... pecteddisease,theriskapproximated15per-centforNegroesand5percentforwhites.Forolderpersons with minimalorsuspectedtuber-culosis,theriskwasapproximately4percentforNegroesand2percentforwhites.Therisk of reactivationamongNegroeswasnotassociated with degree of skinpigmentation.Long-termfollowup of persons with inactivediseaseappearstobeanimportantmeans of detectingactivetuberculosis.NOTE:Personsinterestedinthetablesgivingmoredetails of thelifetableanalysesmayobtainthemfromtheauthor.REFERENCES(1)Puffer,R.R.,Stewart,H.C.,andGass,R.S.:TuberculosisstudiesinTennessee.Subsequentcourse of casesobservedinWilliamsonCounty.Am.J.Hyg.28:490-507(1938).(2)Reisner,D.,andDownes,J.:Minimaltubercu-louslesions of thelung;theirclinicalsignifi-cance.Am.Rev.Tuberc.51:393-412(1945).(8)Lincoln,N.S.,Bosworth,E.B.,andAlling,D.W.:Theafter-history of pulmonary tuberculosis.III.Minimaltuberculosis.Am.Rev.Tuberc.70:15-31(1954).(4)Kallquist,I.:Long-termprognosisin pulmonary tuberculosisdetectedbymassradiography.Acounty-widesurvey with controlsandacom-parisonbetweentwomasssurveys with aseven-yearinterval.Actatuberc.scandinav.,1958,supp.44,p.177.(5)Groth-Petersen,E.,Knudsen,J.,andWilbek,E.:Epidemiologicbasis of tuberculosiseradicationinanadvancedcountry.Bull.WorldHealthOrgan.21:549(1959).(6)Frimodt-M0ller,J.:Acommunity-widetuber-culosisstudyinasouthIndianruralpopula-tion.Bull.WorldHealthOrgan.22:61-170(1960).(7)Comstock,G.W.:TuberculosisstudiesinMuscogeeCounty,Georgia.I.Community-widetuberculosisresearch.Pub.HealthRep.64:259-263(1949).(8)Burke,M.H.,Schenck,H.C.,andThrash,J.A.:TuberculosisstudiesinMuscogeeCounty,Georgia.II.X-rayfindingsinacommunity-widesurveyanditscoverageasdeterminedbyapopulationcensus.Pub.HealthRep.64:263-290(1949).(9)Palmer,C.E.,Shaw,L.W.,andComstock,G.W.:Communitytrials of BCGvaccination.Am.Rev.Tuberc.77:877-907(1958).(10)Edwards,L.B.,andPalmer,C.E.:Isolation of "atypical"mycobacteriafromhealthypersonsAm.Rev.Resp.Dis.80:747-749(1959).(11)NationalTuberculosisAssociation:Diagnosticstandardsandclassification of tuberculosis.NewYork,N.Y.,1961.(12)Newell,R.R.,Chamberlain,W.E.,andRigler,L.:Descriptiveclassification of pulmonary shad-ows.Arevelation of unreliabilityintheroentgenographicdiagnosis of tuberculosis.Am.Rev.Tuberc.69:566-584(1954).(13)Yerushalmy,J.,etal.:Anevaluation of therole of serialchestroentgenogramsinestimat-ingtheprogress of diseaseinpatients with pulmonary tuberculosis.Am.Rev.Tuberc.64:225-248(1951).GraduateTrainingTheConference of BiologicalEditorsatitsannualmeetinginNewOrleansinMarch1962votedto:1.EndorsetheprinciplethatEnglishdepartmentsinsecondaryschoolsandcollegesgivetraininginwritingscientificreports.2.Recommendthatteachers of biologicalsciencesrequirewrittenreportsbystudentsintheircourses.3.Recommendthatteachers of sciencecorrecttheEnglishinsciencereportswrittenfortheircourses.4.Endorseanexploration of possibilities of greatercooperationbetweendepartments of Englishorjournalismanddepartments of biologicalsciences.Anexamplewouldbetopermitatermpaperonasciencesubjecttoservebothdepartments.5.Endorsetheprinciplethatthewritingandpublication of reportsberegardedasanessentialandintegralphase of fulfilledresearch.470PublicHealthReportsUntreatedInactive Pulmonary TuberculosisRisk of ReactivationGEORGEW.COMSTOCK,M.D.,Dr.P.H.THERELAPSINGTENDENCY of pul¬monarytuberculosisiswidelyknownandwelldocumented.In1938,Puffer,Stewart,andGass(1)reportedfromtheWilliamsonCounty(Tenn.)TuberculosisStudythat12percent of whitepersonsclassifiedashavingminimalarrestedtuberculosisand15percent of thosehavinglatentapicaltuberculosishadbe¬comeworseduringa3-yearperiod of observa¬tion.EeisnerandDownes(2)investigatedtherelapserateamongasample of persons with productive,fibrotic,orcalcificminimaltubercu¬losiswhoattendedtheambulatorychestclinics of theNewYorkCityDepartment of Health.Theyfound5percent of whitesand14percent of nonwhiteshaddevelopedactivediseasein5years.Amongagroup of upstateNewYorkpatients,diagnosedbythestaff of HermannM.BiggsMemorialHospitalbetween1937and1947ashavingminimalarrestedtuberculosis,therisk of developingactivetuberculosisdur¬ingthe10yearsfollowingdiagnosiswas13percent(3).Similarstudieshavebeenmadeinothercountries.Kallquist(^),reportingfromSwe¬denontheexperience of 312personsconsideredtohaveinactiveorprobablyinactivetubercu¬losis,notedthat8percenthadshownevidence of activediseasewithinaperiod of 8years.AcomprehensivereportontheDanishTubercu¬losisIndexbyGroth-Petersen,Knudsen,andDr.Comstockis with theTuberculosisBranch of theCommunicableDiseaseCenter,PublicHealthService,Washington,D.C.Wilbek(5)includedobservationson560per¬sonsneverpreviouslyreportedastuberculosiscasesbecausetheirchestroentgenogramswereconsideredtoshowfibrosisonly.Within4years,nearly2percenthaddevelopedactivedisease.InsouthIndia,Frimodt-M^ller(6)foundanaverageannualreactivationrate of 6percentforpersonsclassifiedasprobablyhavinginactivetuberculosisand1percentforthoseinitiallyconsideredtohaveclinicallyin-significant,inactivedisease.Althoughtheforegoingstudieshaveindi¬catedconsiderablevariationintheaveragean¬nualreactivationrate,avariationthatcouldberelatedbothtodifferencesinthedefinitions of acaseandinthelivingconditions of thestudypopulations,allagreedthattherisk of reactivationwassubstantial.Andyetthereissurprisingvariationintheperiod of observa¬tionrecommendedforpersons with inactive pulmonary tuberculosis.Somehealthdepart¬mentsdonotadviseanyfollowup of persons with newlydiagnosedminimalinactivedisease;othersadviseperiodicexaminationsfor ... 5yearsorlonger.Suchvariationinpublichealthpracticesuggestedtheneedforfurtherinfor¬mationontheimportance of relapsesamongpersons with inactivediseaseasasource of activetuberculosis.InformationgatheredbytheMuscogeeCountyTuberculosisStudywasusedtoesti¬matetheprognosis of untreatedinactivepul¬monarytuberculosis.Thediscoveryandpro¬longedobservation of allcases of tuberculosisinthecommunityhasbeenone of themajorVol.77,No.6,June1962461goals of thisstudysinceitsinceptionin1946(7).Consequently,casefindingactivitiesinthisareahavebeenmoreextensivethanisoftenpossible.Communitywidephotofluorographicsurveyswereconductedin1946andagainin1950(8$).Inaddition,thenumber of chestX-rayexaminationsmadebythehealthde¬partmenteachyearapproximatedone-sixth of thetotalpopulation.Themedicalprofessionandalliedagenciesinthecommunityhavebeenhighlycooperativeinreportingtuberculosiscases,usuallyrelinquishingresponsibilityforfollowupsupervisiontothestudy.Asaresult,itisbelievedthatnearlyallknowncaseshavebeenreportedtoandsupervisedbythestudy.Allpersons with definiteorsuspectedtuber¬culosiswereadvisedtoobtainquarterlyorsemiannualexaminationsuntiltheirdiseasehadshownnoevidence of activityforatleast5years.Althoughroutinefollowupwasthendiscontinued,subjectswereadvisedtobeex¬aminedannuallyandweresentreminders of theseexaminationsaslongastheycontinuedtoliveinthearea.Facilitiesforsputumexami¬nationshavealsobeenreadilyavailable, with culturesformycobacteriaaroutinepart of theexaminationsince1947.Sputumspecimenswererequestedatthebeginningandend of theroutinefollowupperiod,andalsowheneverclinicalorroentgenographicevidence of pos¬siblyactivediseasewasnoted.Almostallpa¬tients with ahistory of productivecoughcomplied with therequestsforspecimens.Many,however,deniedthissymptomandfailedtosubmitspecimens.Gastriclavageswererarelyperformed.InitialStudyPopulationThepopulationfromwhichthecases of un¬treatedinactive pulmonary tuberculosisweredrawncomprisesallpersonswhowerefirstreportedtotheMuscogeeCountyTuberculosisStudyashavingdefiniteorsuspectedtuber¬culosisbetweenJanuary1,1946,andJanuary1,1956,andwhoalsometthefollowingcriteria:(a)atthetimethesepersonswerereported,theylivedinMuscogeeCounty,andwere15-65years of age;(b)beforebeingreported,theyhadneverbeenadvisedtobehospitalizedfortuberculosis;(c)2yearsafterthefirstreport,Table1.Yearinwhichmembers of initialstudypopulationwerereportedtotheMuscogeeCountyTuberculosisStudyastuberculosiscasesorsuspectstheywerestillregisteredashavingsuspectedordefinite pulmonary tuberculosisandwerenotknowntohaveextra -pulmonary tuberculosis.Atotal of 1,327personsmetthesecriteria.Theyearinwhichtheywerefirstreportedtothestudyisshownintable1.Morethanone-thirdweredetectedin1946,theyear of thefirstcommunitysurvey.One-sixthcametoatten¬tionin1950,largelyastheresult of thesecondsurvey.Althoughtheproportiondiscoveredineach of theotheryearshasbeenquitecon¬stant,thisactuallyreflectsadeclineinthemor¬bidityrateowingtotheincreaseinpopulation of MuscogeeCounty,Ga.,from95,638personsin1946to158,623in1960.Thecomposition of theinitialstudypopula¬tionandthetype of examinationwhichledtotherecognition of tuberculosisareshownintable2.Almost70percent of thecaseswereinwhitesand30percentinNegroes.However,sinceconsiderablyfewerNegroesthanwhitesresideinthecounty,themorbidityratewasslightlyhigheramongNegroes.Slightlymorethanhalf of thewhitesbutonlyone-third of theNegroeswereovertheage of 45years.Amuchhigherproportion of Negroesthanwhiteswereclassifiedashavingadvanceddisease,44percentcontrasted with 23percent.Forbothraces,theproportion of advancedtuberculosiswaslargeramongtheyoungeragegroups.Almosthalf of thegroupwasbroughtundersupervisionastheresult of thetwomasssur¬veys.Ifpersonsdetectedbysurvey-likepro¬ceduressuchaspreemploymentandfoodhandler462PublicHealthReportsexaminationsarealsoincluded,morethanthree-fifthswerefoundamongpresumablyhealthygroups.Onlyafewwereidentifiedbecausetheyhadbeenincontact with acase of activetuberculosis.Athird of thetotalgroupwasclassifiedassymptomatic,havingbeenreferredforexaminationbyprivatephysiciansorhos¬pitalclinicsorself-referred.Manycontactswereinthecategory of symptomaticreferrals.Thesepersonshadnoevidence of tuberculosisonroutinecontactexaminations,butreturnedforreexaminationwhensymptomsdevelopedratherthanwaitingfortheirnextroutinefol¬lowupexamination.Asmightbeanticipated,ahighpercentage of advancedcasescamefromthegroup of symptomaticreferrals.Althougholderpersonsdidnotparticipatewellinthecommunitysurveys,asurprisinglylargepro¬portion of those with minimalorsuspecteddiseasesamongthemwerediscoveredthisway.ActiveandInactiveTuberculosisArrivingatanappropriatedefinition of ac¬tivetuberculosiswasnotassimpleasmightap¬pearatfirstglance.Primarily,adefinitionwasdesiredwhichwoulddesignatepersonswhoweretrulyillandwhosetuberculosiswassufficientlyseveretorequireamajorchangeintheirlives.Itdidnotseemimportanttostudytheincidence of diseasemanifestedonlybyisolateddemon¬strations of acid-fastbacilliorbyminorroent-genographicchanges.Nosinglecriterionseemedadequatetodes¬ignatesignificantlyactivetuberculosis.Eventhefinding of acid-fastbacilli with theculturalcharacteristics of Mycobacteriumtuberculosiswasfarfromsatisfactory.Inthispart of theworld,acid-fastbacillihavebeenisolated with considerablefrequencyfromcertainhealthypopulationgroups(10).Some of theseor¬ganismscouldhavebeenmistakenforM.tuber-culosis,particularlyintheearlierdays of thestudy.Butthemajorreasonfornotacceptingthissinglecriterionwasthefactthattuberclebacilliwereisolatedfromasizablesegment of thisstudygroupononlyasingleoccasion, with nootherevidence of activediseaseonprolongedbacteriologicalandroentgenographicobserva¬tion.Althoughonewidelyusedstandardwouldautomaticallyclassifythesepersonsashavingactivetuberculosis(ii),theirdiseasedidnotthenandhasnotyetshownanyevidence of be¬comingarealhealthproblem.Theuse of roentgenographicchangealsoseemedinadequateastheonlycriterionforac¬tivedisease,particularlyinview of thedemon¬strateddifficultiesingettingagreementevenamongexpertsinclassifyingtuberculosisfromchestroentgenograms(12,13).AndagainthereTable2.Percentage of tuberculosiscasesininitialstudypopulationdiscoveredbyspecifiedtypes of examination,byrace,stage of disease,andagegroupVol.77,No.6,June1962463wereanumber of persons with roentgeno¬graphicevidence of activetuberculosisbut with nootherevidence of activediseaseevenonpro¬longedfollowup.Aftercarefulconsideration,itappearedthatthemostappropriateweighting of thevariousdiagnosticfactorscouldbeachievedbydefiningtheonset of significantlyactivediseaseasthetimewhenhospitaltreatmentwasfirstrecom¬mended.Thisdefinitionhadseveraladvan¬tages.Firsttherecommendationforhospital¬izationwasaneventwhichcouldbeclearlyrecognizedanddatedfromtherecords.Mostimportant,thedecisiontorecommendhos¬pitalizationindicatedtherecognition of asig¬nificantadversechangeinapatient'scondition.Whilebacteriologicalandroentgenographicfindingsobviouslyinfluencedthisdecisionmuchmorethananyotherfactor,theuse of clinicaljudgmentmanifestedintherec¬ommendationforhospitaltreatmentappearedtobethemostsatisfactorymethod of discount-ingisolatedorinconsistentfindingswhichmightotherwiseinflatetheproportion of in-significantlyactivecases.Kecommendationsforhospitalizationwerenotmateriallyaffectedbytheavailability of hospitalbedsorbyenthu-siasmforambulatorytreatment.Throughouttheperiod of thisstudy,hospitalbedswereavailable with littleornodelayforMuscogeeCountypatients.Furthermore,hospitaltreat¬mentwasalmostalwaysrecommendedforper¬sonsbelievedtohaveactiveprogressivetuber¬culosis,onlyafewexceptionshavingoccurredinrecentyears.Inactivetuberculosiswasdefinedastheab¬sence of significantlyactivetuberculosisforatleast2yearsaftertheindividualwasreportedtothestudyasatuberculosiscaseorsuspect.Theclassification of the1,327personsintheinitialstudypopulation2yearsaftertheyhadfirstbeenreportedasatuberculosiscaseorsuspectisshownintable3.Inthisperiod,314personswerethoughttohaveactivetuberculosis,themajoritywithinafewweeks of initialreport.Fivepersonswithoutevidence of activediseaseareknowntohavediedduringthis2-yearperiod;theremaining1,008comprisetheinac¬tivecasesforthisanalysis.Themajorityhadonlysuspectedorminimaldiseaseinitially.Relativelyfew of thepersons with inactivedis¬easehadadvancedtuberculosisbecausetherewerenotmanyadvancedcasesintheinitialstudygroup.Moreover,70percentwereclassi¬fiedashavingactivediseasebefore2yearshadelapsed.Method of AnalysisTheanalysisincludesobservationsonallper¬sonsinthestudypopulationthroughJune30,Table3.Classification of studypopulation2yearsafterinitialreportastuberculosiscaseorsus¬pect,byrace,initialstage of disease,andagegroup464PublicHealthReports1960, with thetotalperiod of observationrang¬ingfrom4^toltyfayears.Because2yearshadtoelapsebeforeapersoncouldbeclassifiedashavinginactivedisease,thepotentialrange of observationforcases of inactivetuberculosiswas2%to12i/£years.Twomodificationsweremadeinapplyingthelifetablemethod of analysistothefindings of thisstudy.Thefirstmodificationwastheuse of twodifferentas-sumptionsregardingthedevelopment of activetuberculosisamongpersonswithdrawnfromobservationbecausetheymovedawayordis¬continuedexamination.Assumption1isthatthedevelopment of activediseasecouldonlybedeterminedduringtheperiodinwhichthesubjectswereunderobservation.Theperiod of observationwouldtherebybecountedfromthedate of reporttothedate of lastexaminationbeforeJuly1,1960.Thisassumptionunder-statestheperson-years of observationandthusoverestimatestherisk of developingactivedisease.Assumption2isthatsignificantreac¬tivationwouldbecomeknownforsurvivingmembers of thestudypopulationeveniftheydidnotcontinuetobeexaminedbythestudy,andthattheirexperiencecanthusbecountedthroughJune30,1960.Thisassumptionisbasedontheprobabilitythatpersonsdevelop¬ingsignificantlyactivetuberculosisafterdis-continuingobservationwouldseekmedicalcare,andthattheirphysicianswouldthenre¬questtheirpreviouschestroentgenograms.Inthisway,thestudywouldlearnthatreactiva¬tionhadoccurred.Assumption2overstatestheperson-years of observation,mainlybe¬causemostdeathswhichoccurredamongper¬sonswhomovedawaydidnotbecomeknowntothestudy.Itprobablyalsounderstatesthenumber of reactivations.Onbothcounts,as¬sumption2tendstounderestimatetherisk of reactivation.Thesecondmodification of theusuallifetabletechniquewastoapplyitintwosteps.First,thestandardapproachwasusedintheanalysis of thefindingsforthefirst2years of observa¬tion.Forthosewhopassedthroughthis2-yearperiodwithouthavingactivetuberculosisandwhoweretherebydesignatedashavinginactivetuberculosis,thecalculation of time-specificandcumulativeriskswasagainunder¬taken, with thebeginning of thethirdyearasthestartingperiod.Suspectedandminimaltuberculosiswerecombinedintoasinglecategorybecausetheriskwasessentiallythesameforboth of them.Moderatelyandfaradvancedtuberculosiswerecombinedbecause of thesmallnumbersineachgroup.Forpersons with advancedtuberculosis,sofewsurvivedthefirst2yearswithouthavingactivedisease,thatthetwoagegroupshadtobecombined;evenso,thenumbers of advancedinactivecasesaresmall.ResultsThebacteriologicalandroentgenographicfindingsforthe1,008personsclassifiedashav¬inginactivetuberculosisareshownintable4. Of the68forwhomhospitaltreatmentfortuberculosiswasrecommended,60hadbothpositivebacteriologicalandroentgenographicevidence of activetuberculosisatsometime;only1hadneither of thesetwocriteria. Of the940personsforwhomhospitaltreatmentwasneverrecommended,2hadbothbacteriologicalorroentgenographicevidence of activedisease.Vol.77,No.6,June1962465Cumulativeprobability of reactivation of inactive pulmonary tuberculosis,byrace,age,andinitialstage of disease,accordingtotwodifferentassumptions(1and2)regardingfollowupobservationsWHITENEGRO15-44years of ageatinitialreport, with minimalorsuspectedtuberculosis2yearslater.30r.25.20.15.10.05S.004)Oa>.35ru.30K.25.20.15.10.05.0045-64years of ageatinitialreport, with minimalorsuspectedtuberculosis2yearslater15-64years of ageatinitialreport, with advancedtuberculosis2yearslaterJ_LJ_I10121402468Yearsafterinitialreport101214Note:Heavylineindicatesthatthepopulationbaseis50ormorepersons;thinlineindicatesthatthebaseis20-50persons.466PublicHealthReportsTheuse of positivebacteriologicalandroent¬genographicfindingstogetherasacriterion of activetuberculosiswouldhavealteredbutlittletheproportion of personsclassifiedashavingreactivateddisease.Parenthetically,itmaybenotedthatamajorfactorinthedis¬crepancybetweenpositivebacteriologicalfind¬ingsandtheroentgenographicclassification of activetuberculosiswastheconsciousattempttorecordtheinterpretations of chestroent¬genogramswithoutregardtootherfindings.Hadtheseinterpretationsbeencompletelyindependent,it ... 5yearsorlonger.Suchvariationinpublichealthpracticesuggestedtheneedforfurtherinfor¬mationontheimportance of relapsesamongpersons with inactivediseaseasasource of activetuberculosis.InformationgatheredbytheMuscogeeCountyTuberculosisStudywasusedtoesti¬matetheprognosis of untreatedinactivepul¬monarytuberculosis.Thediscoveryandpro¬longedobservation of allcases of tuberculosisinthecommunityhasbeenone of themajorVol.77,No.6,June1962461goals of thisstudysinceitsinceptionin1946(7).Consequently,casefindingactivitiesinthisareahavebeenmoreextensivethanisoftenpossible.Communitywidephotofluorographicsurveyswereconductedin1946andagainin1950(8$).Inaddition,thenumber of chestX-rayexaminationsmadebythehealthde¬partmenteachyearapproximatedone-sixth of thetotalpopulation.Themedicalprofessionandalliedagenciesinthecommunityhavebeenhighlycooperativeinreportingtuberculosiscases,usuallyrelinquishingresponsibilityforfollowupsupervisiontothestudy.Asaresult,itisbelievedthatnearlyallknowncaseshavebeenreportedtoandsupervisedbythestudy.Allpersons with definiteorsuspectedtuber¬culosiswereadvisedtoobtainquarterlyorsemiannualexaminationsuntiltheirdiseasehadshownnoevidence of activityforatleast5years.Althoughroutinefollowupwasthendiscontinued,subjectswereadvisedtobeex¬aminedannuallyandweresentreminders of theseexaminationsaslongastheycontinuedtoliveinthearea.Facilitiesforsputumexami¬nationshavealsobeenreadilyavailable, with culturesformycobacteriaaroutinepart of theexaminationsince1947.Sputumspecimenswererequestedatthebeginningandend of theroutinefollowupperiod,andalsowheneverclinicalorroentgenographicevidence of pos¬siblyactivediseasewasnoted.Almostallpa¬tients with ahistory of productivecoughcomplied with therequestsforspecimens.Many,however,deniedthissymptomandfailedtosubmitspecimens.Gastriclavageswererarelyperformed.InitialStudyPopulationThepopulationfromwhichthecases of un¬treatedinactive pulmonary tuberculosisweredrawncomprisesallpersonswhowerefirstreportedtotheMuscogeeCountyTuberculosisStudyashavingdefiniteorsuspectedtuber¬culosisbetweenJanuary1,1946,andJanuary1,1956,andwhoalsometthefollowingcriteria:(a)atthetimethesepersonswerereported,theylivedinMuscogeeCounty,andwere15-65years of age;(b)beforebeingreported,theyhadneverbeenadvisedtobehospitalizedfortuberculosis;(c)2yearsafterthefirstreport,Table1.Yearinwhichmembers of initialstudypopulationwerereportedtotheMuscogeeCountyTuberculosisStudyastuberculosiscasesorsuspectstheywerestillregisteredashavingsuspectedordefinite pulmonary tuberculosisandwerenotknowntohaveextra -pulmonary tuberculosis.Atotal of 1,327personsmetthesecriteria.Theyearinwhichtheywerefirstreportedtothestudyisshownintable1.Morethanone-thirdweredetectedin1946,theyear of thefirstcommunitysurvey.One-sixthcametoatten¬tionin1950,largelyastheresult of thesecondsurvey.Althoughtheproportiondiscoveredineach of theotheryearshasbeenquitecon¬stant,thisactuallyreflectsadeclineinthemor¬bidityrateowingtotheincreaseinpopulation of MuscogeeCounty,Ga.,from95,638personsin1946to158,623in1960.Thecomposition of theinitialstudypopula¬tionandthetype of examinationwhichledtotherecognition of tuberculosisareshownintable2.Almost70percent of thecaseswereinwhitesand30percentinNegroes.However,sinceconsiderablyfewerNegroesthanwhitesresideinthecounty,themorbidityratewasslightlyhigheramongNegroes.Slightlymorethanhalf of thewhitesbutonlyone-third of theNegroeswereovertheage of 45years.Amuchhigherproportion of Negroesthanwhiteswereclassifiedashavingadvanceddisease,44percentcontrasted with 23percent.Forbothraces,theproportion of advancedtuberculosiswaslargeramongtheyoungeragegroups.Almosthalf of thegroupwasbroughtundersupervisionastheresult of thetwomasssur¬veys.Ifpersonsdetectedbysurvey-likepro¬ceduressuchaspreemploymentandfoodhandler462PublicHealthReportsexaminationsarealsoincluded,morethanthree-fifthswerefoundamongpresumablyhealthygroups.Onlyafewwereidentifiedbecausetheyhadbeenincontact with acase of activetuberculosis.Athird of thetotalgroupwasclassifiedassymptomatic,havingbeenreferredforexaminationbyprivatephysiciansorhos¬pitalclinicsorself-referred.Manycontactswereinthecategory of symptomaticreferrals.Thesepersonshadnoevidence of tuberculosisonroutinecontactexaminations,butreturnedforreexaminationwhensymptomsdevelopedratherthanwaitingfortheirnextroutinefol¬lowupexamination.Asmightbeanticipated,ahighpercentage of advancedcasescamefromthegroup of symptomaticreferrals.Althougholderpersonsdidnotparticipatewellinthecommunitysurveys,asurprisinglylargepro¬portion of those with minimalorsuspecteddiseasesamongthemwerediscoveredthisway.ActiveandInactiveTuberculosisArrivingatanappropriatedefinition of ac¬tivetuberculosiswasnotassimpleasmightap¬pearatfirstglance.Primarily,adefinitionwasdesiredwhichwoulddesignatepersonswhoweretrulyillandwhosetuberculosiswassufficientlyseveretorequireamajorchangeintheirlives.Itdidnotseemimportanttostudytheincidence of diseasemanifestedonlybyisolateddemon¬strations of acid-fastbacilliorbyminorroent-genographicchanges.Nosinglecriterionseemedadequatetodes¬ignatesignificantlyactivetuberculosis.Eventhefinding of acid-fastbacilli with theculturalcharacteristics of Mycobacteriumtuberculosiswasfarfromsatisfactory.Inthispart of theworld,acid-fastbacillihavebeenisolated with considerablefrequencyfromcertainhealthypopulationgroups(10).Some of theseor¬ganismscouldhavebeenmistakenforM.tuber-culosis,particularlyintheearlierdays of thestudy.Butthemajorreasonfornotacceptingthissinglecriterionwasthefactthattuberclebacilliwereisolatedfromasizablesegment of thisstudygroupononlyasingleoccasion, with nootherevidence of activediseaseonprolongedbacteriologicalandroentgenographicobserva¬tion.Althoughonewidelyusedstandardwouldautomaticallyclassifythesepersonsashavingactivetuberculosis(ii),theirdiseasedidnotthenandhasnotyetshownanyevidence of be¬comingarealhealthproblem.Theuse of roentgenographicchangealsoseemedinadequateastheonlycriterionforac¬tivedisease,particularlyinview of thedemon¬strateddifficultiesingettingagreementevenamongexpertsinclassifyingtuberculosisfromchestroentgenograms(12,13).AndagainthereTable2.Percentage of tuberculosiscasesininitialstudypopulationdiscoveredbyspecifiedtypes of examination,byrace,stage of disease,andagegroupVol.77,No.6,June1962463wereanumber of persons with roentgeno¬graphicevidence of activetuberculosisbut with nootherevidence of activediseaseevenonpro¬longedfollowup.Aftercarefulconsideration,itappearedthatthemostappropriateweighting of thevariousdiagnosticfactorscouldbeachievedbydefiningtheonset of significantlyactivediseaseasthetimewhenhospitaltreatmentwasfirstrecom¬mended.Thisdefinitionhadseveraladvan¬tages.Firsttherecommendationforhospital¬izationwasaneventwhichcouldbeclearlyrecognizedanddatedfromtherecords.Mostimportant,thedecisiontorecommendhos¬pitalizationindicatedtherecognition of asig¬nificantadversechangeinapatient'scondition.Whilebacteriologicalandroentgenographicfindingsobviouslyinfluencedthisdecisionmuchmorethananyotherfactor,theuse of clinicaljudgmentmanifestedintherec¬ommendationforhospitaltreatmentappearedtobethemostsatisfactorymethod of discount-ingisolatedorinconsistentfindingswhichmightotherwiseinflatetheproportion of in-significantlyactivecases.Kecommendationsforhospitalizationwerenotmateriallyaffectedbytheavailability of hospitalbedsorbyenthu-siasmforambulatorytreatment.Throughouttheperiod of thisstudy,hospitalbedswereavailable with littleornodelayforMuscogeeCountypatients.Furthermore,hospitaltreat¬mentwasalmostalwaysrecommendedforper¬sonsbelievedtohaveactiveprogressivetuber¬culosis,onlyafewexceptionshavingoccurredinrecentyears.Inactivetuberculosiswasdefinedastheab¬sence of significantlyactivetuberculosisforatleast2yearsaftertheindividualwasreportedtothestudyasatuberculosiscaseorsuspect.Theclassification of the1,327personsintheinitialstudypopulation2yearsaftertheyhadfirstbeenreportedasatuberculosiscaseorsuspectisshownintable3.Inthisperiod,314personswerethoughttohaveactivetuberculosis,themajoritywithinafewweeks of initialreport.Fivepersonswithoutevidence of activediseaseareknowntohavediedduringthis2-yearperiod;theremaining1,008comprisetheinac¬tivecasesforthisanalysis.Themajorityhadonlysuspectedorminimaldiseaseinitially.Relativelyfew of thepersons with inactivedis¬easehadadvancedtuberculosisbecausetherewerenotmanyadvancedcasesintheinitialstudygroup.Moreover,70percentwereclassi¬fiedashavingactivediseasebefore2yearshadelapsed.Method of AnalysisTheanalysisincludesobservationsonallper¬sonsinthestudypopulationthroughJune30,Table3.Classification of studypopulation2yearsafterinitialreportastuberculosiscaseorsus¬pect,byrace,initialstage of disease,andagegroup464PublicHealthReports1960, with thetotalperiod of observationrang¬ingfrom4^toltyfayears.Because2yearshadtoelapsebeforeapersoncouldbeclassifiedashavinginactivedisease,thepotentialrange of observationforcases of inactivetuberculosiswas2%to12i/£years.Twomodificationsweremadeinapplyingthelifetablemethod of analysistothefindings of thisstudy.Thefirstmodificationwastheuse of twodifferentas-sumptionsregardingthedevelopment of activetuberculosisamongpersonswithdrawnfromobservationbecausetheymovedawayordis¬continuedexamination.Assumption1isthatthedevelopment of activediseasecouldonlybedeterminedduringtheperiodinwhichthesubjectswereunderobservation.Theperiod of observationwouldtherebybecountedfromthedate of reporttothedate of lastexaminationbeforeJuly1,1960.Thisassumptionunder-statestheperson-years of observationandthusoverestimatestherisk of developingactivedisease.Assumption2isthatsignificantreac¬tivationwouldbecomeknownforsurvivingmembers of thestudypopulationeveniftheydidnotcontinuetobeexaminedbythestudy,andthattheirexperiencecanthusbecountedthroughJune30,1960.Thisassumptionisbasedontheprobabilitythatpersonsdevelop¬ingsignificantlyactivetuberculosisafterdis-continuingobservationwouldseekmedicalcare,andthattheirphysicianswouldthenre¬questtheirpreviouschestroentgenograms.Inthisway,thestudywouldlearnthatreactiva¬tionhadoccurred.Assumption2overstatestheperson-years of observation,mainlybe¬causemostdeathswhichoccurredamongper¬sonswhomovedawaydidnotbecomeknowntothestudy.Itprobablyalsounderstatesthenumber of reactivations.Onbothcounts,as¬sumption2tendstounderestimatetherisk of reactivation.Thesecondmodification of theusuallifetabletechniquewastoapplyitintwosteps.First,thestandardapproachwasusedintheanalysis of thefindingsforthefirst2years of observa¬tion.Forthosewhopassedthroughthis2-yearperiodwithouthavingactivetuberculosisandwhoweretherebydesignatedashavinginactivetuberculosis,thecalculation of time-specificandcumulativeriskswasagainunder¬taken, with thebeginning of thethirdyearasthestartingperiod.Suspectedandminimaltuberculosiswerecombinedintoasinglecategorybecausetheriskwasessentiallythesameforboth of them.Moderatelyandfaradvancedtuberculosiswerecombinedbecause of thesmallnumbersineachgroup.Forpersons with advancedtuberculosis,sofewsurvivedthefirst2yearswithouthavingactivedisease,thatthetwoagegroupshadtobecombined;evenso,thenumbers of advancedinactivecasesaresmall.ResultsThebacteriologicalandroentgenographicfindingsforthe1,008personsclassifiedashav¬inginactivetuberculosisareshownintable4. Of the68forwhomhospitaltreatmentfortuberculosiswasrecommended,60hadbothpositivebacteriologicalandroentgenographicevidence of activetuberculosisatsometime;only1hadneither of thesetwocriteria. Of the940personsforwhomhospitaltreatmentwasneverrecommended,2hadbothbacteriologicalorroentgenographicevidence of activedisease.Vol.77,No.6,June1962465Cumulativeprobability of reactivation of inactive pulmonary tuberculosis,byrace,age,andinitialstage of disease,accordingtotwodifferentassumptions(1and2)regardingfollowupobservationsWHITENEGRO15-44years of ageatinitialreport, with minimalorsuspectedtuberculosis2yearslater.30r.25.20.15.10.05S.004)Oa>.35ru.30K.25.20.15.10.05.0045-64years of ageatinitialreport, with minimalorsuspectedtuberculosis2yearslater15-64years of ageatinitialreport, with advancedtuberculosis2yearslaterJ_LJ_I10121402468Yearsafterinitialreport101214Note:Heavylineindicatesthatthepopulationbaseis50ormorepersons;thinlineindicatesthatthebaseis20-50persons.466PublicHealthReportsTheuse of positivebacteriologicalandroent¬genographicfindingstogetherasacriterion of activetuberculosiswouldhavealteredbutlittletheproportion of personsclassifiedashavingreactivateddisease.Parenthetically,itmaybenotedthatamajorfactorinthedis¬crepancybetweenpositivebacteriologicalfind¬ingsandtheroentgenographicclassification of activetuberculosiswastheconsciousattempttorecordtheinterpretations of chestroent¬genogramswithoutregardtootherfindings.Hadtheseinterpretationsbeencompletelyindependent,it...