amjph00019-0103

Public Health Then and Now

Public Health Nihilism vs Pragmatism:History, Politics, and the Control ofTuberculosisAmy L. Fairchild, PhD, MPH, and Gerald M. Oppenheimer, PhD, MPH

* '>< .'r ;; .!,* S e ' _L-z.> .X

I,.... ...*~~~~~~~~~~~~~~~~~ ..... .... Tuberculosis (TB)-the 17th century's"captain of all these men of death" and the19th century's "white plague"-began todecline in many countries during the mid- tolate 1800s. In the United States, the diseasesteadily receded until the mid-1980s, whenthat trend was reversed. From 1985 through1991, the number of reported TB cases in theUnited States increased 18%,' provoking gov-ernmental agencies to consider such targetedpublic health interventions as isolation ofinfected persons and directly observed ther-apy. Although directly observed therapy hadbeen used successfully in Madras and HongKong decades before the TB epidemic in theUnited States, public health officials hereremained skeptical of the measure until the1990s.2 However, with the surge in TB cases,the rise in multiple-drug-resistant fonns ofthedisease, and the close association ofthe diseasewith the AIDS epidemic, the political and cul-tural climate ofTB control underwent a radi-cal shift. Congress, reacting to public fear thatuntreatable TB would spread into the generalpopulation, greatly increased funding for con-trol measures.3 Suddenly, directly observedtherapy became the centerpiece of publichealth efforts to stem the disease.4

Although the TB epidemic in the UnitedStates subsided in response to directlyobserved therapy and complementary publichealth interventions,5 it sparked a controversyregarding the relative value of targeted publichealth measures vs broad social reform.6 Somesaw the deterioration of the public health infa-structure since the 1960s-the loss of publicfinancing for TB screening, follow-up, andtreatment-as the predisposing cause of thereemergence of TB in the 1980s.7 The epi-demic was subsequently the tragic result ofHIV disease and insufficient public planningand investment.8 Adherents to this view arguedthat "a structured commitment to find an ade-quate public health infrastructure could control

tuberculosis" and, moreover, could do it "inde-pendent of the need for broader socialchange."9 Others, in contrast, charged thatpublic health campaigns had failed to controlTB effectively because the "control measureshave dealt with individual health habits ratherthan addressing the underlying poverty thatpredisposes individuals to tuberculosis."'"Consequently, such limited interventionscould, by themselves, achieve only a partialand temporary success."

Controversy between supporters of lim-ited, targeted actions and those who urge moresweeping changes is not new, particularly inthe case of TB; here current arguments echoand are reinforced by earlier debates calling forstructural reform in terms of public healthinterventions.'2 Since the mid-1960s, that brieffor broad reform has been further strength-ened by the historical and demographic stud-ies of Thomas McKeown, who maintainedthat incremental secular changes in the popu-lation's standard of living-especially betternutrition-accounted for the bulk of thedecline in TB mortality since the mid-19thcentury. Public health efforts were nugatoryagainst airbome disease. McKeown's influ-ential thesis, which finds little use for clini-cal and public health initiatives, is partiallyreinforced by recent historical works that cast

The authors are with the Program in the History ofPublic Health and Medicine, Division of Sociomed-ical Sciences, Columbia School of Public Health,New York City. In addition, Gerald M. Oppen-heimer is with the Department of Health andNutrition Sciences, Brooklyn College, Brooklyn,NY.

Request for reprints should be sent to Amy L.Fairchild, PhD, MPH, Program in the History ofPublic Health and Medicine, 100 Haven Ave, 17 HTower III, New York, NY 10032-2625 (e-mail: [email protected]).

Note. The authors contributed equally to thisarticle. Their names are arranged alphabetically.

American Journal of Public Health 1105

'M Ir! ...:.,;- N.iHR ,MM15 M It.! .s,: - UP,

Vo.".8-8: ''N 7-7--f-


doubt on the efficacy of public health pro-grams in combating TB.

The belief that broad social refonms (suchas increases in real wages or the standard ofliving), rather than limited public health pro-grams, profoundly reduce rates of disease hasbeen described recently by Ronald Bayer as"public health nihilism."'3 His coinage makesreference to the therapeutic nihilism of theearly-19th-century New Vienna School, whosecore idea was that physicians might diagnoseand describe disease, but they could not cure it.Bayer argues that public health nihilism hasbeen central to debates in such areas as TB,population and family planning,'4 and men-tal health.'5 His term for those who dismisstargeted interventions is evocative, resonat-ing with historical doubt about the efficacy oflimited interventions on disease in individualsand populations.

Are targeted public health approacheswithout historical support? Here we use thehistorical literature to examine whether suchefforts may have had a significant effect onthe historical decline in TB mortality rates inthe United States, Britain, and France. Specif-ically, using the extensive historical studies ofTB, we describe the arguments for and datasupporting the efficacy of2 major public healthinterventions over time: segregation of thoseinfected with pulmonasy TB and eradication ofbovine TB. We find that this literature weakensthe arguments of McKeown and those whoshare his perspective in favor of improvedgeneral well-being as the fundamental cause ofthe fall in TB mortality.

We further argue that, over time, histori-ans framed the issue of the relationshipbetween public health and TB mortality on thebasis of their political sensibilities and values.These sensibilities and values determined thequestions each generation raised about TB andthe solutions they were prepared to entertain.

Our review also suggests the need formultifactorial models, based on local studies,to explain that historical decline. Such modelsmust include factors other than public healthmeasures, including data on nutrition, occu-pational health, and housing. They should alsodifferentiate between historical periods. Tar-geted interventions, especially patient isola-tion, probably became more efficacious as TBendemicity declined and the disease becamemore localized, a problem of smaller socialgroups and families. In addition, the types ofpublic health campaigns mounted in the pre-bacteriological and bacteriological eras weredifferent. Ultimately, a broad series of stud-ies is required to explain the decline of TB.There is always a danger to drawing policyrecommendations regarding disease preven-tion from a single historical study, a limitedperiod of time, or a single variable rather than

from a broad spectrum of studies consideringthe history of a complex disease in differenttimes, places, and populations.

Background: Tuberculosis and theMcKeown Thesis

Thomas McKeown made a signal con-tribution to demographic, epidemiologic, andhistorical studies by attempting to answer achallenging question: Why did mortality frominfectious disease decline dramatically dur-ing the 19th and 20th centuries?'6 The signif-icance of McKeown's question and of thescholarship invested in his answer has had animpact on a generation of social scientists,historians, and policymakers. ' His conclu-sions were most recently recapitulated, forexample, in a new history of the rise anddecline of TB in Japan, whose author noted,"Nutrition is one of the most powerful of allsocially and environmentally determined influ-ences on the development or retardation ofactive tuberculosis.... [N]either medicine norpublic health measures had a significantimpact on mortality from the disease untilafter World War II.-18

McKeown, in studying declining deathrates in England and Wales, pinpointed TBand other airbome disorders. These contributedmore, he asserted, than all other infectiousdiseases combined to the fall in crude mor-tality between 1848 and 1971.'9 To explainthis general reduction in mortality, McKeownlaid out a list of mutually exclusive possiblecauses, starting with changes in the virulenceof microorganisms, continuing on to scientif-ically based clinical interventions, and end-ing with an increase in the quality and level ofnutritional intake. These causes (including thedevelopment of hospitalization, chemotherapy,and vaccination) he successively eliminated,until he was left with 2 hypotheses: publichealth (sanitation) reforms and a generalimprovement in nutrition. McKeown then rea-soned that better sanitation could have affectedonly water- and foodborne diseases, whichcontributed a far smaller fraction to the crudemortality decline. By logical elimination, then,McKeown found that improved host resis-tance, as a consequence of better nutrition,best explained the decline in airbome disease,TB included, in the period under consideration.

McKeown lacked direct evidence to sup-port this last hypothesis; instead, having elim-inated other causes, he initially left the lion'sshare of mortality reduction, by default, tobetter nutrition.20 In another work, he offerednutrition as a hypothesis that, based on cir-cumstantial evidence, should be made to stand"the test of critical examination in a varietyof circumstances over an extended period."'"To buttress his hypothesis, he provided limited

data derived from animal models, inferencesconcerning increased food production in Eng-land and Wales in the 19th century, and currentgeneral experience in developing countries.22

McKeown's opus, although influential,has drawn sharp criticism. He has been takento task for his data analysis and for his thesisthat improvement in nutrition was the pre-dominant cause ofthe decline in TB mortalityin the 19th and 20th centuries. In both hisearly and later work, McKeown treated nutri-tion as "the most significant feature" of a ris-ing standard of living during those centuries;his presumption was that amelioration inhealth status stemmed from a changing econ-omy rather than a set of specific acts engi-neered by human agency. Critics haveresponded by asserting the role of such agency,particularly in the form of broad, sociallybased public health measures.

Simon Szreter, a British historian, findsthat the reasons for the decline in infectious-disease-specific mortality are more complexthan McKeown's hypothesis allows. He arguesthat once airbome diseases are studied indi-vidually, rather than in the aggregate as McK-eown studied them, the disease-specific causesof mortality reduction vary; for example, thestreptococci associated with scarlet fever, byMcKeown's own admission, may havedecreased substantially in virulence duringthe 19th century. Indeed, of all the airbomediseases, only TB could have been affectedby population-level malnutrition.23

By demonstrating that TB began todecline later in the century than McKeownallows, Szreter concludes that in the instanceofTB, as with other diseases, there were mul-tifactorial reasons for the downward trend inmortality. Szreter cites the existence of inter-current infections and occupational hazardsthat weakened host resistance to TB and ofovercrowding and poor ventilation in workand home environments that enhanced trans-mission of the disease.24 He suggests that thesefactors were removed not only by rising realwages and better nutrition but by political andsocial action associated with the public healthmovement, "a complicated history of strug-gle and pressure for relevant clauses in Fac-tories and Workshop Acts, Housing andCrowding Acts, and the enforcement of build-ing regulations and by-laws."25 This is theprebacteriological public health campaign thatattacked the unhygienic environmental andbehavioral detenminants of health.

While Szreter bases his critique and alter-nate hypothesis on an analysis ofMcKeown'soriginal data, local studies ofTB mortality inBritain support the role ofbroad environmentaland political measures in controlling TB.26Anne Hardy finds that housing policy, possiblynutrition, public health education, and pro-

July 1998, Vol. 88, No. 71106 American Journal of Public Health


portionate declines in workers employed inhigh-risk industries help account for the drop-ping rate of TB mortality in London in the19th centuiy.27 In his study ofGlasgow in the20th century, Neil McFarlane contends thatimproved housing, rather than better nutri-tion, helps account for the decline in pul-monary consumption.28

Studies of other locales also support therole of environmental measures in reducinginfectious disease mortality, including thatassociated with TB. Philip Curtin's study ofBritish and French data from military postson the continent and in Algeria, the BritishWest Indies, and southern India documents a"mortality revolution" during the 1850s. Curtinfinds that a sanitary engineering campaignsubstantially rduced mortality within the mil-itary from a number ofwater-bome diseases,first overseas and then at home.29 Significantly,by controlling for nutritional intake and socialclass (limiting the study ofmortality to work-ing-class men in the military), Curtin's workoffers unique insight into the control of TB.

While TB was responsible for the major-ity ofdeaths among troops stationed in Britainand France, the rates ofTB mortality amongtroops abroad were much lower.30 AlthoughCurtin maintains that the cause of the differ-ential decline "remains obscure," he proposesthat improved ventilation-in this instance, inthe troops' barracks-may account for lowerTB-specific mortality rates in overseas posts.He discounts nutrition as a reason for "theenormous differences between death rates[from TB] at home and those overseas,"3' sincediet within the ranks did not vary by posting.

One historian studying the many causesof the decline of TB observes that in "thedebate between McKeown and his critics, thejury is still out."32 Although McKeown's worksuffers from distinct weaknesses, so does thatofhis critics. These critics, who stress the roleofhuman agency in the form ofpublic healthinterventions, lack a body of valid quantita-tive data to support their position. To explaindeclines in TB mortality rates-a quantitativeoutcome-some quantitative measurementand analysis ofthe hypothetical causes wouldstrengthen the critics' arguments. Many pro-pose multifactorial models to explain the pro-gressive fall in TB mortality rates; in theabsence of data, however, they cannot test forthe separate effect ofeach ofthe independentvariables-housing, ventilation, sanitation,working conditions, nutrition, or intercurrentinfections-while controlling for the others.Finally, unlike McKeown, his critics tend tofocus on the 19th century, the prebacteriolog-ical era in particular. Consequently, to counterMcKeown's nutrition theory, they offer publichealth models specific to that era: environ-mental measures targeting mortality in gen-

eral rather than TB in particular. Only a fewcritics have posited public health models thatincorporate the insights of bacteriology: thatTB is a contagious disease that requires casefinding and the disruption of transmissionthrough patient education and isolation. Weturn now to those writers who evaluate McK-eown from the perspective ofthe bacteriolog-ical revolution.

Segregation ofthe Infected:The Sanatorium

First established in Germany (Silesia) inthe mid-19th century, sanatoriums were meantto cure TB through the agency of fresh airand exercise, diet, and graduated labor. In the20th century, sanatoriums incorporated clini-cal interventions such as tuberculin treatmentand pneumothorax. Historians, particularlythe British, have challenged the clinical effi-cacy of these institutions. Nancy Tomes, inher recent review ofTB literature, posits thatBritish scholars, "writing in McKeown'sshadow," feel compelled "to assess the relativeimportance of medical versus non-medicalfactors in the decline in mortality."33 In addi-tion, historians challenge whether sanatori-ums, by isolating tuberculous patients, wereuseful agents ofprimary prevention, as RobertKoch and other contemporary physiciansargued they should be.34 Their skepticism isrooted in chronology.35 Mortality from TBbegan to decline in a continuous fashion inGreat Britain sometime between the 1830sand the 1880s36; in the United States, thatdecline started in the 1860s or 1870s.37 Yetsanatoriums and other special institutions forthe care and isolation ofpatients with TB didnot proliferate until after Robert Koch's dis-covery of the mycobacterium in 1882.38

In Britain, which adopted the sanatoriumlate in the 19th century, there was great con-fidence in the curability ofTB by "open-air"methods.39 With cure as a goal, sanatoriumsand voluntary hospitals carefully selected earlycases of TB over chronic "incurables." Thestate sanatoriums, founded after the NationalInsurance Act of 1911, applied the same rulesof patient selection.40 By the first decade ofthis century, the clinical success rate of thesanatoriums had been found to be far lowerthan expected, with more than half of treatedpatients dead 5 years after discharge.4' Theconsequence of this medical emphasis was,in the view of current British historians such asLinda Bryder and F. B. Smith, a sad waste.Writes Smith: "It was unfortunate that a coterieof medical practitioners captured the tuber-culosis specialty in the late 1880s and there-after set the terms by which the disease wascomprehended. Public money could andshould have been diverted to trying preven-

tive measures and to helping severely impov-erished cases and their families."42

Similarly, American historians haveoften held that sanatoriums, because of theirscope and design, could not have sucessfullyisolated infectious individuals. For example,Sheila Rothman, in describing the coercive,punitive system of treatment and isolationfacilities created by Hermann Biggs in NewYork City shortly after the turn of the cen-tury, argues that institutional organization andphilosophy were fundamentally flawed. AndOtisville, the municipal sanatorium openedin 1906, was "spartan, austere, and some-times punitive."" Patients, accordingly, typ-ically left the institution shortly after arriv-ing."" Those institutions created to holdrecalcitrant, intractable infectious patients byforce, such as the hospital for the tubercularpoor on Blackwell's Island in the East Riverand Riverside Hospital, established on NorthBrother Island, were also failures. Accordingto Rotiman, lacking the medical supplies, thenecessary staff, and the facilities to maintaindiscipline, such "institutions...were too prison-like to be hospitals and too hospital-like to beprisons."'

Barbara Bates, another American histo-rian, bases her conclusions about the ineffi-cacy ofthe relatively extensive system of sana-toriums in Pennsylvania on the limited scopeof institutional care before the 1920s. She con-cludes that "[i]nstitutionalization...was notthe leading cause of Philadelphia's fallingdeath rate. The decline of tuberculosis beganbefore that form of segregation and contin-ued fairly steadily thereafter until 1916. l46According to Bates, the system of charity andreligious hospitals and sanatoriums createdbefore the turn of the century simply did nothouse a significant portion of the tubercular.Those who did receive institutional care usu-ally chose to die at home, undermining thebenefits of segregation.

As these sanatoriums and hospitals weresupplemented and eventually supplanted bystate institutions during the early 20th cen-try,"47 care of the tubercular did increasinglymove out ofthe home, especially after 1910.48Still, sanatoriums fell short on their promise toisolate the infected. Institutions under financialpressure, for example, discharged patientswho seemed improved. Indeed, by 1914Philadelphia's Free Hospital for Poor Con-sumptives Oater the White Haven SanatoriumAssociation) was limiting the maximum stayfor any individual to 14 weeks.49 In addition,the diagnosis was usually made late in infec-tion; once diagnosed, patients continued towait long periods before they could enter asanatorium owing to lack of space.50

Despite her argument concerning its neg-ligible public health effects, Bates' own data

American Journal of Public Health 1107July 1998, Vol. 88, No. 7


allow a reconsideration of the benefits of iso-lation. Bates observes that the rate of decliningmortality usually signals the efficacy of anyparticular intervention. Interestingly, Bates'data show that the rate of decline increasedperceptibly between 1880 and 1900, the periodin which the first sanatoriums opened.5' Shealso notes that the TB death rate fell some-what more rapidly after 1920; in this instance,she believes that "institutionalization may havehad an impact," although segregation "cannotbe separated from other possible causes."52

Ultimately, Bates' primary concer-likethat ofmost American historians-is to docu-ment the patients' experience ofTB, not McK-eown's thesis or even the efficacy of publichealth measures.53 The public health effect ofpatient isolation is, however, central to popu-lation-level, quantitative histories of TB.Leonard Wilson, for example, resurrects andfurther supplements the statistical analyses ofSir Arthur Newsholme, the medical officer ofhealth in Brighton, England, and president ofthe Epidemiological Section ofthe Royal Soci-ety of Medicine. In work published from 1903through 1908, Newsholme, according to Wil-son, found that the segregation of individualssuffenng from advanced pulmonaiy TB (phthi-sis) was responsible for the decline in TB mor-tality in England and Wales, Scotland, Prussia,and New York City. Newsholme's analysesindicated that improvements in nutrition andstandards of living (measured by the courseofwheat prices, among other indices), althoughstrongly correlated with decreasing deathsfrom consumption in Great Britain, were onlymoderately to poorly correlated with decliningdeath rates in Prussia and France; moreover,they were inversely correlated with phthisismortality in Ireland (see Figure 1). Like mod-em demographers, Newsholme observed thatthe pattem of declining TB mortality and risingliving standards in England and Wales wasnot generalizable to other countries.54 If thereexisted strong, universal determinants of thefall in TB mortality rates across nations, thenone had to look elsewhere.

Newshoime's solution, according to Wil-son, was based on his observation that a grow-ing proportion of consumptives were segre-gated in British Poor Law infirmaries andworkhouses. The poor, who were at highestrisk ofTB (and further impoverishment), weresent to those institutions rather than to the rel-atively new and scarce sanatoriums.55 New-sholme reasoned that the segregation of con-sumptives with active disease, by diminishingthe transmission of infection, must have had asignificant effect on TB incidence and mor-tality rates.

A comparison of Ireland with England,Wales, and Scotland provided Newsholmewith a "natural experiment" to test his hypoth-

esis further. In Britain, both typhus-a con-

tagious louse-borne disease-and phthisismortality rates had declined since 1870.Although typhus rates had also decreased inIreland, phthisis mortality rates continued torise. Could this be due to varying social con-

ditions? Newsholme, according to Wilson,found that Ireland, like England, Wales, andScotland, enjoyed improved housing, grow-

ing real income, and lower food costs duringthe last 30 years of the century. Then why theparadoxical differences between Ireland andthe rest?

Newsholme argued that the cause laychiefly in the administration of the poor lawsin Ireland. On one hand, the Irish laws suffi-ciently immobilized the population to preventtyphus-infected paupers from migrating insearch of work and thereby spreading dis-ease.56 On the other hand, a far higher pro-portion of the Irish poor received "outdoorrelief' (relief in the home) rather than "indoor"or institutional benefits. Consequently, most

impoverished Irish consumptives dependedon domestic treatment rather than segregated

confinement as in Britain, thereby considerablyenhancing the transmission of phthisis withinfamilies. Newsholme also found this pattem inNorway, a nation in which only a small pro-

portion of the sick were institutionally treated.Wilson maintains that, by demonstrating

the constancy of the relationship betweendeclining TB mortality and the increased seg-

regation of infected patients and by docu-menting the inconsistent association betweenTB mortality and indices of improved socialconditions across a number of countries, News-hoime successfully eliminated population liv-ing standards as a key determinant in under-standing the decline in TB mortality.

Wilson attempts to further strengthenNewsholme's thesis by analyzing seculartrends in TB mortality for the period 1838 to1961. When Wilson plotted the log ofTB datafor England and Wales, he observed 4 straightlines corresponding, he argues, to distinctphases in the public health program57 (see Fig-ure 2). As programs focused increasingly onisolation or treatment of the infected (asopposed to isolation of only the poor con-


FIGURE 1 -Newsholme's graph comparing phthisis mortality in England andWales, Scotland, and Ireland with the cost of food in the UnitedKingdom (including Ireland) from 1877 to 1901. From ArthurNewsholme, "An Inquiry into the Principal Causes of the Reductionin the Death-Rate from Phthisis During the Last Forty Years, withSpecial Reference to the Segregation of Phthisical Patients inGeneral Institutions," Journal of Hygiene 6 (July 1906): 336.


sumptives), the rate of decline accelerated.Wilson found a similar pattern when he plot-ted data from New York City from 1810 to1960.58 He found a steeper decline in Min-nesota, which by 1918 had achieved the high-est rate of segregation in the nation.59 Wilsonobserved that TB declined in Minnesota as aconsequence of that state's efforts to isolateindex cases and identify contacts, particularlyfamily members.60

To be sure, Wilson's argument is notunassailable. Mortality data, especially 19th-century mortality data, are incomplete.6' Adecline in mortality rates, moreover, does notautomatically translate into a decrease in TBincidence rates, as Wilson's analysis implies.62Neither Wilson nor Newsholme had a directmeasure of the rate of patient segregation.And neither controlled simultaneously forpotentially confounding variables.63

In addition, Wilson simplifies News-holme's position, holding him to a mono-causal explanation for the historical fall ofTB mortality rates.64 Newsholme, however,understood that consumption was a complexdisease whose decline was the consequence ofmultiple factors. Newsholme believed, forexample, that overcrowding within dwellings,urbanization, "well-being" (determined bywages, pauperism, and total cost of living),nutrition, and education about indiscrim-inate spitting and coughing had an impact onthe TB death toll.65 Newsholme did not dis-count the importance ofnutrition, wages, andhousing; he simply argued that such factorslacked a predominant influence. "Segrega-tion in general institutions is the only factorwhich has varied constantly with the phthisisdeath-rate... .It must therefore be regarded ashaving exerted a more powerful influence onthe prevention of phthisis than any of theother factors of which none has varied con-stantly with the phthisis death-rate."66

As a public health official, Newsholmewas particularly concemed about practical andcost-beneficial opportunities to reduce the tollofTB. His research was meant to demonstratethe importance of pragmatic solutions to theepidemic. Poverty, overcrowding, and mal-nutrition, while ofconsequence, were not onlyvariable, relatively weak deterninants of thedecline ofTB but represented difficult issuesfor public health to confront. Institutional seg-regation, especially of open cases, was, toNewsholme, both efficacious and politicallyrealistic. As he wrote in his recollections, pub-lished in 1936, "I drew satisfaction...inas-much as the machinery to provide the segre-gation alrady existed, had been operating to amaterial extent for forty or fifty years, andcould easily be vastly expanded at a far lowerexpense than could be incurred by the con-struction of a new machinery.

At the time Newsholme wrote his auto-biographical recollections on public health,he was serving as advisor to a multisite healthproject in New York State underwritten bythe Milbank Memorial Fund. One of theobjectives ofthe project was to organize com-munity health work to determine the cost andeffectiveness of measures for the control ofTB. One of the sites, Cattaraugus County, arural county of 75 000 persons, became an

epidemiologic and public health laboratory,beginning in 1923."' The work done in Catta-raugus County was influenced by an earlierproject, the Framingham Community Healthand Tuberculosis Demonstration, but gener-ated more sophisticated statistical analyses.69Among the administrative procedures even-

tually evaluated in Cattaraugus County was

the role of the sanatorium in TB control. Aprospective comparative study conducted byJean Downes found that the difference ininfection rates-not mortality rates-for twogroups of families was "probably associatedwith an improvement in the extent to whichsanatoria care was obtained for the infectiouscases and the speed with which they were

hospitlized."70 Significantly, the reduction inmortality was achieved during the height ofthe Depression.

In a later report on her research, "SalientPoints of Attack against Tuberculosis,"Downes offered quantitative support for usingthe "tuberculous family," with an infectionrate 10 to 15 times that of the general com-

munity, as the locus of case finding. In the

same piece, she approvingly quoted an arti-cle by Wade Hampton Frost, the leading epi-demiologist of the period, in which he out-lined a program for the control of TB thatemphasized "the isolation in sanatoria of allknown open cases of pulmonary tuberculo-sis,"7' a recommendation that parallels that ofNewsholme. Yet Frost, who held the first chairof epidemiology at Johns Hopkins Univer-sity, observed that to effectively control TB,societes did not have to isolate all active cases.

For the eventual eradication of tubercu-losis it is not necessary that transmission beimmediate and completely prevented. It is nec-essary only that the rate oftransmission be heldpennanently below the level at which a givennumber of infection spreading (i.e. open) casessucceed in establishing an equivalent numberto cany on the succession. If, in successive peri-ods of time, the number of infectious hosts iscontinuously reduced, the end-result of thisdiminishing ratio, if continued long enough,must be the extennination ofthe tubercle bacil-lus.72

Downes' findings and Frost's stricturesprovide support, based on contemporary data,for Newsholme's retrospective analysis. Iso-lation of active cases ofTB reduces the inci-dence of secondary attacks on those living inclose proximity to the index patients; over

time, such isolation reduces the prevalence ofthe tubercle bacillus and, subsequently, TBincidence and mortality rates in the popula-tion as a whole. In brief, although contempo-rary historians have often denied the value ofsegregation, both original data and the quan-


England and Wales, Pulnonary Tuberculosis Mortahty 1838-1964 Plotted Serni-Logarithmicauylow0 f IPha"t _Ph"* 2 se33

NOW aefect oa Swq, e d ._g d oa pow conaw' a tcggSan of Ar 1t- f we

?S0 pow caAwith- in W-90 c_wtw9ttd awvta.aa trcd pa.

*250 -

1001

a45 Wafid WeeI

J Wan~~~~~~~~~~~~~~~~~~~~~~~~~~~~frdVwea101CL 81 *

6!.

4'18.30 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 96E0

FIGURE 2-Wilson's semilogarithmic graph of phthisis mortality from 1838to 1964. From Leonard G. Wilson, "The Historical Decline ofTuberculosis in Europe and America: Its Causes and Significance,"Joumal of the History ofMedicine andAIlId Sciences 45 (1990): 393.Reproduced by permission of Oxford University Press.

July 1998, Vol. 88, No. 7


titative analyses of earlier historians stronglysuggest that patient isolation did make a dif-ference.

Bovine Tuberculosis

While the sanatorium has figured signifi-cantly in numerous histories of tuberculosis,other bacteriologically based methods for itscontrol have received short shrift in the his-torical literature. Among these are pasteur-ization, tuberculin testing, and eradicationefforts among cattle. The silence of histori-ans, however, should not be interpreted as acommentary on the effectiveness of thesemeasures. Such measures constituted signifi-cant, successful public health initiatives, inter-ventions that produced scientific, economic,and political controversies that have sincebeen forgotten.73

The association between a slow, wastingdisease in cattle and phthisis in humans wassuspected for centuries before Koch's dis-covery of the human tubercle bacillus in1882.74 Sixteen years later, Theobold Smithisolated a bovine mycobacterium, implicat-ing it as the major detenrinant ofTB in cattle.Based on differences between the 2 microbesand his inability to transmit human TB bacillito cattle, Koch concluded that bovine TBposed little threat to humans.75 His positiongenerated considerable dissension from vet-erinarians and led to the appointment of gov-ernment-sponsored commissions in Germany,Britain, andNew York City. In reports issuedbetween 1904 and 1911, all concluded thatMycobacterium bovis was a source ofhumanTB.76

Epidemiologic studies representing vary-ing places and times indicate that bovine-induced TB occurs in humans, although at arate far lower than TB due to M tuberculosis,and is associated pfimarily with nonpulmonaryinfections. For the period from 1911 to 1927,for example, studies found that the overall per-centages ofTB cases due to bovine strains inthe areas around Cambridge, England, wereapproximately 23% for all forms of TB andapproximately 1% for pulmonary TB (Bryderestimates 30% for nonpulmonary TB and 2%for pulmonary TB in England during the1930s77). However, among children 5 years ofage and younger, 68% of cases studied wereattributable to bovine TB. Researchers in NewYork City reported a similar rate in children(66%/o) in 1910.78 The Cambridge regional stud-ies showed that the percentage ofcases due tobovine TB fell with increasing age: 37%among children 5 to 15 years old and 6% inthose 15 years of age and older.79 In general,the prevalence ofTB was higher in rural areasofEngland and Scotland, with overall rates ofbovine (and human) TB 300% greater than

those in the Cambridge area.80In the countries of North America and

westem Europe, the growing suspicion of anassociation between TB in cattle and humansprecipitated social policy responses. In theUnited States, the passage of the Meat Inspec-tion Act of 1906 made inspection ofcarcassesfor TB mandatory across the nation. Yet meatinspection neither eliminated the source ofTB nor aided cattlemen. The increasing rate ofTB in cattle threatened the nation's livestockindustry. One expert estimated that the numberof cattle condemned as unfit for human con-sumption in 1917 could fill a train 15 mileslong. In Chicago in 1916, the packing houses,stockyards, railroads, and stock exchange orga-nized a committee and provided funding toeliminate TB in cattle and hogs. The followingyear, the Tuberculosis Eradication Divisionofthe US Department ofAgriculture's BureauofAnimal Industry spearheaded a long-termTB eradication program that combined tuber-culin testing of cattle, eradication of all react-ing animals with compensation to farmers fortheir losses, certification of "Tuberculosis FreeAccredited Herds," and establishment of"Modified Accredited Tuberculosis FreeAreas."81

Although the division's program did notbegin until 1917, the results were dramatic.Figure 3 compares the rate oftuberculin reac-tors found and removed from cattle popula-tions with TB mortality in the general popu-lation and among severl subgroups. Withoutexception, the initiation ofthe federl TB cat-tle eradication program (1917) and the peak inthe detection and removal of infectious cattle(1918) coincide with sharp declines in pul-monary and exfiapulmonary TB mortality. It isunlikely that the cattle eradication programwas solely responsible for the decline in TBmortality, especially given that its successesdid not precede the decline in human mortal-ity.82The period from 1915 to 1921 was one ofrapid spread of pasteurization in Americancities.83 Together, pasteurization and testingof cattle probably played a major part in theaccelerated decline in exrapuhmonary and per-haps even pulmonary TB, beginning in 1917.

Significantly, in the period after 1940,the resurgence of TB among cattle seems tohave precipitated an increase in extrapul-monary TB in the general population. By1940, every US county had been officiallydesignated a "Modified Accredited Tubercu-losis Free Area," meaning that the number oftuberculin reactors among cattle did not exceed0.5% and that the county was in compliancewith federal testing and disposal rles.? Manycattle owners and legislators regarded this vet-erinary milestone as the final goal, and tuber-culin testing among cattle consequentlydecreased. The rate of tuberculin reactors

found among cattle continued to decline,reaching an all-time low in 1952. In 1955,however, the incidence ofTB among livestockbegan to rise. By 1959, it had increased to 23per 10000 animals tested. As Myers andSteele, who documented the history of thecontrol of bovine TB, concluded, "Resump-tion of the long time standard program ofextensive periodic tuberculin testing againproved its efficacy and the incidence of infec-tion decreased from 23 per 10,000 in 1959 to 6per 10,000 in 1969."85 Contemporary epi-demiologists recognize the need for contin-ued close surveillance of animal and humanreservoirs forM bovis TB.

In the early years of the 20th century,localities in the United States also institutedpasteurization of milk, with the aim of elimi-nating milk as a vehicle for scarlet fever, diph-theria, infant diarrhea, and TB.86 In 1909, theUS Public Health Service reported 500 epi-demics of milkborne disease in the period1880 to 1907.87 That year, Chicago's healthcommissioner, unable to regulate the 12000dairies that supplied the city's residents withmilk, promulgated a compulsory pasurizationordinance, the first in the United States.88 Thefollowing year, New York City required thatall nilk for drinking purposes be pasteurized89By 1922, 90% or more of milk sold in mostlarge US cities was pasteurized.90

Where pasteurization was introduced,tuberculosis declined, particularly among chil-dren.9' In Toronto, which prohibited the sale ofnonpastrized milk begiening in 1918, bovineTB disappeared as an admitting diagnosis inthe local children's hospital for any child bomor raised in the city.92 The introduction ofpas-teuized milk in the Netherlands in 1940 led toa marked decline in TB.93 With the eliminationof milk as the principal vector of transmis-sion of bovine TB, the anatomical distribu-tion ofTB lesions also changed dramatically.Meningeal, bone, joint, and skin TB, previ-ously very common, became rare.94

Westem nations that resisted the elimina-tion of Aeulous cattle or the adoption ofpas-teurization saw TB rates decline more slowly,particularly in children. In Britain, as both Bry-der and Smith explain, fanners, consumers, andlegislators repatedly resisted attmpts to maketuberculin testing of cattle and pasteurizationof milk compulsory. The financial burden oftubrculn testing, the subsequent loss ofinfectedcattle, and concems regarding the safety, taste,and price of pasteurized milk combined withthe scientific controversies to effectively blockenforceable national safety standards. Not untilthe 1950s was meat inspection unifonn, andnot until after 1948 was pasteurization compul-sory.95 Most of the decline in nonpulmonaryTB mortality in Britain occurred in the decadefollowing the outbreak of World War II, once



pasteurization spread quickly across the coun-

try.96 Nonetheless, in 1951, the rate oftubercu-losis attibutable to Mbovis was higher in Britainthan in any other industrial nation.97 The Britishconclusion, then, is that an opportunity to pre-

vent disease was practically wasted: "Pasteuri-sation saved lives and could have saved themearier.998

Conclusion: History as Politics

What caused the decline in TB mortality?Those who have attempted to answer thatquestion often have a set of political valuesthat informs their answers. McKeown, whosework continues to frame so much ofthe debateover TB mortality, had lost faith in clinicalinterventions early in his career. His own train-ing in and experience of medicine left himconvinced that its usefulness was greatly exag-gerated.99 As a professor of social medicine

at Birmingham University during thepost-World War II years, when Britain wasestablishing its National Health Service, hefought against those who would have had theService overcommit to traditional secondaryprevention.100 He argued that the Service mustpromote major efforts to modify environ-mental factors and personal behaviors, "whichare the predominant deteminants of health."'O'For McKeown, history was an appropriatetool for attacking medicine; the social historyof medicine, he noted, "is essentially an oper-

ational approach which takes its terms of ref-erence from difficulties confronting medicinein the present day."'102

At the same time that he deliberatelyattacked medicine, he shared with other aca-

demic members of his generation a certaindisdain for public health. According to Szreter,"Social medicine provided an altogethergrander vision of positive health enhancement

for the populace through deployment of theresources and organization of the state."'03Indeed, social medicine not only represented a

new view of health and medicine butamounted to a "rejection of the narrow focusand complacency of public health."'04 ForMcKeown, public health was a limited, obso-lete method of affecting health, using localadministrative apparatus dating from the Vic-torian era to put into place various technolo-gies that reduced environmental pathogens.'05A certain air of whimsical condescension isevident in his autobiographical accountdescribing his ascendance to the chair of socialmedicine:

I eventually took possession of a room inthe Medical School, shared previously by thepart-time teacher of public health (Dr. G. A.Auden, father of the poet) and the lecturer inforensic medicine, and filled with the drains,waterpipes, contraceptives, and numerous other


10.000

4,00_

Ji.ooi.

VCL

10 0

100

10~~ ~ ~ ~ ~ ~ 0

Source. J. Arthur Meyers and James H. Steele, Bovine Tuberculosis Control in Man and Animals (St. Louis, Mo.: Warren H. Green, 1969);Vital Statistics of the United States; World Health Organization, Annual Epidemiological and Vital Statistics and World Health StatisticsReport.

FIGURE 3-Bovine TB eradication and human TB mortality rates, United States, 1900 to 1967.

July 1998, Vol. 88, No. 7


objects of forensic and nineteenth century pub-lic health interest.'06

Arthur Newsholme's view of publichealth, on the other hand, was that of a pro-

gressive Victorian medical officer of healthdeeply influenced by Koch's discovery of thetubercle bacillus and the doctrine ofTB as a

contagious disease.'07 Like other medical offi-cers of health, such as Arthur Ransome andJames Niven, Newsholme supported and pro-

posed measures to reduce the transmission ofconsumption. He made Brighton, where heserved as medical officer of health, the firsttown that officially established voluntaryreporting of consumption, in 1899.108 In thisendeavor, he faced formidable political andsocial obstacles in a society in which TB was

a stigmatizing disease and the ethical and prac-

tical effects of isolating active cases were seri-ously debated.109 Nonetheless, as medical offi-cer of the Local Government Board,Newsholme successfully introduced compul-sory notification of TB as a national policy,gradually implementing it beginning in1908."°

Newsholme's campaign to isolate activeTB cases was of a piece with his belief incompulsory notification; his epidemiologicalanalyses provided, he believed, scientific sup-port for his position that segregation was an

effective, affordable measure against con-

sumption. For Newsholme and other medical

officers of health, targeted public healthactions-including housing policies and pub-lic education leading to behavioral changes-could effectively contribute toward the declinein TB incidence and mortality; such limitedinterventions were, moreover, politically fea-sible, unlike broader social reform affectingnutrition and poverty."'

Like McKeown, many historians ofmed-icine, especially those affected by the politicsof the 1960s and 1970s, have been skeptical ofthe efficacy claims made by the medical pro-

fession and-by extension-the "new" publichealth, which produced a new cadre of publichealth professionals,"12 justified a new set ofinterventions, and, increasingly, overlappedwith clinical medicine."13 They have oftenreacted strongly against the beliefs ofpreviousgenerations of historians, who tended to cele-brate the successes of physicians, to supportthe rational authority of biomedicine, and tohold to a faith in the historical progress ofmedical science.'14 Instead of the efficacy ofbiomedicine, they have tended to concentrateon "the social causation of health and diseaseand the way in which science is embedded ina society's social relations."''5

Influenced by the civil rights, anti-Viet-nam War, women's liberation, and gay liber-ation movements, these medical historianshave often focused their attention on women,minorities, and the poor within the health care

system. Recognizing that medicine and med-ical institutions could be a means of socialcontrol, they have been suspicious, for exam-ple, of public health actions such as those ofNewsholme and his generation that segregatedthe sick, particularly when those isolated werepoor or otherwise dispossessed."16 In study-ing the history of medical professions andinstitutions, including those of public health,contemporary medical historians have ques-tioned the social and political motives ofthoseinvolved and the objectivity of the diseaseentities that often justified their actions. WritesCharles Rosenberg, a leading historian ofmed-icme:

Critics have tumed the delegitimating toolsof cultural relativism on medicine as they haveon so many other areas in which knowledgeand power are closely linked.... This relativistpoint ofview has sought to undermine not onlythe apparent objectivity of particular diseaseentities but also, by implication, the legitimacyof the social authority wielded by the medicalprofession, which has traditionally articulatedand administered diagnostic categories. Thephysician is not above social interest, but is asocial actor whose mission of defining and treat-ing disease can express and legitimate profes-sional, class, or gender interests."17

The reason for the dearth of interest ofsocial historians in pasteurization and herderadication, in sharp contrast to their exam-ination of sanatoriums, may be that the for-mer measures do not fit easily into the cur-rent historical paradigm. They appear to lackthe political, social, and cultural dimensionsthat make interventions such as segregationof the infected provocative. Issues of the rel-ative roles of clinical medicine and publichealth, class dissension, social control ofknowledge, or construction of disease seemirrelevant here. Measures such as pasteur-ization and tuberculin testing seem minor,rather technical themes relevant to historiesof veterinary science, the spread of technol-ogy, or the milk industry, none ofwhich areofmuch interest to contemporary social his-torians of medicine. The story of pasteur-ization and herd eradication, at least in itscurrent telling, is closer to the older historyof technical progress and success than to thenew social history.

In the essay already cited, Charles Rosen-berg argues that the HIV/AIDS epidemic hasserved notice to society that some diseases-including cholera and TB-are distinctly bio-logical as well as social phenomena."18 Sophis-ticated immunological and virologicaltechniques were required to define HIV/AIDSas a clinical entity and to develop chemother-apies to treat it. Although AIDS and theresponse to it must be placed within a complexcultural context, the biomedical dimensionsof that disease must be recognized as having

July 1998, Vol. 88, No. 7I112 American Journal of Public Health

LIi;ii1: l A Ns1 1l\'1'I !THI lIU IWO (11 LD.b NE

This image of a tubercular woman and her children at work in a tenementunderscores the potential role both of fundamental sanitary and housingreform and of isolation of active cases in staying the spread of tuberculosis.From Sprago, The Bitter Cry of Children (New York: Macmillan, 1906).


profound scientific, clinical, and policy con-

sequences."19AIDS policy successes provide a strong

reason to revisit the role ofpublic health dur-ing past disease episodes. Faced with a new

disease for which cure was unknown and treat-ment limited or unavailable, public health pro-

fessionals cobbled together policies that, with-out eliminating AIDS, probably reduced therisk ofHIV infection and its sequelae. Thesetargeted public health interventions includeheat treatment ofblood products, blood donorscreening, and the institution of needleexchange programs. Such policies, each ofwhich generated opposition, make us more

sensitive to the efficacy of small victories, as

well as the complexities of an incurable infec-tious disease. Consequently, they affirm, inan analogous fashion, that targeted interven-tions incrementally decreased the incidenceof TB. In brief, current experiences duringthe HIV/AIDS epidemic may have served toproduce, ifnot a revolution, then a shift in thecurrent medical history paradigm.

This review ofhistorical literature on TB,we argue, strongly suggests the need for a

careful assessment ofthe role ofpublic healthinterventions in the decline ofTB morbidityand mortality. It supports the hypothesis thatpublic health interventions had an impact onthe falling rates of mortality from infectious

diseases in generl and TB in paricular, begin-ning in the late 19th century. Certainly, publichealth alone could not have effected thischange. Changes in sources of immigration,for example, may have decreased the inci-dence ofTB in Minnesota in the late 19th andearly 20th centuries.'20 Improvements in nutri-tion and a rise in the general population'ssocial conditions may also have contributed tothe decline ofTB mortality. In the 20th cen-

tury, clinical interventions such as chemo-therapy played a substantial role in the drop inTB's toll.

As Nancy Tomes observes, McKeown'swork continues to reinforce scholarly tenden-cies to dismiss all but the broadest socialforces.'2' Within the literature we havereviewed here, the reverse is also true: manyattacking McKeown, bent on toppling an icon,tend to focus on only the narrowest publichealth interventions. 22 Moreover, figures suchas Newsholme and public health organiza-tions such as the Milbank Memorial Fundwere not simply neutral commentators but,rather, agents of public health who favoredrelatively easily engineered targeted inter-ventions-"drawing a bow at a venture andaiming straight at the mark"''23-over politi-cally and technically complex social inter-ventions that would affect a population muchwider than infected individuals and their

immediate contacts.The findings of Bayer et al., presented in

this issue of the Journal, underscore the wayin which a relatively easily engineered tar-geted intervention can produce impressiveresults.'24 Directly observed therapy played acrucial role in improving the effectivenessofTB drug treatment in cities that had faredmiserably before 1993, prior to an influx infederal funding for such therapy. Yet successsparks the risk that advocacy for a targetedintervention can take on the qualities of anorthodoxy resistant to a more complex per-spective. Bayer et al. also found that localesthat had achieved high treatment completionrates without universal directly observed ther-apy improved those rates only marginally,and at very high costs. However, those whohad advocated for directly observed therapyas a solution were loath to acknowledge thatanything less than universal directly observedtherapy could achieve the desired publichealth results. Accordingly, the work byBayer et al. developed a complex history ofits own as it traveled from one major medicaljournal to the next, receiving starkly mixedreviews that foreshadowed a storm of con-troversy. Understanding that institutionalcommitments often drive methodological cri-tiques that undercut heterodox findings, oneanonymous reviewer equated DOT (directlyobserved therapy) with "Dogma on Tuber-culosis."

On the one hand, then, we urge that anuanced historical and epidemiologicalapproach to the question ofTB be restored. Ifthe relative contribution of different inter-ventions and factors is to be sorted out, pur-suit of monocausal explanations for theretreat of TB, like monotypic intervention,is insufficient. On the other hand, while wemight call for more complexity in bothdebate and scholarship, our review suggeststhat this will be problematic. In addition tothe institutional, ideological, and politicalcommitments that individuals bring to theiranalyses and arguments, we are left with theharsh reality that TB was and is a diseaseinequitably distributed in society; immigrants,the poor, the homeless, and individualsinfected with HIV are its primary carriers.'25

Targeted interventions are Janus-faced:by definition, they visit upon the sick theburdens of intervention; at the same time,those very burdens can effectively reducethe toll of disease. And herein lies the fun-damental dilemma: in bringing together his-tory, policy, and advocacy in the case ofTB,we have created a false choice between broadsocial change and one or another targetedpublic health intervention, between nihilismand pragmatism. It is not only the case, then,that we must bring a diversity of methods

American Joumal of Public Health 11 13

Lantern slide warning of tubercular cows, New Jersey. Courtesy of the StateHistorical Society of Wisconsin, Visual Materials Archive.

July 1998, Vol. 88, No. 7


and evidence that embrace the specificity oftime, place, and person to bear on the prob-lem of TB. We must also acknowledge thefundamentally political nature of the prob-lem and of the solutions we offer as historiansand policy advocates. O

AcknowledgmentsAn earlier version of this paper was presented at the124th Annual Meeting of the American PublicHealth Association, November 1996, New YorkCity.

Ronald Bayer, Gerard E. Carrino, Daniel M.Fox, Barron Lerner, Richard Neugebauer, MaryNorthridge, David Rosner, Anne Stone, and mem-bers of the Colloquium on the History of Medicinegave us invaluable comments for which we aregrateful. We also thank the 3 anonymous reviewersfor their insightful reading and criticism of the man-uscnpt.

Endnotes1. Ronald Bayer and Laurence Dupuis, "Tubercu-

losis, Public Health, and Civil Liberties," AnnualReview ofPublic Health 16 (1995): 309.

2. Ronald Bayer and David Wilkinson, "DirectlyObserved Therapy for Tuberculosis: History of anIdea," Lancet 345 (June 17, 1995): 1545.

3. Money available to the Centers for Disease Con-trol and Prevention for TB control rose from $25million in 1991 to $104 million in 1993. Bayerand Wilkinson, "Directly Observed Ther-apy,"1547.

4. Ibid.5. Ibid., 1545-1548; C. Patrick Chaulk et al.,

"Eleven Years of Community-Based DirectlyObserved Therapy for Tuberculosis," Journal ofthe American Medical Association 274 (1995):945-95 1; S. E. Weis et al., "The Effect ofDirectly Observed Therapy on the Rates ofDrugResistance and Relapse in Tuberculosis," NewEngland Journal ofMedicine 330 (1994):1 179-1184; R. L. Hotchkiss, "Directly ObservedTreatment of Tuberculosis," New England Jour-nal ofMedicine 329 (1993): 135.

Interestingly, when discussed in isolation,directly observed therapy was not disavowed infavor of broader social interventions involvinghousing and institutions. Rather, it was initiallyopposed as too costly and labor-intensive. Amongpolicymakers striving to devise immediate, prag-matic policy solutions, directly observed ther-apy was challenged as an intrusion on autonomy,as a violation of the constitutional requirement ofdue process and the requirement that the leastrestrictive alternative be used, and as contrary tolegal requirements that restricting the behaviorof individuals with disabilities be based on indi-vidualized assessments. Bayer and Dupuis,"Tuberculosis, Public Health, and Civil Liber-ties," 318-319; "Appendix A: Dissents to Rec-ommendations," in The Tuberculosis Revival.Individual Rights and Societal Obligations in aTime ofAIDS (New York: United Hospital Fund,1992) 30-32; George J. Annas, "Control ofTuberculosis-The Law and the Public's Health,"New England Journal ofMedicine 328 (1993):587-588.

6. Deborah Wallace, "Roots of Increased HealthInequality in New York," Social Science andMedicine 31 (1990): 1219-1227; Karen Brud-ney and Jay Dobkin, "Resurgent Tuberculosisin New York City: Human ImmunodeficiencyVirus, Homelessness, and the Decline of Tuber-culosis Control Programs," American Review ofRespiratory Diseases 144 (1991): 745-749;Bayer and Dupuis, "Tuberculosis, Public Health,and Civil Liberties," 307-326; J. Lederberg, R. E.Shope, and S. C. Oaks, eds. Emerging Infections:Microbial Threats to Health in the United States(Washington, D.C.: National Academy Press,1992); Recommendations of the MassachusettsCommission for the Elimination of Tubercu-losis (Boston: Massachusetts Department of Pub-lic Health, 1992).

7. Brudney and Dobkin, "Resurgent Tuberculosis,"747-748; Sheldon Landesman, "Commentary:Tuberculosis in New York City-The Conse-quences and Lessons of Failure," American Jour-nal ofPublic Health 83 (1993): 766-768.

8. Landesman, "Commentary," 767.9. Ibid., 766.

10. Barron H. Lerner, "New York City's Tubercu-losis Control Efforts: The Historical Limitationsof the 'War on Consumption,"' American Jour-nal ofPublic Health 83 (1993): 758.

11. Ibid.12. Georgina D. Feldberg, Disease and Class: Tuber-

culosis and the Shaping ofModern North Amer-ican Society (New Brunswick, N. J.: RutgersUniversity Press, 1995).

13. Ronald Bayer, "Does Anything Work? PublicHealth and the Nihilist Thesis" (paper presentedat the 124th Annual Meeting of the AmericanPublic Health Association, New York City,November 1996).

14. James McCarthy, "Broad versus TargetedApproaches to Population Control" (paper pre-sented at the 124th Annual Meeting of the Amer-ican Public Health Association, New York City,November 1996); Steven W. Sinding, John A.Ross, and Allan G. Rosenfield, "Seeking Com-mon Ground: Unmet Need and DemographicGoals," International Family Planning Perspec-tives 20 (March 1994): 23-32.

15. Richard Neugebauer, "The Uses of Psychoso-cial Epidemiology in Promoting Refugee Health,"American Journal ofPublic Health 87 (1997):726-728.

16. Major works by Thomas McKeown and col-leagues include T. McKeown and R. G. Record,"Reasons for the Decline of Mortality in Englandand Wales during the Nineteenth Century, Popu-lation Studies xvi (1962): 94-122; T. McKeown,R. G. Record, and R. D. Tumer, "An Interpreta-tion of the Decline of Mortality in England andWales during the Twentieth Century," Popula-tion Studies xxix (1975): 391-422; McKeown,The Modern Rise ofPopulation (New York: Aca-demic Press, 1976); and McKeown, The Role ofMedicine: Dream, Mirage, or Nemesis? (Prince-ton, N.J.: Princeton University Press, 1979).

17. Older work in the tradition ofMcKeown includesthat of John B. McKinlay and Sonja M. McKin-lay, "The Questionable Contribution of MedicalMeasures to the Decline of Mortality in the UnitedStates in the Twentieth Century," Milbank Memo-rial Fund Quarterly (Summer 1977): 405-428;0. R. Omran, "The Epidemiological Transition,"Milbank Memorial Fund Quarterly 49 (1971):

509-538; Omran, "Epidemiological Transitionin the United States," Population Bulletin 32(1977): 3-42; and Omran, "A Century of Epi-demiologic Transition in the United States," Pre-ventive Medicine 6 (1977): 30-51. More recentwork includes that of Sumit Guha, "The Impor-tance of Social Intervention in England's Mor-tality Decline: The Evidence Reviewed," SocialHistory ofMedicine (1994): 89-113; and S. Diaz-Briquets, "Determinants ofMortality Transition inDeveloping Countries before and after the SecondWorld War," Population Studies 35 (1981):399-411. Historical works, in addition to thosedescribed in this paper, that were influenced byMcKeown include S. C. Farrow, "McKeownReassessed," British Medical Journal 294 (1987):1631-1632; J. M. Winter, "The Decline of Mor-tality in Britain, 1870-1950," in T. Barker andM. Drake, eds., Population and Society in Britain,1850-1980 (New York: New York UniversityPress, 1982); and Allan Mitchell, "An InexactScience: The Statistics of Tuberculosis in LateNineteenth Century France," Social History ofMedicine 3 (1990): 387-403.

Among social science publications in whichMcKeown is influential or prominently cited areworks on social inequality and health and workson theories ofhealth status change in populations.These include Richard G. Wilkinson, UnhealthySocieties: The Afflictions ofInequality (New York:Routledge, 1996): 30-33; Denny Vager6,"Inequality in Health-Some Theoretical andEmpirical Problems," Social Science and Medi-cine 32 (1991): 367-371; and Christopher J. L.Murray and Lincoln C. Chen, "In Search of aContemporary Theory for Understanding Mor-tality Change," Social Science and Medicine 36(1993):143-155.

Policy papers apply McKeown's work toproblems of developing countries. See, for exam-ple, John C. Caldwell, "Health Transition: TheCultural, Social and Behavioral Determinants ofHealth in the Third World," Social Science andMedicine 36 (1993): 125-135, and World Bank,World Development Report 1993: Investing inHealth (New York: Oxford University Press,1993).

18. William Johnston, The Modern Epidemic: A His-tory ofTuberculosis in Japan (Cambridge, Mass.:Council on East Asian Studies, Harvard Univer-sity, 1995), 34.

19. See McKeown, The Role of Medicine, 33;McKeown calculated that 40% of the reduction inmortality between 1848 and 1971 was due to adecline in airborne diseases; 21%, to water- andfoodbome diseases; 13%, to other conditions;and 26%, to a decrease in conditions not asso-ciated with microorganisms.

20. This critique of the McKeown thesis is fromSimon Szreter, "The Importance of Social Inter-vention in Britain's Mortality Decline c.1850-1914: A Re-interpretation of the Role ofPublic Health," The Societyfor the Social HistoryofMedicine 1 (1988): 7-10.

21. McKeown, The Role ofMedicine, 59.22. Ibid., 59-65.23. Szreter, "Social Intervention," 13, 16-17. Using

McKeown's own data, Szreter argues that risingbronchitis rates in the second half of the 19thcentury weaken McKeown's assertion that thefall of respiratory rates in general or TB rates inparticular was critical to the decline in that cen-

July 1998, Vol. 88, No. 71 114 American Journal of Public Health

tury's mortality rates. Szreter also suggests that asustained decline in TB rates did not occur untilafter 1867. He then argues:

Improvement in respiratory TB would, then,no longer appear to have been either thechronologically prior or the quantitativelypredominant feature of the nineteenth cen-tury mortality decline in England and Wales.According to the logic of McKeown's ownarguments, the foregoing would indicate aprimary role for sanitary reform and publichealth measures, rather than rising nutri-tional levels or living standards. The chang-ing incidence of mortality from respiratoryTB in Victorian Britain, rather than beingcast in the role of a leading and determininginfluence, can be seen as a dependent func-tion of the general intensity and frequency ofother debilitating diseases. (16-17)

For further discussion of the relationship betweenintercurrent infections and TB, see Alex Mer-cer, Disease, Mortality and Population in Tran-sition, Epidemiological-Demographic Changein England since the Eighteenth Century as Partof a Global Phenomenon (Leicester: LeicesterUniversity Press, 1990), 97-123.

24. Szreter, "Social Intervention," 13.25. Ibid. For other studies arguing that the public

health movement in Great Britain was a signifi-cant cause of the general mortality decline, seeAnthony S. Wohl, Endangered Lives: PublicHealth in Victorian Britain (Cambridge, Mass.:Harvard University Press, 1983); George Rosen,A History ofPublic Health (New York: MD Pub-lications, 1958); G. T. Griffith, Population Prob-lems in the Age ofMalthus (New York: Cam-bridge University Press, 1926); M. C. Buer,Health, Wealth and Population in the Early Daysofthe Industrial Revolution (London: Routledge,1926); P. E. Razzell, "An Interpretation of theModem Rise of Population in Europe: A Cri-tique," Population Studies 28 (1974): 5-17; andJ. C. Riley, The Eighteenth-Century Campaign toAvoid Disease (London: Macmillan, 1987).

26. Local studies include Anne Hardy, The EpidemicStreets: Infectious Disease and the Rise ofPre-ventive Medicine, 1856-1900 (Oxford, England:Clarendon Press, 1993); Neil McFarlane, "Hos-pitals, Housing, and Tuberculosis in Glasgow,"Social History ofMedicine 2 (1989): 59-85; andGillian Cronje, "Tuberculosis and MortalityDecline in England and Wales, 1851-1910," inUrban Disease and Mortality in Nineteenth Cen-tury England, ed. Robert Woods and John Wood-ward (London and New York: Batsford Aca-demic and Educational and St. Martin's Press,1984).

27. Hardy, Epidemic Streets, 249-250, 265.28. McFarlane, "Hospitals, Housing, and Tubercu-

losis," 59.29. Curtin finds that the greatest mortality gains

between 1839 and 1870 were overseas, wheretargeted efforts to improve sanitary conditionswere implemented. Over the next 20 years, assanitary engineering became more common inEurope, the greatest gains in mortality took placeat home. Philip Curtin, Death by Migration:Europe 's Encounter with the Tropical World inthe Nineteenth Century (Cambridge, England:Cambridge University Press, 1989): 80, 159.

30. Between 1837 and 1846, TB caused 50% of alldeaths among the troops in Britain. During the1860s, it caused 38% of the deaths of Britishtroops and 22% of the deaths of French troopsposted in their respective countries. However, itcontributed to only 7% of all deaths among troopsin Madras, Algeria, and the Windward and Lee-ward islands during the 1840s. Curtin, Death byMigration, 76. In the 1820s and 1830s, less than10% of deaths among British troops abroad wereattributable to TB, as compared with 55% ofdeaths among British troops in England. In theperiod from the 1840s to the 1860s, although TBmortality dropped 75% among troops in Britain,the mortality rate from TB continued to be higherin Britain (63.38 deaths per 1000 troops in Britainvs 7.53 deaths per 1000 in the Windward andLeeward Command and 6.99 per 1000 in Madrasfor the period 1837 to 1847; 3.16 deaths per 1000troops in Britain vs 1.6 per 1000 in the Wind-ward and Leeward Command and 2.86 per 1000in Madras for the period 1859 to 1867). Curtin,Death by Migration, 30, 35.

31. Curtin, Death by Migration, 42, 76. See also S.Guha, "Nutrition, Sanitation, Hygiene and theLikelihood of Death: The British Army in India,"Population Studies 47 (1993): 385-401.

32. Mitchell, "Inexact Science," 387.33. Nancy Tomes, "The White Plague Revisited,"

Bulletin of the History ofMedicine 63 (1989):467, 468.

34. Linda Bryder, Below the Magic Mountain: ASocial History of Tuberculosis in Twentieth-Century Britain (Oxford, England: ClarendonPress, 1988), 30; Barbara Bates, BargainingforLife: A Social History of Tuberculosis,1876-1938 (Philadelphia: University of Penn-sylvania Press, 1992).

35. See also Lemer, "New York City's TuberculosisControl Efforts," 758-766.

36. The beginning of the decline in TB mortality inEngland is highly contested. McKeown arguesthat the decline began in the 1830s, if not earlier,whereas Szreter counters that the decline did notbegin before 1867, if not later. Anne Hardy sup-ports Szreter's figures and writes that "the begin-ning of the decline may be pushed even furtherforward, towards the 1880s." Hardy, EpidemicStreets, 214. Scholars consistently maintain thatthe decline in the United States began in the1870s, although there are no data comparable tothe Registrar General's mortality statistics forEngland and Wales supporting this estimate.

37. The overall annual mortality rate from TB in 1900was approximnately 200 per 100 000 population inthe United States. Among Blacks, the rate was400 per 100 000. Barbara Gutmann Rosenkrantz,"Introductory Essay: Dubos and Tuberculosis,Master Teachers," in Rene Dubos and Jean Dubos,The White Plague: Tuberculosis, Man, and Soci-ety (New Brunswick, N.J.: Rutgers UniversityPress, 1952, 1987), xiv-xv, note 1. In urban areas,the mortality rate was higher. In New York City,for example, the mortality rates were 428 per100 000 in 1870 and 256 per 100 000 in 1890.The New York City mortality rate varied from 49per 100 000 on the upper West Side to 776 per100000 in the tenement district of lower Manhat-tan. Sheila M. Rothman, Living in the Shadow ofDeath: Tuberculosis and the Social ExperienceofIllness in American History (New York: Basic


Books, 1994), 184. In Philadelphia, the mortalityrate in 1870 was 350 per 100000; this rate haddropped to less than 200 per 100 000 by 1910.By 1925, the mortality rate was less than 100 per100 000. Bates, Bargainingfor Life, 3 15.

In Pennsylvania, the TB mortality rate firstdropped to below 10 deaths per 100000 in 1956.In England and Wales, for all forms of TB, themortality rate was close to 400 per 100 000 in1865; it declined to less than 200 per 100000 inthe decade before 1900 and to 100 per 100 000 in1930. Bryder, Below the Magic Mountain, 7. Thevast majority of cases (98% of all pulmonarycases and 70% ofnonpulmonary cases) were dueto infection with the human TB bacillus. Theremaining 2% of pulmonary cases and 30% ofnonpulmonary cases were attributable to bovineinfection caused by infected meats and, moreoften, infected milk. Bryder, Below the MagicMountain, 3.

38. In Great Britain, there were approximately 17hospitals for the treatment of TB (representingabout 1 100 beds) by 1893. By 1907, there were96 institutions for the "open-air treatment" ofTB (4081 beds). Not counting Poor Law beds,there were, in 1911, 5500 beds in 259 local andvoluntary sanatoriums and isolation hospitals (97local institutions, 79 voluntary institutions, and 83isolation hospitals; 1300 beds were provided bylocal authorities and 4200 by voluntary institu-tions). By 1916, the number of beds hadincreased to nearly 12 000. Of the 11 893 beds,6072 were provided by local authorities and 5821by voluntary institutions. In 1920, there were145 781 beds, 8845 provided by local authori-ties and 6936 by voluntary institutions. Bryder,Below the Magic Mountain, 23, 32, 44. In theUnited States, in comparison, there were 34 sana-toriums (representing 4485 beds) in 1900; by1925, there were 536 sanatoriums with more than700 000 beds. Rothman, Shadow ofDeath, 198.In Pennsylvania-the state with the most exten-sive system of TB institutions-there wereapproximately 2000 beds by 1910 and slightlyless than 4000 by 1920. Bates, BargainingforLife, 28 1.

39. Bryder, Below the Magic Mountain, 24.40. Ibid., 36-41.41. Ibid., 68-69.42. F. B. Smith, The Retreat of Tuberculosis,

1850-1950 (London: Croom Helm, 1988),244-245.

43. Sheila M. Rothman, "The Sanitorium Experi-ence: Myths and Realities," in The Tuberculo-sis Revival: Individual Rights and Societal Oblig-ations in a Time ofAIDS (New York: UnitedHospital Fund, 1992); 72.

44. Rothman, Shadow ofDeath, 209.45. Rothman, "Sanitorium Experience," 73.46. Regarding the TB control movement as a whole,

Bates similarly asserts that "[n]o effects of thetuberculosis movement can be seen in the fallingmortality rates at least until the 1 920s and, eventhen, they are not clear." Bates, BargainingforLife, 331.

47. Ibid., chaps. 2, 5, and 17.48. Bates observes that the proportion of patients

dying in sanatoriums increased from 2% to morethan 18% from 1909 to 1916. Bates, Bargain-ingforLife, 165-166.

49. Ibid., 157-158.50. Ibid., 318.

July 1998, Vol. 88, No. 7 American Journal of Public Health 11 15


51. Ibid., graph 17-2, 319.52. Ibid., 318, 321-325. Among the other causes,

Bates identifies an increase in per capita income,improvements in ventilation in a wide range ofinstitutions, improvements in the conditions andspeed of immigrant transatlantic travel, and nutri-tional advances that would have come hand inhand with innovations in preservation and trans-portation.

53. Tomes points out that, although their publicationsconveniently dovetailed with the resurgence ofTB, scholars turned their attention to the socialhistory of TB well before the resurgence gainedthe attention of public health officials and poli-cymakers. Tomes, "White Plague Revisited," 467.Although these works, which were published afterpolicy debate over how to control the resurgenceof TB began, resonate with contemporary andhistorical debate over the retreat of the disease,they were neither inspired nor driven by suchdebate.

54. Leonard G. Wilson, "The Historical Decline ofTuberculosis in America: Its Causes and Signif-icance," Journal ofthe History ofMedicine andAllied Sciences 45 (1990): 366-396; Samual H.Preston, Mortality Patterns in National Popula-tions (New York: Academic Press, 1976).

55. In 1911, more than 60% of TB beds in Britainwere located in Poor Law infirmaries.

56. Specifically, the Irish Poor Law of 1847, passedin the wake of the potato famine, prohibited wan-dering abroad, moving from one locale toanother, or begging in public. Newsholme,"Poverty and Disease, as Illustrated by the Courseof Typhus Fever and Phthisis in Ireland," Epi-demiological Section, Proceedings ofthe RoyalSociety ofMedicine 1 (1907-1908): 14.

57. An exponential phenomenon plotted logarithmi-cally will produce a straight line, the slope ofwhich corresponds to the rate of change.

58. Wilson, "Decline of Tuberculosis," 389-395.59. Leonard G. Wilson, "The Rise and Fall of Tuber-

culosis in Minnesota: The Role of Infection,"Bulletin of the History ofMedicine 66 (1992):16-52.

60. Ibid., 50.61. Linda Bryder, "'Not Always One and the Same

Thing': The Registration of Tuberculosis Deathsin Britain, 1900-1950," Social History ofMedi-cine 9 (August 1996): 253-265.

62. Neither Newsholme nor Wilson took into accountthe lag time between an intervention (segrega-tion) and the decline in TB-specific mortalityrates as a consequence of reduced infection inci-dence rates. In addition, mortality rates are afunction of incidence and case fatality rates dur-ing the periods in question; better care mightconceivably have reduced the case fatality rate.

63. For further criticism of Newsholme's studies,see John M. Eyler, Sir Arthur Newsholme andState Medicine, 1885-1935 (Cambridge, Eng-land: Cambridge University Press, 1997):182-186.

64. Wilson, "Decline of Tuberculosis," 367.65. Newsholme, "The Notification of Phthisis

Pulmonalis," The Practitioner 67 (1901): 27;Newsholme, "An Inquiry into the PrincipalCauses of the Reduction in the Death-Rate fromPhthisis during the Last Forty Years, with SpecialReference to the Segregation of PhthisicalPatients in General Institutions," Journal ofHlygiene 6 (July 1906): 320; Newsholme, The

Last Thirty Years in Public Health: Recollec-tions and Reflections on My Official and Post-Official Life (London: George Allen & Unwin,1936): 130-13 1. Newsholme also held thatimproved sanitation (including drainage andimproved ventilation and lighting) and improve-ment in the milk supply interdicted the trans-mission of the tubercle bacillus. Newsholme,"An Inquiry into the Principal Causes," 350-353;Newsholme, "The Relative Importance of theConstituent Factors Involved in the Control ofPulmonary Tuberculosis," Transactions of theEpidemiological Society of London 25(1905-1906): 39.

66. Newsholme, "An Inquiry into the PrincipalCauses," 374, 375. Significantly, while News-holme touted the benefits of segregation, he didnot credit sanatoriums themselves or knowledgeof the infectiousness ofTB with the retreat of thedisease: "The decrease of phthisis where it hasoccurred cannot have been due to improved edu-cation as to the infectivity of the disease or to theintroduction of Sanatoria, both of these havingoccurred after well-marked decrease had set in,and the Sanatoria up to the present time havingbeen insignificant in number relative to the amountofthe disease."

67. Newsholme, The Last Thirty Years in PublicHealth, 127.

68. For example, see Edgar Sydenstricker, "TheDecline in the Tuberculosis Death Rate in Cat-taraugus County," Milbank Memorial FundQuarterly 61 (1928): 41-50; Jean Downes, "AStudy of the Effectiveness of Certain Adminis-trative Procedures in Tuberculosis Control," Mil-bank Memorial Fund Quarterly 14 (1936):317-327; Downes, "The Risk of Mortality amongOffspring of Tuberculosis Patients in a RuralArea in the Nineteenth Century" American Jour-nal ofHygiene 26 (1937): 557-569; Downes,"The Effect of Tuberculosis on the Size ofFam-ily," Milbank Memorial Fund Quarterly 17 (July1939); 274-287; Downes, "Salient Points ofAttack against Tuberculosis," Milbank MemorialFund Quarterly 18 (1940): 44-60.

Cattaraugus County was of particular inter-est to investigators. Despite steady declines inthe TB mortality rate in New York State as awhole and in comparable rural counties amount-ing to a 32% decrease between 1913 and 1923,the mortality rate of Cattaraugus County hadremained constant since about 1913. MilbankMemorial Fund Quarterly 2 (April 1924): 1. TheMilbank Memorial Fund selected "typical"American communities "to demonstrate...whether by intensive application ofknown healthmeasures the extent of sickness in the UnitedStates can be further and materially diminishedand mortality rates further and substantiallyreduced, and whether or not such practical resultscan be achieved in a relatively short period oftime and at a per capita cost which communi-ties will willingly bear." Quoted in C. E. A.Winslow, Health on the Farm and in the Village:A Review and Evaluation of the CattaraugusCounty Health Demonstration with Special Ref-erence to Its Lessonsfor Other Rural Areas (NewYork: Macmillan, 1931). See also "ControllingTuberculosis in Cattaraugus County," MilbankMemorial Fund Quarterly 2 (January 1925): 5.

69. The Framingham study sought to demonstratethat public health interventions such as case find-

ing, clinical treatment, public health nursing, andhealth education reduced TB-specific mortalityover time. Framingham Monograph No. 10. Gen-eral Series IV. Final Summary Report,1917-1923 Inclusive (Framingham, Mass.:National Tuberculosis Association, 1924).

70. Downes, "Effectiveness of Certain AdministrativeProcedures," 323. Specifically, Downes found a15% to 25% reduction in infection rates amongfamilies of individuals with TB who receivedsanatorium care from 1931 to 1935 relative to agroup of families followed from 1923 to 1930,before implementation of the sanatorium pro-gram. Downes concluded: "The difference in theinfection rate for the two groups of families (20.0±4.56) is more than four times its probable errorand may be considered as statistically significantsince such a difference will arise by chance lessoften than once in a hundred times." Downes,"Effectiveness of Certain Administrative Proce-dures," 322-323.

71. Downes, "Salient Points of Attack," 59.72. Wade Hampton Frost, "How Much Control of

Tuberculosis?" American Journal of PublicHealth and The Nation's Health 27 (1937): 763.

73. The turn-of-the-century American bacteriologistwho confirmed German findings that tuberculincould be used to detect TB in cattle saw littlescientific consensus when it came to bovine TB:"There is scarcely a subject related to agricul-ture or public health that has occasioned as muchor as bitter discussion, or has led to the expressionof so many divergent views as this one of tuber-culosis in cattle." Quoted in J. Arthur Myers andJames H. Steele, Bovine Tuberculosis Controlin Man and Animals (St. Louis, Mo.: Warren H.Green, 1969): 57.

74. J. M. Grange, "Human Aspects of Mycobacteri-umbovis," in Mycobacteriumbovis, ed. Charles 0.Thoen and James H. Steele (Ames: Iowa StateUniversity Press, 1995): 29.

75. Grange, "Mycobacteriumbovis," 30. Speakingat the International Tuberculosis Congress inLondon in 1901, Koch argued that bovine TBmerited little attention from public health author-ities: "I should estimate the extent of infectionby the milk and flesh of tubercular cattle and thebutter made of their milk as hardly greater thanthat of hereditary transmission, and I thereforedo not deem it advisable to take any measuresagainst it." Quoted in Myers and Steele, BovineTuberculosis, 57.

76. Grange, "Mycobacteriumbovis," 31.77. Bryder, Below the Magic Mountain, 133.78. Myers and Steele, Bovine Tuberculosis, 60.79. Smith, Retreat of Tuberculosis, 185.80. Ibid., 186.81. Myers and Steele, Bovine Tuberculosis, 75-84.82. Although the downturn in TB mortality coin-

cides with the influenza pandemic in the UnitedStates, Alfred Crosby offers little to suggest acausal relationship. Writes Crosby, "Spanishinfluenza was not a typical communicable dis-ease in its choice of people to infect and kill.Unlike tuberculosis, typhoid fever, and venerealdisease, it did not show a clear preference forthe poor, the ill-fed, ill-housed, and shabbilyclothed. Sometimes there was a discernible cor-relation between flu, pneumonic complications,and crowded living conditions. . .but by and largethe rich died as readily as the poor." Immigrants,Crosby notes, had a higher death rate from

11 16 American Journal of Public Health July 1998, Vol. 88, No. 7

influenza, however. Alfred W. Crosby, Ameri-ca 's Forgotten Pandemic: The Influenza of1918(Cambridge, England: Cambridge UniversityPress, 1989): 227. Decreases in TB mortality fol-lowing the federal bovine eradication programalso occurred despite decreases in breast-feed-ing and increased use of infant formula. RimaD. Apple, "'To Be Used Only Under the Direc-tion of a Physician': Commercial Infant Feed-ing and Medical Practice, 1870-1940," Bulletinofthe History ofMedicine 54 (1980): 402-417.

83. Pasteurization had become widespread in majorAmerican cities (those with populations of morethan 75 000) by 1921. Small cities (with popula-tions of less than 10000) also pasteurized exten-sively by 1921. S. Henry Ayers, The Present Sta-tus of the Pasteurization ofMilk, United StatesDepartment of Agriculture Bulletin No. 342, Jan-uary 8, 1916; revised October 19, 1922, 5-7.

84. Myers and Steele, Bovine Tuberculosis, xiv.85. Ibid., xiv.86. George Rosen, Preventive Medicine in the United

States, 1900-1975, Trends and Interpretations(New York: Prodist, 1977): 30.

87. Ibid., 30.88. Ibid., 29.89. Rosen, Public Health, 336.90. Ayers, Pasteurization ofMilk, 5-7.91. Myers and Steele, Bovine Tuberculosis, 70.92. George Jasper Wherrett, The Miracle ofthe Empty

Beds: A History of Tuberculosis in Canada(Toronto: University of Toronto Press, 1977): 47.

93. Thoen and Steele, Mycobacteriumbovis, xx.94. Grange, "Mycobacteriumbovis," 36, 38.95. Smith, Retreat ofTuberculosis, 175-194; Bryder,

Below the Magic Mountain, 137-138, 246-247.96. Smith, Retreat of Tuberculosis, 175-194; Bry-

der, Below the Magic Mountain, 137-138,246-247.

97. Smith, Retreat of Tuberculosis, 186. From 1951to 1965, Paris scientists estimated that approxi-mately 3% of cases studied were caused by thebovine strain. Myers and Steele, Bovine Tuber-culosis, 65.

98. Smith, Retreat of Tuberculosis, 193.99. McKeown, The Role ofMedicine, xii.

100. Szreter, "Social Intervention," 33.101. McKeown, The Role ofMedicine, xvi.102. McKeown, "A Sociological Approach," in Med-

ical History and Medical Care (London: OxfordUniversity Press, 1971): 4.

103. Szreter, "Social Intervention, " 33.104. Charles Webster, "The Origins of Social Medi-

cine in Britain," Society for the Social History

ofMedicine Bulletin 38 (1986): 53.105. Szreter, "Social Intervention," 33.106. McKeown, The Role ofMedicine, xiii.107. Eyler, Newsholme.108. Ibid., 148-158.109. Hardy, The Epidemic Streets, 258, 262.1 10. Ibid., 264.111. Newsholme, The Last Thirty Years in Public

Health, 128; Hardy, The Epidemic Streets, 26.112. Prior to the bacteriological revolution of the

1870s and 1880s, public health in the UnitedStates was largely the domain of the sanitarians,reformers, and, occasionally, physicians devotedto eliminating the filth and squalor believed tocause disease via miasmas. Although socialreformers and sanitarians maintained a presencein the public health movement after the bacteri-ological revolution, public health was increas-ingly dominated by a new set of public healthprofessionals, many educated in northeasternmedical schools and further trained in Europeanlaboratories. John Duffy, The Sanitarians: A His-tory ofAmerican Public Health (Urbana: Uni-versity of Illinois Press, 1990), and Paul Starr,The Social Transformation ofAmerican Medi-cine: The Rise ofa Sovereign Profession and theMaking of a Vast Industry (New York: BasicBooks, 1982).

1 13. The question ofwhether the "new public health"differed sharply from the old-long at the centerof historical debate-hinges on different under-standings of the relationship between publichealth and clinical medicine in the bacteriologicalera. One contingent of historians concludes thatbacteriology helped to refine and focus the effortsof those 19th-century reformers who had devel-oped a sense of public health as a communalconcern. See Bruno Latour, The Pasteurization ofFrance (Cambridge, Mass.: Harvard UniversityPress, 1988). Duffy, Sanitarians, chap. 14; Rosen,Public Health, chaps. 7 and 8; Lloyd Stevenson,"Science Down the Drain: On the Hostility ofCertain Sanitarians to Animal Experimentation,Bacteriology, and Immunology," Bulletin oftheHistory ofMedicine 29 (1995): 1-26; BarbaraGutmann Rosenkrantz, Public Health and theState: Changing Views in Massachusetts,1842-1936 (Cambridge, Mass.: Harvard Uni-versity Press, 1972); Starr, Social Transforma-tion ofAmerican Medicine; and Elizabeth Fee,Disease and Discovery: A History of the JohnsHopkins School ofHygiene and Public Health,1916-1939 (Baltimore: Johns Hopkins UniversityPress, 1987).


However, some scholars and social theo-rists, representing a second contingent, argue thatbacteriology narrowed the scope ofpublic healthtoo much: the sanitarians had maintained a holis-tic concern about the individual in his or herenvironment that embraced health, moral status,and living conditions. In such analyses, the pub-lic health movement became closely associatedwith clinical medicine, which, in the 1960s and1970s, was under sharp attack. Rosenkrantz, Pub-lic Health and the State, 179; Edward S. Golub,The Limits ofMedicine: How Science ShapesOur Hope for the Cure (New York: TimesBooks, 1994); Daniel M. Fox, Power and Ill-ness: The Failure and Future ofAmerican HealthPolicy (Berkeley: University of California Press,1993); Rene Dubos, Mirage ofHealth: Utopias,Progress and Biological Change (NewBrunswick, N.J.: Rutgers University Press, 1959,1993); and Dubos and Dubos, The White Plague.

114. Elizabeth Fee, "Public Health, Past and Present:A Shared Social Vision," introduction to GeorgeRosen, A History ofPublic Health, expandededition (Baltimore: Johns Hopkins UniversityPress, 1993): x.

115. Susan Reverby and David Rosner, "Beyond 'theGreat Doctors,"' in Health Care in America,Essays in Social History, ed. Reverby and Rosner(Philadelphia: Temple University Press), 1979: 4.

116. David Musto, "Quarantine and the Problem ofAIDS," in AIDS: The Burdens ofHistory, ed.Elizabeth Fee and Daniel M. Fox (Berkeley: Uni-versity of California Press, 1988): 67-85.

117. Charles E. Rosenberg, "Disease and Social Orderin America: Perceptions and Expectations," inExplaining Epidemics and Other Studies in theHistory ofMedicine (Cambridge, England: Cam-bridge University Press, 1992): 259-260.

118. Ibid.119. Ibid., 260.120. Wilson, "Minnesota," 19-21.121.Nancy Tomes, "The Private Side of Public

Health: Sanitary Science, Domestic Hygiene,and the Germ Theory," Bulletin of the HistoryofMedicine 64 (1990): 509-539.

122. See, for instance, McFarlane, "Hospitals, Hous-ing, and Tuberculosis," 59.

123. Newsholme, "Public Health Authorities," 456.124. Ronald Bayer et al. "Directly Observed Therapy

and Treatment Completion for Tuberculosis inthe United States: Is Universal Supervised Ther-apy Necessary?" American Journal of PublicHealth 88 (1998): 1052-1058.

125. See note 37.

July 1998, Vol. 88, No. 7 American Joural of Public Health 1117

amjph00019-0103

Documents