DE

BA

TE

DE

BA

TE

1701

Anthropology and Epidemiology:learning epistemological lessons through a collaborative venture

Antropologia e Epidemiologia: aprendendo lições epistemológicasatravés de um empreendimento colaborativo

Dominique Pareja Béhague 1,2

Helen Gonçalves 2

Cesar Gomes Victora 1,2

1 Department ofEpidemiology andPopulation Health, LondonSchool of Hygiene andTropical Medicine, KeppelStreet, London, WC1E7HT,United Kingdom.Dominique.Behague@lshtm.ac.uk2 Department of SocialMedicine, Federal Universityof Pelotas, Pelotas, RS,Brazil.

Abstract Collaboration between anthropology andepidemiology has a long and tumultuous history.Based on empirical examples, this paper describesa number of epistemological lessons we have learnedthrough our experience of cross-disciplinary co-llaboration. Although critical of both mainstreamepidemiology and medical anthropology, ouranalysis focuses on the implications of addressingeach discipline’s main epistemological differences,while addressing the goal of adopting a broadersocial approach to health improvement. We be-lieve it is important to push the boundaries of re-search collaborations from the more standard formsof “multidisciplinarity,” to the adoption of theo-retically imbued “interdisciplinarity.” The morewe challenge epistemological limitations and mod-ify ways of knowing, the more we will be able toprovide in-depth explanations for the emergenceof disease-patterns and thus, to problem-solve. Inour experience, both institutional support and theadoption of a relativistic attitude are necessaryconditions for sustained theoretical interdiscipli-narity. Until researchers acknowledge that metho-dology is merely a human-designed tool to inter-pret reality, unnecessary methodological hyper-spe-cialization will continue to alienate one field ofknowledge from the other.Key words Anthropology, Epidemiology, Multi-disciplinary, Epistemology

Resumo A colaboração entre Antropologia e Epi-demiologia tem uma longa e confusa história. Ba-seado em exemplos empíricos, o artigo descreve li-ções epistemológicas aprendidas ao longo da nossatrajetória de colaboração transdisciplinar. Apesarde críticos da Epidemiologia e da AntropologiaMédica tradicional e crermos que ambas possuemo objetivo de adotar uma abordagem social à pro-moção da saúde populacional, enfocamos as im-plicações de confrontar suas principais diferençasepistemológicas. É importante avaliar os limitesda colaboração padrão (“multidisciplinaridade”)e destacar a relevância de adotar a “interdiscipli-naridade.” Quanto mais profissionais de diversasdisciplinas convergir e modificar seus modos desaber e desafiar suas posturas epistemológicas, maispoderão dar explicações aprofundadas aos padrõesde doenças e solucionar problemas concretos. Emnossa experiência, ambos o suporte institucional ea adoção de uma abordagem realista das limita-ções epistemológicas das disciplinas são necessári-os à manutenção do grupo e interdisciplinaridadeteórica. Até que se perceba que a metodologia éuma ferramenta humana criada para interpretara realidade, a desnecessária hiperespecializaçãometodológica das disciplinas continuará a alienarum campo do conhecimento do outro.Palavras-chave Antropologia, Antropologia Mé-dica, Epidemiologia, Multidisciplinar, Epistemo-logia

1702B

éhag

ue,

D. P

. et a

l.

Introduction

As applied medical anthropology has become amore central discipline in public health over thelast 30 years, challenges and debates relating tocross-disciplinary collaboration, particularly withepidemiologists, have come to the fore1 . The de-mand for anthropological input in public healthemerged less from academic epidemiology andmore from the limitations that professionals inleading organizations, such as the World HeathOrganization and bilateral donor agencies, expe-rienced when relying solely on epidemiology forimproving program development, evaluation andimplementation2. As a result, a certain templatefor expected ways of collaborating developed.Most commonly, applied epidemiological projectsnow generally incorporate sub-studies based onuse of “qualitative methods,” usually in a so-called“formative phase,” to be carried out by anthro-pologists. In this template, the anthropologist’srole has consisted largely of helping epidemiolo-gists adapt standardized measurement tools tospecific contexts, providing a descriptive narra-tive of patients’ subjective experiences, or explain-ing the reasons for the failure of a particular pro-grammatic initiative.

While this form of collaboration is generallydeemed useful, in a general sense, the inclusion ofqualitative components in public health researchinitially proved difficult. The merits of includingqualitative studies have not always been widelyendorsed, and more than once, the authors ofthis paper noted comments made by colleaguesthat multidisciplinarity has been engaged in moreout of lip-service to donors than real analyticalneed. For many anthropologists, in turn, the de-velopment of applied medical anthropology with-in the larger public health and epidemiologicalframework belittled the potential of anthropo-logical contributions, a view that polarized theanthropological community. While many anthro-pologists embraced the challenge of working inpublic health, others rightly highlighted that ap-plied anthropology has more often than not beenused in a superficial way, devoid of theory andreduced to an over-simplified methods “tool-kit”3,4. Fundamental and insurmountable episte-mological differences between the two disciplineshave rarely been acknowledged outside academicsettings, to the extent that collaboration betweenthe two disciplines has occurred in a parallel, rath-er than cross-fertilizing, fashion5.

Under the pressure of study funders, the au-thors of this paper began to collaborate in multi-

disciplinary projects in the beginning of the 1990s,at a time when the inclusion of anthropology andother disciplines in public health had become anentrenched expected norm, even while the debatesdescribed above had also reached a disharmoni-ous peak. In many regards, we were directly influ-enced by these debates and certainly experienced adegree of disharmony. From the epidemiologist’sperspective, anthropological works were found tobe verbose, excessively anecdotal, inappropriatelybased on small sample sizes selected according toconvenience rather than random allocation. Insum, anthropology was seen to besubjective andun-scientific. The anthropologists in turn, oftenfelt frustration with epidemiology’s biological bias,reductionism, tendency to homogenize and sim-plify reality, lack of theoretical sophistication, andblack-boxing of the culture concept, referred toonly when needing to explain unexpected or atyp-ical epidemiological findings.

With time and many discussions, however,we discovered that these perceptions of each oth-ers’ approaches were inaccurate depictions of ouractual epistemological positions. Often, they re-sulted from a misunderstanding of what eachother’s discipline was seeking to achieve, emerg-ing from too much attention technical differenc-es relating to methodology and not enough at-tention to conceptual and analytical similarities.Indeed, critics within both anthropology andepidemiology have highlighted the futility of ap-plying methods without a strong theoreticalframework, arguing that the predominant focuson methodological specialization in both fieldshas inhibited the use of shared conceptual mod-els and theoretical interests6,7. A recent study ofprofessionals engaged in multidisciplinary pub-lic health research demonstrates the emergenceof parallel debates occurring in both anthropol-ogy and public health which call for more criticalexamination of biomedicine and its dominantrole in conceptualization of public health. Thesedevelopments represent potential commonground that could be capitalized upon in movingtowards greater conceptual – rather than meth-odological - collaboration between disciplines8.

Using empirical examples from our work, thispaper aims to describe the epistemological les-sons we have learned through our collaboration.Our analysis focuses on the methodological andtheoretical implications of actively addressingassumptions regarding our disciplines’ main dif-ferences, while focusing on the mutual goal ofadopting a broader social approach to popula-tion health improvement. Although the analysis

1703C

iência & Saúde C

oletiva, 13(6):1701-1710, 2008

presented below is at times critical of both main-stream epidemiology and medical anthropology,such critiques are not gratuitous but aimed ulti-mately at enriching multidisciplinary collabora-tion for the purposes of improving public healthpractice. As our paper will demonstrate, we be-lieve it is important to push the boundaries ofresearch collaborations from the more standardforms of “multidisciplinarity,” or the processwhereby professionals of different disciplinesparticipate in a single project, to “interdiscipli-narity,” or the process where by professionalsfrom distinct disciplines work together to gener-ate novel concepts and integrate different levelsand forms of explanation9.

Empirical examples of the collaboration:lessons learned

Using examples from our own research, this sec-tion will describe a number of areas where wehave found convergence between our respectivedisciplines to be particularly fruitful and focusedon theoretical, rather than simply methodologi-cal, exchange. Underlying all the forms of ex-change we highlight below is an iterative process,whereby lines of inquiry, methodological devel-opments, conceptual models, and the analysis andinterpretation of data used in both disciplinesfeed into one-another.

Questionnaire design and improvingacceptability of epidemiological surveys

A traditional form of collaboration betweenthe two disciplines consists of using ethnographicinsight to better develop questionnaires for quan-titative surveys, primarily by improving the word-ing and social suitability of questions as dictatedby formative ethnographic research exploring lo-cal taxonomies and illness categories10. For ex-ample, our research in Northeast Brazil showedthat the word “canseira” appropriately capturedthe “rapid or difficult breathing” concept that char-acterizes pneumonia in small children with a res-piratory infection. Incorporation of this term inthe questionnaire – instead of the more complexexpression “respiração rápida ou difícil” — helpedimprove its validity. Throughout our collabora-tion, we also discovered that anthropological in-sight can also increase local understanding of theobjectives of the research, as well as respondent’sacceptability and compliance with the surveyors’requests.

In our work, we took this form of collabora-tion further, and used it to refine our theoreticaland interpretive understanding of phenomenonin question by using ethnographic insight to in-troduce unexpected questions into our researchproposals. Frequently, for example, we have de-veloped new questions to be used in our quanti-tative designs to help explain larger well-knownepidemiological associations. For example, stud-ies of teen pregnancy have frequently found thatif the mothers of young girls were themselvesteen-mothers, their children are more likely tothemselves become pregnant as teens. Using anethnographic approach, we explored the role ofthe family further and in a more detailed fash-ion, and found that some parents actively “pres-sure” their children to engage in serious roman-tic relationships at an early age. In a subsequentcase-control study, we then included this phe-nomenon as a quantitative measure and foundthat young girls who had experienced this sort ofpressure, as subjectively described by them, weremore likely to become pregnant as teens11. Ourethnographic research also suggested that if a girl’smother had children by more than one father,the risk of teenage pregnancy was increased. Thishypothesis, which was unlikely to have arisen fromepidemiological investigation, was confirmed inthe quantitative analyses of the whole sample12.

Without such an iterative process in whichone discipline influences the thematic content ofthe other, we are unlikely to have been able toadvance our theoretical and explanatory under-standing of particular phenomenon in question.However, in our experience, cross-disciplinaryfertilization did not simply occur at the level ofinfluencing or improving the content and mechan-ics of the research process. Rather, as the follow-ing sections will demonstrate, we found that the-oretical interdisciplinarity was more likely to un-fold if we actively challenged each disciplines’ epis-temological assumptions and limitations.

Use of epidemiological resultsto frame anthropological studies

Medical anthropologists often focus their re-search on describing “discourses” and “world-views” that develop around a particular concept,disease or widespread phenomenon. Such stud-ies are often based implicitly on the underlyingaim of representing the “voices” of marginalizedsub-populations, and thus, have tended to pro-duce descriptive results that aim to explore local-ly salient cultural “constructs.” While interesting

1704B

éhag

ue,

D. P

. et a

l.

and providing a useful starting point, these stud-ies can prove to be analytically direction-less, asthey do not tend to be structured around a spe-cific set of focused questions. Also, because epi-demiology has a strong biomedical bias, and be-cause biomedicine is typically interventionist,many epidemiologists have little patience withwhat they see as unfocussed research that willnot lead to concrete public health improvement.

In response to this critique, we found thatdesigning an anthropological study around ques-tions emerging from epidemiological results addsa qualitatively different kind of analytical slant toour ethnographic approach. While applied an-thropology has not routinely engaged in the useof comparison groups for the purposes of ex-plaining the reasons for health phenomena (andanthropologists’ attempts to do so have oftenbeen brandished “reductionistic” by their peers),we found that working alongside epidemiologistspushed us towards focused exploration and ex-planation. To better address explanation throughthe use of comparison groups, we increased ourethnographic sample sizes and used randomlyselected sub-samples from the larger epidemio-logical studies within which we conducted ourresearch (specifically, the 1993 and 1982 cohorts).Random stratified sampling was deemed neces-sary not to conduct probabilistic analyses, but tolearn about the range of experiences among whatis in Brazil, a socially and economically highlyheterogeneous population and through this, toidentify locally salient subgroups for analyticalcomparison. This sampling scheme also led tothe inclusion of both introverted and sociallysheltered participants, as well as those who spentvery little time in or near their homes, somethingthat certainly a convenience sample, which tendsto favor extroverted informants, would not havecaptured13-16.

Using a comparative analytical frameworkfacilitated by this methodological approach, oneof our studies, for example, focused on trying toexplain the reasons for Brazil’s world-renownhigh cesarean section rates, as found in epidemi-ological studies, by comparing subgroups ofwomen experiencing and actively seeking both c-sections and normal births17,18. Similarly, anoth-er study explored why certain subcategories ofmothers identified in epidemiological studies –namely, white women and those with male in-fants19 – wean their children substantially soonerthan other women. This study revealed an unex-pected set of influences, emerging from the med-ical establishment’s use of growth charts, that

push women to wean their children early becauseof heightened concerns that their infants werenot growing well20.

The concurrent use of comparison groups ininter-linked epidemiological and anthropologi-cal studies has proved particularly useful. Forexample, one study on the relationship betweenteen pregnancy and both employment and edu-cation used an epidemiological case-control studyof teen pregnancy by 18-19 years of age, togetherwith an allied ethnographic study using the samecomparative groups. Ethnographic results dem-onstrate that youth who became pregnant hadhad school difficulties and failure before fallingpregnant. In addition, young girls were found tovalue forging a lasting bond with their partners,for this also provided them with social statusand upward mobility, without having to submitto the rules of educational institutions11,21. Interms of practical implications, these analysessuggested that it is not merely the lack of accessto or information on contraception that leads tohigh rates of teen pregnancy, pointing to the needto take the broader sociocultural dimensions ofpregnancy in youth into account. Even if sexualeducation were effective, this is usually providedbetween the 5th and 8th grades, and our studyshowed that girls who are likely to get pregnantare either dropping out before reaching this gradeor are retained in lower grades due to repeatedschool failure. Therefore, targeting sexual educa-tion classes by age rather than grade would defi-nitely make more sense12.

Use of ethnography for the interpretationof epidemiological results

Epidemiologists are becoming more and moreaware of their discipline’s analytical and inter-pretive limitations22. Although epidemiology ismeant to be rigorously based in hypothesis test-ing, whereby the associations to be tested arebased on clearly postulated and plausible hypoth-eses derived either from the literature or clinicalknowledge, this is not always the case. Rather,epidemiologists often test multiple hypotheseswithout being able to fully explain the reasonsfor the association in question. This is particu-larly so as epidemiology moves away from study-ing only biological phenomenon and towardsexploring societal patterns and contexst, wherecausal pathways are likely to be longer, morecomplex, diverse, and even cyclical23.

In this regard, the use of ethnography for theinterpretive enrichment of epidemiological results

1705C

iência & Saúde C

oletiva, 13(6):1701-1710, 2008

has proved to be fruitful form of interdiscipli-nary cross-fertilization. Unexpected epidemiolog-ical findings particularly benefit from additionalanthropological insight when it come to analyz-ing subsidiary hypotheses and conducting sub-group analysis. In a recent study, for example, weused ethnography to explore and explain possi-ble reasons for the associations establishedthrough statistical analysis. Looking specificallyat quantitative results from two studies, one ondeterminants of mental morbidity and the otheron age of sexual initiation, we used ethnographicdata to elucidate the way that statistical associa-tions are comprised of multiple pathways of in-fluence that correspond to the unique experienc-es of specific subgroups. In exploring these path-ways, we highlight the importance of an addi-tional set of mediating factors that account forepidemiological results relating to both types ofoutcomes; these include the awareness and expe-rience of inequities, young men and women’s re-actions to the role of violence in everyday life,traumatic life events, increasing social isolationand introversion as a response to life’s difficul-ties, and differing approaches towards socio-psy-chological maturation24. Although these factorsare difficult, if not impossible, to capture in aquantitative survey, they represent key aspects thatshould be included in public health initiativesaiming to mental health. In many ways, this goescontrary to the norm in public health, which tendsto structure interventions according to quantita-tive indicators that have been “proven” to be caus-ally salient in epidemiological studies, often with-out explicitly exploring the underlying phenom-enon such indicators are likely to indicate.

Another study similarly explored epidemio-logically-established gender differences in physi-cal activity in the 1993 cohort, which showed thatgirls are more sedentary than boys. In-depth eth-nographic research found that for a number ofsocial and cultural reasons, the frequency withwhich young men socialised outside the schooland home settings was higher than for girls. Thesebehaviours were subsequently asked about in thequantitative survey, and analyses shows that thisforms of sociability was associated with a seden-tary lifestyle. Further ethnographic work enrichedour explanatory understanding of the reasonsfor this gender difference. Young men are encour-aged by their parents to engage in a number ofphysical activities outside the home in order todevelop their masculinity, maturity, and identity.In contrast, young girls’ behaviour are activelycontrolled, to the extent that they are kept inside

the home as a way of protecting them against themany dangers that they are perceived to be par-ticularly at risk of experiencing, including physi-cal violence25.

Hypothesis raising, sub-group analysisand effect modification

The ethnographer’s ability to discern differ-ences and patterns according to subgroups rep-resents an underutilized yet important area ofconvergence with epidemiology. In one of ouranthropological studies, we used both qualita-tive and quantitative analysis to explore how somewomen living in shantytowns resist the negativedepictions widely made of their social class stand-ing by actively rejecting antenatal care providedin their local primary health care centre. Thesewomen considered such services to be a poorsubstitute for what the wealthy take for granted.Being particularly attuned to the values of up-ward mobility, several of these women investedtheir household’s scarce resources in travelling topublic primary level facilities situated in richerneighbourhoods or tertiary level facilities, and attimes, in paying for private sector care26. In manyways, this study lay the theoretical ground-workfor distinguishing between subgroups of shanty-town dwellers according to key attitudes relatingto economic inequities, upward mobility, andnormative society.

The ethnographic focus on discerning socialpatterns according to subgroups also holds greatpotential for the epidemiological exploration ofeffect modification. A recent analysis using eth-nographic insight, for example, found that thestatistical association between early teen pregnancyand mental morbidity in adulthood not only holdsafter controlling for confounders, but is modifiedby social class, such that the negative impact ofpregnancy on mental morbidity is significantlymore pronounced amongst the poor than the rich.The use of ethnographic case-studies in this anal-ysis enabled us to clearly explore the mechanismsthat account for these epidemiological findings.The ethnographic study found that the associa-tion between pregnancy and mental morbidity ismore pronounced amongst a subgroup of poorwomen who feel marginalized from mainstreamsociety. These women are highly politicised andparticularly sensitive to the social stigma associ-ated with teen pregnancy in poor youth. Becauseof this, some reject what they identify as upperclass values, which includes the view that teen-pregnancy should be avoided. For these girls, teen

1706B

éhag

ue,

D. P

. et a

l.

pregnancy represents a desired state that reaffirmstheir pride as members of the working class. Evenso, these girls suffer considerable psychologicalstrain ensuing from the social prejudice that theirreproductive decisions stimulate27.

Similarly, another study on the relationshipbetween pacifier use and breastfeeding durationused focused comparative ethnography to guideepidemiological subgroup analysis and consid-eration of effect modifiers. Ethnographic insightproved fundamental to teasing out whether ornot pacifier use was causality related to breast-feeding duration, and to discern in which sub-groups of women the relationship proved to bemore strongly associated28. Although the ethno-graphic study was based on a small sample, thesocial patterning it uncovered guided the epide-miological identification of effect modifiers, bring-ing a level of analytical nuance to the epidemio-logical analysis that would have otherwise goneunnoticed. In this cases, it is particularly interest-ing to note that – whereas epidemiological stud-ies alone showed a strong association betweenpacifier use and short breastfeeding duration –the combination with ethnographic analyses sug-gested that association was not causal. After thisstudy was published, a randomized controlledtrial supported the lack of a causal association29.

Contrary to common depictions of the limi-tations of ethnographies based on small sample-sizes, these studies show the power that in-depthanthropological analysis has for discerning pat-terns within a small sample size, which can laterbe confirmed in quantitative survey work. In ourwork, we have found that a remarkably high pro-portion of the effect modifications (or interac-tions) suggested by small ethnographic studiesare likely to be confirmed as statistically signifi-cant in epidemiological analyses.

These experiences demonstrate the impor-tance of having in-depth knowledge of local con-ditions, practices and realities for improving theinterpretive caliber of epidemiological research.Indeed, a growing concern amongst social epide-miologists no longer working within a biologicalconceptual framework is the relative lack of a rig-orous conceptual basis upon which epidemio-logical hypotheses and analyses are conducted.Related to this is the need some social epidemiol-ogists identify to ground their work more fullywithin local contextual knowledge7,30-32.

Discussion

Our experiences with the above collaborative ven-tures has not always been easy, nor has it devel-oped without some degree of professional risk.While a template of sorts was developed in earlydays for the combined use of qualitative andquantitative methods, those seeking to engage inconceptual interdisciplinarity have been requiredto chart new ground. This not only brings with ita number of communicative difficulties, it holdsimportant implications for publication. Althougha number of epidemiological and public healthjournals now accept and encourage publicationof studies using qualitative components, this iscertainly a minority occurrence, and still heavilyfocused on the addition of “qualitative methods,”rather than true disciplinary theoretical exchange.Furthermore, many public health journal requirestudies reporting on ethnographic findings tostarkly separate a “neutral” description of resultsfrom their discussion, even though this is anti-thetical to the analytical and interpretive frame-works most often required in anthropology. Jour-nals publishing in medical and social anthropol-ogy, in turn, have as of yet failed to fully explorethe benefits of considering anthropological stud-ies that include quantitative analyses. Remark-ably few journals are rigorous on both accounts,to the extent that in our experience, reviewers’comments can often be quite polarized, depend-ing on their disciplinary orientation. In an amus-ing albeit somewhat offensive review, one of ourpaper combining ethnographic and epidemio-logical results was described as follows “this ma-terial is thrown together as a Spanish paella inthat ‘anything that ever swum the seas’ can go init: it is tasty; but no one knows what is in it or itsnutritional value.”

To remedy this constraint, we believe greaterattention should be given to the conceptual andtheoretical bases for interdisciplinary exchange,a type of exchange that requires an explicit con-sideration of the epistemological boundaries ofeach of our respective disciplines. In our experi-ence, engaging with substantive analytical ques-tions centered our discussion around conceptualmodels, and charting the hierarchical pathwaysof influence from one set of social determinantsto various health outcomes in question33. Episte-mological convergence was facilitated through athree specific foci which were in actuality facili-tated, rather than inhibited, by the “applied” na-ture of our work.

1707C

iência & Saúde C

oletiva, 13(6):1701-1710, 2008

First, the need in public health is to not sim-ply describe phenomena, but to explain the rea-sons for their appearance and to propose waysto change them. This requirement of developingresearch questions within a problem-solvingframework demands a more sophisticated andexplicit theoretical orientation, one that is gearedtowards generating understandings of causalityand mechanisms of change. To fulfill this objec-tive, both methodological and epistemologicalmodifications are required34, to which both epi-demiology and anthropology have the potentialto contribute. In our view, so central is this typeof interdisciplinary cross-fertilization that mostof our studies – even those headed by anthro-pologists — now comprise both an epidemio-logical and anthropological approach.

Second, while discussions around methodsoften constituted the starting-point of our col-laboration, our mutual interest in explanationlead us to focus less on methodological refine-ment and specialization, and more on challengesin the interpretation of data and meaning attrib-uted to our analytical conclusions. As some au-thors have claimed, debates around the limita-tions of statistical measurements have severelyrestricted theoretical developments. Authors cri-tique the excessive amount of attention that isoften given to developing more specific and so-phisticated measurement techniques at the ex-pense of using good but simple “summary mea-sures” to advance theoretical premises and hy-potheses35,36. As Frohlich et al. argue, the funda-mental barrier to better exploring the relation-ship between structure, local context and ill-healthis the dominance of black box “risk factor” epide-miology that couches theoretical limitations un-der methodological sophistication37. Similar cri-tiques have been leveled at applied anthropolo-gists, who have developed highly sophisticatedmethodological how-to manuals for conductingformative anthropological research within spe-cific disease programs, often putting forth intri-cate interviewing techniques (e.g. pile sorting orranking) that in fact do not originally belong toanthropology and that do little to advance un-derstandings of social change4.

Third, as the focus moves away from describ-ing the differences between the two disciplinessolely in terms of methods, unhelpful dichoto-mies are deconstructed. As some authors havepointed out, stereotyped dichotomies often putforth in debates on interdisciplinary collabora-tion – including deductive-inductive, natural-ar-tificial, specific-generalisable – are not necessari-

ly (or simplistically) determined by methodsused38,39. Hammersely, for example, has shownhow in-depth case study methods and large-scalemultivariate survey research can – depending onhow they are used — share the similar underly-ing aim of developing a conceptual model on howvariables are related, taking time and place intoconsideration40. In other words, open-ended qual-itative methods can be just as reductionistic asquantitative surveys, and a cross-sectional sur-vey can be equally as inductive as participant-observation.

Conclusion

Our concern with highlighting the interdiscipli-nary basis upon which epistemological assump-tions can be challenged and collaboration im-proved is not an academic exercise, but one thathas significant consequences for public health re-search and practice. The more disciplines can con-verge and modify standardized ways of knowing,the more they will be able to provide in-depth andcontextually sensitive explanations for the emer-gence of disease-patterns and thus, to problem-solve.

Despite many advances in multidisciplinarycollaboration, the relationship between epidemi-ology and anthropology has yet to develop fully.Today, there are a great number of useful publi-cations on how anthropology can better contrib-ute to epidemiology, than vice versa1. This prob-ably reflects the excessive focus that exists on theexchange of methods, the relative subordinateposition of anthropology to epidemiology with-in public health, and the subsequent need an-thropologists have to demonstrate and provetheir discipline’s relevance. To modify this andother power balances, many questions regard-ing cross-disciplinary fertilization still need an-swering. What contributions can epidemiologystill make to anthropology? How would an an-thropology influenced by epidemiology be dif-ferent from an epidemiology influenced by an-thropology? What professional and institutionalconditions would be necessary for further devel-oping a conceptual and interdisciplinary – ratherthan simply multidisciplinary — approach toresearching health problems?

In our experience, the importance of institu-tional support for in-depth and sustained inter-disciplinary collaboration cannot be under-esti-mated. At both the Department of Social Medi-cine (DSM) at the Federal University of Pelotas

1708B

éhag

ue,

D. P

. et a

l.

and the London School of Hygiene and TropicalMedicine (LSHTM) (two institutions that haveinfluenced each-other positively with regards toour own interdisciplinary collaboration), signifi-cant efforts have been made to include and main-tain marginal disciplines such as anthropology.At the DSM, the only opening for a junior facultyposition in several years was allocated to an an-thropologist, quite a development in a depart-ment hitherto restricted to medical doctors andepidemiologists. At the LSHTM, in turn, the num-bers of anthropologists who have joined and re-mained at the school has quadrupled since thelate 1980s.

In addition to institutional factors, a funda-mental quality contributing to the success of ourcollaboration has been the adoption of a relativ-istic and open stance regarding the epistemolog-

ical limitations of both disciplines. In our experi-ence, such relativism initially manifested itself ina mutual interest in devoting considerable timeto learning about the methods and terminologyof each-others’ disciplines, but with time, thislearning process quickly lead to the developmentof a fruitful and yet critical perspective on thelimits of disciplinary specialization for generat-ing new knowledge. As Van der Geest has argued,disciplinary specialization and “ethnocentrism” isat the core of inhibited cross-disciplinary re-search41. Until each discipline demonstrates great-er humility and realizes that disciplines are mere-ly humanly-designed tools to study and inter-pret and explain reality, unnecessary hyper-spe-cialization will continue to alienate one field ofknowledge from another.

Collaboration

DP Béhague conducted the literature search andwrote the paper. H Gonçalves commented on adraft of the paper, and wrote certain sub-sectionsof the paper. CG Victora commented on a draftof the paper, and edited sections of the paper.

1709C

iência & Saúde C

oletiva, 13(6):1701-1710, 2008

Acknowledgements

The research upon which our collaboration hasdeveloped has been supported by the EuropeanUnion, the National Program for Centers of Ex-cellence - PRONEX (Brazil), Conselho Nacionalde Desenvolvimento Científico e Tecnológico(CNPq), and the Ministry of Health (Brazil), thePanAmerican Health Organization, the WorldHealth Organization, Rio Grande do Sul StateResearch Foundation (FAPERGS), and TheWellcome Trust. At earlier points, training fel-lowships were provided by The US National Sci-ence Foundation (for DP Béhague) and the Co-ordenação de Aperfeiçoamento de Pessoal de Nív-el Superior (for H Gonçalves). D Béhague wassupported by a Wellcome Trust post-doctoralfellowship during the completion of some of theresearch cited in this paper, as well as during thepreparation of this manuscript.

References

Trostle JA. Epidemiology and Culture. Cambridge:Cambridge University Press; 2005.de Zoysa I, Habicht J, Pelto G, Martines J. Re-search steps in the development and evaluation ofpublic health interventions. Bulletin of the WorldHealth Organization 1998; 76(2):127-133.Scrimshaw S, Hurtado E. Rapid Assessment Proce-dures for Nutrition and Primary Health Care: Ap-proaches to Improving Programme Effectiveness. LosAngeles: UCLA Latin American Center; 1987.Bennett FJ. Qualitative and Quantitative Methods:In-Depth or Rapid Assessment? [editorial]. Soc SciMed 1995; 40(12):1589-1590.Dunn F, Janes C. Introduction: Medical Anthro-pology and Epidemiology. In: Jaines C, Stall R, Gif-ford S, editors. Anthropology and Epidemiology: In-terdisciplinary Approaches to the Study of Health andDisease. Dordrecht: D. Reidel Pub Company; 1986.p. 3-34.Agar M. Recasting the “Ethno” in “Epidemiology”.Medical Anthropology 1996; 16:361-403.Krieger N. Epidemiology and the Web of Causa-tion: Has Anyone seen the Spider? Soc Sci Med 1994;39(7):887-903.Napolitano D, Jones C. Who needs ‘pukka anthro-pologists’? A study of the perceptions of the use ofanthropology in tropical public health research.Tropical Medicine and International Health 2006;11(8):1264-1275.Rodgers I, Scaife M, Rizzo A. Interdisciplinarity:An Emergent or Engineered Process? In: Derry S,Gernsbacher, MA, Schunn, CD, editor. Interdisci-plinary Collaboration: An Emerging Cognitive Sci-ence. London: Routledge; 2005. p. 265-86.Lombardi C, Kendall C. Brazil: a RAP survey. Dia-logue on Diarrhoea 1998; 39:4.Gonçalves H, Gigante D. Work, schooling, and re-productive health: an ethno-epidemiological studyof adolescent women belonging to a birth cohort.Cad Saúde Pública 2006; 22(7):1459-1469.Gigante D, Victora C, Goncalves H, Lima R, BarrosF, Rasmussen K. Risk factors for childbearing dur-ing adolescence in a population-based birth cohortin southern Brazil. Rev Panam Salud Publica 2004;16(1):1-10.Béhague D, Victora C, Barros F. Consumer demandfor caesarean section deliveries in Brazil: informeddecision-making, patient choice or social inequal-ity? A population-based birth cohort study linkingethnographic and epidemiological methods. BMJ2002; 324(20): 942-944.Béhague D. The Shaping of Adolescent Pathology inthe Wake of Brazil’s New Democracy [doctoral]. Mon-treal: McGill University; 2004.Victora CG, Barros FC, Lima RC, Behague DP,Gonçalves H, Horta BL, Gigante DP, Vaughan JP.The Pelotas (Brazil) Birth Cohort Study, 1982-2001.Cad Saúde Pública 2003; 18(5):1241-1256.Victora C, Araújo C, Menezes A, Hallal P, VieiraM, Neutzling M, Gonçalves H, Valle NC, Lima RC,Anselmi L, Behague D, Gigante DP, Barros FC. Meth-odological Aspects of the 1993 Pelotas (Brazil) BirthCohort Study. Rev. Saúde Pública 2006; 40(1):39-46.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

1710B

éhag

ue,

D. P

. et a

l.

Béhague D. Beyond the simple economics of cesar-ean section birthing: women’s resistance to socialinequality. Culture, Medicine and Psychiatry 2002;26(4):473-507.Barros F, Vaughan J, Victora C, Huttly S. Epidemicof Caeserean sections in Brazil. Lancet 1991;338(20):167-169.Barros F. Breastfeeding and socioeconomic status inSouthern Brazil. Acta Pediatr Scand 1986; 75:558-562.Béhague D. Growth Monitoring and the Promotionof Breastfeeding. Soc Sci Medicine 1993; 37(12):1565-1578.Goncalves H. Aproveitar a vida: um estudo antropoló-gico sobre valores, juventudes e gravidez em uma cidadedo interior [tese]. Porto Alegre (RS): Departamentode Antropologia Social, UFRGS; 2004.Black N. Evidence based policy: proceed with care.BMJ 2001; 323(4):275-279.Victora C, Habicht J, Bryce J. Evidence-based pub-lic health: moving beyond randomized trials. Am JPublic Health 2004; 94(3):400-405.Béhague D, Gonçalves H. Explorando mecanismosde causalidade: colaboração entre Antropologia eEpidemiologia. Rev. Saúde Pública. In review 2008.Gonçalves H, Hallal PC, Amorim TC, Araujo CL,Menezes AM. Efeito de fatores sócio-culturais so-bre o nível de atividade física na adolescência. RevPanam Salud Publica 2007; 22(4): 246-256.Béhague D, Gonçalves H, Dias da Costa J. Themaking of medicine for the poor: primary healthcare interpretations of professionals, patients andpoliticians in Pelotas, Brazil. Health Policy and Plan-ning 2002; 17(2):131-143.Béhague D, Goncalves H, Kirkwood B, Gigante D,Victora C. Teen pregnancy and mental morbidity inearly adulthood: stigma and class struggles amongstyouth in Pelotas, Brazil. British Journal of Psychiatry.In review 2008.Victora C, Behague D, Barros F, Olinto M, Weider-pass E. Pacific Use and Short Breastfeeding Dura-tion: Cause, Consequence, or Coincidence? Pediat-rics 1997; 99(3):445-453.Kramer M, Barr R, Dagenais S, Yang H, Jones P,Ciofani L, Jane F. Pacifier use, early weaning, andcry/fuss behavior: a randomized controlled trial.JAMA 2001; 286 (3): 322-326.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

Davey Smith G, Ebrahim S. Epidemiology — is ittime to call it a day? International Journal of Epide-miology 2001; 30:1-11.Krieger N. Commentary: Society, biology and thelogic of social epidemiology. International Journalof Epidemiology 2001; 30:44-46.McPherson K. Commentary: Epidemiology? Keepit broad and deep. International Journal of Epidemi-ology 2001; 30:48.Victora C, Huttly S, Fuchs S, Olinto M. The role ofconceptual frameworks in epidemiological analy-sis: A hierarchical approach. International Journalof Epidemiology 1997; 26(1):224-227.Baum F. Researching Public Health: Behind theQualitative-Quantitative Methodological Debate. SocSc Medicine 1995; 40(4):459-468.Williams G. The determinants of health: structure,context and agency. Sociology of Health and Illness2003; 25 (Sp):131-154.Prandy K. Class, stratification and inequalities inhealth: a comparison of the Registrar General’s So-cial Classes and the Cambridge Scale. Sociology ofHealth and Illness 1999; 21:466-484.Frohlich KL, Corin E, Potvin L. A theoretical pro-posal for the relationship between context and dis-ease. Sociology of Health and Illness 2001; 23(6):776-797.Hammersley M. Deconstructing the Qualitative-Quan-titative Divide. What’s Wrong With Ethnography?London: Routledge; 1992.Johnson A, Johnson OR. Quality into Quantity: Onthe Measurement Potential of Ethnographic Field-notes. In: Sanjek R, editor. Fieldnotes: The Makingsof Anthropology. Ithaca: Cornell University Press;1990. p. 160-86.Hammersley M. The Logic of Theory-Testing in CaseStudy Research. What’s Wrong With Ethnography?London: Routledge; 1992.Van der Geest S. Overcoming Ethnocentrism: HowSocial Science and Medicine Relate and ShouldRelate to One Another. Soc Sci Medicine 1995;40(7):869-872.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.