have become quite popular in scientific publications 48 , although they are not capable of apprehending (biopsychosocial) suffering (and they are not used much in real health planning practice).Specifically, their use has been ideological: to contribute to justifying a bias in health planning of public services towards disease control.
In conclusion, life, need and language are thus mirrored in the world of human sciences by function (e.g.explored by psychology), conflict (e.g.studied by sociology) and meaning (e.g.represented by philology).These categories have corollaries -norms, rules and systems of signs, respectively -which are transparent ('not given') to consciousness 49 (not easily objectifiable) but can be easily represented.This facility, together with the ideological and political functions of conservative HSR, probably explains the long lasting predominance of this public health paradigm.For instance, responsiveness, transparency, and rule of law are values often used in contemporary policy speeches.These terms rubbed off on scientific studies although they were deemed invalid and unverifiable criteria 50 . 45Morin, E. 1979.p.227, Le paradigme perdu: la nature humaine, Paris, France, Editions du Seuil.Lancet, 363, (9402) 83 available from: http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(03)15191-4/fulltext. 47Disease control programs in LMIC public services -probably the largest bureaucracy ever invented by mankind -were not questioned on their economical function (to prevent a subsidized public sector to compete with private interests in the domain of health care delivery) and rarely on their biological mass effectiveness.Other concepts were systematically misused, as is the case of allocative efficiency often mistaken for technical efficiency. 51Only the underlying conflicts (health care privatization, restrictions on public services functions to avoid subsidized competition with the private sector) allow deciphering of this paradox.
Norms, rules and systems represent a border of human sciences.Knowledge and exploration of these borders is an issue of human sciences 52 .However, scientists cannot study norms, rules and systems without concurrently studying functions, conflicts and meaning.Consider the following examples: -The role of mental health services cannot be studied without defining norms in mental health and in behavioral deviance; -Health services management cannot be studied without scrutinizing the underlying social, political, geostrategic and economic conflicts; -And governance and stewardship, for instance, cannot be researched without clarifying the semantic ambiguities and ideological valence of its florid terminologies.
In HSR-E, the absence of such balance not only reveals the traditional but acknowledgedly variable dominance of biomedical sciences upon human sciences.It also reveals the scientists' internalization of dominant policy discourses and the continued dependence of public research in applied sciences to private funds and interests.To detect and limit the consequences of this material dependence, the "Praxeology" concept of Bourdieu opens an interesting avenue to those academics who are keen to remain relatively free to search and teach. V.
Bourdieu's praxeology applied to HSR-E The epistemological architecture of public health schools: a mirror of the social world?
Implicit theories of practice 53 , says Bourdieu, are correlative of neglecting the social conditions that permitted the theory to exist.In the case of ethnology, researchers are interpreters in essence, leading them to privilege communication relations and symbolic exchange relationships.In the 70's, several authors belonging to the French school of ethnology (Robert Jaulin 54 , Georges Balandier 55 , Pierre Clastres 56 and others) questioned the 51 Segall, M.M. 2003.District health systems in a neoliberal world: a review of five key policy areas.International Journal of Health Planning and Management, 18, (Suppl 1) S5-S26 available from: http://www.ncbi.nlm.nih.gov/pubmed/14661938. 52 the sciences which are the last to host the concept of representation, the field of episteme: "On dira donc qu'il y a "science humaine" non pas partout où il est question de l'homme, mais partout où on analyse, dans la dimension propre à l'inconscient, des normes, des règles, des ensembles signifiants qui dévoilent à la conscience les conditions de ses formes et de ses contenus ».Michel Foucault.Philosophie : Anthologie.Collection Folio essais (No 443), Paris: Les Editions Gallimard, 2004 (p.288). 53Examples of theories of practice in public health are those related to the regulation of health care delivery, the process of policy making, the design of disease control programs, the motivation of health care providers with pay-for-performance, etc.They almost always are part of the social practice of a complex set of institutions.
political function of ethno-science.By contrast, questioning of the social conditions and function of public health and medical sciences, however, emerged only occasionally 575859 -perhaps because of their high economic weight of the health sector (up to 16 % of GDP in the USA) hampered truly contextualized research and to prevent journalists' findings to feed descriptive HSR.
How did pro-market health policies and their determinants impact on HSR-E knowledge?In the academic world, these social and political factors made their mark on two sciences that barely interact, namely medicine and public health.Medicine responds mainly to individual demand for care; Public health responds to collective necessities or needs defined by professionals.The quasi absence of interactions between these sciences meant that individual health care delivery did not have to echo preoccupations for collective health and consequently, public health did not have to address individual care delivery -in spite of the wealth of evidence pleading for the opposite (mentioned above).To address collective health problems, the disciplines allied to traditional public health were generally mobilized in a standard pattern.This recurring epistemological architecture had already been described in 1965 60 in terms of the following sequence: professional identity of health professionals and its influence upon motivation; and the management of integrated services networks.
In conclusion, the social conditions of public health theory production deterred many researchers from questioning the traditional epistemological functions in public health, although the underlying policies were shown to be unable to achieve their limited disease control objectives in LMICs.What other route could they have taken?
VI. Acquiring a critical, theoretical knowledge relevant to HSR-E In this section, we examine: why HSR-E needs to take as its object the relations between internalized (linguistic, ideological, behavioral) conditions of social system reproducibility and its (economic, social, institutional and political) structures; why such a study needs a reflexive concern for the possibility conditions of for the research itself, as a surrogate for a quality check.
How can methodological concerns make human sciences amenable to yielding critical concepts relevant to HSR-E and help some scientists evade the determinism described above?Bourdieu proposes a typology of approaches to theoretical knowledge production in human sciences: 1.The phenomenological approach to knowledge Historically, the phenomenological methodologies overcame difficulties linked to the application to human sciences of a positivist approach typifying the natural sciences -which excluded all methods not solely reliant on sensory observations.Phenomenological methods, empirical in essence, acknowledged the validity of psychological experience -and were thus the first in introducing a reflexive concern in science.Notice that phenomenology admits modes of enquiry in sociology but also history and economics.This approach makes explicit the truth of a direct, social experience (interviews, observations), as was the case with early ethnological studies and is now with most epidemiological research into social determinants of health.The phenomenological approach to qualitative research does not require that the conditions making this knowledge possible be made explicit.It does not rule out triangulation of informant sources to improve qualitative data validity.Together these sources are often deemed sufficient to derive knowledge of social process and of the role of actors.Although phenomenology admits modes of enquiry in different disciplines, mono-disciplinarity has often been historically and logically attached to phenomenological approaches to the social sciences.
Bourdieu criticized the validity of direct, social experience on the notion that belief are independent of the will of an individual and are rather collective (cultural, ideological, linguistic) phenomena.Notice that two other French sociologists explored this concept: -In political sciences, Gilles Deleuze and Felix Guattari 63 offered an original concept of pluralism whereby the individual is not conceived as the basis of any social organization.Social subjectivities are above individuals, and they structure and dismantle any collectivity.
-In social psychology, the existence of a collective dimension in perception had also been studied on experimental design by Pierre Moscovici. 64 The objectivist / structuralist approach to knowledge Beliefs (opinions, statements, discourses) as detected in qualitative research are objects of interpretations and explanations that attempt to understand their economic, symbolic, and linguistic environments, which structures the actors' practices and representations.Therefore, objectivist scientists stress the need for distance between researcher and interviewee.Some even plead that a relative a priori ignorance of the environment's features optimizes the researcher's distance and therefore objectivity.This approach addresses the objective relationships which structure praxis and actors' representation of praxis, at the cost of some rupture with the truth of the direct social experience.Only when these underlying relationships are explored, can objectivist knowledge organize the features of the social world.However, the objectivist approach, says Bourdieu, stresses the message structure (for instance the structure of a health policy, of a hospital organization, of a regulation) to the detriment of its function.Speech merely treated as a product of language leads to decoding studies that favor sign structures -the relations they have amongst themselves -over their practical functions (which never vanish).To the contrary, these practical functions (which are the key concerns of public health and HSR -more than of ethnology) are at the core of the praxeological study of health systems and policies, which Bourdieu introduces (see below).
The objectivist analysis denies that symbolic forces and relationships depend on political ones.Such analysis does not recognize that the knowledge of a code (say, of a control procedure) only imperfectly permits understanding the relevant interactions.At the end of the day, it does not oppose the phenomenological analysis in that both establish the postulate that immediate comprehension -an unconscious decoding process -is possible.The How does this apply to HSR-E, e.g. in policy assessment?These collective phenomena materialize: in the social strategy of the interviewee, limiting his/her sincerity, his/her organization rationale, affecting his/her identity, the country or region's degree of freedom of expression, shaping his/her thoughts and expression, and his/her limited access to information.
Finally, the lack of a priori knowledge used in interviews limits the relevance of the interviewer's questions; the researcher's false ingenuity limits any complicity with interviewees; and it permits avoiding important but politically sensitive issues.This explains why so few policy research published in 'scientific' health and medical journals are contextualized and it so often fail to refer to key government and administration features: because it is phenomenological in essence.
objectivist knowledge differs from the phenomenological in the type of envisaged reflexivity.Phenomenologists are compelled to observe their psychology while objectivists limit the domain of validity of their research with assumptions related to the similarity in experience of observer and observed: such similar social and cultural experiences would be a requisite to interpret peoples' representation and their determinants.But, concludes Bourdieu, if one ignores all the other important possibility conditions of any science, one then studies bodies of practice merely as symbolic facts (and is thus liable to personal opinions).
The acquaintance of HSR-E with biomedical sciences led many public health researchers to rely heavily on repertoires of representations.Unfortunately, these hardly capture the ('idiosyncratic') practices of social interaction's symbolism.These repertoires are theories of theories (e.g.theory of good governance, of staff motivation with managed care techniques, disease specific programs selection with cost-effectiveness analysis,…), which closely mirror those used by health bureaucracies worldwide.The material conditions of the study often lead researchers to turn a blind eye to their key political function and neglect other options.Furthermore, predetermined programs of discourses and actions are adapted to some 'cultures' but not to others.The difficulty in translating into Arabic or Asian languages concepts such as 'transparency', 'accountability', 'rule of law', simply reveals that local social hierarchies use different concepts to reproduce themselves ('European values', 'European way of living together'65 , the Chinese government led 'person-centered policy' and 'quest of social harmony', etc.).
In conclusion, the objectivist approach introduced a distance between the observer and the social world but persistently failed to recognize that researchers remain part of it in spite of these efforts.Consequently, they rarely felt compelled to assume the ethical and logical consequences of this belonging, namelythat researchers have a responsibility to: • improve the state and conditions of their study objects; • learn from the successes, failures and conditions of this experience; and • make explicit their underlying philosophical and political position as well as the study conditions (e.g. the researchers' status and funds origin).
Only such a reflective effort would put readers in a position to appraise their topic and methods selection, their results and interpretation.

The praxeological approach to knowledge
Praxeological knowledge thus introduces a rupture with objectivist knowledge by scrutinizing the dialectic relationships between the economic, linguistic, social, institutional, political structures -and the internal, individual features (habitus, representations, ways of doing) that permit institutions such as health systems to reproduce themselves.As seen above, the two-sided, interwoven nature of these relationships often justifies the use of interdisciplinary methods to study them.
Praxeology thus aims at studying, beyond statistics, one such relationship: the production modes of practices, the dialectic of practices and their conditions.To avert any structure realism and reification of concepts such as social classes and parenthood structures, says Bourdieu, there is a need to move from statistical regularity to the production principle of the observed order.Unlike phenomenological approaches, both objectivist and praxeological knowledge view the object of science as possibly being captured by a construct, against the evidence of common sense and pre-established classifications (and official definitions).For instance, in the health sector, the public health discourse has a practical impact on health care practice, which researchers should deconstruct (in the terms of Derrida) instead of elaborating it.The question is: how?
In HSR-E, 'praxeologist' scientists can access primary and secondary data on the impact (output and outcome) of health policies on health care delivery and management, and on peoples' health status and access to care.They can also access discourses of decision makers, health professionals and bureaucrats with interviews and focus groups.The key contribution of the praxeology approach to such qualitative research is that it assumes that somehow, social, political, economic and linguistic structures featuring the prevailing social hierarchy are mirrored in structures internalized by people.These internal features are the conditions of social structures and hierarchy reproducibility.
Bourdieu names these internal features: habitus, ways of doing and being, and representations.Social structures determine these internal features with education, information, propaganda, advertisement, organized in markets which all convey a symbolic capital.These 'substrates' convey discourses which were deciphered by Guy Debord and Roland Barthes66 , who both studied the implicit objects discourses and their social functions.Likewise, Jean Baudrillard studied how consumption objects and their exhibition contribute to structure modern social relationships.This is how objects acquire a social symbolic function largely exceeding their usage value67 .
In conclusion, an insight into the actors' representations, habitus and ways of doing can shed a light on how the social structure (e.g.health systems) tends to reproduce itself -with which symbolic, organizational and cultural mechanisms.This study object is key to produce a critical theory on health systems while adopting a praxeological approach to qualitative research.However, this is not sufficient per se: quantitative and qualitative technical evidences on health care delivery, services and systems are needed to cross check and prompt opinions.

VII. Contributions of biomedical and human sciences to HSR-E
"Feuerbach does not see that the "religious feeling" (public health, medicine) is a social product and that the individual (the health system, the individual patient) under scrutiny belong(s) to a specific social type."K. Marx, 7 th thesis on Feuerbach, German Ideology 68 (parentheses are of the authors).
Deeds of actors import as much as thoughts to 'praxeologist' researchers, as well as their mutual, dialectic adjustment.To detect deeds, in HSR-E, quantitative studies of -and nested qualitative research into -disease and care distribution, allow assessment of their effects, while singling out undesirable effects, contrasts between In HSR-E, we contend that health systems should be studied as social constructions.Their representations (which are thought objects) should not, therefore, automatically be treated as the cause of praxis.Objective structures (e.g. the system's productivity, its efficiency and people's access to quality health care) should be determined (and possibly measured) before studying the mechanisms linking them to praxis (and praxis representation) 75 .Such nested biomedical (and other quantitative) studies will yield paradoxical observations ('problem based research'), which are a prerequisite for exploration of the social, political and economic determinants of health policies and systems, a precondition that allows questioning of official speeches aimed at justifying policies, that is, self-apologetic discourses based on alleged functionality and contributing thereby to the system reproducibility.For instance, since it has been shown that Costa Rica spent nine times less on health care per capita than the USA and scored as well on access to quality care and on epidemiological and demographic indicators 76 , economic and political factors (and not alleged health benefits) can now be seen as key determinants of the promotion of the US health system model by multilateral agencies 77 .
Comparisons of health policies which disentangle their effects from other health systems determinants permit plausible representation of the regularity conditions of poor policies design -their determinants.This is why HSR-E, like any applied science, cannot be treated as being exclusively a human science.

VIII. Conclusion
The relevance of the knowledge acquired in the realm of HSR-E lies in its subsequent use e.g.equipping all social actors aiming at equity in health in policy decisions, from mutual aid associations to a MoH general director, with relevant and good quality knowledge.HSR-E knowledge can roughly be categorized as scientific or professional, with an accent respectively put on description and prescription.While this paper did not address the latter, professional knowledge should also be an object of research and teaching in public health academic institutions 78 .Unlike the WHO 'building blocks', HSR-E should treat medical care and access as key components for strengthening health systems as they address the key determinant of health status and are powerful political pacifiers.Therefore, access as defined here could usefully become a new MDG in health.
Researchers in publicly funded institutions should make explicit how their HSR are linked to decision making and action to make them, in Popperian terms, 'prone to validation'.Those making theories of practices should specify which groups authorize the language they use and give it authority if they do not want to merely contribute to the existing social and symbolic order.In other words, one way of explaining how a study possibly contributes to decision making is by requiring that it makes explicit how it attempts to:  .2006.099598v1. 77 In this kind of study, the technical challenge consists in attributing epidemiological and demographical features of a country to the health services characteristics and to disentangle them from the impact of other determinants -confounding factors. 78Notice that while Walter Benjamin criticized the exclusion of science from those faculties where professional knowledge was taught (W.Benjamin.Oeuvres 1. La vie des étudiants.Collection Folio Essais.Paris: Les Editions Gallimard, 2000, p. 126), symmetrically, it is regrettable that professional knowledge has often been barred from public health faculties and publications.
improve decision making; reveal the social functions of decision makers discourse, deeds, policies and health system structureswhile contrasting them with field achievements; understand its host social structure and how it determines health systems features and actors behaviors.
By contrast, phenomenological knowledge does not question the research possibility conditions of the experience it attempts to explain.Therefore, it universalizes an experience of the social world, while producing a doxa, a corpus of concepts associated with a given set of social and economic conditions (e.g. the expansion of health markets in LMIC) and their reproduction conditions (e.g.pro-market health policies and their "scientific" ground).If one considers that the opinion addresses what is explicitly questionable, says Bourdieu, than it is in the interest of the oppressed social classes to shrink the doxa domain.In HSR-E, as in other applied sciences, phenomenological knowledge has a conservative function in that it prevents questioning the doxa79 -the dominant policy and its methodological alibis.
Research methodologies in human and consequently in applied sciences do thus have a political valence.If HSR-E were unambiguously recognized as an applied science with a human sciences component, and if the related research were entitled to question the established order, scientists would examine the relationships between culture and language, between social order and policy discourse, between politics and health system evaluation.It is not and it does not, or does so only rarely.The praxeological approach can also fruitfully be applied to the analysis of action research outputs in health policies and management.It requires that health systems and policies be explored in the light of their outputs and outcomes, and that the dialectic between economic and political actors' roles, relationships and health decisions be explored in the light of their demonstrated weaknesses.The most important feature in these evaluations is the link between the problem study -the conditions structure (in HSR-E, an issue of health sciences, management and policy) and the study of the praxis (an issue of sociology, history and political sciences).
Merely decoding policy substance and structures does not permit analysis of the complex interactions between actors (e.g. between a regulation institution and health professionals).Far from phenomenological traditions, critical researchers should thus consider hypotheses generated by the ex-ante context knowledge and consistent with a critical epistemology of human sciences.If researchers aim at contributing to the debate on human rights and essential needs (access to care, teaching, food, water, air and justice), they should make explicit the possibility conditions of their studies -e.g.their own system of values, what is known of the prevailing 'epistémé', and their research conditions.The researcher's commitment to introduce change in the social world and to learn from his/her strategies and conditions of successes and failures are essential to the praxeological approach to HSR.In public universities, research independence should be conceived as an academic goal and an evaluation standard.
48but not in real world policies, which reveals the mere ideological function of these indicators.49Foucault, M. 2004.Philosophie : anthologie, Collection Folio essais (No 443) ed.Les Editions Gallimard.