THE WILLIAMSON COUNTY TUBERCULOSIS STUDY A Twenty-Four—Year Epidemiologic Study

Item

Title
THE WILLIAMSON COUNTY
TUBERCULOSIS STUDY
A Twenty-Four—Year Epidemiologic Study
extracted text
THE WILLIAMSON COUNTY
TUBERCULOSIS STUDY
A Tweiilv-F our-Year Epidemiologic Study

L. D. ZEIDBERG, M.D., M.P.II.
Director of tuberculosis Research, Division of Tu­
berculosis Control, Tennessee Department of Pub­
lic Health; Professor of Kpideniiology, Vanderbilt
University School of Medicine;

K. S. GASS, M.I). (Deceased)
Director, Division of Tubei culosis (’ontrol, Tennes­
see Department of Public; I h alt h

ANN DILLON, M.S.P.IL
Director, Statistical Service, Tennessee Department
of Public Health

IL II. IIUrCIIESON, M.D., AI.P.IL
Commissioner, Tennessee Department of Public
Health, Nashville, Tennessee

With the support of

the J cnnessec Tuberculosis and Health Association, Inc.
and the

Local Affiliated Tuberculosis Associations of Tennessee

THE WILLIAMSON COUNTY
TUBERCULOSIS STUDY
A Twenty-Four—Year Epidemiologic Study

L. D. ZEIDBERG, M.D., M.P.II.
Director of Tuberculosis Research, Division of Tu­
berculosis Control, Tennessee Department of Pub­
lic Health; Professor of Epidemiology, Vanderbilt
University School of Medicine

R. S. GASS, M.D. (Deceased)
Director, Division of Tuberculosis Control, Tennes­
see Department of Public Health

ANN DILLON, M.S.P.H.
Director, Statistical Service, Tennessee Department
of Public Health

R. H. HUTCHESON, M.D., M.P.IL
Commissioner, Tennessee Department of Public
Health, Nashville, Tennessee

With the support of

The Tennessee Tuberculosis and Health Association, Inc.
and the
Local Affiliated Tuberculosis Associations of Tennessee

FOREWORD

i

Tuberculosis in Tennessee in the early nineteen thirties was an affliction of the people
sufficient to challenge the thought and attention of any who could see its enormity and the
opportunity to combat it from a fresh approach. It was realized that the factors underlying
the problem were numerous and required elucidation. The State was fortunate in having a
Commissioner of its Department of Health who accepted the responsibility of initiating the
effort in a methodical and carefully planned way. It seemed appropriate and promising to
attempt to apply epidemiologic principles, notwithstanding some great difficulties inherent
in a disease of such great chronicity and long duration. The time, the circumstances, and
the mutual inspiration and imagination of Bishop and Frost combined to set off a closely
linked study and campaign which, during the next quarter century, bore their fruit richly
and abundantly.
The attitudes of these two men were distinguished by their honesty, by a certain intel­
lectual humility reflecting the limited experience of themselves and others, and by the per­
ception of their disciplined minds. Bishop knew he needed help and he knew where to go
when he turned to Frost. Their initial achievement was in posing the right questions. These
revealed freedom from prejudice and an insight and prescience which were truly remarkable,
setting the tone which characterized the work throughout. The questions were clear and
simple, and justified the expectation that some of the answers would be of a similar sort.
With this solid footing the project was started in a community which was ideally suitable,
not only because of its relatively stable population and prevalence of tuberculosis, but also
because of its reasonable proximity to the university and the seat of state government,
where interest and skill would be readily available and sustained. Unlike many long under­
takings of this kind it continued consistently to its completion.
With the subsequent interposition of socioeconomic and other changes, especially
specific drug therapy for tuberculosis, it is improbable that a study with this content can
be conducted again. Its uniqueness, therefore, promises to be historic. But its merit, above
all else, is in the clarity of its purpose, the faithfulness of its pursuit, the richness of its
findings, and the honest objectivity of its interpretations.
Through the years, despite grievous losses, the work was carried on steadily by compe­
tent and devoted people representing the various disciplines which were brought to bear.
This complete integration and report of the study is a testimonial to the greatness of their
contributions and a fine service to all students of tuberculosis.
Each in his own way will judge and apply the various aspects and findings of the study
as they fit with his particular interest and experience. One cannot avoid fascination with
the byproduct of revelations of histoplasmosis—the weed discovered among the crops of
tuberculosis or, to speak more prosaically, the calcifications among those insensitive to
tuberculin. In general, there is much to clarify, substantiate, and add to our store of knowl­
edge—be it in epidemiology, sociology, public health, clinical medicine, or other disciplines
related to the understanding of tuberculosis. We may place strong confidence in the original
conception of the study, its conduct, and its interpretations. All will appreciate the smoothly
logical way in which it is presented.
J. Burns Amberson

i

II

i

II

DEDICATION
The untimely death of Roydon Simpson Gass on September 4, 1962 brought to a
close more than thirty years dedicated to the control of tuberculosis. The co-authors of
this report, in recognition of his contributions to the Williamson County Tuberculosis
Study during its entire course, and in sorrow over the loss of an esteemed colleague and
devoted friend, wish to dedicate this volume to his memory.
A. D.
R. H. H.
L. D. Z.

li

[I

il

!

ii

TABLE OF CONTENTS
V-

1

Foreword: J. Burns Amberson. .

i

Dedication

ii

I. History of the Study

1

II. Description of Williamson County

5

III. Definitions

6

IV. Statistical Methodology

7

V. Objectives

9

VI. Plan of the Study

9

VII. Composition of the Study Group. .

11

VIII. Household Studies...........................

12

IX. Study of Cases

27

X. Community Studies

32

XI. Summary

38

XII. References
XIII. Appendix
A. Study Forms
B. Tables I-XXIIIB

..............

40
42
52

1

I.

I

I

I

THE AMERICAN REVIEW
OF RESPIRATORY DISEASES
Clinical' and Laboratory Studies of Tuberculosis and Respiratory Diseases

VOLUME 87

March 1963

PART 2 OF TWO PARTS

History of the Study

NUMBER 3

Tennessee problem. The stage was set for the
development of special studies.

A. The Tuberculosis Problem in Tennessee

13. Pilot Studies

As far back as the first ollicial records were
kept it was known that Tennessee had a higher
tuberculosis death rate than the national aver­
age. Throughout the years, as the tuberculosis
mortality curve for the nation showed a more
or less regular decline, the curve for Tennessee
showed a proportionate drop, so that in almost
any year the tuberculosis death rate in this state
was approximately 50 per cent higher than the
nation’s. It has been only in the last ten years
or so that the tuberculosis mortality rate for
Tennessee has dropped a little more sharply
than the rate in the country as a whole.
Tennessee’s excessive resident tuberculosis
mortality was undoubtedly due in part to its
relatively large nonwhite population. That was
not the whole story, however. In the white popu­
lation there was a marked rise in tuberculosis
death rates beginning at age 50 (figure 1). Al­
though not shown in the chart, death rates for
females were higher than for males, while in the
nation the rates were higher in males. The mor­
tality rates in white rural residents were con­
siderably higher than for those living in urban
areas. This is exactly the reverse of the situa­
tion in the country as a whole. On the other
hand, the experience of the Tennessee Negro
did not differ appreciably from that of Negroes
in the United States generally (figure 1). Al­
though the rates for Negroes in Tennessee were
slightly higher than the national average, they
were not quite as excessive as the rates for the
state’s white population. The pattern of the
resident Tennessee and United States tubercu­
losis mortality curves for Negroes was similar,
and in Tennessee the urban rates slightly ex­
ceeded the rural.
All of these distinctive characteristics of the
tuberculosis mortality in Tennessee made it ap­
parent that studies done in other areas of the
country, and control measures undertaken else­
where, would very likely be inapplicable to the

As Commissioner of the Department of Pub­
lic Health of Tennessee, Dr. E. L. Bishop was
deeply concerned with the unusual tuberculosis
problem in the state. With the support of the
Rosenwald Fund he established an epidemio­
logic study of tuberculosis in Trenton, Gibson
County, in 1930. This study was largely ex­
ploratory in nature and it quickly demonstrated
that the techniques employed in the study of
acute communicable disease could not be ap­
plied without modification to the study of a
chronic communicable disease such as tuberculo­
sis.
Dr. Bishop called upon his former teacher,
Dr. Wade II. Frost, professor of epidemiology
at the Johns Hopkins School of Hygiene and
Public Health, for guidance. Dr. Frost had a
personal interest in tuberculosis because he had
had the disease. Frost confessed his own lack of
experience and the absence of guides or prece­
dents in the epidemiologic study of a chronic
disease. In a conference with Dr. Bishop on
July 10, 1930, following a review of the work in
Gibson County, he pointed up the deficiencies
of the study, and made constructive suggestions
concerning the clear definition of purposes and
methods.
The chief objective of the Gibson County
Study seems to have been the study of methods.
In his discussion with Dr. Bishop on July 10,
1930, Frost outlined in some detail the kind of
study he visualized. He stated:
The basic material for the study should be an un­
selected series of cases of tuberculosis which have
come to the notice of the State Department of
Health through official morbidity reports or
through examinations made in the chest clinic.
Mortality records would seem to be useful in this
series only if the death has been fairly recent and
if the circumstances are such that the investigator
can obtain approximately the same information

1

2

WILLIAMSON COUNTY TUBERCULOSIS STUDY
NON- WHITE

WHITE

400

350-

I<

I
I
I

300"

I{

TENN.RURAL

o 250o
o
oo

I
I

200-

5a.
UJ

■I

<

//
/
/ //

150-

X

;|

o

TENN.
URBAN

-------' I
I
I

100-

__ I s'
us.

50-

///

\
AGES

0

10

20

30

40

50

60

70

10

20

30

40

50

60

70

80 90

Fig. 1. Resident tuberculosis death rates by race and age group: average for 1933-1935 in
rural and urban Tennessee, and in the United States in 1934.

■i

that would have been available during the illness
of the deceased.1
1
!
■‘I

I

i

The verbatim statement is repeated here be­
cause the plan was followed exactly in the Wil­
liamson County Study later on.
Frost also outlined in detail the need to obtain
the antecedent history of tuberculosis in present
or former household members and to keep an
exact record of duration and intimacy of contact.
In addition, he recommended that economic,
social, and industrial histories of the tuberculous
patient be obtained. The clinical status and
progress of disease were to be recorded. He listed
the information it would be desirable to have
for each household contact, and suggested the
follow-up intervals for continued observation of
the contacts. He also emphasized the importance
of studying a group of control households for
comparison with the tuberculosis households,
and of tuberculin surveys in school children as a
'Y basis for comparison with similar age groups in
• families with tuberculosis.
In 1930 a second study, in reality more of a
survey, was initiated in Kingsport, Sullivan

1 Memorandum by Dr. Bishop of the conference
with Dr. Frost.
(

80 90

County, Tennessee, with the support of the
Rockefeller Foundation, of the International
Health Division of which Dr. Frost was one of
the scientific directors. The survey included 132
Negro families, almost all such families in Kings­
port, whether or not they had a known case of
tuberculosis. It was intended as a study to test
new methods in the field. Here the modified life­
table method was used to describe the risk ex­
perienced by contacts of known cases of tuber­
culosis. This concept is described in the section
on Statistical Methodology.
In a letter to Dr. Bishop dated January 12,
1931, Dr. Frost reviewed the work of the Kings­
port Study. He appeared pleased that the defi­
ciencies of the Gibson Study had been largely
corrected, and he recommended at least one
more year of study, to be followed by analysis of
the collected data, which he anticipated would
require still another year of intensive work by a
competent epidemiologist. With all of this, how­
ever, he realized that:
It certainly will not give a final answer to any
considerable number of the questions in the epi­
demiology of tuberculosis, but I am confident that
it will give us a better basis for answering a cer­
tain number of questions and, in addition to facts

WILLIAMSON COUNTY TUBERCULOSIS STUDY

that are added as a direct result of this study, I
think we will have gained a good deal in experi­
ence and appreciation of studies of this general
kind. I think it is the kind of study to which we
must look more and more in the future, and that
you are doing an excellent service in pointing the
way.2

3

dilations. It is clear from the memorandum that
Fiost had a study plan worked out with great
precision and in considerable detail. His plan
was followed closely in establishing the William­
son County Study, which was activated one
month later, in December, 1931.
Soon after this letter there appears to have
Following his analysis of the Kingsport Study
been some serious consideration of a plan to de­ data (1), Frost saw the need for establishing a
velop still a third study that would embody the substitute for the “primary” case. His develop­
lessons learned in Gibson and Sullivan counties, ment of the “index” case concept is described in
and that would have a longer duration. For the the section on Statistical Methodology.
new study certain basic requirements were
In the early planning of the Williamson
stated. A county with a full-time health depart­ County Study a decision was made to use the
ment was considered essential; a stable popula­ Opie classification for pulmonary tuberculosis
tion, characteristic of Tennessee, numbering
(2) because his system seemed to be best suited
about 25,000 people was desirable; and prefer­ to the study’s needs in this regard. Dr. Opie
ably, the county was to be close to Nashville served as consultant, and his schema for the
where the consultative services of personnel of graphic presentation of a household’s experience
Vanderbilt University Medical Center and of with tuberculosis was later adopted. This schema
the State Department of Public Health would proved to be most useful to the study personnel.
be available. Williamson County fulfilled these
The Rockefeller Foundation, through its In­
requirements.
ternational Health Division, having supported
the Kingsport Study, now undertook the partial
C. The Williamson County Study
support of the Williamson County Study, and
In a long, detailed memorandum dated No­ continued its contributions without interruption
vember 19, 1931, Frost first stated the objectives throughout the 24 years during which the study
of the Williamson County Study (see the sec­ actively collected data. The Foundation also
tion on Objectives), and snelled
spelled out thn
the organ- rendered invaluable professional services to the
ization and operation of the study. Much of study through periodic consultation of its expert
what he wrote was similar to his recommenda­ personnel with the study personnel. The mag­
tions for the Kingsport Study, but now he pre­ nificent financial and professional contributions
sented operational plans rather than suggestions. °f the Rockefeller Foundation cannot be over­
At this early date he still wrote of the first case estimated, and this opportunity is taken to ex­
as the “primary” case. He emphasized the in­ press the gratitude of the Tennessee Department
vestigation of the evolution of tuberculosis from of Public Health.
infection to the development of disease, and a
At the time of its inception, and for about
systematic epidemiologic and clinical study of eight years after, the Williamson County Study
tuberculosis of old age. He advocated setting up operated without the benefit of county or state
the study within the framework of the county tuberculosis hospital beds for its patients. Fur­
health department with resident personnel on thermore, specific antituberculosis therapy was
hand to carry out the work. He suggested that still a dream in 1931 and did not become a
records of all deaths from tuberculosis and of reality until about 17 years later. During the
all known cases be assembled and subjected to early years of the study, therefore, it was possi­
statistical study. All of the families of these pa­ ble to observe the natural history of tuberculosis
tients were to be included in the study, with
uninfluenced, in general, by man-made forces
priority of visits given to those families with
for the prevention and control of the disease.
present cases or recent deaths. He outlined the
As time went on, the focus of the study
investigative procedure for household contacts
changed somewhat. Although no definite time
and the develoimiUnt of a family schedule that
would record the comings and goings of each limit was set, some of the forms that were de­
family member for the purpose of life-table cal- vised and used allotted space for recording only
five years of collected data. The Rockefeller
2 Letter to Dr. Bishop dated Juty 12, 1931.
Foundation’s support was practically on a year-

4

I
1

;i

I
I

1

1

I
I

WILLIAMSON COUNTY TUBERCULOSIS STUDY

to-year basis but, when it became apparent that phasized, although it was not originated in the
Frost’s concept of the "longitudinal section” of study, was the principle that factors influencing
study of a chronic disease was sound and fruit­ the occurrence of a communicable disease may
ful, the period of the study was extended again operate long before any contact is established.
and again until it covered 24 years. The findings In her analyses of the study data, Dr. Ruth R.
were reported in 19 papers and two books (3-23). Puffer showed that there was a familial suscepti­
The termination of the study at the end of 1955 bility to tuberculosis apart from familial con­
was a painful decision, but it was based on the tact (22). She cited the original sources of this
marked decline in the number of new cases in concept in her book, and showed how it was ap­
plied to the study data. In a sense this goes be­
the, community and households.
The longitudinal study of tuberculosis de­ yond Frost’s treatment of antecedent history,
stroyed one concept and established another in since he was most concerned with historical evi­
its place. It was generally believed in the past dence of past contact with tuberculosis. The
that it was not necessary to observe tuberculosis familial susceptibility is based on heredo-concontacts for more than two years after contact stitutional factors that may operate to increase
was broken because, if they did not develop the the risk of disease wholly apart from the influ­
disease in that time, they were not apt to later. ence of contact on such risk.
Over the years there were many changes in
The study showed this concept to be in error. It I
the
personnel of the study. Considering the span
will be demonstrated below during the presenta­
of
time,
there was remarkably little turnover in
tion of the data that the interval between ex­
posure and onset of disease is governed by some of the key personnel. Dr. Bishop was fol­
factors that are not limited by temporal con­ lowed as Commissioner by Dr. W. Carter Wil­
siderations except as they affect the age of the liams, who was succeeded by Dr. R. H. Hutche­
contact. The study demonstrated the need for son, the incumbent. All maintained a keen
prolonged observation of contacts through "crit­ interest in the details of operation and in ac­
complishments of the study, while they made
ical” ages of life4
sure
that there were sufficient state appropria­
Rather early in the study, when the first
analysis was made of tuberculin reactions and tions for its support. Dr. Frost served as con­
pulmonary calcification (5), another sacred sultant epidemiologist until his untimely death
cow” was destroyed. It was shown that calcifica­ in 1938. He paid many visits to Williamson
tion was as often associated with tuberculin County, and participated directly in the collec­
nonreactors as with reactors, the true meaning tion and analysis of data. He was succeeded by
of this development was not understood at the Dr. James A. Doull, who served from 1940 to
time, but it became apparent when Christie and 1949. Dr. R. S. Gass, the clinical consultant, was
Peterson demonstrated the association of pul­ with the study through its entire fife. Dr. Ruth
R. Puffer served as statistician and consultant
monary calcification and histoplasmin sensitivity
(24). Their discovery then led to the incorpo­ for all but the last few years of the study. The
ration of histoplasmosis studies in the Tubercu­ first epidemiologist, the late Dr. H. C. Stewart,
losis Study (25-35). Two other developments functioned in that capacity for a total of 15
stemmed from the unexpected finding of tuber­ years, although not consecutively. Miss Ann
culin-negative pulmonary calcification. One was Dillon was statistician for 14 years, then suc­
the alteration of the Opie classification by the ceeded Dr. Puller as consultant when the latter
abandonment of "latent childhood tuberculosis
left to take up new duties with the World Health
as the designation of uncomplicated pulmonary Organization. There were many others, public
calcification. The other result, perhaps more sig­ health nurses, technicians, secretarial and cleri­
nificant, was the use of "blind” studies to evalu­ cal personnel, who served the study devotedly
ate the relationship of the tuberculin reaction for many years. Two of the nurses, Mrs. Condon
and the presence of calcium in the lungs. Al­ W. Taylor and Miss Zetta Comer, were on the
though this may not have been the first "blind” study staff for the entire period. An incomplete
study, it was certainly among the early ones, and
listing of the professional personnel is appended
deserves mention as an established epidemiologic
below. A complete fisting would require a re­
method of study.
search project of unwarranted magnitude, neAnother epidemiologic concept that vas em­

WILLIAMSON COUNTY TUBERCULOSIS STUDY

cessitating a painstaking search of the files of
the State Personnel Office. Omissions are not in­
tentional, but are rather the result of blunted
memory. The accomplishments of the study are
Commissioners of Health
*Dr. E. L. Bishop
Dr. W. C. Williams
Dr. R. H. Hutcheson

i -

Consultant Epidemiologists
*Dr. Wade H. Frost
Dr. Kenneth F. Maxey
Dr. James A. Doull

Consultant Clinician
*Dr. R. S. Gass
Consultant Statisticians
Dr. R. R. Puffer
Ann Dillon

Directors of Division of Preventable Diseases
Dr. James A. Crabtree
*Dr. Grit Pharris
Dr. C. B. Tucker

5

a tribute to all of these people. The final salute
belongs to the people of Williamson County
whose cooperation was far beyond normal ex­
pectations.
Rockefeller Foundation Consultants
Dr. John Farrell
Dr. Eugene L. Opie
*Col. Fred Russell
Dr. George K. Strode
Dr. Hugh II. Smith

Epidemiologists
Dr. Jesse Ellington (Kinsgport Study)
*Dr. H. C. Stewart
Dr. R. L. Gauld
Dr. Willi am J. Murphy
Dr. C. W. Wells
Dr. L. D. Zeidberg

Clinicians

Dr. William D. Hickerson
*Dr. Virginia Hickerson
Dr. E. F. Harrison
Dr. John Cunningham

*Deceased

Description of Williamson County

Williamson County is a rural, agricultural
area of 593 square miles in the central part of
Tennessee. About 350,000 acres (84 per cent of
the land area) are devoted to farms. The popu­
lation is largely native-born; many of the fami­
lies date their residence from the first settle­
ment of this section in about 1785. The county
was organized in 1799 and named for the 'Wil­
liamson family. The county seat, Franklin,
honored the name of Benjamin Franklin.
The population of the county has followed the
■ trend of most agricultural areas. It has been
decreasing slowly over the years from 26,321 in
1890 to 24,307 in 1950. The town of Franklin
has grown in population to a small degree and
numbered 5,475 inhabitants in the 1950 census.
Negroes comprise approximately 22 per cent of
the county population.
Tobacco and crop farming were the principal
sources of farm income in the past, but in recent
years dairy and cattle farming have become im­
portant. Tobacco is still the chief cash crop.
The town of Franklin has several relatively
small industries. The largest of these, a stove

factory, employed about 200 men.3 In 1950 the
effective buying income of the people of Wil­
liamson County was 8622 per capita, compared
with the statewide average of $874.
In the past, more than 20 physicians practiced
in W illiamson County. Although the largest
number of them was concentrated in Franklin,
practitioners were located in many areas of the
county. 1 he trend has been toward a decrease
in the number of physicians, and almost all are in
Franklin. However, good roads and a bus line
from Franklin to Nashville, 18 miles distant,
bring the medical facilities of the city within
reach of Williamson County residents.
Three small private hospitals with a combined
capacity of 40 to 50 beds were owned and oper­
ated by local physicians until recently, when a
modern 50-bed county hospital was constructed.
The Williamson County Health Department
is the third oldest full-time local health depart­
ment in the state, having been activated in 1921.
Until a few years ago it had a full-time health
officer, but is currently operating with part-time
services. rl he staff included a nurse supervisor
and eight public health nurses at the height of
3 This factory was closed recently.

6

WILLIAMSON COUNTY TUBERCULOSIS STUDY

individuals, usually but not necessarily with blood
or legal ties, who occupy a dwelling unit.
IIousehold member: A household member is an
individual who lives and sleeps in a household
unit and shares a common table with other mem­
bers of the household.
Household associate: A household associate is
hospital.
a household member other than the index case
who was in the household with the index case at
Definitions
the time of his illness.
Close relative: A close relative is a parent, sibFor the sake of clarity and to avoid misunder­
ling, or child of an index case.
standing, the following terms used in the study
Other member: All other household members
who are not close relatives of an index case are
are defined:
classified as “other members.” This group includes
Index case: The index case is the person with more distant relatives, such as grandparents,
known or suspected tuberculosis in a household, grandchildren, aunts, uncles, cousins, et cetera, of
who first came to the attention of the study and an index case; relatives by marriage such as a
motivated the investigation of that household, d his spouse and “in-laws”; and others who have no
does not imply that the index case wTas the first to relationship to the index case at all. It is a very
have occurred in that household, nor that it was heterogeneous group.
Prevalence: As used in the study, prevalence is
the only one present at the time of discovery. The
use of “case” is deplored in many quarters because always spoken of in connection with “prevalence
it is an inanimate term. However, as used herein, of infection” or “prevalence of disease.” It refers
“index case” designates a person and the term has to the proportion of persons who have either con­
now been used and accepted to such an extent dition at a given moment of time. The lapse of
time, such as is required in defining a rate, is not
that to abandon it would be to complicate greatly
the presentation of study material and the discus­ inherent in the meaning of the term prevalence.
It is not a rate, but a ratio. Because tuberculrsis
sion that follows it.
Prevalence case: A prevalence case is a house­ is a chronic disease that frequently has an in- dious onset, infection or disease that wras present
hold associate of an index case who was found to
have latent apical or manifest tuberculosis on first six months before or six months after the house­
examination within the period from six months hold was first investigated was classified as “prev­
prior to six months after the opening of the house­ alent” for the purposes of the study, unless more
precise information was available.
hold for investigation. In analyses of the study
Incidence: Prevalence and incidence often ap­
published previously, the latent apical cases ^eie
pear
in medical literature as synonymous terms,
presented separately from the manifest.
but
they
are not. Incidence is a rate of occurrence
Incidence or “new” case: A household associate
of an index case ■who developed latent apical or of some specific event among a specified popula­
manifest tuberculosis under observation. Latent tion in a stated period (per year, month, day,
apical cases were not classified as “new” in previ­ week, hour, et cetera).
Person-year: One person observed for one year
ous analyses because their onset could not be de­
termined, due to the absence of symptoms and or two persons observed for one-half year each
signs. Their inclusion in the present analysis is constitute one person-year; one person observed
based on the observation that many of them be­ for five years has five person-years of experience,
came manifest eventually and, when they did, a et cetera. It is a convenient device for taking ac­
problem of classification arose. To call them “new” count of the time element in a chronic disease of
cases at that time would have appeared to ignore a sometimes long and variable incubation period.
“Sputum-positive” household: A household in
the knowledge that roentgenologically demonstra­
ble disease had existed previously. The combina­ which the index case has or has had sputum posi­
tion of latent apical and manifest cases in the tive for tubercle bacilli at some time during the
present analysis produced results that differed
period of observation. Households in which the
from previously published study data. Because of index case died before investigation are also in­
the much greater frequency of latent apical cases cluded in this category.
among whites compared with that among non­
Other - than - “sputum - positive” household: A
whites in the study, differences in attack rates be­ household in which the index case has never had
tween the two races are washed out to some de­ tubercle bacilli in the sputum, or one in which the
gree. When comparisons in the past showed
index case has not had sputum examined. In 47
significantly higher attack rates in nonwhite
white
and eight nonwhite households, the index
households with sputum positive for M. tubercu­
case
started
out with a noninfectious sputum, but it
losis compared with the rates in whites, the cur­
later
became
infectious. The household was then
rent analysis reveals differences that are only of
reclassified
as
of the date of the finding that the
borderline significance.
sputum
contained
tubercle bacilli and the houseHousehold: A household is an aggregation of

the study activity, but normally includes five
nurses including the supervisor. A sanitarian,
typist, and clerk round out the staff. The Health
Center in which they serve is a recently con­
structed building adjacent to the new county

il

I
i

/

WILLIAMSON COUNTY TUBERCULOSIS STUDY

hold experience was grouped with that of the in­
fectious households.
Classification oj tuberculosis: The classification
used in the study is the one proposed by Opie (2).
The Opie schema fits into the official classification
of the National Tuberculosis Association as set
forth in its publication Diagnostic Standards and
Classification oj Tuberculosis, 1931 edition, but
it also differentiates “manifest” and “latent” dis­
ease. Manifest tuberculosis denotes evidence of
disease by history and physical examination and
includes manifest childhood tuberculosis. Al­
though presumably no other confirmatory evi­
dence is required, in actual practice roentgeno­
graphic confirmation was obtained in all but a
few patients who were too ill to have chest films
or who died. Latent apical tuberculosis indicates
roentgenographic lesions usually—but not exclu­
sively—limited to the apical regions of the lungs,
without associated history, symptoms, or physical
signs that could be considered characteristic of
tuberculosis. Activity of the disease is neither in­
herent nor implied in the Opie terminology of
manifest and latent apical disease. It is possible
for manifest disease to be inactive, and active
latent apical disease is conceivable, although it
may be rare. Opie also describes latent childhood
tuberculosis which denotes a lesion of the first
infection type demonstrable roentgenographically,
but without symptoms or physical signs. In the
beginning of the study a diagnosis of latent child­
hood tuberculosis was made with great frequency
because of the high prevalence of pulmonary cal­
cifications in this area. Later, when it was shown
that such calcification is often associated with a
negative tuberculin reaction, the diagnosis was sel­
dom made. Subsequently, it was shown that many
of these calcifications were probably the result of
histoplasmal infections.

Clinical Classification of Opie

Latent tuberculosis:
A. Latent tuberculosis with hypersensitiveness
to tuberculin demonstrated by the intracutaneous test, but with no lesions of the
lungs demonstrable by roentgenographic
examination .
B. Calcified nodules in the lung
C. Latent infiltration of childhood type:
(a) Soft or flocculent (potentially pro­
gressive)
(b) Strandlike (healed)
D. Tuberculosis of tracheobronchial lymph
nodes:
(a) Calcified
(b) Massive caseous
E. Latent apical tuberculosis (adult type of
latent tuberculosis):
(a) Supraclavicular (see further subdivi­
sion)
(b) Supra- and infraclavicular (see fur­
ther subdivision)

7

Manifest tuberculosis:
F. Manifest pulmonary tuberculosis with sputurn negative for tubercle bacilli:
(a) Childhood type (primary type)
(b) Adult type (reinfection)
G. Manifest pulmonary tuberculosis with tu­
bercle bacilli in sputum:
(a) Childhood type
(b) Adult type
H. Tuberculosis of organs other than the lungs
I. Suspected tuberculosis
J. Miliary tuberculosis
Latent apical tuberculous lesions should be
classified as follows:
I. Scant apical lesions indicated by salients
or spots below the margin of the second
rib
II. Apical lesions extending over half or
more of the space above the clavicle
III. Lesions above and below the clavicle of
the extent designated by the Standards
of the American Sanatorium Association
as minimal (but with no symptoms or
physical signs significant of tuberculosis)
Stage of tuberculosis: The classification proposed
by the National Tuberculosis Association in its
publication Diagnostic Standards and Classifica­
tion of Tuberculosis (1931 edition) is used.
Statistical Methodology

The story of Dr. Wade Hampton Frost’s in­
terest and participation in the study has been
told in the Introduction. Dr. Kenneth F. Maxey,
as editor of a collection of Frost’s papers pub­
lished posthumously (36), describes him as the
man who made the transition from descriptive
to analytical epidemiology. To meet the needs
of the new approach, he needed to find, adapt,
or develop new analytical methods.
Frost’s previous experience had been chiefly
in the epidemiologic study of acute infectious
diseases. When confronted with the special prob­
lems inherent in the study of a chronic disease,
he realized that methods would have to be
changed. He said that
.. studies of tuberculo­
sis must be carried out in longitudinal section
rather than cross section as are studies of acute
infections.” It was necessary to develop a
method, comparable to Chapin’s secondary at­
tack rate for acute infectious diseases, that
would express the risk to associates of persons
with tuberculosis. Such expression of risk could
then be compared with the experience of persons
in a control population. However, whereas in an
acute disease the issue may be settled within a
few days or weeks, for tuberculosis it may not
be resolved for years. Frost stated that 11.. . ob­
servation of the exposed group must extend

8
I!

I

I
I

WILLIAMSON COUNTY TUBERCULOSIS STUDY

over a sufficient number of years to define the Frost designated such an individual the “index”
rates of morbidity and mortality prevailing in case because the discovery of this person led
successive periods throughout the usual span of to the investigation of his household associates.
The index case may or may not be the very first
life”(l).
Person-years concept and the life-table person with tuberculosis in a specific household,
method: In order to take account of the time but it is the first known to the investigator.
factor, Frost modified methods he found in pa­ Often, investigation turns up other diseased per­
pers by Elderton and Perry on mortality of sons in the same household, and it may be im­
people treated in sanatoriums (37), and par­ possible to establish the chronology br sequence
ticularly in Weinberg’s report of his study of of disease occurrence, and therefore, impossible
mortality of children born of tuberculous par­ to designate the “source” case. Without a pri­
ents (38). Guided by these he developed a modi­ mary case it is not proper to speak of “second­
fication of life-table construction which allowed ary cases” or “secondary attack rates” in the
him to define risk and experience in terms of sense used by Chapin in his studies of acute
“person-years.” He ’used this method in the communicable diseases. Furthermore, by us­
Kingsport Study to calculate age-specific mor­ ing the index case to identify the people at risk,
tality rates for comparison with rates in the the whole universe of such people is included.
general population (1). The method was later This would not be true in those instances in
utilized in the Williamson County Study, in which a “primary” case of tuberculosis had died
which its application to the analysis of retro­ and no current case of the disease was present.
The index case becomes the pivotal or focal
spective data (4), course of disease (7), and
person
in a household, the frame of reference,
observational morbidity and mortality data
(10) was described. Of course the method has so to speak. Everything of epidemiologic import
been used throughout the final analysis of the that happens in the household is related to the
date of investigation of the household (which
data presented in this paper.
In order to obtain the data required for the coincides closely with the date of discovery of
application of this method, it was necessary to the index case). The individual members of the
know the date of establishment of the household, household are identified in terms ol their rela­
the former members present before the date of tionship to the head of the household and to
investigation but not currently in the household, the index case. Antecedent experience is meas­
and retrospective historic data concerning the ured back in time from the date of investiga­
occurrence of tuberculosis and deaths from the tion. Prevalence data are related to the date of
disease among former household members. Spe­ investigation, and observational or incidence
cial forms were therefore devised; they were data start with that date.
The use of the index case served another, more
tested in the Kingsport Study, modified as in­
subtle,
purpose. The Kingsport Study showed
dicated, and then used throughout the William­
son County Study, with other modifications that up a hidden bias or “joker,” as Frost liked to
became necessary. The special forms were used designate it. When a comparison was made of
for recording the information listed above. Re­ antecedent mortality in the study households
productions of the forms used in this study are and in the control population, the former group
unexpectedly showed lower rates for the ages 20
shown in the Appendix, pages 42-51.
Concept of the index case: The onset of a dis­ to 49 years. The “joker” was discovered by
ease like tuberculosis is difficult to pinpoint be­ Frost who showed that in the study households
cause symptoms may be vague or altogether ab­ the antecedent history was given usually by a
sent for a considerable period of time following parent who had been present at the time of es­
the development of a lesion. It is often impossi­ tablishment of the household, and was obviously
ble, therefore, to determine the “primary” case alive. In the control families, however, the “in­
of tuberculosis in a household in which more formant” was not necessarily a parent, but was
than one case has occurred. The first person in occasionally a survivor of a family that was
a household who comes to the attention of medi­ broken by the death of a parent. 1 his situation
cal or health authorities because of tuberculosis tended to swell the antecedent mortality rates
is a definite entity, around which the experience in the coirtrol families. By eliminating the adult
of other household members can be observed. informant from the study household population

WILLIAMSON COUNTY TUBERCULOSIS STUDY

at risk, the adjusted rates came into line with
the expected rates.
By casting the index case in the role of the
excluded “informant,” bias in the calculation of
antecedent mortality was eliminated. Frost also
pointed out that the index case, since he already
had tuberculosis, should also be eliminated from
any calculation of future risk of morbidity and
mortality; otherwise, bias in the direction of
increasing the rates would be introduced. By the
ingenious development of his concept of the
index case, Frost was able to overcome several
serious methodologic obstacles to the epidemio­
logic study of chronic disease.
The use of the index case in the present study
was described in two early papers that reported
on study findings (4, 11).
Although both the index case and modified
life-table concepts were developed and tested to
some degree before the Williamson County Tu­
berculosis Study was initiated, their full utiliza­
tion and exploitation were not achieved until
Frost, as consultant to the study, built them into's
its program and personally observed their ap­
plication to the longitudinal study of the chronic,
infectious, communicable disease, tuberculosis.

Objectives

In the initial planning of the study, five gen­
eral objectives were stated:
1. The systematic study of the familial in­
cidence of tuberculosis
2. Investigation into the factors and cir­
cumstances related to the “breakdown”
of an individual with the adult type of
tuberculosis
3. Investigation into the evolution of tuber­
culous infection in childhood, especially
of children who arc in close contact with
a tuberculous parent
4. An epidemiologic and clinical study of
tuberculosis of old age
5. The development of a program of tuber­
culosis control designed to have practical
state-wide application.
As the study developed, more emphasis was
placed on the epidemiologic aspects of tubercu­
losis and less on its pathogenesis and clinical fea­
tures. The latter were by no means neglected,
but the total lack of adequate hospital facilities
in the early part of the study and the shortage
of hospital beds later on considerably hampered
the originally intended clinical investigations.

9

The emphasis on the epidemiology of tuber­
culosis was clearly indicated five years after the
study began in a memorandum by Dr. Wade
Hampton Frost on December 28, 1936. Dr.
Frost posed the following questions, which have
become the heart of the study:

Given a case of tuberculosis of designated
clinical type brought to the attention of the
clinic because of known or suspected tubercu­
losis (not because of discovery in the course
of examination of a known contact):
1. What may ope expect to find in the
household associates of this individual examined at this time with respect to: fre­
quency of infection as demonstrated by
the tuberculin test; frequency of tubercu­
lous lesions of various kinds as demon­
strated by physical and x-ray examination?
0$ What may one expect to find in the
antecedent history of this household with
respect to the rate of occurrence of: mani­
fest tuberculosis; death from tuberculosis;
death from other causes?
3. What may one expect to find in sub­
sequent years with respect to the rate of:
occurrence of manifest tuberculosis; mor­
tality from this cause and from other
causes?
x
Ahich of the material that has been reported
in the past, and that will be presented here,
actually contains answers to these questions.

Plan of the Study
The study was concerned with the experience of
people as members of a household in which they
were exposed to an individual with known tuber­
culosis, who was termed the “index case.” The
study was planned on a household, rather than on
a familial basis, because it was believed that the
inclusion of all household associates of a known
tuberculous individual would provide information
concerning their relative risk of developing the
disease when the exposure factor was held more or
less constant while relationship to the index case
varied.
Source oj cases: At the outset the households
of all reported cases of tuberculosis carried on the
register of the Williamson County Health De­
partment, as well as the households of all those
reported to have died of tuberculosis during the
preceding five years, were included in the study.
This plan produced an immediate large body of
households for investigation and observation. As
new cases of tuberculosis were discovered in the
community, they and their households were added
to the study group. In addition, known tubercu­
lous persons who moved into the county were
taken up for study.

10

1

[/

WILLIAMSON COUNTY TUBERCULOSIS STUDY

New cases of tuberculosis became known to the hold, and included information concerning the
type of dwelling, ownership or rental status; type
study in any one of a number of ways. Many
of community; number of people in the dwelling,
were referred by local physicians who had estab­
number of rooms, number of sleeping rooms, num­
lished a diagnosis, or suspected the presence of
ber of beds, and the sleeping arrangements; source
the disease. Others presented themselves to the
study for roentgenographic examination either be­ and type of water supply, bathroom and toilet fa­
cause they had symptoms, because they were cilities; condition of screening if present; source
motivated .by health education activities even of milk supply, whether pasteurized or not, and
quantity consumed daily; tuberculin status of
though they had no symptoms, or because they
cows; evaluation of personal hygiene of household
were required by regulation to have periodic chest
films in order to qualify for positions (teachers, members; and summary of former residences in
• comparison with present quarters.
food handlers).
At the initial investigation of the household,
In the beginning, people with tuberculosis,
arrangements were made for the examination of
whether the disease was active or inactive, were
accepted by the study as index cases and their all household members. The head of the house­
hold was urged to bring all members to the study
households were brought under investigation. In
offices, located in the county Health Center,
time, however, it was deemed advisable, because
where chest films could be taken and examina­
of the great load carried by the study staff, to
tions and tuberculin skin tests could be performed.
limit acquisition of new households to those of
persons with active, manifest tuberculosis. It was Records of examinations were made on Forms
believed that this change would in no way alter 984 to 987 (Appendix, pp. 45-50). The initial exam­
ination of all household members soon after the
the achievement of the original objectives, since
discovery of the index case was of course essential
inactive cases ordinarily would not be expected to
in order to establish the baseline status of the
give rise to other cases.
household with respect to tuberculous infection
Initial investigation of the household: As soon
as possible after the index case was discovered, and disease. In this way a large body of preva­
lence data was accumulated. These will be pre­
the study epidemiologist and a public health
nurse visited the patient’s household. At this visit, sented in a following section.
Coiitinued observation oj the household: At
in addition to the usual public health services
rendered to a tuberculous patient and his family, yearly intervals after a household was brought
under investigation, the public health nurse visited
the collection of study data began. A roster of
to check the household roster. (When an active
current members of the household was recorded
case of tuberculosis was present, the nurse visited
on Form 981 (Appendix, p. 42) which included,
the patient quite frequently; but for the purposes
besides identifying data for each individual and
his relationship to the head of the household, in­ of the roster check she made a special annual
visit.) The nurse carried her own record of the
formation concerning the date of establishment of
roster for reference, and recorded removals from
the household, date of investigation, and date and
result of the first examination of each person. Al­ or additions to the household since the previous
check, with dates. Because the experience of
though, in the actual analysis of data, relationship
household members was to be counted only dur­
was always classified with the index case as the
point of reference, on the roster, relationship re­ ing their residence in the household, it was essen­
tial to have precise dates (month and year) of re­
ferred to the head of the household because his
moval
or entry.
was the more stable position in the household.
At the time of the household check the nurse
Relationship to the index case could then be de­
termined from the recorded data. The current inquired concerning the health of each member
household roster form was subsequently used for during the preceding year, recorded any reported
illness by code on Form 988 (Appendix, p. 51),
periodic household checks (see below). A second
and wrote a note on the Nursing Notes sheet in
form that was completed on the first investigation
the
person’s record (Form 987, Appendix, p. 50).
of a new household recorded information con­
At
this
time she also reminded the household
cerning former members of the household living
elsewhere or dead (Form 982, Appendix, p. 43). members that they were due to return to the
study offices for their periodic check which in­
Included were all members who had ever lived in
the household at any time since its establishment, cluded at least a chest film, and at times a Man­
toux test. White household associates less than
but who had moved or died before the household
five years of age were re-examined every six
was opened for investigation. Here again, identify­
months, and adults were examined at yearly
ing data for each former household member were
intervals. At first this plan was followed for
listed, together with information concerning the
date and length of residence in the household, nonwhite associates also, but later they were
history of tuberculosis and date, current address re-examined roentgenographically at six-month
intervals regardless of age, because it had been
if living, or cause and date of death if deceased.
found that, if there was a longer interval, some
A third form (Number 983, Appendix, p. 44)
completed at this time was the Environmental developed tuberculosis and progressed rapidly to
Record. This was a record of the socioeconomic an advanced stage before they were discovered.
Households continued to be observed periodiand environmental hygienic status of the house-

r<

WILLIAMSON COUNTY TUBERCULOSIS STUDY

cally as long as the members remained a fairly
intact unit and continued to reside in the county.
Individual members were observed only during
their residence in the household. If they removed,
they were dropped from observation for the pur­
poses of the study even though they might con­
tinue to live in the county. If they returned to the
household at any time, observation was resumed
as of the date of re-entry. Similarly, if a house­
hold removed from the county, observation was
discontinued, but if and when it returned, obser­
vation was resumed. In some instances households
became fragmented through marriage and re­
moval of children after parents had died, so that
the family ceased to exist as an integral unit. In
such circumstances it was deemed profitless to at­
tempt to follow all the separate splinters of a
former household, and all the members were
dropped from observation with the exception of
the index case, if still living.
In time, as observation of household members
continued over the years, new cases of tubercu­
losis developed among them. Some improved and
achieved inactive status, others progressed to more
advanced stages of the disease, and some died.
These experiences form the basis of the incidence
studies to be presented later.
In a community that is predominantly rural
and agrarian, but in which approximately onefourth of the population lives in a rural town, it
is difficult to apply standards or criteria with
which to measure the socioeconomic status. Nu­
trition is of recognized importance as a factor in
tuberculosis. Poverty and malnutrition may be
synonymous in urban environments, but such a
relationship will not necessarily hold in an agrar­
ian economy in which many of the people raise
their own food and do not need much cash to
supply this necessity. At the outset, therefore,
one may question the importance of the socio­
economic factors in determining the occurrence of
tuberculosis in a rural society. However, much has
been written about this factor, and to neglect it
entirely in a study of this kind is to court valid
criticism. Besides, poverty means other things in
addition to malnutrition, such as delayed and in­
adequate medical care, crowding, poor education
and hygiene, inadequate rest and recreation, and
so forth. It was deemed advisable, therefore, to
investigate the influence of the socioeconomic
factor in this study, recognizing meanwhile that
it would be difficult to establish universally ac­
ceptable criteria for evaluating the socioeconomic
status of the study households.
It was decided to set up only three major socio­
economic classifications, i.e., upper, middle, and
lower. When each household was originally opened
for investigation, an Environmental Record
(Form 983) was completed by the epidemiologist,
aS described on page 10 of the text. Information
was taken from this form, coded, weighted, and
scored. The total score determined the classifica­
tion according to standards that were set up arbi­
trarily by the study after consultation with a

11

number of responsible people in the community
and after deliberation by the study staff. No ef­
fort was made to. devise standards for general use
in other communities. The study was concerned
only with the local situation, and attempted to
determine what standards applied in the county.
The following is an itemization of the various at­
tributes considered, together with the score given
each variant:
Item

Variant

Score

Rooms per person

2 or more
1.0 to 1.9
Less than 1

1
2
3

Persons per bed

Less than 2
2
More than 2

1
2
3

Persons per bed­
room

2
2.1 to 2.9
3 or more

Water supply .

In the house
Outside—near house
Outside—distant

1
2
3

Toilet facilities

Water closet in house
Sanitary privy
Open privy
None

1
2
3
4

Bathing facilities

Present in house
None in house

1
3

Screening

Good
Defective
None

1
2
3

I

1
2
3

Classification
Upper socioeconomic class—town
rural

7-8
7-9

Middle socioeconomic class—town
rural

9-14
10-15

Lower socioeconomic class—town
rural

15+
16 +

The same standards were used for both white
and nonwhite households.
Composition of the Study Group

During the 24 years of the study, 828 house­
holds (680 white and 148 Negro) were under ob­
servation for periods of time ranging from a few
months to the entire 24 years. In these house­
holds during this time there were 4,214 persons
other than index cases (3,446 white and 768 Ne­
gro), distributed by age, sex, and race, as shown

12

I

4

WILLIAMSON COUNTY TUBERCULOSIS STUDY

in table 1. There were 110 more males than fe­
males in the white households, while in the Isegio
households there was an equal distribution by
sex.
Household members were under observation for
a total of 32,175.25 person-years, or an average of
7.6 years per person. The average length of ob­
servation was similar for whites and nonwhites,
and the range was from a few months to the en­
tire 24 years of the study. The distribution of
the experience of household associates of the in­
dex case is shown in Table I‘ by age, sex, race,
and relationship to the index case. Males contrib­
uted proportionately more to the total experience
than females, particularly among the whites.
A comparison of the age, sex, and race distri­
bution of the Williamson County population (es­
timated as of January 1, 1944, the mid-period of
the study) and the household associates popula­
tion is shown in Table II. In the households 81.8
per cent were white, while in the county the
whites comprised 78.2 per cent of the population.
The households had slightly more males, 51.8 per
cent compared with 50.3 per cent in the county
population. In both the white and Negro house­
holds there were comparatively fewer subjects
less than five years of age, and more in the age
group of 55 years and older than in the county
population.
The average number of subjects in association
with the index case at one time or another in each
household was quite the same for both races (5.1
for white and 5.2 for nonwhite).
The study households were divided socioeco­
nomically as follows (a few households could not
be classified because of insufficient information):

be added by the inclusion of relatively meager
later data.
In the presentation of the study data, findings
are considered not significant if observed dif-.
ferences are less than twice their standard er­
rors; of borderline significance if the differenceis between 2.0 and 2.4 times its standard error;
and significant if the difference is as much as.
or greater than 2.5 times its standard error.
Whenever feasible, observed rates were ad­
justed by using the age and sex distribution of
the total household population as standard.
Household Studies
In an effort to answer the questions posed by
Frost, the study material will be presented
under the headings of “Antecedent History/'
“Prevalence Data,” “Attack Rates,” and “Death
Rates.”

Antecedent History

Historic data obtained from lay people are
notoriously unreliable in most circumstances,
and there is understandable hesitancy in attach­
ing much validity to them. Special effort was
made in the study, therefore, to substantiate the
history of tuberculosis or death from the disease*
by recourse to existing official and nonofficial
records and by a search of the memory and files
of the local practicing physicians. In a small,
relatively stable community such as Williamson
Negro
White
County, the personal history of its inhabitants
Num­ Per
Num­ Per
Cent
ber
Cent
ber
is generally pretty well known. Physicians have
1.4
2
Upper
68 10.2
Upper
treated several generations of the same family
38.1
Middle 56
Middle 294 44.2
in many instances. Records of vital statistics are
60.5
89
Lower
45.6
Lower 304
frequently kept in the family Bible. From all of
these sources, sufficient verification of historic
100.0
147
Total
666 100.0
Total
data was obtained to allow the study personnel
Results
to place more confidence in this material than
might ordinarily be accorded to it. That this
Earlier results of the study have been reported in nineteen papers and two books. A list
confidence was not misplaced may be inferred
of these will be found in the references. The
from a comparison of historic and observational
data in a previously reported analysis of the
present report generally covers the same mate­
rial that has been presented in the past, but
experience of children of an infectious parent
much of it now has the complete data for the 24
(21).
years. In some instances the material is handled
By the application of life-table methods, the
differently than it was when originally reported.
annual and cumulated morbidity and mortality
A few of the present analyses have never been
from tuberculosis in study households of white
reported before. Some previously published data
persons was calculated for the ten years preced­
are included as originally reported for the sake

ing
the onset of disease in the index case, for the
of completeness and because nothing new could
year of onset, and for the ten subsequent years.
The results of the analyses are shown in Tables
‘ All tables with roman numerals are appendix
III and IV. These data were reported early in
tables.

WILLIAMSON COUNTY TUBERCULOSIS STUDY

13

TABLE 1
Distribution of Household Associates of Index Cases by Age*, Sex, and Race
Williamson County, Tennessee, 1931-1955
Total

White

Nonwhite

Age Group
Total

Male

Female

Total

Male

Female

Total

Male

Female

Total

4,214

2,162

2,052

3,446

1,778

1,668

768

384

384

<5
5-14
15-24
25-34
35-44
45-54
55-64
65-74
75 +
Unknown

743
1,003
845
455
372
282
245
173
93
3

366
553
416

377
450
429
210
176
143
124
93
49
1

613
808
687
374
305
236
193
143
84
3

298
448
333
206
166
119
96
68
42
2

315
360
354
168
139
117
97
75
42
1

130
195
158
81
67
46
52
30
9

68
105
83
39
30
20
25
12
2

62
90
75
42
37
26
27
18
7

*

215
196

139
121
80
44
2

Age at beginning of observation.

the history of the study (4), but they include a
rather large portion of the total study popula­
tion because the greatest accumulation of house­
holds and individuals occurred at the beginning
of the study, as described above. Actually in­
volved were 468 index cases, 1,965 household
associates present at the first investigation of
the households, 777 former household contacts
who were still alive at the time of first investiga­
tion but were living elsewhere, and 427 former
contacts who had died before their households
were brought under observation. These house­
hold associates contributed 16,928.25 personyears of experience. For both the infectious and
other households, there was slightly more tuber­
culosis morbidity and mortality during the years
following the onset of disease in the index case
than before. What is of particular importance
is the recognition of the fact that the households
of the index cases had considerable experience
of cases of tuberculosis and resulting deaths well
before the discovery of the index case. The data
also show the relatively greater risk of disease
and death among associates of ^sputum-positive,”
compared with other, index cases.
The antecedent history of 913 known tuber­
culous subjects revealed that in 63.5 per cent
of those less than 45 years of age, and in 52.8
per cent of those more than 45, there had been
tuberculosis in parents and/or siblings (16). This
offers no surprise since the familial character
of the disease had been recognized long before
the bacteriologic era of medicine. It is note­

worthy, however, that in each group referred
to above, more than 22 per cent of the tubercu­
lous subjects who reported familial disease had
not had household contact with other diseased
members of the family. This observation adds
weight to the evidence suggesting a familial sus­
ceptibility to tuberculosis. This concept was de­
scribed in the Introduction and developed in a
monograph (22).
The foregoing is an anwscr to the second ques­
tion put by Dr. Frost in the memorandum
quoted above concerning the antecedent history
of tuberculosis in the study households.

Prevalence of Infection
As stated previously in the description of the
study plan, it was important to know at the out­
set how many of the household associates of a
newly discovered index case were already in­
fected. Infection was measured by the reaction
to the intradermal injection of Old Tuberculin.
The tuberculin used throughout the study was
from a lot obtained from and standardized by
the Phipps Institute in Philadelphia. It was
tested from time to time to be sure that it had
lost none of its potency. Reactions counted as
significant varied with the dosage used. Thus,
when 0.01 mg. (1 tuberculin unit) was used, any
induration of 5 mm. or more was considered pos­
itive. When 0.1 mg. of OT (10 tuberculin units)
was used, only induration of 10 mm. or more was
deemed significant. Although initially many per­
sons who had negative reactions to 0.01 mg. were

*

14

I

t

i

I

WILLIAMSON COUNTY TUBERCULOSIS STUDY

tum-positive” index cases the prevalence of in­
fection ranged from 46.2 per cent in those less
than five years of age to 95.0 per cent in those
55 years of age and older. In the Negro house-;
holds the respective percentages were 52.2 and
100.0. The increase of tuberculin reactors with
age is not a manifestation of a greater susceptibil­
ity to infection in the older age groups, but is
rather an expression of the cumulative effect of
long-term exposure to a chronic communicable
disease that may be present but undetected for
a considerable period of time.
In a number of households the members were
exposed not only to the index case but also to
others, the “prevalent” cases, that had not been
discovered before the household was opened for
investigation. However, multiple exposure to
more than one case of tuberculosis apparently
had no effect on the prevalence of infection
among the household associates. It is shown in
Table VII that the adjusted percentage of reac­
tors among those exposed to both index and
prevalence cases is almost identical with the
percentage among those exposed to an index
case only (54.4 and 54.8, respectively).
Apart from the influence of exposure and
dosage, infection occurred with equal frequency
in males and females, and in white and Negro
associates. It is important to bear this in mind,
particularly when attack and death rates are
considered. Whatever differences may be noted
by sex and race in the risk of developing tuber­
culosis or dying of it, such differences are not
based on any dissimilarity in the susceptibility
to tuberculous infection.
Relationship to the index case had little in­
fluence on the risk of becoming infected. In
the white households containing an infectious
JOOindex case, the percentage of reactors was 73.8
90SPUTUM
POSITIVE
for close relatives compared with 67.6 for other
80
household members. When adjusted for age the
70OTHER
UJ
o
difference is of borderline significance, but com­
2 60
parison at each age shows a consistently higher
5o 50
percentage of reactors for close relatives, with
UJ
40
but one exception, the 15- to 24-year age group.
30
In all other households, both white and Negro,
20
relationship to the index case did not influence
10the prevalence of infection among associates.
0.
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
The socioeconomic status of the households
AGE IN YEARS
appeared to have a bearing on the frequency of
Fig. 2. Prevalence of tuberculous infection in infection only in the white households containing
white household associates, by age and sputum an infectious index case (Table VIII), but
status of the index case, Williamson County, Ten­
not in the others. In the former, the adjusted
nessee, 1931-1955.

retested with 1.0 mg. of antigen, these results
have not been included in the final analysis be­
cause 6f the strong possibility of false positive
reactions as shown by Palmer and his associates
(39). Not all household associates were tested
with tuberculin. However, it was assumed that
the experience of the tested subjects was repre­
sentative of the group.
By the time the index case had been discov­
ered and the household members had been
brought under observation, many of them were
already infected. Infection appeared to depend
on the amount of exposure and on the dosage of
organisms. The influence of dosage on the risk
of becoming infected is shown in comparing the
percentage of reactors in “sputum-positive and
other households (Tables VA and B, VIA and
' B). When household associates were exposed to
a “sputum-positive” index case, 3 of 4 became
i infected (71.4 per cent for white, and 70.2 per
: cent for nonwhite associates). On the other
hand, when there was exposure to an index case
in whom no sputum positive for M. tuberculo­
sis had been demonstrated, slightly more than
one of every 3 became infected (37.0 per cent
for white, and 37.1 per cent for Negro house­
hold members). The risk of infection appeared
to be almost twice as high in the “sputum-posi­
tive” households as in the others (figure 2).
Most striking was the extraordinarily high fre­
quency of infection among those less than 15
I years of age in the “sputum-positive” households.
The percentage of tuberculin reactors in­
creased with age in both “sputum-positive and
in “other-than-sputum-positive” white and
Negro households. For white associates of “spu­

WILLIAMSON COUNTY TUBERCULOSIS STUDY

percentage of reactors was 65.9 for the upper
and middle socioeconomic classes, and 81.5 for
the lower socioeconomic classes. Among children
less than 15 years of age in the lower class
households, 67 per cent were infected compared
with 46.0 per cent in the combined upper and
middle class households. This age group con­
tributed most heavily to the higher frequency
of infection among associates in lower class
households.
In summary, it may be said that in these tu­
berculous households there was a high degree of
susceptibility to infection that was not signifi­
cantly influenced by sex, race, or socioeconomic
factors. Age was of importance chiefly as a
measure of the increasing opportunity to become
| infected over the years through repeated expo­
sures in the household and in the community.
Relationship to the index case appeared to have
some significance in the white households con­
taining an infectious index case. The most
important factor seemed to be that of exposure
within the household to an individual with spu1 turn positive for M. tuberculosis. When this oc­
curred, the risk of becoming infected was sub­
stantially greater than in a household in which
no bacilli had been demonstrated in the sputum
of the index case. This might be termed the
“dosage factor,” but it can be applied only
loosely because of difficulty in quantitating it.

Prevalence of Disease
Because of the insidious nature of tuberculo­
sis, undetected cases occur. From a public health
standpoint it is extremely important to discover
them. For the purposes of the study, their de­
tection was imperative because it was essential
to differentiate household members with and
without disease in order to establish the identity
of the group at risk. In white households 188
cases of previously undetected tuberculosis were
found on first examination. In nonwhite house­
holds 41 prevalent cases were discovered.
The diagnosis of existing disease at the first
investigation of the household depended largely
on roentgenographic findings, supported by tu­
berculin tests and sputum examinations. How­
ever, early in the study a few patients were diag­
nosed on clinical criteria alone because they were
too ill to be brought to the X-ray unit, and no
portable equipment was available at that time.
Although the objective was to examine every
household member, that goal was never achieved.

15

In calculating the prevalence of tuberculosis in
the study households a problem arose. If the
population denominator was considered to be
only those examined, the calculated prevalence
was extraordinarily high because the examined
group was heavily weighted with subjects who
had symptoms. If, on the other hand, the entire
household population present on first investiga­
tion was used as the denominator, the calculated
prevalence was too low because this group im
chided those with undiagnosed, asymptomatic
disease who had not been examined physically or
roentgenographically. After weighing the prob­
lem it was decided to use the whole household
population as the denominator because it was
believed that under continued observation most
if not all of the undetected cases would even­
tually be discovered. Actually, several persons
who had not been examined at the time when
their household was first investigated were later
discovered to have disease of such character
that it was possible for the clinician to estimate
its duration and include such cases among the
11 prevalent” rather than the “new.” In the data
presented below the whole household popula­
tion was used as the denominator in calculating
the prevalence of disease. This procedure differs
from that followed in previously published anal­
yses.
Since disease, when it occurs, must necessarily
follow upon infection, parallel prevalence curves
might be expected for both were it not for the
fact that the interval between the two events
may be long and variable in a chronic disease
like tuberculosis. It is not surprising to find,
therefore, that, although there is a general in­
crease in the prevalence of tuberculosis with age
among household associates, just as there was an
increase in the percentage of those infected, the
two curves are not exactly parallel (figure 3).
As with the prevalence of infection, the “dos­
age factor” appeared to have some influence on
the prevalence of tuberculosis among household
associates. In the white “sputum-positive” house­
holds 9.8 per cent (adjusted) of the members
were found to have tuberculosis on first exami­
nation compared with 6.0 per cent in the “otherthan-sputum-positive” households (Tables IXA
and B). The difference was significant at the
<0.01 level. In general the prevalence of dis­
ease tended to rise steadily with age up to 55
years. It was consistently higher at each age,

WILLIAMSON COUNTY TUBERCULOSIS STUDY

16

OTHER-THAN-SPUTUM-POSITIVE INDEX CASE

SPUTUM - POSITIVE INDEX CASE
lOOOr
90.0

INFECTION

80.0
700

I

INFECTION

600
50.0
400
300

>
20.0

DISEASE

DISEASE
10.0-

/

z

Ixl
O

/

(t
Ixl

/

/
/

/
/
/
/
/
I
I

\/

81
0.7-

I

0.5 ■

0.40.3-

I
I
I
L_________

0.2-

0.1--

0

5

10

15 20 25 30 35 40 45 50 55 60 65 70 0
AGE IN

I

/

I
I
I
I

0.6-

I

/
/

5

,

10 15 20 25 30 35 40 45 50 55 60 65 70

YEARS

Fig. 3. Prevalence of tuberculous infection and disease among white household associates,
by age and by the sputum status of the index case, AA illiamson County, Tennessee, 1931 1955.
except 55 years of age and older, in the “sputum­
positive” households than in the others.
Among Negro associates, the comparison of
“sputum-positive” households with other house­
holds revealed similar findings to those described
for the whites (Tables XA and B). Prevalence of
disease was significantly higher in the “sputum­
positive” households than in the others; the levels
in the former were consistently higher at each
age; and they tended to rise regularly with age.
Perhaps the most significant finding was the com­
paratively high prevalence of disease in Negro as­
sociates less than 15 years of age (5.2 per cent)
compared with that in white households contain­
ing an infectious index case (2.3 per cent). Alter
15 years there was considerably more disease in
the white associates, 15.0 per cent compared with
9.6 per cent in the non white associates.
\ The higher prevalence of disease in the “spu­
tum-positive” households was not surprising,
since in them there was substantially more infec' tion than in other households, and therefore a
larger group with the potential of developing tu­
berculosis.

The sex factor appeared to have only a little
influence on the prevalence of tuberculosis. In
the ‘'sputum-positive” households of white per­
sons, adjusted percentages of 8.8 for males and
10.9 for females were not different (Table IXA).
In the other white households the females, with
an adjusted prevalence of 7.3 per cent, were only
slightly higher than the males, with 4.9 per cent
(Table 1X13).
Among Negro associates there was also little
difference in the prevalence of tuberculosis by
sex. In the “sputurn-positive” households the ad­
justed percentages of 7.1 for males and 9.5 for
females were not significantly different (Table
X/1). In the other Negro households the num­
bers were too small for comparison.
It should be noted that, although the differ­
ences in the prevalence of disease in males and
females were not large, they tended to be uni­
formly higher in toto in the female household
associates, and were also generally higher at each
age with the exception of those less than five
years old. In that age group the males had more
disease on first examination, and it will be shown

WILLIAMSON COUNTY TUBERCULOSIS STUDY

later that they also developed disease more fre­
quently under observation.
A comparison of the prevalence of disease by
race showed little difference in the totals, with
percentages somewhat lower for Negro asso­
ciates than for whites in both the “sputum-posi­
tive” and other households. Among the Negroes,
the most significant finding was the relatively
high prevalence of tuberculosis in those less
than 15 years of age in “sputum-positive” house­
holds'. This has been mentioned previously in the
presentation of age-specific data for the preva­
lence of disease.
Considering that tuberculosis has been a more
virulent and fatal disease in Negroes than in
whites, it may be surprising at first glance that
not more disease was found in the Negro house­
holds under study. It must be emphasized that
these are prevalence data representing a count of
survivors. It will be shown later in the discus­
sion of the course of tuberculosis that it was gen­
erally a rapidly fatal disease in the Negro. As a
result there were not as many survivors to be
counted in the Negro households when they were
brought under observation for the first time. In
mass surveys done throughout the country it was
subsequently demonstrated that there was a
lower prevalence of tuberculosis in the Negro
population, which was in accord with the study
findings.
The chronicity of tuberculosis in the white
household population was manifested also by the
accumulation of prevalence cases in the older age
group. Of the 188 white associates who had tu­
berculosis at the time the index case was discov­
ered, 134 (71.3 per cent) were more than 35
years of age, compared with 20 of 41 nonwhite
cases (48.8 per cent) in the same age group. By
itself such a concentration of prevalence cases in
older whites would not necessarily indicate
chronicity unless it could be shown that cases
developed at a much earlier age. This was in­
deed so, as will be reported below, because the
highest attack rates among whites occurred in
those between 15 and 34 years of age, and the
highest death rates were observed in those more
than 55 years of age.
An analysis of the prevalence of tuberculosis
among household associates by their relation­
ship to the index case showed generally more
disease in close relatives than in other members.
In nonwhite households containing an infectious
index case and in white households containing

(

17

other than “sputum-positive” cases, the differ­
ences were of borderline significance. In the white
“sputum-positive” households the prevalence was
slightly higher than in other associates. The con­
sistency of the several observations gives weight
to the differences, imparting to them a signifi­
cance that is lacking when they are considered in­
dividually. A striking observation was made in
comparing the age-specific prevalence by rela­
tionship. In both the white and nonwhite “spu­
tum-positive” households, the close relatives less
than five years of age had a high prevalence of
tuberculosis, 7.7 and 5.3 per cent, respectively,
while no disease was present on first examination
of other household associates of the same age.
This may be taken as a measure of the greater
risk experienced by very young persons who were
closely related to the index case. Although it is
quite possible that a “dosage” factor was re­
sponsible at least in part, it is also probable that
hereditary and constitutional factors may have
been operating.
1 he influence of the socioeconomic factor on
the prevalence of tuberculosis was suggestive but
not clean cut. In the households of white “spu­
tum-positive” cases the adjusted percentage with
disease varied from 10.3 in the upper and mid­
dle socioeconomic classes to 11.4 in the lower
socioeconomic class (Table XI/1). The magni­
tude of the differences was not sufficiently great
to be significant. In the other white households
the prevalence was 5.4 and 7.4 per cent, respec­
tively (Table XI/3) ; here too, the variations
could have been accounted for by chance alone.
In both instances, however, the increase in prev­
alence was in the expected direction. The Negro
households had too few persons of upper class
socioeconomic status to make comparisons mean­
ingful. There was no difference in the prevalence
of disease in the combined upper and middle
compared with lower class households (Table
XII). In the other Negro households, the break­
down produced such small numbers that analy­
sis was not feasible.
Briefly, in siimmary, it may be said that the
sputum status of the index case exerted an ap­
preciable influence on the prevalence of tuber­
culosis in household associates; prevalence /
tended to increase with age; females generally I
had consistently but not significantly more dis- /
ease than males; there was slightly less disease in/
Negro associates than in white; and close rela­
tives in general had more tuberculosis than other

II

18

WILLIAMSON COUNTY TUBERCULOSIS STUDY

household members, but the most significant dif­
ference was noted in close relatives less than five
years of age. The influence of the socioeconomic
factor was suggestive in the white but not in the
Negro households.
Frost's first question, previously quoted, has
been answered by these data concerning the fre­
quency of tuberculous infection and disease
among household associates at the time of dis­
covery of the index case.
Attack Rates

I

Having established the baseline prevalence
of infection and disease, it was possible to
identify new cases of tuberculosis as they de­
veloped in the households under observation.
Over the 24 years of the study, 105 new cases—
68 white and 37 nonwhite—were discovered. At
this point it may be well to emphasize again the
definition of a “new” case as presented in the
section “Definitions.” For this final analysis of
the study data, latent apical disease was included
among the “new” cases, while in previous re­
ports only the manifest cases were considered.
The reasons for this change have been set forth
previously. The purpose in again mentioning the
change in the treatment of new cases is to indi­
cate how and where it affected the results.
Aside from the effect of the change in defini­
tion on the attack rates, the factor of time ex­
erted its influence also. In the last five years of
the study very few new cases were discovered
while the number of subjects under observation
increased. Furthermore, all previously discov­
ered new cases that survived and remained in
their households continued to add to the accu­
mulation of person-years of experience without
adding to the number at risk. The result was’ a
substantial population denominator increase of
more than 35 per cent in the white “sputum­
positive” households while the case increase was
only 20 per cent. In the nonwhite “sputum­
positive” households, only one new case was
added in the last four years while the population
(person-years) figures rose more than 16 per
cent. The effect was to reduce the attack rate.
The marked reduction in tuberculosis morbidity
in later years is concealed in the total rates, but
it was precisely because of that reduction that
the study was finally terminated, having reached
the point of “diminishing returns.”
As was the case with both the prevalence of
infection and disease in household associates, the

attack rate was also markedly influenced by the
“dosage” factor. In the “sputum-positive” house­
holds of whites, the adjusted rate of 4.5 per 1,000
person-years was significantly larger than the
rate of 1.7 in other households (Tables XIIIA
and B). Similarly, in the “sputum-positive”
households of nonwhites the attack rate was
greater than in the other households, 8.2 com­
pared with 2.9 per 1,000 person-years (Tables
XIVA and B). In both races, the rates in the
“sputum-positive” households were higher at
practically every age than the corresponding age­
specific rates in “other-than-sputum-positive”
households (figures 4 and 5).
It may be argued that there should be sub­
stantially more new cases in the “sputum-posi­
tive” households because they harbor a much
larger group of infected persons with the poten­
tial of developing tuberculosis. While this is un-^
doubtedly true, there must be other factors op­
erating because there is no true arithmetic
relationship between the number infected and the
attack rate. Among the whites, not quite twice as
many were infected in the “sputum-positive”
households as in the others, but the attack rate
was almost three times as high in the former. The
difference was even more marked in the Negro as­
sociates, in whom 1.6 times as many were infected
and more than three times as many developed tu­
berculosis in the “sputum-positive” households
compared with the others.
If there had been a true arithmetic relation­
ship between infection and the development of
disease, there would have been a steady rise in
new cases with age. Such was not the case. In
the white households containing an infectious
index case, three peaks of incidence were noted:
in those less than five years of age, 15 to 34
years, and older than 55 years (figure 6). In
the nonwhite households containing an infec­
tious index case, the incidence was also high in
the less than five-year group, but the pubertal
peak occurred earlier, at 10 to 14 years of age.
The attack rate was agam high after 55 years of
age (figure 7). In the other households, whites
showed the highest rates at 15 to 24 years, 13 of
32 cases occurring in that age group. In non­
white oilier households 2 of the total of 6 cases
occurred at ages five to nine.
The sexes showed no difference in total at­
tack rates, just as they showed none for preva­
lence of infection. However, there was a decid­
edly unequal contribution to the incidence of

WILLIAMSON COUNTY TUBERCULOSIS STUDY

19

10.0
tn

9.0

bJ

8.0

Z

7.0

LU
0.

6.0

o
o
o
or

5.0

cr
o
cn
cr

bJ
Q.

\
\

4.0
3.0

LU

<

2.0

q:

1.04

\SPUTUM
\POSITIVE

/
/

W\^
OTHER

y

0-10

5

15

10

20 25 30 35 40 45 50 55 60 65 70 75
AGE IN YEARS

Fig. 4. Tuberculosis attack rate per 1,000 person-years among white household associates,
by age and sputum status of the index case, \\ illiamson County, Tennessee, 1931—1955.
20.0

15.0bJ

< 10.0-

tr

\

1

5.0-

'

\

x

, A, , V-'"

0__ ->
0 5

b,

r-1-!

10

,

T

SPUTUM
POSITIVE

x

f^-rT

T

OTHER

t

r

15 20 25 30 35 40 45 50 55 60 65 70 75
AGE IN YEARS

E «>
“Cmmty,
•« —
.—
a ’
case, Williamson
Tenne^eeV
WSl-WSs

new disease by males and females at each of the
three critical levels described above. All of the
new cases of persons less than five years of age
were male. In the white “sputum-positive” house­
holds, attack rates were higher in the females 15
to 34 years of age than in males of corresponding
age (15 to 24 years, 12.5 per 1,000 person-years
compared with 5.9,’ 25 to 34 years, 14.8 compared
with 4.7). At 55 years of age and older, the at­
tack rate for males was 5.4, and for females 3.2
(figure 6). In Negro “sputum-positive” house­
holds the highest age- and sex-specific attack
rate in the entire study was observed in Negro
females 10 to 14 years of age. The rate of 31.4
per 1,000 person-years means that 3 per cent of
the Negro females in this age group in the “spu-

turn-positive” households developed tuberculosis
e\erj year. In these households there was an
appreciably higher incidence of tuberculosis in
males more than 55 years of age compared with
that of females of the same age (figure 7).
The high incidence of disease at both ends of .
the span of life may be explained by poor re- 1
sistance to infectious diseases during these pe- j
nods. It is not clear, however, why the resistance '
of males should be poorer than that of females.
Morbidity and mortality for all causes is higher
in male infants than in female. The experience of
the study infants appears to indicate a similar
variability by sex for tuberculosis morbidity.
There is nothing in the study data that offers

I

WILLIAMSON COUNTY TUBERCULOSIS STUDY

20
20.0

15.0I

LlI

< 10.0
(T

5.0-

----- MALE

I
I
I

\
\
—i------ t--------- --------- 1

0

0

10

5

^-'"’female

^-r

5. 55 60 65 70 75
15 20 25 30 35 40 45 50
AGE IN YEARS

Fig. 6. Tuberculosis attack rate per 1 000 person-vears among associates of white 1 sputumpositive” index cases, by age and sex, Williamson County, Tennessee, 1931-1955.
40.0

30.0-

I
I
I
I
I
I

Ul

§ 20.0-

I
I
I
I

t

10.0-

\
\
\
\
\
\
\
\
\

MALE

/>

//
0-M-r
0

5

FEMALE

10

15 20 25 30 35 40 45 50 55 60 65 70 75

AGE IN YEARS

- ,vwv
---- ——associates of -tnm
nonwhite
spuUCJl a
.• ------- ---------- o
inEE
Fig. 7. Tuberculosis attack rate per
1,000
and
sex,
Williamson
County,
Tennessee,
1931-1955.
turn-positive” index cases, by age i—

I

any explanation for this finding. The higher at­
tack rate for pubertal and young adult females
is also difficult to explain, despite the fact that
this age period has been known to be ciitical foi
tuberculosis since Hippocrates first called atten­
tion to it.
It has been suggested that the decelerative
phase of growth following the onset of the
menses in the pubertal female is associated with
a diminished ability to retain calcium and nitro­
gen, and that this may lower her resistance to
tuberculosis (40). When the action of corticos­
teroids on tuberculosis became known, it was as­
sumed by some that they might be responsible
for the rise in incidence of and death from tu­
berculosis at puberty. However, there is cur­
rently no evidence to support such a view. It has
been suggested by numerous writers that preg­
nancy may be an important factor in the occur­
rence of tuberculosis among young females. Rich

(41) discusses this at some length. While preg­
nancy may be a factor in some situations, it was
found to have no influence on tuberculosis in the
study population, as will be shown below. In­
deed, there is quite as much support for the be­
lief that pregnancy, with its anabolic processes,
actually may have a beneficial effect on tuber­
culosis, provided that the pregnant woman takes
reasonably good care of herself. The question
must be left there, for the study data do not
provide an answer. They do provide, however,
an index for tuberculosis control in emphasizing
the peculiar susceptibility or poor resistance of
the pubertal and young adult female to tuber­
culosis.
In the study households, white and Negro as­
sociates started out on equal terms with respect
to the prevalence of infection at the time of in­
vestigation. It could be inferred that members
of both races had similar risk of and susceptibil-

WILLIAMSON COUNTY TUBERCULOSIS STUDY

21

ity to infection. The subsequent experience of
10.0
the two races was quite different, however. The
incidence of tuberculosis was higher in Negro CD
□ white
□r
associates than in white, whether in “sputum­
□ NONWH IT&
positive” or other households (figure 8). The ad­ Lu 8.0justed incidence of 7.9 per 1,000 person-years in V
z
o
sputum-positive” households of nonwhites had a CD
standard error 2.2 greater than the rate of 4.5 in Ct 6.0whites. The difference in white and nonwhite UJ
Q_
households containing other than “sputum-posi­ O
tive cases was not quite as great. As was pointed O
4.0out in the section “Definitions,” the inclusion of
(t
latent apical cases influenced the white much
Id
more than the Negro associates. This is one indi­ CL
cation of the greater severity of the disease that Ld
did develop in Negroes. Had these cases been ex- H 2.0<
.eluded as in the past, the difference in white and Ct
Negro attack rates would have been much
0
greater. It will be shown later that the stage of
disease on first diagnosis was also more advanced
SPUTUM-POSITIVE
OTHER-THAN
in Negroes than in whites, and the death rates
SPUTUM-POSITIVE
for the former were also higher.
Fig. 8. Tuberculosis attack rate per 1,000 per­
The explanation for the greater frequency and son-years, by race and sputum status of the index
severity of disease in the Negro has been the case, Williamson County, Tennessee, 1931-1955.
subject of considerable discussion and contro­
versy over the years. The literature abounds index case, the adjusted incidence of 9.8 per
with reports and studies that are contradictory 1,000 person-years for close relatives, although
and nullifying. There is sufficient material to almost twice as high as the rate of 5.0 for other
support any contention that may be made, if members, was nevertheless not significantly
one is willing to ignore the evidence on the other higher, but was in the expected direction. In the
side. There are proponents of the belief that the white “other-than-sputum-positive” households
generally poorer socioeconomic ’ status of the the differences by relationship were small, and
Negro is responsible. Others contend that he has there were too few cases in other Negro house­
holds to make comparisons. Perhaps of the
greater susceptibility or poorer resistance to tu­
greatest
significance is the high incidence of dis­
berculosis as a racial characteristic. Still others
ease
in
close
relatives less than 15 years of age
support the belief that all can be explained on
the basis of natural selection and the epidemic in both white and Negro “sputum-positive”
curve, the Negro having had only a scant 450 to households, with little or no disease in other
500 years of experience with the disease while members of similar age. Tuberculosis is not in­
the white man has had perhaps 5,000 years in herited. It has been suggested that an heredowhich to weed out the most susceptible and pro­ constitutional factor may, however, determine
duce a more resistant population. The study susceptibility or resistance to the disease. The
material provides no answer, but some data on data presented here cannot be said to prove that
the influence of the socioeconomic factor will be such a factor is of any great significance in in­
fluencing the occurrence of tuberculosis. The
presented below.
study findings do fall in line with the studies of
Relationship to the index case appeared to in­
animals by Lurie (42) and of human twins by
crease the risk of developing tuberculosis in the Kallmann and Reisner (43).
white households containing an infectious index
Analysis of attack rates by socioeconomic
case. The adjusted rate of 6.8 per 1,000 personstatus did not yield consistent results. In the
years for close relatives was significantly higher white sputum-positive” households there was
than the rate of 1.8 for other members. In the
more new disease in the middle and lower socio­
nonwhite households containing an infectious
economic classes than in the upper, but the in-

I
II

WILLIAMSON COUNTY TUBERCULOSIS STUDY

22

quire. Poverty and malnutrition do not neces­
crease in rates was not progressive from the up- <
sarily go hand in hand m such a situation, as is
per to the lower class (Table XVA). All of the f
from the data presented below.
new cases of those less than 15 years of age de- evident
<
Nutrition
studies were undertaken in William­
veloped in the lower class households. The only
son County under the guidance of Dr. William J.
new case of that age group in the white “other- :
than-sputum-positive” households was also in the Darby, Professor and Head of the Department
of Biochemistry, and Director of the Division of
lower class (Table XVB). In Negro households,
Nutrition of Vanderbilt University School of
all cases of those less than 35 years of age were
found among the lower class group (Tables XVIA Medicine. These were studies of the hemoglobin,
erythrocyte count, packed cell volume, and total
and B). In the “sputum-positive” households the
lower socioeconomic class experienced signifi­ serum protein of randomly selected household
cantly more tuberculosis than the combined members and other residents of the county. Al­
upper and middle classes. There were only 2 though analyses of all the data were completed,
only^the data on hemoglobin will be discussed,
cases in the latter, and, had all 31 cases been
proportionately distributed, there would have and the discussion will be limited to white house­
hold associates who lived in rural sections of the
been more than 7 cases.
that is, outside the single urban com­
A comparison of the attack rates in the lover county,
munity’ in the county. In all, 229 persons were
class “sputum-positive” white and nonwhile
tested. Their average hemoglobin level for each
households showed the respective rates to be
age group and sex showed no significant varia­
4.8 and 10.1 per 1,000 person-years. The differ­
ence was only of borderline significance. For the tion by socioeconomic class (table 2). Values
“other-than-sputum-positive” households of the for males were almost uniformly higher than for
females, but this is not considered unusual. Only
lower socioeconomic class, there was no difference
26 Negro household associates living in rural
in the incidence of new tuberculosis by race. It is
areas were tested, and the number was too small
perhaps worth repeating at this point that the
for meaningful analysis. Similarly, the number
criteria for classifying the socioeconomic status
of white and Negro urban household residents
were identical for both races. These findings sug­
tested was too small for consideration. It might
gest that, if such variables as exposure, dosage,
be
added parenthetically that a comparison of
and socioeconomic status were held constant, ad­
justed attack rates did not vary significantly in values on all tests between tuberculosis house­
hold members and other nonhousehold members
white and Negro household associates of known
cases of tuberculosis, although the rates in Negro of the community showed no essential differences
associates were uniformly higher than in whites. between them. In any case, the foregoing data
indicate that among the rural study population
This uniformity may in itself be significant.
The singularly small impact of socioeconomic of Williamson County, the socioeconomic status
factors on the incidence of tuberculosis in the did not influence the hemoglobin levels, which
study population needs explanation, particularly may be considered an index of nutrition.
The influence of pregnancy on the incidence
in view of the almost universal agreement among
investigators that these factors weigh heavily in of tuberculosis was also studied. It has been con­
determining the occurrence and distribution of tended by some in the past that the actual preg­
the disease. In the first place, when many factors nancy undermined the resistance of the pregnant
woman, and placed her at greater risk of de­
are operating, some may have considerably moie
influence than others. It appears that the ex­ veloping tuberculosis or of breaking down an
existing quiescent disease. Others have claimed
posure and dosage factors in the study house­
that the postpartum period, with all the physi­
holds had an overwhelming impact, an impact
so great that they may reasonably have dwarfed cal and emotional stress associated with the care
of a newborn infant, rather than the pregnancy
the effect of the other factors. This is not sur­
itself, produced the greater risk of tuberculosis.
prising in a communicable disease. I1 urthermore,
in considering a factor such as the socioeconomic For the purposes of the current analysis, the
pregnancy and the postpartum periods were
status, it is important to think in terms of time,
considered together and comprised the two-year

X’

A*ZXV

...

WILLIAMSON COUNTY TUBERCULOSIS STUDY

23

TABLE 2
Hemoglobin Values Among White Household Members (Including Index Cases) Living in
Rural Areas of Williamson County, by Age, Sex, and Socioeconomic Status, 1948-1951
Total

Male

Female

Age Group

Number Tested

Average
Hemoglobin

Number Tested

Average
Hemoglobin

Number Tested

Average
Hemoglobin

Total
Total

229

14.4

101

15.1

128

13.9

10-14
15-44
45+

42
139
48

14.2
14.5
14.5

24
53
24

14.2
15.6
15.0

18
86
24

14.3
13.8
13.9

Upper Socioeconomic Class
Total

6

14.8

3

15.5

3

14.1

10-14
15-44
45+

0
5
1

15.0
13.9

0
2
1

16.3

14.1

13.9

0
3
0

Middle Socioeconomic Class

Total

73

14.2

26

15.0

47

13.8

10-14
15-44
45+

9
41
23

14.0
14.2
14.4

3
13
10

14.3
15.2

14.9

6
28
13

13.8
13.7
13.9

Lower Socioeconomic Class
Total

150

14.5

72

15.2

78

13.9

10-14
15-44
45+

33
93
24

14.3
14.6
14.6

21
38
13

14.2
15.8
15.2

12
55
11

14.5
13.7
13.9

livery. Only actual household observation was
counted, so that for those who entered a house­
hold when already pregnant their experience
was counted as of the date of entry. Similarly, if
a woman left the household before delivery or
before the postpartum year was completed, she
was counted as having had household experience
only for the period of her residence there. A
woman who had a miscarriage was arbitrarily
given three months of pregnancy experience, but
the full postpartum year was counted. The re­
sults of the analysis are shown in table 3. In
white households 25 new cases developed in
women between the ages of 15 and 44. Of these,
only 2 occurred during the pregnancy-post­
partum period, and 23 at other times. The ad­

justed attack rates of 2.6 for pregnancy, and 5.3
for nonpregnancy periods, while not in the ex­
pected direction, are not significantly different.
Among the Negro women, although the rates
were higher during pregnancy, 9.5 compared
with 2.6 adjusted, these too were not signifi­
cantly different.
Thirteen white women had had tuberculosis
before they became pregnant. During preg­
nancy and the postpartum period there was no
evidence of progression of the disease in any of
them. One woman actually improved and the dis­
ease became arrested during pregnancy. One
died about two years post partum. The other 11
patients showed no change in the stage or ac­
tivity of their disease. Among Negro females

WILLIAMSON COUNTY TUBERCULOSIS STUDY

24

TABLE 3
Tuberculosis Attack Rates per 1,000 Person-Years During Pregnancy and Nonpregnancy
Periods among 15- to 44-Year-Old Female Household Associates of Index Cases,
by Age and Race
Williamson County, Tennessee, 1931-1955

Age Group

New Cases

New Cases

Person-Years
Experience

Nonpregnancy

Pregnancy

Total

Person-Years
Exerience

Number Rate

New Cases
Person-Years
Experience

Number Rate

Number Rate

White Households
Total adjusted rate*

5,122.00

15-24
25-34
35-44

2,409.50
1,362.00
1,350.50

25

G74.50

2

0

284.00
284.00
106.50

1
1
0

4.9
4.9
17 7.1
8 5.9

5.2
5.3
16 7.5
7 6.5
0

23

3.0
2.6
3.5
3.5

4,447.50

9.5
9.5
16.0
7.6

941.50

3

386.50
250.75
304.25

0
0
3

2,125.50
1,078.00
1,244.00

Nonwhite Households

Total adj usted rate *

1,258.50

6

15-24
25-34
35-44

511.75
382.25
364.50

2
1
3

4.8

317.00

3

125.25
131.50
60.25

2
1
0

44
3.9
2.6
8.2

* Adjusted by age, using the percentage
standard.

3.2
2.6
9.9

distributions of the total white household population as

there were 9 who had had tuberculosis at some
time before they became pregnant. Seven of
them showed no progression of disease dining
the pregnancy-postpartum period, while 2 pro­
gressed to more advanced disease. These data do
not indicate that pregnancy or the postpartum
period was associated with any unusual risk of
developing tuberculosis or of worsening an exist­
ing disease among female household members of

child-bearing age.
Part of Frost’s third question, relative to the
occurrence of tuberculosis in the household pop­
ulation under observation, has been answered by
these studies. The answer may be sum'inarized
as follows: The factors that influenced the risk
1 of developing tuberculosis in a household in
which a known case was present were those of
race, age, sex, sputum status, of the index case
and, to a minor extent, relationship to the index
case, and the socioeconomic status of the house­
hold. Pregnancy did not affect the attack rate in
women of child-bearing age.

Death Rates

Deaths from all causes are considered here in
addition to deaths from tuberculosis in order to

determine whether the tuberculous households
under study had a mortality experience that dif­
fered from that of the general population.
The tuberculosis data are presented first.
There were 23 deaths in white households and 29
in Negro households. Here again the dosage
factor brought an appreciable weight to bear.
Among white associates in “sputum-positive
households the mortality rate was 1.7 per 1,000
person-years (Table XVIIA). This differed from
the rate of 0.5 in other white households (Table
XVIIB) by a standard error of 2.3, with a value
of p = >0.01 and <0.05. In the Negro house­
holds the respective rates of 6.6 and 1.9 for
associates of “sputum-positive” and other index
cases (Tables XV1IIA and R) were different at
the level of 2.4 S.E. In view of the higher prev­
alence of infection and higher incidence of tuber­
culosis in “sputum-positive” households com­
pared with other households, it is to be expected
that higher death rates would also obtain.
Deaths did not occur uniformly at all ages. In
the “sputum-positive” households of whites there
were three peaks of almost equal height: in those
less than 5 years, at 25 to 34 years, and in those
older than 55 years of age (figure 9). These

WILLIAMSON COUNTY TUBERCULOSIS STUDY

25

25.0
1

(D
(E

UJ

20.0-

I

z
o

(D
(T.
Ld
CL

O
O
O

15.0-

10.0-

MOD VVH ITE

a:

LlJ
0Ld

5.0-

<
(E
0__
0

WHITE

5

io 15 20 25 30 35 40 45 50 55 60 65 70 75
AGE IN YEARS

Fig. 9. Tuberculosis death rate per 1,000 person-years among associates of “sputum-positive”
index cases, by age and race, Williamson County, Tennessee, 1931-1955.

peaks corresponded almost identically with the incidence of tuberculosis at 55 years and older,
peaks of incidence for this same household but females had a higher death rate at this age.
group. There were no deaths in the five- to 14- In the nonwhite “sputum-positive” households
year age group, the so-called “golden” years. there were 3 deaths of persons less than five years
These years were, however, not “golden” for of age. Two of these were males. Females had a
the Negro associates in “sputum-positive” house­ higher death rate than males at 15 to 24 years,
holds in which there were 4 deaths with a rate of 13.5 compared with 8.3. The males had a higher
5.0 per 1,000 person-years. The highest rate rate at 55 years and older.
among Negroes (23.4) occurred in those less than
Racial differences in death rates were marked,
five years of age. Two other high points were nonwhite associates showing significantly higher
noted, 10.6 at 15 to 24 years, and 8.8 at 55 years rates than whites in both the “sputum-positive”
and older. These peaks differed slightly from the and other households. In the former, the death
high points of incidence. It may be recalled that rate for Negroes was 6.G per 1,000 persons per
the highest attack rate was observed at 10 to year, and 1.7 for the whites. In the other house­
14 years of age. In the “other-than-sputum-posi- holds, the respective rates were 1.9 and 0.5. The
tive” households no deaths were recorded among most marked difference racially was in the age­
whites less than 35 years of age. The peak in specific death rates. Among whites in “sputum­
the Negro “other-than-sputum-positive” house­ positive” households only one of the 23 deaths oc­
holds occurred at 15 to 24 years when the rate curred in persons less than 15 years of age, while
was 5.2 per 1,000 persons per year. There was also among the Negroes there were 8 deaths in this
one death in the five- to 14-year group, but none age group of a total of 29 for associates of all ages.
It may be significant that, during the last five
among the whites.
There was no difference in total death rates years of the study, there were only 2 deaths from
by sex for white or nonwhite, for “sputum-posi­ tuberculosis in white households and none in
tive” or other households. There was, however, Negro households.
Relationship to the index case did not appre­
considerable variation in the rates by sex at spe­
ciably
influence tuberculosis mortality. In both
cific ages. The only death of an individual less
white
and
nonwhite “sputum-positive” house­
than five years of age in white “sputum-positive”
holds the death rate was higher among close
households was in a male. At 25 to 34 years, and
relatives than among other members, but the
at 55 years and older, females showed the higher
differences were not significant. The only death
rates, 4.2 and 4.3, respectively. Thus, the peaks
that occurred in an associate loss than five years
of death rates by sex differed slightly from the
of age in white “sputum-positive” households
peaks for attack rates. Males had a higher was that of a close relative of the index case. In

I

I

26

WILLIAMSON COUNTY TUBERCULOSIS STUDY

the Negro “sputum-positive” households, 6 of the formly higher in Negro households than in white,
7 deaths that occurred in persons less than 15 for associates of both “sputum-positive” and
other index cases. In the former, the rates for
years of age were those of close relatives.
In white households of “sputum-positive” Negro and white were 15.6 and 10.9 per 1,000 per­
cases, tuberculosis mortality did not vary sig­ son-years, respectively; in the latter, the rates
nificantly with socioeconomic class. Only one were 12.5 and 9.3. In every instance, except one,
death was recorded in upper class households, in death rates were also higher in each age group.
an individual more than 55 years of age. The The exception was in “other-than-sputum-posionly death that occurred in an associate less tive” white households, in which the death rate
than five years of age was in a lower class house­ for other causes among associates 55 years and
hold. In the nonwhite households of “sputum­ older was higher than in the nonwhite households.
positive” cases, however, the socioeconomic sta­ In “sputum-positive” households, rates for
tus appeared to have a marked influence on the Negro males were about 50 per cent higher than
tuberculosis mortality. There were so few upper- rates for white males, while Negro females had
class households that they contributed little to rates about three times as high as white females.
the total picture. Only one death occurred in Practically the same comparisons were noted in
upper and middle class households compared “other-than-sputum-positive households,
Total death rates for causes other than tuber­
with 24 in lower class households. The respective
culosis
were not appreciably influenced by socio­
mortality rates were 1.1 and 8.4 per 1,000 per­
economic
factors. However, in white households
son-years (Table XXA).
all
deaths
except one in those less than 15 years
A comparison of “sputum-positive” households
in the lower socioeconomic class by race showed of age occurred in lower class households. The
the Negro tuberculosis mortality rate of 8.4 to be one exception was in a middle class household,
more 1than five times as high as'the white rate of In Negro households all deaths of those less than
1.6 (Tables XIXA and XX4). Since, in this five years of age were in the lower class.
In comparing white and Negro mortality
comparison, the factors of sputum status of
rates
for causes other than tuberculosis in each
the index case and the socioeconomic class were
socioeconomic
class individually, the rates for
held “constant,” the differences observed would
have to be explained on some other basis. This Negro associates were almost uniformly higher
matter will be discussed more fully after the than those for the white associates. The reason
data for mortality from other causes have been for this is not apparent and can only be specu­
lated upon. It might be assumed that persons of
presented.
A consideration of mortality from other, non- the same socioeconomic class, regardless of race
tuberculous causes in the study households pro­ (the same criteria for classification were used for
duced some interesting results. In both white both races), would be exposed to similar experi­
and nonwhite “sputum-positive” households, ences with respect to medical care, preventive
death rates for other causes were slightly but not medical services, health education, and so forth,
significantly higher than rates in other-than- and might be expected, therefore, to react simi­
“sputum-positive” households. In all households, larly to those forces in the environment that
death rates were high in the younger than five- affect health, all other things being equal. The
year age group, fell from five to 14 years, then fact that the Negro has a much higher mortality
generally rose steadily with age as expected. In than the white, class for class, indicates that all
the nonwhite “sputum-positive” households the other things are not equal. For those who claim
rate of 23.4 for associates less than five years of that the inequality is based on the physical in­
feriority of the Negro, the athletic records estab­
age was quite high.
In general there were no significant differences lished by members of that race offer the most
in death rates for other causes by sex, except in eloquent denial. It is more likely that the stand­
the white “sputum-positive” households in which ards or criteria used for socioeconomic classifica­
the rate of 12.9 for males was significantly higher tion cannot be applied equally to both races be­
than the rate of 8.7 for females after adjustment cause of inherent and incalculable social and
cultural differences.
for age.
In summary, tuberculosis mortality rates were
By race, death rates for other causes were uni-

I

f
WILLIAMSON COUNTY TUBERCULOSIS STUDY

&

27

higher in “sputum-positive” households than in fection tuberculosis. Some of the analyses of
others; they varied with age and paralleled fairly study data show this temporal relationship.
Experience of children of “sputum-positive”
well the critical ages for attack rates; there were
parents:
These studies, which combined historic
age-specific differences in death rates by sex;
mortality rates were four and a half times as and observational data, were reported in detail
high in nonwhite “sputum-positive” households in 1954, one year before the study was terminated
as in white; relationship to the index case ap­ (21). There was not sufficient additional material
peared to exert an influence in producing higher after the report to warrant the revision of
death rates in close relatives than in other mem­ tables of graphs. Some of the original data are
bers in the younger age groups; and socioeco- . Reprinted with permission.
.Although several different kinds of analyses
nomic status seemed to .show an effect (‘hieflyin
the jLionwhite households with an infectious in- were made of the material in this group of
children of “sputum-positive” parents, the data
dex case.
Deaths for causes other than tuberculosis in relating to the interval between exposure and on­
study households generally followed an expected set of disease are of prime importance here. The
pattern by age; rates were higher in “sputum­ children were divided into four groups according
positive” households compared with those in to their age at first exposure to a parent with
others, but not significantly so; no difference by open tuberculosis: less than one year, one to
sex was noted except in white “sputum-positive” four years, five to fourteen years, and fifteen
households where the rate for males was signifi­ years and older. Cumulative probabilities of
cantly higher than the rate for females; rates developing tuberculosis were calculated (Table
were almost consistently higher in Negroes than XXI), and showed that in children first exposed
in whites in toto at all ages and, socioeconomi­ during the first year of life or after 15 years of
cally, class for class. The greatest difference by age, tuberculosis tended to develop rapidly.
race was noted for the females. Negro females When first exposure occurred between one and
had death rates about three times as high as white 15 years the production of disease tended to be
females. Socioeconomic factors did not appreci­ delayed until after 15 years of age (figure 10). In
ably affect death rates except in the younger age some instances, the interval between exposure
groups.
and disease development was 15 years or more,
j These findings tend to substantiate the data
Study of Cases
previously presented with respect to attack
rates. They indicate that there are critical ages,
Interval between Exposure and
infancy and puberty, during which first infec­
Onset of Disease
tion is extremely hazardous and often progresses
quickly
to manifest disease. Puberty also ap­
The interval between exposure to “open” tu­
berculosis and the development of the primary pears to be a critical period for the activation of
focus of infection is probably very short. The previously contained infections, some of them
sequence of events following infection may be so apparently well controlled for upward of 15
variable that it may range from a lifelong con­ years. These findings tend to support the pro­
tainment of the primary focus (the most com­ ponents of the “autogenous reinfection” school of
mon sequel) to the development of progressive, thought. I
Experience of ‘‘new” cases: The “new” cases
manifest tuberculosis. The temporal relation­
were
those that developed during observation in
ship between infection and disease may also vary
considerably. There may be no demonstrable the study households. For these cases there
interval between these two events, as in infants, was not only a record of observed exposure to
in whom the primary infection may progress tuberculosis, but also a history of contact with
immediately by local extension and by lympho­ tuberculosis before the household was brought
genous and hematogenous dissemination. In under investigation. In many instances there
most instances, however, the usual picture is had been exposure to several household mem­
that of initial local containment of the primary bers and a mixture of contact to “open” and
infection, followed after a variable interval that “closed'’ disease. The interval between first
may be many years by the development of rein- exposure and onset of disease by type of ex-

28

WILLIAMSON COUNTY TUBERCULOSIS STUDY

20

15

J

/

UJ
F-

10

15 AND
OVER

i//

O
UJ

/under
.• ONE YEAR

/
5’111

YEARS

/
5

0
0

5

10

15

20

25

30

35

40

AGE IN YEARS

Fig. 10. Cumulative probability per 100 of developing tuberculosis, by age, for children of
“sputum-positive” index cases, according to age at first exposure, Williamson County Tuber­
culosis Study.

i ■

posure and by age and race is shown in table 4.
For the whites whose first exposure was to a
person known to have tubercle bacilli in the
sputum, the general pattern was one of a short
interval between exposure and disease develop­
ment for those less than one year of age and for
the age groups of 15 to 34 and 55 years and
older. These were also the critical ages during
which the incidence of disease was high. In the
age group of one to 14 years, the interval was
generally more than 10 years. As a rule, inter­
vals were somewhat shorter when first exposure
was to an individual with infectious sputum than
when initial exposure was not of that character.
The “new” cases in Negroes showed a different
pattern. In them intervals were shorter in
almost every age group. Only 4 of 30 Negro
“new” cases whose first exposure was to “open”
tuberculosis had intervals of more than 15 years
compared with 10 of 29 among the whites.
Whereas, as noted above, most of the white
subjects in the one- to 14-year age group with
initial exposure to a known infectious case had

intervals between exposure and onset of dis­
ease of more than 10 years, Negro subjects in
the same age group usually developed disease
in less than five years.
Course of Disease

In the early years of the study the course of
tuberculosis in the household population was
practically unaffected by any man-made miti­
gating forces. Those were the years before the
state tuberculosis hospital construction program
had begun, and long before the current specific
antituberculosis therapy became available. In
effect, the natural course or history of tubercu­
losis was observed. In 1939 the first state tuber­
culosis hospital beds came into use, and after
1949 specific therapy was offered to patients.
Although tuberculosis mortality rates had
been declining steadily over the years even
before control measures came into being, the
influence of such measures, including hospitaliza­
tion and therapy, could not be discounted. With
this consideration in mind, the study period was

)

29

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE 4
Interval between First Exposure to Tuberculosis and Onset of Disease for
by Age at First Exposure, by Type of Exposure, and by Race

New”Cases,

Williamson County, Tennessee, 1931-1955

Age at First
Exposure

White

Nonwhite

Interval in Years beween Exposure and Onset

Interval in Years between Exposure and Onset

Number
of New
Cases

1-4

i

5-9

10-14

15-19

20+

Un­
known

Number
of New
Cases

<1

1-4

5-9

10-14

13

3

6

15-19 20+

Exposure to Infectious Sputum at First Contact

Total

29

2

9

<1
1-4
5-14
15-24
25-34
35-54
55 +

3
4
5
G
5
2
4

1

1

1

6

2
3

2

1

1
3

1

8
4
2

2

1

1

1

2

3

1

30
2
1
13
4
2
G
2

4

1

3

G
3

1

2
1

1

3

1

1
1

2

1

1
2
1

2

1

1

1

1

1
1
1

Exposure to Infectious Sputum with Other Exposure First

Total

1G

2

<1
1-4
5-14
15-24
25-34
35-54
55+
Unknown

5

1

6

1

2

4

1

3

1

1

1
4

3

1

3
2

1

2
1

1

1

1
1

1

I

i

I

No Exposure to Infectious Sputum
Total

23

<1
1-4
5-14
15-24
25-34
35-54
55+

2
2
4
4
1
7
3

1

5

5

2

5

1

2
1
1

1
1

2
1
1
1

2

3

5

1

1
2

1

2

2

1

1

1

1

1

1

4
1

divided into three segments: 1931 to 1939, 1940
to 1948, 1949 to 1955. The cumulative probabil­
ity of dying of tuberculosis was determined by
using the life-table method, and was calculated
by successive years for each period. Since mor­
tality was greatly influenced by the stage of
disease on first diagnosis, separate analyses were
done by stage. Only among those with moder­
ately advanced tuberculosis on first diagnosis,

1
1

1

1

1

1

1

however, were the numbers sufficiently large
to permit a comparison during the three periods.
The results of the analysis are shown in figure
11. A marked decrease in the mortality risk
was noted during the second period (19401948), but practically no change occurred in
the third period. It must be borne in mind that
no adjustment for age has been made and that
the comparisons are based on crude data. In

!
I

30

WILLIAMSON COUNTY TUBERCULOSIS STUDY

5O-,
1932-1939

40-

»Ld
O

cr

30-

••1940-1948

Ld
Q_

-------- 1949-1955

20-

10-

°-r
o

2

T

i

3

4

5

6

7

YEARS FOLLOWING DIAGNOSIS

Fig. 11. Cumulative probability of dying in white persons diagnosed as having moderately
advanced tuberculosis, for three periods of time, Williamson County, Tennessee, 1932-1955.
the earliest period a large portion of the deaths
were among young people. In the last period
deaths w^re predominantly in the older age
groups. In all likelihood, a large share of the
decline in deaths during the three periods may
be attributed to the same causes that have
brought about a steady decline in tuberculosis
death rates over the last 60 years or more. The
part played by hospitalization and drug therapy
in reducing the mortality of study patients is
difficult to assess. The absence of an appreciable
difference in mortality in the second and third
periods would suggest that the influence of
hospitalization and drugs was not marked.

Stage of Disease on Diagnosis

u

r

Index, prevalence, and new cases: In all of
the study households there were, in time, 1,124
cases of tuberculosis. Of these subjects, 790 (ex­
cluding 38 who died of tuberculosis just prior
to opening the household) were index cases, 229
were prevalence cases, and 105 were new cases.
Because tuberculosis is an insidious disease,
there is a marked tendency to delay in seeking
medical care. Physical signs may be lacking in
the earlier stages of tuberculosis. As a result the
disease is often in an advanced stage when the
diagnosis is first made. Delay in diagnosis means
greater opportunity for infection of contacts. It
is chiefly for these reasons that case-finding
programs have been developed. The study plan

offered an excellent opportunity to measure the
effectiveness of case finding by comparing the
stage of disease on diagnosis of index cases
(which were usually discovered because of
symptoms); of prevalence cases (discovered
during contact investigation); and of new cases
(discovered by periodic examination during ob­
servation). The results of this analysis are shown
in table 5.
The most striking comparisons are noted in
the “sputum-positive” households (figure 12).
For the white index cases, 97.9 per cent were in
the moderately or far advanced stage on first
diagnosis compared with 100.0 per cent for
Negro index cases. Only 19.4 per cent of white
new cases were in the advanced stages of dis­
ease on diagnosis. Among the nonwhite new
cases, on the other hand, the disease in 61.3 per
cent was moderately or far advanced when dis­
covered. Although this represented a consider­
able reduction in comparison with the status of
the index cases, it was still inordinately high, and
it pointed up the necessity for re-examining
Negro household contacts more frequently than
the yearly intervals that were apparently suffi­
cient for the white study population.
As expected, index cases that did not have
infectious sputum were generally in an earlier
stage of disease on first diagnosis than those who
had sputum positive for M. tuberculosis.
Prevalence cases were also, as a rule, diag-

31

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE 5
Number and Per Cent of Index Cases, Prevalent Cases, and New Cases by Stage of Disease
on Diagnosis, by Sputum Status of the Index Case, and by Race
Williamson County, Tennessee, 1931-1955
While

Stage of Disease

Index Cases
Number

Per
Cent

Prevalent Cases

Number

Per
Cent

Nonwhite

New Cases
Number

Index Cases

Per
Cent

Prevalent Cases

New Cases

Per
Cent

Number

Per
Cent

Number

Per
Cent

100.0

36

100.0

31

100.0

8

22.2

3

9.7

Number

Index Cases with Infectious Sputum
Total

194

Manifest childhood . .
0
Latent apical and
minimal................
4
Moderately advanced 64
Fatal and far ad­
vanced .................
126

100.0

92

100.0

36

100.0

63

3

3.3

3

8.3

0

2.1
33.0

65
11

70.7
12.0

26
3

72.2
8.3

0
18

21
4

58.3
11.1

9

28.6

29.0
22.6

64.9

13

14.1

4

11.1

45

71.4

3

8.3

12

38.7

5

100.0

6

100.0

1

20.0

3

50.0

2
1

33.3
16.7

I

Index Cases with Other than Infectious Sputum

Total

481* 100.0

Manifest childhood. .
1
Latent apical and
minimal...............
346
Moderately advanced 127
Fatal and far ad­
vanced .................
7

89 f 100.0

311 100.0

50

0.2

0

1

3.2

0

71.9
26.4

70
15

78.7
16.9

16
12

51.6
38.7

38
11

76.0
22.0

3
0

60.0

1.5

4

4.5

2

6.5

1

2.0

1

20.0

100.0

!

0

*

Does not include 2 cases of undetermined classification.
t Does not include 7 cases of undetermined classification.
J Does not include 1 case of undetermined classification.

ft

a

nosed in the minimal stage of tuberculosis. It
must be remembered that these^ persons were
discovered when a household was opened for
investigation and contacts of the index case
were examined for the first time. Most of these
subjects either had no symptoms or had symp­
toms that were so vague or mild that they did
not seek medical care. It was not surprising,
therefore, to find that their disease was often
minimal.
These study findings strongly support the em­
phasis that has been placed on case finding as a
keystone of tuberculosis control. They demon­
strate that not only has the unsuspected case
of tuberculosis been discovered through such a
program, but also that the stage of disease on
discovery has been less advanced than that
found in the symptomatic person who seeks
medical care. Early diagnosis has led to more
effective therapy, particularly since the avail-

WHITE

INDEX CASES
PREVALENT
CASES
NEW CASES

NOhWHITE

INDEX CASES
PREVALENT I
CASES
R

NEW CASES I

1
0

10

20

30

40

60

50

70

80

90

100

PER CENT
H FAR ADVANCED

I

MODERATELY ADVANCED
[^MINIMAL & LATENT APICAL
[J] MANIFEST CHILDHOOD

Fig. 12. Percentage distribution of index cases,
prevalent cases, and new cases in “sputum-posi­
tive households,” by stage of disease on first diagnosis and bv’ race, Williamson County, Tennessee,
1931-1955.

1UG15

AilS<

32

::

i)

r

WILLIAMSON COUNTY TUBERCULOSIS STUDY

ability of the antituberculosis drugs. It has also
led, in many instances, to effective termination
of contact between the tuberculous individual
and his household associates.
Influence of socioeconomic status on stage of
disease of index cases: It is often said that one
of the concomitants of poverty is a delay in
seeking medical care. The study data provided
an opportunity to determine whether the stage
of disease on diagnosis of index cases varied with
the socioeconomic status. The tabulation of the
data is shown in table 6. Index cases with infec­
tious sputum in general were diagnosed in a
late stage of disease regardless of their socio­
economic status. In fact, 83.3 per cent of per­
sons in the upper socioeconomic class had far ad­
vanced tuberculosis when first diagnosed, com­
pared with 54 per cent in the middle class and
71.2 per cent in the lower class. None of the upper
class index cases were in the minimal stage,
while two cases in each of the middle and lower
classes were in the early stage of disease. Little
difference by sex was noted.
For the “other-than-sputum-positive” index
cases there was also little variation in the stage of
disease on first diagnosis by socioeconomic status.
Among upper class subjects, 77.3 per cent had
latent apical or minimal disease, compared with
69.1 per cent in the middle class and 73.9 per cent
in the lower class. There were no index cases
with far advanced disease in the upper class
households, while four with late disease were
present in the middle class and three in the
lower class. The disease in females appeared
to have been discovered earlier than in males.
If this were a reflection of the greater tendency
of the male to delay seeking medical care, it is
difficult to understand why it was not also appar­
ent in the “sputum-positive” households.
It does not appear, therefore, that socioeco­
nomic status influenced the speed with which
the index cases sought medical care. Perhaps the
insidious nature of tuberculosis is the much
more important factor in influencing delay in
obtaining medical attention.
Data for the Negro index cases are not pre­
sented because there were too few in the upper
socioeconomic class and because the total num­
ber for all classes was small. A comparison of
the stage of disease on diagnosis of index cases
in lower class “sputum-positive” white and Negro
households revealed no difference.

Community Studies
Although the study was designed primarily
to observe the impact of tuberculosis on per­
sons exposed in a household with a known case
of the disease, it did not ignore the influence of
the disease on the community. Community stud­
ies were done to provide comparisons with the
tuberculous households. These studies included
tuberculin and roentgenographic surveys and
analyses of tuberculosis mortality and general
mortality in the whole population.

Tuberculin Surveys

After the occurrence of pulmonary calcifica­
tion in negative tuberculin reactors had been
observed in the early years of the study, tubercu­
lin and roentgenographic surveys of school chil­
dren were undertaken to quantitate the relation­
ship between calcium deposition and tuberculin
sensitivity and to determine its significance.
It will suffice to summarize results of five sur­
veys done at two-year intervals between 1937
and 1945. These have been reported previously
(5, 8, 15, 19).
In all, 5,828 tuberculin tests were done on
white, and 1,478 tests on Negro school children
between six and 20 years of age. The totals repre­
sented two or more tests completed on 1,482
white and 339 Negro subjects. Roentgenograms
were obtained at the same time, so that eventu­
ally a series of sensitivity tests and chest films
was available for a fairly large group of sub­
jects. As a result of these studies it was possible
to demonstrate that 83 per cent of the white
and 72 per cent of the Negro children had
pulmonary or tracheobronchial calcification, or
both, while less than half of them manifested
tuberculin sensitivity. Over the eight-year pe­
riod of the surveys it was shown that the preva­
lence of sensitivity declined from 35.8 per cent
to 20.0 per cent in the five- to nine-year group,
from 48.6 to 33.3 in the 10- to 14-year group,
and from 49.4 to 41.8 in the 15- to 19-year group.
It was believed by the investigators that this
decline was a reflection of the falling tubercu­
losis mortality rate during the same eight-year
period. It was also established that a rather
large proportion of the children reacted only
to 1.0 mg. of tuberculin, and that the size of
the reaction was frequently less than 10 mm.
Some doubts were expressed concerning the
significance of such a reaction, particularly since

TABLE 6
Number and Per Cent of White Index Cases by Stage of Disease on Diagnosis, by Sputum Status, by Sex, and by
Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Total

Stage of Disease

Total

Num­
ber

Male

Per Num­
Cent
ber

Upper Socioeconomic Class

Female

Per
Cent

Num­
ber

Total

Male

Middle Socioeconomic Class

Female

Per Num­ Per Num­ Per Num­ Per
Cent ber Cent
ber Cent ber Cent

Total

Num­
ber

Male

Lower Socioeconomic Class

Female

Per Num­ Per Num­
Cent ber Cent
ber

Total

Per Num­
Cent
ber

Male

Per Num­ Per
Cent ber Cent

Female

Num­
ber

Per
Cent

s

Infectious Sputum
Total

191* 100.0 100 100.0

Latent
apical
and minimal...
4
Moderately ad­
vanced .........
G3
Fatal and far ad­
vanced
124

91 100.0 24 100.0 14 100.0 10 100.0

U)

o
87 100.0 44 100.0

43 100.0

80 100.0 42 100.0

38 100.0

2.1

2

2.0

2

2.2

0

33.0

34

34.0

29

31.9

4

64.9

64

64.0

60

G5.9 20

0

1G.7

4

83.3 10

0

2

2.3

1

2.3

1

2.3

2

2.5

1

2.4

1

2.G

0

38

43.7 17

38.6

21

48.8

21

26.2 13

31.0

8

21.1

71.4 10 100.0

47

54.0 2()

59.1

21

48.8

57

71.2 28

6G.7

29

76.3

28.6

72.1 119

61.3 222

79.6 34

77.3 14

73.7 20

80.0 143

69.1 43

55.1 100

77.5 164

73.9 62

63.9 102

81.6

22.7

26.3

20.0

29.0 34

43.6

20.2

24.8 33

34.0

22

17.6

26.4

72

37.1

53

19.0 10

1.5

3

1.5

4

1.4

0

5

5

60

a
w
w
w

Q

d

r
o

473* 100.0 194 100.0 279 100.0 44 100.0! 19 100.0 25 100.0 207 100.0 78 100.0 129 100.0 222 100.0 97 100.0 125 100.0
I

Latent
apical
and minimal... 341
Moderately ad­
vanced .........
125
Fatal and far ad­
vanced .........
7

Q
O

a

Other than Infectious Sputum
Total

F

26

55

co
w
co

d

o

0

0
4
1.9 1
1.3
3
2.3
3
1.4 2
2.1
1
0.8
*r
These add up to 664, which does not jibe with the figure of 680 given under “Composition of the Study Group.” Two index cases served in two
different households. In addition 14 households could not be classified socioeconomically because of insufficient information.

co

WILLIAMSON COUNTY TUBERCULOSIS STUDY

smaller dose would also have reacted to the
larger. Inasmuch as the reverse is not necessarily
true, the comparison will show minimal differ­
ences. Any decline in tuberculin reactors would
therefore be more significant than if comparisons
had been possible using the same strength of anti­
gen in all three periods. The data show a regular
decrease in the adjusted percentage of reactors
over the entire period of the study, so that the
final prevalence of 6.2 per cent is less than half
of 12.7 per cent observed during the first period.
For the white population, the decline was regu­
lar and steady. For the Negro, there was no
change in the first two periods, but a sharp
drop was noted in the third.
The prevalence of infection in household as­
sociates five to 19 years of age is shown by age
and race for the three periods in Table XXIII/3.
Relatively few tests were done during the second
and third periods, chiefly because the acquisition
of new households during those years was quite
small. Only 148 new households were brought
under observation between 1940 and 1947, and
80 during 1948 to 1955. The small numbers
make for some irregularity in the data, but on
the whole the same decline observed in the
clinic population was noted in the households.
For the whole tested household population the
prevalence of infection dropped from 50.8 per
cent, adjusted, in the first period, to 27.9 per
cent in the last. Among white household associ­
ates the decline was steady over the three pe­
riods. In the nonwhite it was irregular, but the
smallness of numbers here makes an interpreta­
tion hazardous. Comparison of the clinic and
household populations five to 19 years of age
points up the markedly greater risk of infection
among household associates (figure 15). Among
the latter, the adjusted prevalence of 48.4 per
cent was more than five times as high as that in
the general population (9.2 per cent).
The decline in community and household in­
fection is not surprising. In the later period the
use of antituberculosis drugs produced relatively
rapid reversal of infectiousness, with a reduction
of the reservoir of organisms in the community.
In addition, there was a swing of tuberculosis to
the older age groups, and this may have resulted
in the exposure of fewer children in the fiveto 19-year-old group, as has been suggested by
Robins (44).
From these observations it may be predicted

Q

35

70-i
rx

60-

77

50-

—-

UJ

o

40-

UJ

30-

£

o

II

;■

UJ

2010-


0

r—n

::

r-’

I

1932- 1940- 1948- 1932- 1940- 1948- 1932- 1940- 19481939
1947 1955 1939 1947 1955 1939 1947 1955

5-9

10-14

15-19

AGE GROUPS IN YEARS

■ CLINIC POPULATION

E] TUBERCULOSIS HOUSEHOLDS

Fig. 15. Percentage of reactors to Old Tuber­
culin among persons five to 19 years of age in the
clinic population and in associates of tuberculous
households, by three periods of time, in William­
son County, Tennessee.

that there will be a future decline in tuberculosis
morbidity in this community, since tomorrow’s
new cases may be expected to arise largely from
today’s infected population.
Mortality Studies
Comparison of county and household popula­
tions: Mortality rates for all causes and for
tuberculosis were compared for the county pop­
ulation and for the household associates of index
cases, as shown in table 7. Death rates for
tuberculosis were higher among household as­
sociates, as expected. Mortality rates for all
causes were similar in white household and
county populations, but among Negro subjects
the rate was 50 per cent higher in study house­
holds than in the county. In comparing the
death rates for all causes in the nonwhite and
white groups, the ratio was 1.4 to 1 for the
county, and 2.0 to 1 for the adjusted rates in the
study group. Tuberculosis mortality rates in
the Negro and white showed a ratio of 2.3 to 1
for the county, and G.2 to 1 for the household
associates (adjusted). The excessive mortality
for all causes in the Negro household group
was in part due to the higher tuberculosis mor­
tality rates in these persons. It was not the whole
story, however, because, after the tuberculosis
deaths were subtracted from the total, the
death rates for all other causes were again higher



i

I

r

)

)

co
o

TABLE 7
_
_______
2ounty
________
• All Causes and for Tuberculosis by Race, for Three
Average Annual Deaths and Death Rates in Williamson C
for
D
eath
R
ates
for
A
ll
Causes and for Tuberculosis among Household
Periods of Time, and Total Deaths and
Associates, by Race
Williamson County, Tennessee, 1932-1955

Pd
02

Time Period

Population Group

Annual
Average
Number

Rate Per
1,000

Annual
Average
Number

Rate Per
1,000

Annual
Average
Number

Rate Per
1,000

Annual
Average
Number

Rate Per
1,000

23*

87.4

Tuberculosis

Total

Tuberculosis

Total

Tuberculosis

Total

o

Nonwhite Deaths

White Deaths

Total

Annual
Average
Number

Rate Per
1,000

114*

19.5

67.3
76.5
68.1
57.0

18.8
12.4
14.0
12.5
11.1

Annual
Average
Number

Rate Per
1,000

29*

495.2

Q
O

d

kJ

d

W

Total
Total
1932-1939
1940-1947
1948-1955

Household associ­
ates
Adjusted]
County
County
County
County

393*
244.2
247.1
248.4
237.1

12.2

114
9.3
10.2
W.O
9.7

* Deaths for whole study period.
f Using race-specific county population as standard.

r

52*

14.7
22.5
14.4
7.2

161.7
176.9
59.1
92.9
57.8
29.6

279*

10.6
9.3

176.9
170.4
180.2
180.1

9.1
9.1
9.3
9.3

9.0
12.2
9.5
5.2

82.5
46.3
65.4
48.9
27.1

54.9

5.7
10.2
4.9
2.0

105.0
187.2
89.7
39.0

H

o
d
r
o
02

d

o
kJ

37

WILLIAMSON COUNTY TUBERCULOSIS STUDY

in the household associates compared with
those in the general county population.
Death rates for all causes and for tuberculosis
during three periods of time, 1932 to 1939, 1940
to 1947, and 1948 to 1955, were compared in the
county population. A steady decline waj noted in
each category through the three periods. How­
ever, if the tuberculosis death rates were sub­
tracted from the total, the death rates for all
other causes rose slightly in each period for
white county residents, while the rates fell
slightly for the Negroes.
The decline in tuberculosis death rates was
striking in both races, but was more marked and
more dramatic in the Negro population. The
whites showed a drop in tuberculosis mortality
from 65.4 per 100,000 in the first period to
27.1 in the third period, or a reduction of 58.5
per cent. In the Negro population, tuberculosis
mortality rates declined from 187.2 per 100,000
in the first period to 39.0 in the last, or a drop
of 79.1 per cent. Although in the 1932 to 1939
period the ratio of Negro to white deaths was
2.9 to 1, in the years from 1948 to 1955 it was
only 1.4 to 1. These changes quantitate the clini­
cal impression that the Negro began to handle
his tuberculosis much better during the later
years of the study than in the earlier period. The
gap in tuberculosis mortality between white and

Negro was narrowed remarkably during these
years.
Comparison of Tennessee and United States
mortality: In the ^Introduction” it was shown
that tuberculosis mortality was considerably
higher in the older white population in Tennes­
see than in the United States. The reason for
this phenomenon has been the subject of dis­
cussion and speculation for many years. The
fact that the Tennessee pattern of mortality is
now becoming apparent in the United States
(at least for the males) has not clarified the
situation.
A comparison of total and tuberculosis mor­
tality for the population older than 55 years of
age for two periods of time in Tennessee and the
United States is shown in table 8. For the
Tennessee population, the years 1933 and 1953
were compared. For similar population groups
in the United States, the years 1934 and 1952
were used, because the data for these years were
readily at hand and they were close enough to
the other years to provide comparability.
First, a comparison of the two periods in
Tennessee shows that total and tuberculosis
mortality declined for all older than 55 and for
each age group. The decline in tuberculosis
mortality was the more striking. Although in the
earlier year 5.4 per cent of all deaths were tuber-

TABLE 8
Death Rates for All Causes and for Tuberculosis, and Proportion of Tuberculosis Deaths
to All Deaths, in White Population More than 55 Years Old, Tennessee and the
United States, for Two Periods of Time
Tennessee

United States

Age Group in Years
Death Rate per
1,000, All Causes

Tuberculosis
Tuberculosis
Death Rate per Deaths. Per Cent Death Rate per
1,000, All Causes
100,000
of All Deaths

1933

Tuberculosis
Death Rate per
100,000

Tuberculosis
Deaths,
Per Cent
of All Deaths

1934

Age 55 and older...

38.6

208.6

5.4

44.7

83.4

1.9

55-64...........
65-74...........
75 and older

18.9
41.1
119.1

170.2
226.1
333.4

9.0
5.5
2.8

22.3
48.8
128.6

78.0
88.1
94.7

3.5
1.8
0.7

1953

1952

Age 55 and older.. .

37.1

70.5

1.9

38.9

38.2

1.0

55-64...........
65-74...........
75 and older

15.1
35.8
108.7

41.4
84.3
132.2

2.7
2.4
1.2

17.7
39.1
108.6

31.0
43.5
50.8

1.8
1.1
0.5

38

ii

X
1

J

WILLIAMSON COUNTY TUBERCULOSIS STUDY

culosis deaths, in 1953 only 1.9 per cent of the
total mortality was caused by tuberculosis.
In the United States there was also a decline
in total and tuberculosis mortality in the 18-year
interval between the two periods. Here too, the
decrease was noted not only in the totals but
in each age group. The percentage of tubercu­
losis deaths was 1.9 in 1934, and fell to 1.0 in
1952.
Comparison of Tennessee and United States
data shows several interesting facts. Mortality
for all causes was lower in Tennessee than in the
country at large in both periods, while tubercu­
losis mortality was considerably higher in the
state. In the interval between the two periods
the percentage of tuberculosis deaths in the
state decreased 66 per cent, while the reduction
was only 54 per cent in the United States.
The foregoing indicates that the higher mor­
tality from tuberculosis in Tennessee was in
place of, and not in addition to, mortality from
other causes. It means, in effect, that older peo­
ple died of tuberculosis rather than of other
causes, such as heart disease or malignancy. In
fact the mortality from these latter two causes
in Tennessee was lower than the national rate.
The question also arises whether tuberculosis
among the older white population was of long
duration or of recent origin. An analysis of the
experience of the white study population older
than 55 years reveals that no uniform pattern of
tuberculosis occurrence prevailed. Although it
was often difficult to establish the date of onset
of the disease because of its vague symptoms and
insidious nature, according to the best judgment
of the study clinicians, it appeared that in many
instances the disease had been present for a
considerable period of time in this group of
elderly persons. Many showed a remarkable
ability to live with the tubercle bacillus in an
almost symbiotic relationship for more than 20
years—and in a few instances, for more than 30
years—before succumbing. On the other hand, it
was evident that a number of older persons con­
tracted their disease very shortly before diag­
nosis. A decline in resistance to infection, com­
monly seen in older people, may have been
responsible for the activation of a previously
quiescent infectious focus. In either case, the
eventual death from tuberculosis of these older
persons tended to raise the age-specific mortality
rate to its unusually high level compared with
the nation’s rate.

Comparison of urban and rural mortality: In
the “Introduction” it was noted that one of the
oddities of Tennessee tuberculosis mortality
was the higher rate in rural residents compared
with that in urban residents. It would have been
most desirable to have included the investigation
of urban as well as rural tuberculous households
in the study plan. But, because the study was
limited to Williamson County, which is rural,
no comparable data for an urban population
were available. Of course, one could speculate
that differences might be based on the com­
parative quality and availability of medical care,
or on a variety of other environmental or host
factors. It would be pure speculation, however,
and as such would contribute nothing toward a
definitive reason for the intriguing observation
of the higher rural tuberculosis mortality in
Tennessee.

Summary
Resident tuberculosis mortality rates for
Tennessee differed from the national average in
several respects: (a) year by year Tennessee
rates were about 50 per cent higher than the
Unites States average; (5) Tennessee age­
specific rates for whites showed a marked rise
after 50 years of age; (c) in the Tennessee white
population the mortality rates were slightly
higher in females than in males; and (d) the
rates in the rural areas of Tennessee were higher
than in the urban areas.
After preliminary pilot studies, the Tennessee
Department of Public Health established a tu­
berculosis study in Williamson County, Tennes­
see, in 1931 in order to: conduct a systematic
study of the familial incidence of tuberculosis;
investigate the factors and circumstances related
to the “breakdown” of an individual with the
adult type of tuberculosis; investigate the evolu­
tion of tuberculous infection in childhood, par­
ticularly in children in close contact with a
tuberculous parent; undertake an epidemiologic
and clinical study of tuberculosis of old age; and
develop a program of tuberculosis control for
state-wide application to cope with the particu­
lar tuberculosis problem in Tennessee.
The study was organized on a household basis
for the observation of members who were ex­
posed to an “index” case of tuberculosis in the
household. During the 24 years of the study,
828 households comprising 4,214 persons other

t

WILLIAMSON COUNTY TUBERCULOSIS STUDY

t

>

39

than index cases were observed for 32,175.25
Tuberculosis mortality rates were higher in
person-years.
sputum-positive” households than in others;
Antecedent history revealed that in the ten they varied with age and closely paralleled criti­
years prior to the discovery of the index case cal ages for attack rates; death rates differed
there had been considerable tuberculosis mor­ by sex at different ages; they were four and onebidity and mortality in the households. Both half times as high in nonwhite households as in
morbidity and mortality were appreciably higher white; relationship appeared to be a factor in
in households in which the index case had spu­ the younger age groups; and socioeconomic
tum positive for M. tuberculosis than in other status appeared to exert some influence in the
households. Among known tuberculous patients, nonwhite households containing “sputum-posi­
more than half had a history of tuberculosis in tive” persons.
parents and/or siblings.
Deaths from causes other than tuberculosis
Prevalence of infection among household as­ were not significantly higher in “sputum-posi­
sociates varied with age and the sputum status tive” than in other households; the expected age
of the index case. Infection increased with age. pattern was noted; white males had higher death
In “sputum-positive” households, more than 3 of rates than white females in “sputum-positive”
4 household associates were already infected when households only, with no difference noted in other
the index case was discovered. In other house­ households or among Negro associates; rates in
holds less than 50 per cent of the associates were Negroes were generally higher than in whites,
infected. Sex, race, relationship to the index in toto, at all ages, and socioeconomically, class
case, and socioeconomic status did not appear to by class. The greatest difference by race was
influence the prevalence of infection.
noted for the females.
On first examination of household associates
The interval between exposure and the de­
after discovery of the index ease, a high fre­ velopment of tuberculosis varied in children of
quency of previously undetected tuberculosis “sputum-positive” parents according to the age
was found. The prevalence of disease was sig­ of first exposure. The probability of developing
nificantly influenced by the “dosage” factor; it tuberculosis soon after exposure was high in in­
tended to increase with age; females generally fancy and in children 15 years of age and older.
had more disease than males; less tuberculosis When first exposure occurred between one and
was found in Negro associates than in white;
14 years of age, tuberculosis did not develop until
close relatives in general had more disease than after 15 years of age. The interval between ex­
other members, notably so in the age group posure and onset of disease was generally
younger than five years; and the socioeconomic shorter among Negro associates than among
factor was of suggestive importance among white.
white but not among Negro associates.
The probability of dying of tuberculosis
The development of new cases of tuberculosis among white persons with moderately advanced
in the households under observation was in­ tuberculosis on first diagnosis declined during the
fluenced by a variety of factors. Three periods
three consecutive eight-ycar periods of the
of life appeared to be critical: infancy, puberty, study.
and old age. Pubertal females had higher attack
The value of case finding was shown by a
rates than males of similar age. In infancy and
comparison
of the stage of disease on diagnosis
old age the rates were higher for males than for
females. Attack rates were higher in “sputum­ of index and new cases. Index cases were pre­
positive” than in other households. Rates were dominantly in the advanced stage while more
also higher among Negro associates than among than half of the new cases were in the minimal
white. Close relatives had a slightly higher attack or latent apical disease stage. Socioeconomic
rate than other members, the difference being status did not appear to influence the stage of
most marked in those less than 15 years of age. disease on diagnosis.
Tuberculous infection of the community and
The socioeconomic status also influenced the at­
tack rate in the younger age. groups. Pregnancy the household population five to 19 years of age
and the postpartum period did not appear to be showed about a 50 per cent decline during the
a factor in the attack rate for females between three successive eight-year periods of the study.
15 and 44 years of age.
The level of infection was more than five times

*
I-

40

I

WILLIAMSON COUNTY TUBERCULOSIS STUDY

as high in the household population as in the
community.
Death rates for all causes and for tuberculosis
were higher in the study population than in the
community. In the general population the rates
declined over the three successive eight-year
periods, the most marked decline being noted
among Negro residents, particularly for tuber­
culosis mortality, which dropped about 80 per
cent. The ratio of Negro to white tuberculosis
deaths declined from 2.9 to 1, to 1.4 to 1.
Tennessee tuberculosis mortality rates in those
55 years of age and older declined about 66 per
cent from 1933 to 1953, while in the United States
the decline was 54 per cent between 1934 and
1952. Mortality rates for all causes were lower,
and for tuberculosis higher in the Tennessee pop­
ulation compared with the rates in the United
States.
High tuberculosis mortality rates in those in
the study population who were more than 55
years of age were about equally attributable to
disease of long duration and to disease of recent
onset.
Nutrition levels, using hemoglobin values as
an index, did not vary with the socioeconomic
status in the rural study and clinic populations.

REFERENCES
(1) Frost, W. H.: Risk of persons in familial
contact with pulmonary tuberculosis, Amer
J Public Health, 1933, 23, 426.
(2) Opie, E. L.: The epidemiology of tuberculo­
sis in relation to the pathological anatomy
and pathogenesis of the disease, in Ema?iuel Libman Anniversary Volume, New
York International Press, 1932, 3, 901.
(3) Stewart, H. C., Gass, R. S., and Puffer, R.
R.: Tuberculosis studies in Tennessee; A
clinic study with reference to epidemiology
within the family, Amer J Public Health,
1936, 26, 689.
(4) Stewart, H. C., Gass, R. S., Gauld, R. L.,
and Puffer, R. R.: Tuberculosis studies in
Tennessee: Infection, morbidity and mor­
tality in the families of the tuberculous,
Amer J Hyg, 1937,00, 527.
(5) Gass, R. S., Gauld, R. L., Harrison, E. F.,
Stewart, H. C., and Williams, W. C.: Tu­
berculosis studies in Tennessee: Roent­
genological evidence of tuberculous infec­
tion in relation to tuberculin sensitivity in
school children, Amer Rev Tuberc, 1938,
35,441.
(6) Puffer, R. R., Stewart, H. C., and Gass, R.
S.: Tuberculosis studies in Tennessee: Sub­
sequent course of cases observed in Wil-

liamson County, Amer J Hyg, 1938, 28,
490.
(7) Puffer, R. R., Stewart, H. C., and Gass, R.
S.: Analysis of the subsequent course of
diagnosed cases of tuberculosis, Amer J
Public Health, 1939,^, 894.
(8) Gass, R. S., Murphy, W. J., Harrison, E. F.,
Puffer, R. P., and Williams, W. C.: Tu­
berculosis infection in relation to tuber­
culin sensitivity in school children: Roent­
genological evidence—second report, Amer
J Public Health, 1941,37,951.
(9) Puffer, R. R., Gass, R. S., Murphy, W. J.,
and Williams, W. C.: Tuberculosis studies
in Tennessee: Prevalence of tuberculous in­
fection and disease in white and colored
families as revealed at the time of investi­
gation, Amer J Hyg, 1941, 34, 71.
(10) Puffer, R. R., Gass, R. S., Murphy, W. J.,
and Williams, W. C.: Tuberculosis studies
in Tennessee: Morbidity and mortality in
colored families during the period of ob­
servation, Amer J Hyg, 1942, 35, 367.
(11) Puffer, R. R., Doull, J. A., Gass, R. S.,
Murphy, W. J., and Williams, W. C.: Use
of the index case in the study of tubercu­
losis in Williamson County, Amer J Public
Health, 1942, 32, 601.
(12) Stewart, H. C., Gass, R. S., Puffer, R. R.,
and Williams, W. C.: Tuberculosis studies
in Tennessee: Morbidity and mortality in
white households during the period of ob­
servation, Amer J Hyg, 1943, 37, 193.
(13) Stewart, H. C., Gass, R. S., and Williams,
W. C.: Prevalence and incidence of tuber­
culosis among household associates accord­
ing to age and sex of index case, Amer J
Public Health, 1943, 33, 379.
(14) Puffer, R. R., Stewart, H. C., Gass, R. S.,
and Williams, W. C.: Accuracy of tubercu­
losis death rates in Williamson County,
Tenn., Amer J Public Health, 1943, 33,
370.
(15) Gass, R. S„ Harrison, E. F., Puffer, R. R.,
Stewart, H. C., and Williams, W. C.: Pul­
monary calcification and tuberculin sensi­
tivity among children in Williamson
County, Tenn., Amer Rev Tuberc, 1943,
47, 379.
(16) Puffer, R. R., Stewart, H. C., and Gass, R.
S.: Tuberculosis according to age, sex,
family history and contact, Amer Rev Tu­
berc, 1945, 51, 295.
(17) Puffer, R. R., Stewart, II. C., and Gass, R.
S.: Tuberculosis in household associates:
The influence of age and relationship, Amer
Rev Tuberc, 1945, 52, 89.
(18) Puffer, R. R., Stewart, H. C., and Gass, R.
S.: Studies of tuberculosis considering re­
lationship and family history: Summary of
analyses from Williamson County Tuber­
culosis Study, Tuberculology, 1946, 8, 14.
(19) Puffer, R. R., Stewart, H. C., Gass, R. S.,

WILLIAMSON COUNTY TUBERCULOSIS STUDY

>

and Harrison, E. F.: Serial tuberculin
tests and stability of the tuberculin reac­
tion, Amer Rev Tuberc, 1946, 54, 541.
(20) Puffer, R. R., Zeidberg, L. D., Dillon, A.,
Gass, R. S., and Hutcheson, R. H.: Tu­
berculosis attack and death rates of house­
hold associates, Amer Rev Tuberc, 1952,
65, 111.
(21) Zeidberg, L. D., Dillon, A., and Gass, R. S.:
Risk of developing tuberculosis among chil­
dren of tuberculous parents, Amer Rev
Tuberc, 1954, 70, 1009.
(22) Puffer, R. R.: Familial Susceptibility To
Tuberculosis, Harvard University Press,
Cambridge, Massachusetts, 1946.
(23) Puffer, R. R.: Practical Statistics in Health
and Medical Work, McGraw-Hill Book
Co., Inc., New York, 1950.
(24) Christie, A., and Peterson, J. C.: Pulmonary
calcification in negative reactors to tuber­
culin, Amer J Public Health, 1945, 35,
1131.
(25) Zeidberg, L. D., Dillon, A., and Gass, R.
S.: Some factors in the epidemiology of
histoplasmin sensitivity in Williamson
County, Tennessee. Amer J Public Health,
1951,47,80.
(26) Ajello, L., and Zeidberg, L. D.: Isolation of
Histoplasma capsulatum and Allescheria
boydii from soil, Science, 1951, 113, 662.
(27) Zeidberg, L. D., Ajello, L., Dillon, A., and
Runyon, L. C.: Isolation of Histoplasma
capsulatum from soil, Amer J Public
Health, 1952, 42, 930.
(28) Gordon, M. A., Ajello, L., Georg, L., and
Zeidberg, L. D.: Microsporum gypseum
and Histoplasma capsulatum spores in soil
and water, Science, 1952,116, 208. •
(29) Zeidberg, L. D.: A theory to explain the geo­
graphic variations in the prevalence of
histoplasmin sensitivity, Science, 1954, 119
654.
(30) Gass, R. S., and Zeidberg, L. D.: Diagnostic
aids in histoplasmosis, Amer Rev Tuberc,
1954, 70, 360.
(31) Zeidberg, L. D., and Ajello, L.: Environ­
mental factors influencing the occurrence
of Histoplasma capsulatum and Micro­
sporum gypseum in soil, J Bact, 1954, 68,
156.
(32) Zeidberg, L. D.: A theory to explain the geo­
graphic variations in the prevalence of

41

histoplasmin sensitivity, Amer J Trop Med,
1954, 3, 1057.
(33) Zeidberg, L. D., Ajello, L., and Webster, R.
H.: Physical and chemical factors in rela­
tion to Histoplasma capsulatum in soil,
Science, 1955,122, 33.
(34) Gass, R. S., Zeidberg, L. D., and Hutcheson,
R. H.: Chronic pulmonary histoplasmosis
complicated by pregnancy and spontaneous
pneumothorax, Amer Rev Tuberc, 1957, 75,
111.
(35) Zeidberg, L. D., Gass, R. S., and Hutcheson,
R. H.: The placental transmission of histo­
plasmosis complement-fixing antibodies,
AMA J Dis Child, 1957, 94, 179.
(36) Maxcy, K. F. (Editor): Papers of Wade
Hampton Frost, 1941, The Commonwealth
Fund, New York, 1941.
(37) Elderton,. W. P., and Perry, S. J.: Studies
in National Deterioration: VI. A third
study of the statistics of pulmonary tu­
berculosis; the mortality of the tuberculous
and sanatorium treatment, 1910. VIII. A
fourth study of the statistics of pulmonary
tuberculosis; the mortality of the tubercu­
lous, sanatorium and tuberculin treatment,
1913, Draper’s Company Research Memoirs,
Cambridge University Press, London.
(38) Weinberg, W.: Die Kinder der Tuberkulosen,
S. Hirzel, Leipzig, 1913.
(39) Palmer, C. E., Ferebee, S. H., and Petersen,
O. S.: Studies of pulmonary findings and
antigen sensitivity among student nurses:
VI. Geographic differences in sensitivity to
tuberculin as evidence of nonspecific al­
lergy, Public Health Rep, 1950, 65, 1111.
(40) Johnston, J. A.: Nutritional requirements of
the adolescent and its relation to the de­
velopment of disease, Amer J Dis Child
1947, 74, 487.
(41) Rich, A. R.: The Pathogenesis of Tubercu­
losis, ed. 2, Charles C Thomas, Springfield,
Illinois, 1951, p. 189 et seq.
(42) Lurie, M. B.: Heredity, constitution and tu­
berculosis: An experimental study, Amer
Rev Tuberc, 1941, 44 (Supplement, p. 1).
(43) Kallman, F. J., and Reisner, D.: Twin stud­
ies on the significance of genetic factors
in tuberculosis, Amer Rev Tuberc, 1943,
47, 549.
(44) Robins, A. B.: The age relationship
x of cases
of pulmonary tuberculosis and their associ­
ates, Amer J Public Health, 1953, 43, 718.

WILLIAMSON COUNTY TUBERCULOSIS STUDY

42

CO
Ch

*
<

(A

>1

c J

U

H UJ
<n X

in O
<n «i
O CD
J

c

o

Uj O

m

o

X

3 H

V
>-

X

22

o --------

Q UJ

o
uj

2

- ------------

O
in

V)

g__
X

uj
uj

X «n
< «n
— uj
u z

in

z

— UJ

sg __
X

o ____

</>

3
<o

Ul
t-_______

o

o

<

UJ

cc

z

u.

UJ

o
o

£

2

<
2o

o

!

S ll

UJ

5<n

v>
V)

u

z

o

u.

oc

§

Id
CO

2

2

2(n

u
ce

(X

Z

U-

X

in
tn

<

<

I
I

O

UJ
F-

FU.

uj

2a

Uj

X

<
a

H
z

1

s

e!

I
UJ

X
Id

a

X
in

U

uj

F“



in

J

UJ

UJ

UJ

2 o .

<

r 2 x
<

X

X

UJ
a.

O
H

UO

z
UJ

X

“TT

D
O

_____ <n
x

UJ

m

n
UJ
V)

H

2

g
X

o

<D

tt
UJ
CD

3
UJ

DC

X
X

o

o

O

g
o
UJ

tn

2
u.
o
UJ

5

UJ

<

3
da:

UJ
in

ix

in
sc

<

g
<

a:

UJ

-

z

43

WILLIAMSON COUNTY TUBERCULOSIS STUDY

»

s
0>

5S
X-

>■?

ih

O J

H W
<n X

4 X kJ

t<o
»3

«n U

<n _i
O m
3 3

o

LU O

S »-

H Z

u

X

> X

O Q

O UJ
U
O

3S

M
UJ

>« u• o<

— w

»>-

H " X
•» H H

£

£

$
a
a
o
a

LJ

o

£

2

«n

-*5•»

UJ

£UJ

a

X

<n
UJ

a

I

5 ?<n

V)
o

2

o
tn
a:
UJ

5.t
IJJO

a:



s
s
UJ

o

» 5

£o
o

2

UJ
V)

8X

a:

o

w

v>

UJ

UI
X
<

X
w
X

o

cn
CE

UJ

s

2 o

w

ENVIRONMENTAL RECORD

>4^
HH. NO.

SOCIAL. ECONOMIC AND HYGIENIC STATUS OF THE HOUSEHOLD

DATE

RURAL

COMMON I TV:

SEMI-RURAL

URBAN
B.

C.

SINGLE
DUPLEX
APARTMENT

RESIDENCE:

PROSPEROUS
MEDIOCRE

PREDOM I NAThNG WHITE

AGRICULTURAL
INDUSTRIAL
RESIDENTIAL

PREDOMINATING COLORED
MIXED. WHITE AND COLORED

OWNED. FREE
OWNED. ENCUMBERED
RENTED
RENTAL VALUE

NO. OF ROOMS
NO. OF SLEEPING ROOMS
NO. BEDS
SLEEPING ARRANGEMENTS

RUN DO VW
NO. OF YEARS IN COMMUNITY
NO. OF YEARS IN HOUSE

NO. PEOPLE IN HOUSE

SANITATION:
WATER SUPPLY:

CISTERN
WELL
SPRING OR STREAM
MAIN CONNECTIONS
BATHING FACILITIES:
SHOWER
TUB
NONE

IN HOUSE
NEAR HOUSE
DISTANT

SCREENING:
GOOD
DEFECTIVE
NONE

TOILET FACILITIES:
WATER CLOSET
SANITARY PRIVY
OPEN PRIVY
NONE

MILK SUPPLY:
COWS TBCLN. TESTED
COWS NOT TESTED
MILK PASTEURIZED
MILK DEALER
AMT. MILK USED

REMARKS ON EVIDENCE OF PERSONAL HYGIENE AND CLEANLINESS OBSERVED:

Fd
>
CP

o
O
O
d

d
w
H
W
o
d
d

o
CP
CP
CP

H
d

o

SUMMARY OF FORMER RESIDENCES:

SIGNED

WILLIAMSON COUNTY TUBERCULOSIS STUDY
TENNESSEE DEPARTMENT OF PUBLIC HEALTH - 983

I

45

WILLIAMSON COUNTY TUBERCULOSIS STUDY

INDIVIDUAL RECORD
SEX

COLOR

NAME

OCCUPATION

AGE

HH. AND
IND. NO.

DATE OF RECORD
CONTACT WITH TUBERCULOSIS
RELATION

NAME

YEAR
ENDED

YEAR
BEGUN

TYPE OF CASE

PAST ILLNESSES

REMARKS

(GIVE DATES)

OTHER
RESPIRATORY

INFLUENZA

PNEUMONIA

PLEURISY

TYPE CONTACT
(HH..VISIT.ETC)

OTHER
DISEASES

ONSET

HI STORY OF TBC.

FAMILY MEMBERS
NAME

SEX

ADDRESS

YEAR
B I RTH

YR.LAST STATUS
OBSERV. (W. I . D. )

CAUSE

HISTORY OF TUBERCULOSIS'
(WITH OATES)

PARENTS

SIBLINGS

CONSORT

CHILDREN

W. I .0. MEANS WELL. ILL. DEAD - USE REVERSE SIDE FOR LISTING ADDITIONAL SIBLINGS, CONSORTS. AND CHILDREN.

WILLIAMSON COUNTY TUBERCULOSIS STUDY
TENNESSEE DEPARTMENT OF PUBLIC HEALTH

984

WILLIAMSON COUNTY TUBERCULOSIS STUDY

46

CLINICAL RECORD

COLOR

NAME

SEX

AGE

HH. AND
IND. NO.

B IRTHPLACE

MARITAL STATUS____________

DIRECTIONS FOR FINDING HOUSE

MAILING ADDRESS

I .

2.

3.
ADDRESS

PHYSICIAN
p REV1OUS_ D I AGNOSIS OF TUBERCULOSIS (GIVE DATE)

PREVIOUS MEDICAL CARE

CURRENT SYMPTOMS AND ILLNESS
DATE DATE DATE DATE DATE

DATE

PLEURISY

FATIGUE

PNEUMONIA

COUGH

INFLUENZA

EXPECTORATION

BRONCHITIS

PA I N IN CHEST

OTHER (SPECIFY)

LOS$ OF WEIGHT
NIGHT SWEATS

HOARSENESS
SHORTNESS BREATH
FEVERISH
AMENORRHEA

TEMPERATURE

HEMOPTYSIS

PULSE

DATE

WEIGHT

AMOUNT

HEIGHT (IN INCHES)

WHY DID PATIENT SEEK
X-RAY EXAMINATION?

ONSET OF ILLNESS (DATE AND MANNER OF ONSET)

CHANGES IN SYMPTOMS AT REEAXAMI NATI ON

MARKEDLY UNSATISFACTORY.
CODE: 0, SATISFACTORY; 1.2.3. SL IGHTLY. MODERATELY OR
WILLIAMSON COUNTY TUBERCULOSIS STUDY
TENNESSEE DEPARTMENT OF PUBLIC HEALTH

985

WILLIAMSON COUNTY TUBERCULOSIS STUDY

47

A

NAME

HH.AND IND.NO.
X-RAY EXAMINATIONS

DATE
CLASSIFICATION

CODE

DATE
CLASS IFICATION
CODE

SPUTUM EXAMINATIONS
DATE
RESULT

DATE

RESULT

SKIN TESTS

TUBERCULIN

OTHER (SPECIFY)

DATE
AMOUNT
RESULT
SUMMARY OF HOSPITALIZATION

DATE
ADMITTED

HOSP ITAL

TREATMENT

OTHER TREATMENT AND CARE
TYPE

(WITH DATES

INSTITUTED AND DISCONTINUED)

CHANGES (WITH dates)
CONTACT
OCCUR AT ION
ENVIRONMENT

ECONOMIC STATUS

REHABILITATION
DATE

REFERRED

DATE RETURNED TO WORK

TYPE TRAINING RECEIVED

TYPE POSITION

DATE
DISCHARGED

RECOMMENDATIONS

WILLIAMSON COUNTY TUBERCULOSIS STUDY

48

4

FIRST EXAMINATION

COLOR

NAME

HH. AND
IND. NO.

AGE

ADDRESS

SEND REPORT TO DR.

DATE

SEX

PHYSICAL EXAMINATION
GENERAL APPEARANCE

RETARDED MOTION

THYROID

DEPRESSIONS

GLANDS

ANEMIC

TONSILS

HEART

TEETH
LEFT LUNG

SUMMARY OF FINDINGS

RIGHT LUNG

DIAGNOSIS (BY HISTORY AND
PHYSICAL EXAMINATION)

X-RAY EXAMINATION
RECOMMENDED
TECHNIQUE

FILM TECHNIQUE USED
OBJECTIVE DESCRIPTION

FILM NO.

DIAGNOSIS (ON ALL
AVAILABLE DATA)
RECOMMENDATIONS

EXAM.
BY

WILLIAMSON COUNTY TUBERCULOSIS STUDY
TENNESSEE DEPARTMENT OF PUBLIC HEALTH

986

49

WILLIAMSON COUNTY TUBERCULOSIS STUDY



REEXAMINATION

NAME

COLOR

DATE

SEND REPORT TO DR.

SEX

AGE

HH. AND
IND. NO.
EXAM.
NUMBER

ADDRESS

CHANGES IN PHYSICAL FINDINGS

TENTATIVE DIAGNOSIS

RECOMMENDED
TECHNIQUE

FILM TECHNIQUE USED
X-RAY • OBJECTIVE DESCRIPTION - FILM NO.

RECOMMENDATIONS
EXAM.
BY

DIAGNOSIS
DATE

AGE

SEND REPORT
TO DOCTOR

EXAM.
NUMBER

ADDRESS

CHANGES IN PHYSICAL FINDINGS

TENTATIVE DIAGNOSIS

RECOMMENDED
TECHNIQUE

FILM TECHNIQUE USED
X-RAY - OBJECTIVE DESCRIPTION

DIAGNOSIS

FILM NO.

RECOMMENDATIONS

EXAM.
BY

50

WILLIAMSON COUNTY TUBERCULOSIS STUDY

A

NURSING NOTES

HH.AND IND.NO.

NAME

DATE

WORKER

NOTES



y

WILLIAMSON COUNTY TUBERCULOSIS STUDY
TENNESSEE DEPARTMENT OF PUBLIC HEALTH - 987

HOUSEHOLD RECORD OF ILLNESS
HEAD OF HOUSEHOLD

IND.
NO.

NAME

H.H. NO.
DATE
OF
Bl RTH

MAR.
SEX STAT

DATE OF HOUSEHOLD CHECK

RELATION
TO HEAD
OF H.H.

d
d

>
%
O?
O
Q

O
d

A
d
td

w
w

Q
d

r
o
OQ

CH

| nJ

03

H
d

c
*1

RECORD DATES OF ENTRY OF NEW ASSOCIATES AND DATES OF REMOVAL OF THOSE WHO LEAVE.
IF NO ILLNESS OCCURRED SINCE DATE OF
PREVIOUS CHECK. INDICATE WITH "o"
USE REVERSE
IF ILLNESS OCCURRED. SPECIFY BY CODE AND GIVE DETAILS ON INDIVIDUAL RECORD.

SIDE FOR INFORMATION PERTAINING TO THOSE *HO HAVE NO INDIVIDUAL RECORD.

CODE
1. MEASLES
2.

WHOOPINC COUCH

FOR

ILLNESS

3.

PNEUMONIA

5.

INFLUENZA

7.

PREGNANCY

•. OPERATION

4.

PLEURISY

6.

SEVERE COLD (3+ WEEKS)

9.

SERIOUS INJURY

10.

OTHER (CONFINED TO BED)

■ILLI

CaWMTV TUBIRCULOIIS STUDY
e>kiT«iar of public hcalth

988

CH

52

WILLIAMSON COUNTY TUBERCULOSIS STUDY

A

TABLE I
Person-Years Experience of Household Associates of Index Cases, by Age, Sex,
Race, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955

Age Group

Total

Close
Relatives

Female

Male

Total

Other
Members

Total

Close
Relatives

Other
Members

Total

Close
Relatives

Other
Members

Total
Total

32,175.25 16,705.50 15,409.75 16,654.00 8,851.50 7,802.50 15,521.25 7,914.00 7,007.25

<5
5-14
15-24
25-34
35-44
45-54
55-64
65-74
75+
Unknown

918.50 501.50 417.00
863.25 514.75 348.50
765.50
1,781.75 1,016.25
3,244.002,072.75
1,171.25
3,712.25
2,491.25
1,221.00
6,956.25 4,564.00 2,392.25
6,676.75 4,559.25 2,117.50 3,755.502,671.75 1,083.75 2,921.25 1,887.50 1,033.75
3,372.50 1,819.25 1,553.25 1,628.25 980.50 647.75 1,744.25 838.75 905.50
3,427.50 1,287.50 2,140.00 1,712.50 620.50 1,092.00 1,715.00 667.00 1,048.00
3,493.75 1,158.75 2,335.00 1,766.25 487.25 1,279.00 1,727.50 671.50 1,056.00
3,009.50 1,075.75 1,933.75 1,465.00 472.75 992.25 1,544.50 603.00 941.50
799.75 1,460.25 1,172.00 388.00 784.00 1,088.00 411.75 676.25
2,260.00
615.75 260.25 355.50
574.50 224.75 349.75
705.25
485.00
1,190.25
2.50
2.50
4.50
4.50
7.00
7.00

White
Total
<5
5-14
15-24
25-34
35-44
45-54
55-64
65-74
75+
Unknown

26,318.50 13,577.00 12,741.50 13,679.00 7,246.25 6,432.75 12,639.50 6,330.75 6,308.75

828.25
1,492.25
5,666.75 3,841.50
5,444.75 3,835.75
2,620.00 1,444.50
2,766.75 1,004.00
854.00
2,902.00
787.75
2,520.75
568.25
1,862.50
413.00
1,035.75
7.00

772.50 414.00
719.75 414.25 305.50
664.00
2,611.00
1,741.00
955.25
3,055.75
2,100.50
1,825.25
1,609.00 3,035.25 2,215.50 819.75 2,409.50 1,620.25
1,175.50 1,258.00 776.00 482.00 1,362.00 668.50
1,762.75 1,416.25 533.50 882.75 1,350.50 470.50
2,048.00 1,494.00 405.00 1,089.00 1,408.00 449.00
1,733.00 1,246.00 354.25 891.75 1,274.75 433.50
908.50 304.25
954.00 264.00 690.00
1,294.25
540.25 229.75
495.50 183.25 312.25
622.75
2.50
4.50
4.50
7.00

358.50
870.00
789.25
693.50
880.00
959.00
841.25
604.25
310.50
2.50

Non white

Total
<5
5-14
15-24
25-34
35-44
45-54
55-64
65-74
75+
Unknown

5,850.75 3,188.50 2,668.25 2,975.00 1,605.25 1,369.75 2,881.75 1,583.25 1,298.50
289.50
1,289.50
1,232.00
752.50
660.75
591.75
488.75
397.50
154.50

188.00
722.50
723.50
374.75
283.50
304.75
288.00
231.50
72.00

101.50
567.00
508.50
377.75
377.25
287.00
200.75
166.00
82.50

143.50
656.50
720.25
370.25
296.25
272.25
219.00
218.00
79.00

100.50
390.75
456.25
204.50
87.00
82.25
118.50
124.00
41.50

43.00
265.75
264.00
165.75
209.25
190.00
100.50
94.00
37.50

146.00
633.00
511.75
382.25
364.50
319.50
269.75
179.50
75.50

87.50
331.75
267.25
170.25
196.50
222.50
169.50
107.50
30.50

58.50
301.25
244.50
212.00
168.00
97.00
100.25
72.00
45.00

WILLIAMSON COUNTY TUBERCULOSIS STUDY

53

TABLE II
Composition of the Williamson County Population* and the Household Associates Population,
by Age, Sex, and Race
Williamson County, Tennessee, 1931-1955
Williamson County

Age Group

Total
Num­
ber

Per
Cent

Male
Num­
ber

Households

Female

Per
Cent

Num­
ber

Total

Per
Cent

PersonYears

Male
Per
Cent

PersonYears

Female

Per
Cent

PersonYears

Per
Cent

Total

Total
<5
5-14
15-24
25-34
35-54
55+
Unknown

24,855 100.0 12,501 50.3 12,353 49.7 32,175.25 100.0 10,654.00 51.8 15,521.25 48.2

2,630
5,221
4,397
3,390
5,378
3,839

10.6
21.0
17.7
13.6
21.6
15.4

1,362 5.5
2,654 10.7
2,222 8.9
1,687 6.8
2,619 10.5
1,957 7.9

1,268 5.1
2,567 10.3
2,174 8.7
1,704 6.9
2,759 11.1
1,881 7.6

1,781.75
6,950.25
0,676.75
3,372.50
6,921.25
6,459.75
7.00

5.5
21.6
20.8
10.5
21.5
20.1

863.25 2.7
3,712.25 11.5
3,755.50 11.7
1,628.25' 5.1
3,478.75 10.8
3,211.50 10.0
4.50 —

918.50 2.9
3,244.00 10.1
2,921.25 9.1
1,744.25 5.4
3,442.50 10.7
3,248.25 10.1
2.50

White

Total

19,430

78.2

9,808 39.5

<5
5-14
15-24
25-34
35-54
55+
Unknown

1,992
4,062
3,415
2,604
4,247
3,110

8.0
16.3
13.7
10.5
17.1
12.5

1,039
2,071
1,734
1,311
2,072
1,581

4.2
8.3
7.0
5.3
8.8
6.4

9,020 38.7 20,318.50

81.8 13,079.00 42.5 12,039.50 39.3

953
1,991
1,680
1,293
2,175
1,528

4.0
17.0

3.8
8.0
6.8
5.2
8.8
6.1

1,492.25
5,606.75
5,444.75
2,020.00
5,668.75
5,419.00
7.00

16.9

8.1
17.6
16.8

719.75
3,055.75
3,035.25
1,258.00
2,910.25
2,095.50
4.50

2.2
9.5
9.4
3.9
9.0
8.4

772.50
2,611.00
2,409.50
1,362.00
2,758.50
2,723.50
2.50

2.4
8.1
7.5
4.2
8.6
8.5

Nonwhite
Total

5,425

21.8

<5
5-14
15-24
25-34
35-54
55+

038
1,159
982
786
1,131
729

2.6
4.7
4.0
3.2
4.6
2.9

* Estimated as

2,093 10.8
323
583
488
376
547
376

1.3
2.3
2.0
1.5
2.2
1.5

2,733 11.0

315
576
494
411
584
353

1.3
2.3
2.0
1.7
2.3
1.4

5,850.75

18.2

2,975.00

9.2

2,881.75

9.0

289.50
1,289.50
1,232.00
752.50
1,252.50
1,040.75

0.9
4.0
3.8
2.3
3.9
3.2

143.50
656.50
720.25
370.25
568.50
516.00

0.4
2.0
2.2
1.2
1.8
1.6

146.00

633.00
511.75
382.25
684.00
524.75

0.5
2.0

of January 1, 1944 (mid-period of study)

1.6
1.2

2.1
1.6

54

WILLIAMSON COUNTY TUBERCULOSIS STUDY

4,

<

/

TABLE Ill*
Annual and Cumulated Mortality and Morbidity from Tuberculosis in Households of
Manifest “Sputum-positive” Index Cases during a 21-Year Period Bordering on the Onset
of the Index Case
Williamson County, Tennessee
Cases (Tuberculosis)

Deaths (Tuberculosis)
Year Relative
to Onset of
Index Case

Prior
10
9
8

Life Experience,
Person-Years

Number

Rate per 100

1

0.48

1
3

0.42
1.18

6
5
4
3
2
1

186.00
196.25
207.25
213.00
221.25
239.25
255.00
262.00
275.25
285.50

Year of onset

339.00

Subsequent
1
2
3
4
5
6
7
8
9
10

300.00
261.50
212.25
184.00
166.75
141.75
121.75
107.00
100.50
95.50

2
1

0.67
0.38

1

0.54

3

2.64

Total

4370.75

12

0.27

* Reprinted with permission, Amer J Hyg, 1937, 26, 552

Number

Rate per 100

Cumulative
Percentage

1
2

0.51
0.97

2
2
2
1
1

0.90
0.84
0.78
0.38
0.36

0.51
1.48
1.48
2.38
3.22
4.00
4.38
4.74
4.74

2.08

7

2.0G

6.80

2.75
3.13
3.13
3.67
3.67
3.67
6.31
6.31
6.31
6.31

4
3
1
2

1.33
1.15
0.47
1.09

1

0.71

2

1.87

1

1.05

8.13
9.28
9.75
10.84
10.84
11.55
11.55
13.42
13.42
14.47

6.31

32

0.73

14.47

Cumulative
Percentage

0.48
0.48
0.48
0.90
2.08
2.08
2.08
2.08

WILLIAMSOxX COUNTY TUBERCULOSIS STUDY

55

TABLE IV*
Annual and Cumulated Mortality and Morbidity iFROM
.
1 UBERCULOSIS IN HOUSEHOLDS OF
Manifest “Other-than-Sputum-positive” Index Cases
i
-J DURING A 21-V EAR PERIOD BORDERING
on the Onset of the Index Case
______________________ Williamson County, Tennessee
Year Relative
to Onset of
Index Case

Deaths (Tuberculosis)
Life Experience,
Person-Years

Prior
10
9
8
7
6
5
4
3
2
1

583.50
619.50
646.00
662.75
687.75
713.50
742.50
772.50
803.00
840.25

Year of onset

848.00

Subsequent
1
2
3
4
5
6
7
8
9
10

742.00
626.50
540.75
492.75
459.00
420.50
377.25
344.75
324.75
310.00

Total
*

12,557.50

Cases (Tuberculosis)

Number

Rate per 100

Cumulative
Percentage

Number

Rate per 100

Cumulative
Percentage

1

0.17

0.17
0.17
0.17
0.32
0.32
0.46
0.59
0.59
0.71
0.71

2
2
2
4
2
2
2
2
2

0.32
0.31
0.30
0.58
0.28
0.27
0.26
0.25
0.24

0.32
0.63
0.93
1.51
1.79
2.06
2.32
2.57
2.81

0.95

2

0.24

3.05

0.95
0.95
0.95
1.15
1.37
2.08
2.08
2.37
2.68
2.68

2
3
1
3
2
1

0.27
0.48
0.18
0.62
0.44
0.24

4

1.68

3.32
3.80
3.98
4.60
5.04
5.28
5.28
6.96 ’
6.96
6.96

38

0.30

1

0.15

1
1

0.14
0.13

1

0.12

2

0.24

1
1
3

0.20
0.22
0.71

1
1

0.29
0.31

14

0.11

Reprinted with permission, Amer J Hyg, 1937, 26, 553

2.68

6.96

!
i

WILLIAMSON COUNTY TUBERCULOSIS STUDY

56

I

i

TABLE VA
Prevalence of Infection among White Household Associates of “Sputum-positive
Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955

Number
Tested

Number
Tested

Number
Tested
Number

Positive

Positive

Positive

Age Group in Years

Female

Male

Total

Per Cent

Index

Number

Per Cent

Number | Per Cent

Total

Total

350

250

71.4

175

120

G8.G

175

130

74.3

<5
5-14
15-24
25-34
35-54
55+

52
137
G3
32
46
20

24
90
51
2G

4G.2
G5.7
81.0
81.2
87.0
95.0

24

12
45
24
13
17
9

50.0
58.4
75.0
92.9
89.5
*

28
GO
31
18
27
11

12
45
27
13
23
10

42.9
75.0
87.1
72.2
85.2
90.9

70.5

102

79

77.5

50.0
G7.2
75.0
*

8
35
19
7
7
3

53.3
77.8
82.G
*

*
*

15
45
23
8
8
3

■10

19

32
14
19
9
Close Relatives

Total

214

158

73.8

112

79

<5
5-14
15-24
25-34
35-54
55+

29
10G
47
13
12
7

15
7G
37
12
11
7

51.7
71.7
78.7
92.3
91.7
*

14
61
24
5
4
4

41
18
5
4
4

*
*

Other Members
Total

13G

92

G7.G

G3

41

G5.1

73

51

69.9

<5
5-14
15-24
25-34
35-54
55+

23
31
16
19
34
13

9
14
14
14
29
12

39.1
45.2
87.5
73.7
85.3
92.3

10
16
8
9
15
5

5
4
6
8
13
5

50.0
25.0
*

13
15
8

4
10
8
G
1G
7

30.8
GG.7
*
GO.O
84.2
*

* Less than 10 persons; no percentage calculated

*

8G.7
*

10

19
8

57

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE VB
Prevalence of Infection among White Household Associates of “Other-than-Sputum-positive”
Index Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955
Total

Age Group in Years

Male
Positive

Number
Tested

Positive

Number
Tested
Number

Per Cent

Female

Number
Tested

Number

Per Cent

Positive
Number |

Per Cent

Total
Total

519

192

37.0

249

93

37.4

270

99

3G.7

<5
5-14
15-24
25-34
35-54
55+

81
243
97
35
39
24

10
67
50
22
25
18

12.4
27.6
51.6
62.9
64.1
75.0

37
130
46
9
17
10

9
37
24
4
12
7

24.3
28.5
52.2
44.4
70.6
70.0

44
113
51
26
22
14

1
30
26
18
13
11

2.3
26.6
51.0
69.2
59.1
78.6

Close Relatives
4

I

Total

358

99

27.7

170

51

30.0

188

48

25.5

<5
5-14
15-24
25-34
35-54
55+

67
201
67
12
7
4

6
51
30
7
3
2

9.0
25.4
44.8
58.3
*
*

28
104
30
5
3
0

5
29
13
2
2
0

17.9
27.9
43.3
*
*

39
97
37
7
4
4

1
22
17
5
1
2

2.6
22.7
46.0
*
*
*

Other Members

Total

161

93

57.8

79

42

53.2

82

51

62.2

<5
5-14
15-24
25-34
35-54
554-

14
42
30
23
32
20

4
16
20
15
22
16

28.6
38.1
66.7
65.2
68.8
80.0

9
26
16
4
14
10

4
8
11
2
10
7

*
30.8
68.8
*
71.4
70.0

5
16
14
19
18
10

0
8
9
13
12
9

50.0
64.3
68.4
66.7
90.0

* Less than 10 persons; no percentage calculated
/’

WILLIAMSON COUNTY TUBERCULOSIS STUDY

58

i>

TABLE VIA
Prevalence of Infection among Nonwhite Household Associates of “Sputum-positive” Index
Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955

Age Group in Years

Female

Male

Total

Number

Number
Tested

Number
Tested

Per Cent

<

Positive

Positive

Positive

Number
Tested

Number

Per Cent

Number

Per Cent

Total
Total

218

153

70.2

98

66

G7.3

120

87

72.5

<5
5-14
15-24
25-34
35-54
55 +

46
83
43
11
21
14

24
56
31
9
19
14

52.2
67.5
72.1
81.8
90.5
100.0

22
39
22
4
6
5

10
27
15
3
6
5

45.5
69.2
68.2
*
*
*

24
44
21
7
15
9

14
29
16
6
13
9

58.3
65.9
76.2
*
86.7
*

Close Relatives
Total

149

105

70.5

69

48

69.6

80

57

71.2

<5
5-14
15-24
25-34
35-54
55 +

31
62
29
5
13
9

16
42
22
4
12
9

51.6
67.7
75.9
*
92.3
*

16
27
17
1
3
5

7
20
12
1
3
5

43.8
74.1
70.6
*
*
*

15
35
12
4
10
4

9
22
10
3
9
4

60.0
62.9
83.3
*
90.0
*

Other Members
Total

69

48

69.6

29

18

62.1

40

30

75.0

<5
5-14
15-24
25-34
35-54
55+

15
21
14
6
8
5

8
14
9
5
7
5

53.3
66.7
64.3
*

6
12
5
3
3
0

3
7
3
2
3
0

*

9
9
9
3
5
5

5
7
6
3
4
5

*
*
*
*
*
*

*
*

* Less than 10 persons; no percentage calculated

58.3
*
*
*

WILLIAMSON COUNTY TUBERCULOSIS STUDY

59

4*

3

b

TABLE Vl/i
Prevalence of Infection among Nonwhite Household Associates cof “
"2
____ -than-SputumOther
positive” Index Cases, by Age, Sex, and Relationship to the Index Case
__ _____________
Williamson County, Tennessee, 1931-1955

I

Total

Age Group in Years

Positive

Number
Tested

t

Male
Positive

Number
Tested
Number

Per Cent

Female

Positive

Number
Tested

Number

Per Cent

42.5
*

Number

Per Cent

22

6

27.3

0
2
0
3
1
0

*

1

*
*
*

6
9
2
4
1
0

22

9

40.9

12

5

41.7

4
15
2
0
1
0

0
6
2
0
1
0

3
4
2
2
1
0

0
2
0
2
1
0

44.4

10

1

10.0

*

3
5
0
2
0
0

0
0
0
1
0
0

*

Total

I


i

Total

62

23

37.1

40

17

<5
5-14
15-24
25-34
35-54
55+

11
37
5
5
3
1

1
11
3
4
3
1

9.1
29.7
*

5
28
3
1
2
1

1
9
3
1
2

*
*
*

32.1
*

*

*

Close Relatives

<

Total

34

14

<5
5-14
15-24
25-34
35-54
55+

19
4
2
2
0

0
8
2
2
2
0

41.2

42.1
*
*
*

40.0
*
*

*
*
*

Other Members

*

*

Total

28

9

32.1

18

8

<5
5-14
15-24
25-34
35-54
55+

4

1
3
1
2
1
1

*

1
13

1
3

1

1
1

18

1
3
1
1

16.7
*
*
*
*

Less than 10 persons; no percentage calculated

1
1
1

1
1

23.1
*
*
*
*

60

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE VII
Prevalence of Infection among White Household Associates according to Exposure to Index
Cases and Prevalent Cases, or Index Cases Only, by Age and Relationship
Williamson County, Tennessee, 1931-1955
Associates of Prevalence Cases
and Index Cases

Total

Associates of Index Cases Only

/

Age Group in Years

Positive

Positive
Number
Tested

Number | Per Cent

Positive

Number
Tested

Number
Tested
Number

Per Cent

117

61.3

Number

Per Cent

613

275

44.9
54-5

22
97
60
32
37
27

19.8
35.3
59.4
65.3
77.1
93.1

I

*

Total

Total
Adjusted]

804*

<5
5-14
15-24
25-34
35-54
55+

131
374
154
58
56
31

392

33
151
95
40
44
29

48.8
54.7

191

25.2
40.4
61.7
69.0
78.6
93.5

20
99
53
9
8
2

^4-4
11
54
35
8

55.0
54.5
66.0
88.9

2

90.0

111
275
101
49
48
29

90

62.5

402

145

36.1

10

66.7
57.0
66.7

10

t
t

79
223
68
15
12
5

77
35
10
8
5

12.7
34.5
51.5
66.7
66.7

57.4

211

130

61.6

t

32
52
33
34
36
24

12
20
25
22
29
22

37.5
38.5
75.8
64.7
80.6
91.7

Close Relatives

Total

546

235

43.0

144

<5
5-14
15-24
25-34
35-54
55+

94
302
110
22
12
6

20
122
63
16
8
6

21.3
40.4
57.3
72.7
66.7

15
79
42

0
1

i

45
28
6
0
1

i

Other Members

Total

258

157

60.9

47

27

<5
5-14
15-24
25-34
35-54
55+

37
72
44
36
44
25

13
29
32
24
36
23

35.1
40.3
72.7
66.7
81.8
92.0

5
20
11
2
8

1
9

1

2
7
1

45.0
63.6

i

t
i

fl

* This total differs from that in Table VA and B because prevalence cases that had been tuberculin
tested are not included here
f For age and relationship, using total white household population present on investigation, as
standard
J Less than 10 persons; no percentage calculated

«■

c
WILLIAMSON COUNTY TUBERCULOSIS STUDY

61

<•

!

TABLE VIII
Prevalence of Infection among White Household Associates of Index Cases, by Age, Socio­
economic Status, and Sputum Status of the Index Case
Williamson County, Tennessee, 1931-1955
Upper and Middle
Socioeconomic Classes

Total

Age Group in Years

Positive

Number
Tested

Positive

Number
Tested

Number

Per Cent

Lower Socioeconomic Class
Positive

Number
Tested

Number

Per Cent

Number

Per Cent

“Sputum-positive” Index Cases
>

Total
Adjusted

348

251

72.1
75.7

129

88

68.2
65.9

219

1G3

74.4
81.5

<5
5-14
15-24
25-34
35-54

50
136
64
32
46
20

24
90
52
26
40
19

48.0
66.2
81.2
81.2
87.0
95.0

11
39
29
11
26
13

3
20
24
9
20
12

27.3
51.3
82.8
81.8
77.0
92.3

39
97
35
21
20

21
70
28
17
20
7

53.8
72.2
80.0
81.0
100.0
100.0

55+

“Other-than-Sputum-positive” Index Cases

Total
Adjusted

511

190

37.2
^8.5

132

64

48.5
51.6

379

126

33.2
48.2

<5
5-14
15-24
25-34
35-54
55 +

79
244
92
35
37
24

10
68
48
22
24

12.7
27.9
52.2
62.9
64.9
75.0

10
46
35
15
12
14

2
15
20
8
9
10

20.0
32.6
57.1
53.3
75.0
71.4

69
198
57
20
25
10

8
53
28
14
15
8

11.6
26.8
49.1
70.0
60.0
80.0

18

WILLIAMSON COUNTY TUBERCULOSIS STUDY

62

A
■u

TABLE 1XA
Prevalence of Tuberculosis among White Household Associates of “Sputum-positive” Index
Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955

Age Group in Years

Number of
Household
Members

Female

Male

Total

Cases
Number

Per Cent

Number of
Household
Members

Cases
Number

Per Cent

Number of
Household
Members

Cases
Number

Per Cent

Total
Total
Adjusted

877

92

10.5
9.8

448

39

8.7
8.8

429

53

12.4
10.9

<5
5-14
15-24
25-34
35-54
55+

92
218
179
87
175
126

3
4
12
12
39
22

3.3
1.8
6.7
13.8
22.3
17.5

43
125
97
45
82
56

2
1
4
7
16
9

4.7
0.8
4.1
15.6
19.5
16.1

49
93
82
42
93
70

1
3
8
5
23
13

2.0
3.2
9.8
11.9
24.7
18.6

Close Relatives

Total
A djusted

500

43

8.6
10.1

263

18

6.8
9.2

237

25

10.6
11.0

<5
5-14
15-24
25-34
35-54
55+

57
163
125
41
60
54

3
3
9
3
13
12

5.3
1.8
7.2
7.3
21.7
22.2

28
95
69
20
27
24

2
1
2
3
5
5

7.1
1.1
2.9
15.0
18.5
20.8

29
68
56
21
33
30

1
2
7
0
8
7

3.4
2.9
12.5

185

21

,11.4
8.3

192

28

15
30
28
25
55
32

0
0
2
4
11
4

20
25
26
21
60
40

0
1
1
5
15
6

24.2
23.3

Other Members
Total
Adjusted

377

49

<5
5-14
15-24
25-34
35-54
55+

35
55
54
46
115
72

0
1
3
9
26
10

13.0
9.5
1.8
5.6
19.6
22.6
13.9

7.1
16.0
20.0
12.5

14.6
10.7
4.0
3.8
23.8
25.0
15.0

<

4

63

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE IXB
Prevalence of Tuberculosis among White Household Associates of “Other-than-Sputumpositive” Index Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955
Total

>

Age Group in Years

Number of
Household
Members

Male

Cases
Number |

Per Cent

Number of
Household
Members

Female
Cases

Cases

Number of
Household
Members

Number |

861

57

0.49
7.48
7.14
14.63

106
237
178
95
135
110

0
2
6
6
20
23

2.9

514

26

8.93
10.26
24.24

81
184
133
39
37
40

0
2
5
3
9

5.2

347

31

25
53
45
56
98
70

0
0
1
3
13
14

Number

Per Cent

1001

39

3.9
4-9

109
291
203
107
168
123

0
0
1
8
12
18

Per Cent

Total
Total
Adjusted

1862

<5
5-14
15-24
25-34
35-54
55+

215
528
381
202
303
233

96

5.2
6.0

0
2
7
14
32
41

0.38
1.84
6.93
10.56
17.60

6.6
7.3

0.84
3.37
6.32
14.81
20.91

Close Relatives
Total
Adjusted

1094

<5
5-14
15-24
25-34
35-54
55+

151
410
289
95
76
73

43

0
2
5
8
11
17

3.9
7.0

580

17

5.9

0.49
1.73
8.42
14.47
23.29

70
226
156
56
39
33

0
0
0
5
4
8

5.1
8.1

1.09
3.76
7.69
18.92
22.50

Other Members

i

Total
Adjusted

768

<5
5-14
15-24
25-34
35-54
55+

64
118
92
107
227
160

53

0
0
2
6
21
24

6.9
4-7

2.17
5.61
9.25
15.00

421

22

3.6
39
65
47
51
129
90

0
0
1
3
8
10

2.13
5.88
6.20
11.11

8.9
6.0

2.22
5.36
13.27
20.00

1

!

WILLIAMSON COUNTY TUBERCULOSIS STUDY

64

$

TABLE XA

----3
N
onwhite Household Associates of “Sputum-positive”
Prevalence of Tuberculosis among
Index Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955

Age Group in Years

Number of
Household
Members

Female

Male

Total

Cases

Number

Per Cent

Number of
Household
Members

Cases
Number

Per Cent

Cases

Number of
Household
Members

Number

Per Cent

Total

K
11

Total
A djusted

464

36

7.8
8.3

228

15

6.6
7.1

236

21

8.9
9.5

<5
5-14
15-24
25-34
35-54
55+

66
128
98
40
72
60

3
7
6

4.5
5.5
6.1
7.5
9.7
16.7

31
59
53
23
32
30

2
3
2
1
3
4

6.5
5.1
3.8
4.3
9.4
13.3

35
69
45
17
40
30

1
4
4
2
4
6

2.9
5.8
8.9
11.8
10.0
20.0

143

15

10.5

18

1
3
3
2
3
3

5.6
6.0
11.5
20.0
11.5
23.1

6.5

3
7

10

Close Relatives

•I

Total
Adjusted

286

27

9.4
10.5

1 13

12

8.4

<5
5-14
15-24
25-34
35-54
55+

39
90
69
15
39
34

3
4
5
2
6
7

7.7
4.4
7.2
13.3
15.4
20.6

21
40
43
5
13
21

2
1
2
0
3
4

9.5
2.5
4.7

23.1
19.0

50
26
10
26
13

85

3

3.5

93

6

10
19
10

0
2
0
1
0
0

17
19
19
7
14
17

0
1
1
0
1
3

Other Members
Total
Adjusted

178

9

5.1
4.5

<5
5-14
15-24
25-34
35-54
55+

27
38
29
25
33
26

0
3
1
1
1

7.9
3.4
4.0

3

11.5

3.0

18

19
9

10.5
5.6

5.3
5.3

7.1
17.6

WILLIAMSON COUNTY TUBERCULOSIS STUDY

66

TABLE XIA
Prevalence of Tuberculosis among White Household Associates of “Sputum-positive” Index
Cases by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Upper and Middle
Socioeconomic Classes

Total

Age Group in Years
Number of
Household
Members

Cases
Number

Per Cent

Number of
Household
Members

Lower Socioeconomic Class

Cases

Cases

Number of
Household
Members

Number |

Per Cent

Number

Per Cent

397

47

11.8
10.3

467

44

9.4
11.4

28
60
83
44
88
94

1
1
5
7
17
16

3.6
1.7
6.0
15.9
19.3
17.0

60
154
95
41
85
32

1
3
7
5
22
6

1.7
1.9
7.4
12.2
25.9
18.8

8.5

242

21

8.7

1
1
3
3
10
3

3.8
1.2
5.9
13.6
23.3
18.8

Total
Total
Adjusted

864

<5
5-14
15-24
25-34
35-54
55+

88
214
178
85
173
126

91

10.5
10.7

2
4
12
12
39
22

2.3
1.9
6.7
14.1
22.5
17.5

Male

Total
Adjusted

442

38

8.6
9.6

200

17

<5
5-14
15-24
25-34
35-54
55+

41
124
96
44
81
56

1
1
4
7
16
9

2.4
0.8
4.2
15.9
19.8
16.1

15
40
45
22
38
40

0
0
1
4
6
6

2.2
18.2
15.8
15.0

26
84
51
22
43
16

Female

Total
A djusted

422

53

12.6
12.0

197

30

15.2

225

23

10.2

<5
5-14
15-24
25-34
35-54
55+

47
90
82
41
92
70

1
3
8
5
23
13

2.1
3.3
9.8
12.2
25.0
18.6

13
20
38
22
50
54

1
1
4
3
11
10

5.0
10.5
13.6
22.0
8.5

34
70
44
19
42
16

0
2
4
2
12
3

2.9
9.1
10.5
28.6
18.8

67

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XI/?
Prevalence of Tuberculosis among White Household Associates of “Other-than-Sputumpositive” Index Cases by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Upper and Middle
Socioeconomic Classes

Total

Age Group in Years

Number of
Household
Members

Cases
Number |

Per Cent

Number of
Household
Members

Lower Socioeconomic Class

Cases
Number Per Cent

Cases

Number of
Household
Members

Number

Per Cent

1125

48

4.3
7.4

0
2
5
8
10
23

0.5
2.3
8.7
7.0
28.8

Total

Total
Adjusted

1836

<5
5-14
15-24
25-34
35-54

212
525
373
200
296
230

55+

94

5.1

711

46

39
107
154
108
153
150

0
0
2
6
22
16

1.3
5.6
14.4
10.7

173
418
219
92
143
80

368

18

4.9

619

21

3.4

19
60
79
61
74
75

0
0
1
1
7
9

1.3
1.6
9.5
12.0

89
229
120
44
90
47

0
0
0
7
5
9

15.9
5.6
19.1

343

28

8.2

506

27

5.3

20
47
75
47
79
75

0
0
1
5
15
7

1.3
10.6
16.7
9.3

84
189
99
48
53
33

0
2
5
1
5
14

1.1
5.1
2.1
9.4
42.4

6.0

0
2
7
14
32
39

0.4
1.9
7.0
10.8
17.0

6.5
5.4

Male

Total
Adjusted

987

<5
5-14
15-24
25-34
35-54
55+

108
289
199
105
164
122

39

4.0

4+
0
0
1
8
12
18

0.5
7.6
7.3
14.8

Female
Total
A djusted

849

<5
5-14
15-24
25-34
35-54
55+

104
236
174
95
132
108

55

6.5
7.5

0
2
6
6
20
21

0.8
3.4
6.3
15.2
19.4

68

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XII
Prevalence of Tuberculosis among Nonwiiite Household Associates of “Sputum-positive”
Index Cases, by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Upper and Middle
Socioeconomic Classes

Total
Age Group in Years
Number of
Household
Members

Cases
Number |

Per Cent

Number of
Household
Members

Lower Socioeconomic Class

Cases

Cases

Number of
Household
Members

Number

Per Cent

342

25

7.3
8.0

3
6
5
3
4
4

5.6
5.6
6.3
10.3
9.3
13.8

Number

Per Cent

122

11

9.0
7.0

12
20
19
11
29
31

0
1
1
0
3
6

10.3
19.4

54
108
79
29
43
29

60

6

10.0

168

9

5.4

5
9
9
7
11
19

0
0
1
0
2
3

18.2
15.8

26
50
44
16
21
11

2
3
1
1
1
1

7.7
6.0
2.3
6.2
4.8
9.1

62

5

8.1

174

16

9.2

7
11
10
4
18
12

0
1
0
0
1
3

28
58
35
13
22
18

1
3
4
2
3
3

3.6
5.2
11.4
15.4
13.6
16.7

Total

Total
Adjusted

464

<5
5-14
15-24
25-34
35-54
55+

66
128
98
40
72
60

36

7.8
7.8

3
7
6
3
7
10

4.5
5.5
6.1
7.5
9.7
16.7

5.0
5.3

Male

Total
Adjusted

228

<5
5-14
15-24
25-34
35-54
55+

31
59
53
23
32
30

15

6.6
6.1

2
3
2
1
3
4

6.5
5.1
3.8
4.3
9.4
13.3

*

ji

Female
|i

Total
Adjusted

236

<5
5-14
15-24
25-34
35-54
55+

35
69
45
17
40
30

21

8.9
9.4

1
4
4
2
4
6

2.9
5.8
8.9
11.8
10.0
20.0

* Less than 10 persons; no percentage calculated

9.1

5.6
25.0

69

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XIIL4
Tuberculosis Attack Rates per 1,000 Person-Years among White Household Associates OF
“Sputum-positive” Index Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955
Total

Age Group in Years

Male
New Cases

New Cases

Person-Years

*

Female

Person-Years

Number

Rate

New Cases

Person-Years
Number

Rate

Number

Rate

5.2
4.8

Total
Total
Adjusted

7,476.00

36

4.8
4.5

3,803.75

17

4.5
4-1

3,672.25

19

<5
5-14
15-24
25-34
35-54
55+

333.75
1,351.50
1,489.00
903.50
1,729.00
1,669.25

3
1
13
9
3
7

9.0
0.7
8.7
10.0
1.7
4.2

196.00
772.00
850.25
429.75
814.50
741.25

3
0
5
2
3
4

15.3
5.9
4.7
3.7
5.4

137.75
579.50
638.75
473.75
914.50
928.00

0
1
8
7
0
3

1.7
12.5
14.8
3.2

Close Relatives

Total
Adjusted

4,018.75

29

7.2
6.8

2,175.00

13

6.0

1,843.75

16

8.7

<5
5-14
15-24
25-34
35-54
55+

176.50
861.75
1,004.75
533.75
690.00
752.00

3
1
13
7
1
4

17.0
1.2
12.9
13.1
1.4
5.3

111.50
519.75
621.50
278.75
304.75
338.75

3
0
5
2
1
2

26.9
8.0
7.2
3.3
5.9

65.00
342.00
383.25
255.00
385.25
413.25

0
1
8
5
0
2

2.9
20.9
19.6

1,628.75

4

2.5

1,828.50

3

1.6

84.50
252.25
228.75
151.00
509.75
402.50

0
0
0
0
2
2

3.9
5.0

72.75
237.50
255.50
218.75
529.25
514.75

0
0
0
2
0
1

4.8

Other Members
Total
Adjusted

3,457.25

7

<5
5-14
15-24
25-34
35-54
55+

157.25
489.75
484.25
369.75
1,039.00
917.25

0
0
0
2
2
3

2.0
1.8

5.4
1.9
3.3

9.1

1.9

WILLIAMSON COUNTY TUBERCULOSIS STUDY

70

TABLE XIIIB
Tuberculosis Attack Rates per 1,000 Person-years among White Household Associates of
“Other-than-Sputum-positive” Index Cases, by Age, Sex, and Relationship to the
Index Case
Williamson County, Tennessee, 1931-1955

Person-Years

Person-Years

Number

New Cases

New Cases

New Cases

Age Group in Years

Female

Male

Total

Number |

Rate

9,870.75

15

1.5
1.6

523.75
2,283.75
2,185.00
828.25
2,095.75
1,954.25

0
0
4
1
5
5

5,071.25

6

302.75
1,580.75
1,594.00
497.25
633.75
462.75

Number

Rate

8,964.75

17

1.9
1.9

1.8
1.2
2.4
2.6

634.75
2,031.50
1,770.75
888.25
1,844.00
1,795.50

0
1
9
1
2
4

0.5
5.1
1.1
1.1
2.2

1 .2

4,487.00

8

1.8

0
0
2
1
2
1

1.3
2.0
3.2
2.2

349.00
1,399.00
1,237.00
413.50
534.25
554.25

0
1
6
0
0
1

4,799.50

9

1.9

4,477.75

9

2.0

221.00
703.00
591.00
331.00
1,462.00
1,491.50

0
0
2
0
3
4

285.75
632.50
533.75
474.75
1,309.75
1,241.25

0
0
3
1
2
3

5.6
2.1
1.5
2.4

Rate

<

Person-Years

Total
Total
Adjusted

18,835.50

32

<5
5-14
15-24
25-34
35-54
55+

1,158.50
4,315.25
3,955.75
1,716.50
3,939.75
3,749.75

0
1
13
2
7
9

1.7
1.7

0.2
3.3
1.2
1.8
2.4

Close Relatives

Total
Adjusted

9,558.25

14

<5
5-14
15-24
25-34
35-54
55+

651.75
2,979.75
2,831.00
910.75
1,168.00
1,017.00

0
1
8
1
2
2

1.5
1.5

0.3
2.8
1.1
1.7
2.0

0.7
4.9

1.8

Other Members

I

|

Total
A dj usted

9,277.25

18

<5
5-14
15-24
25-34
35-54
55+

506.75
1,335.50
1,124.75
805.75
2,771.75
2,732.75

0
0
5
1
5
7

1.9
1.9

4.4
1.2
1.8
2.6

3.4

2.1
2.7

AX

71

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XIVA
Tuberculosis Attack Rates per 1,000 Person-Years among Nonwhite Household Associates OF
“Sputum-positive” Index Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955
Total

Age Group in Years

Male

New Cases

Female

New Cases

Person-Years

Person-Years
Number

Rate

New Cases
Person-Years

Number

Rate

Number

Rate

9.7
9.7

Total
Total
Adjusted

3,780.00

31

8.2
7.9

1,821.50

12

6.6
6.0

1,958.50

19

<5
5-9
10-14
15-24
25-34
35-54
55+

128.25
337.50
464.75
851.00
518.75
801.75
678.00

1
2
9
6
1
6
6

7.8
5.9
19.4
7.1
1.9
7.5
8.8

54.75
144.00
210.25
481.50
254.00
365.50
311.50

1
0
1
4
0
2
4

18.3

5.5
12.8

73.50
193.50
254.50
369.50
264.75
436.25
366.50

0
2
8
2
1
4
2

10.3
31.4
5.4
3.8
9.2
5.5

1,151.00

9

7.8

1,143.25

15

13.1

34.50
89.00
135.75
349.00
160.00
138.25
244.50

0
0
1
4
0
0
4

16.4

41.00
109.75
146.00
196.50
132.25
301.75
216.00

0
2
8
2
0
2
1

4.8
8.3

Close Relatives
Total
Adjusted

2,294.25

<5
5-9
10-14
15-24
25-34
35-54
55+

75.50
198.75
281.75
545.50
292.25
440.00
460.50

24

10.5
9.8

0
2
9
6
0
2
5

10.1
31.9
11.0
4.5
10.9

7.4
11.5

18.2
54.8
10.2
6.6
4.6

Other Members

Total
Adjusted

1,485.75

7

4.7
5.0

670.50

3

4.5

815.25

4

4.9

<5
5-9
10-14
15-24
25-34
35-54
55+

52.75
138.75
183.00
305.50
226.50
361.75
217.50

1
0
0
0
1
4
1

19.0

20.25
55.00
74.50
132.50
94.00
227.25
67.00

1
0
0
0
0
2
0

49.4

32.50
83.75
108.50
173.00
132.50
134.50
150.50

0
0
0
0
1
2
1

7.5
14.9
6.6

4.4
11.1
4.6

8.8

72

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE X1VB
Tuberculosis Attack Rates per 1,000 Person-Years among Nonwhite Household Associates
of “Other-than-Sputum-positive” Index Cases, by Age, Sex, and Relationship to
the Index Case
Williamson County, Tennessee, 1931-1955

Person-Years

Person-Years

Person-Years
Number |

New Cases

New Cases

New Cases

Age'Group in Years

Female

Male

Total

Rate

Number

Rate

Number

Rate

Total

Total

2,076.75

6

2.9

1,153.50

4

3.5

923.25

2

2.2

<5
5-9
10-14
15-24
25-34
35-54
55+

161.25
222.00
265.25
381.00
233.75
450.75
362.75

1
2
0
0
0
2
1

6.2
9.0

88.75
127.00
175.25
238.75
116.25
203.00
204.50

1
2
0
0
0
1
0

11.3
15.7

72.50
95.00
90.00
142.25
117.50
247.75
158.25

0
0
0
0
0
1
1

4.0
6.3

2.3

4.4
2.8

4.9

Close Relatives

Total

894.25

2

2.2

454.25

1

2.2

440.00

1

<5
5-9
10-14
15-24
25-34
35-54
55+

112.50
120.00
122.00
178.00
82.50
148.25
131.00

1
0
0
0
0
1
0

8.9

66.00
76.50
89.50
107.25
44.50
31.00
39.50

1
0
0
0
0
0
0

15.2

46.50
43.50
32.50
70.75
38.00
117.25
91.50

0
0
0
0
0
1
0

699.25

3

4.3

483.25

1

2.1

22.75
50.50
85.75
131.50
71.75
172.00
165.00

0
2
0
0
0
1
0

26.00
51.50
57.50
71.50
79.50
130.50
66.75

0
0
0
0
0
0
1

15.0

6.7

8.5

Other Members

Total

1,182.50

4

<5
5-9
10-14
15-24
25-34
35-54
55+

48.75
102.00
143.25
203.00
151.25
302.50
231.75

0
2
0
0
0
1
1

3.4

19.6

3.3
4.3

39.6

5.8

73

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XV/1
Tuberculosis Attack Rates per 1,000 Person-Years among White Household Associates of
“Sputum-positive” Index Cases, by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Total
Age Group
in Years

Upper Socioeconomic Class

New Cases
PersonYears

Number

Rate

Middle Socioeconomic Class

New Cases
PersonYears

Number

Rate

Lower Socioeconomic Class

New Cases
PersonYears
Number

Rate

New Cases

PersonYears
Number Rate

Total

Total

7,375.50

36

4.9

696.00

1

1.4 2,898.75

17

5.9 3,780.75

18

4.8

<5
5-14
15-24
25-34
35-54
55+

317.75
1,310.00
1,473.50
900.00
1,707.50
1,666.75

3
1
13
9
3
7

9.4
0.8
8.8
10.0
1.8
4.2

16.75
35.75
104.50
89.00
164.00
286.00

0
0
1
0
0
0

121.00
362.25
474.00
398.00
700.25
843.25

0
0
8
5
0
4

16.9
12.6

180.00
912.00
895.00
413.00
843.25
537.50

3
1
4
4
3
3

16.7
1.1
4.5
9.7

9.6

4.7

3.6
5.6

Male
Total

3,757.25

17

4.5

226.75

0

1,462.00

7

4.8 2,068.50

10

4.8

<5
5-14
15-24
25-34
35-54
55+

186.00
754.00
845.00
425.25
806.00
741.00

3
0
5
2
3
4

16.1

11.75
18.25
46.00
20.75
46.00
84.00

0
0
0
0
0
0

76.50
236.00
259.25
179.50
357.00
353.75

0
0
3
2
0
2

3
0
2
0
3
2

30.7

11.6
11.1

97.75
499.75
539.75
225.00
403.00
303.25

4.7

5.9
4.7
3.7
5.4

5.7

3.7
7.4
6.6

Female
Total

3,618.25

19

<5
5-14
15-24
25-34
35-54
55+

131.75
556.00
628.50
474.75
901.50
925.75

0
1
8
7
0
3

5.3

1.80
12.7
14.74
3.2

469.25

1

2.1 1,436.75

10

7.0 1,712.25

8

5.00
17.50
58.50
68.25
118.00
202.00

0
0
1
0
0
0

44.50
126.25
214.75
218.50
343.25
489.50

0
0
5
3
0
2

82.25
412.25
355.25
188.00
440.25
234.25

0
1
2
4
0
1

17.1

23.3
13.7
4.1

2.4
5.6
21.3

4.3

I •

I

74

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XV/3
Tuberculosis Attack Rates per 1,000 Person-Years among White Household Associates of
“Other-than-Sputum-positive” Index Cases, by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Upper Socioeconomic Class

Total

Age Group
in Years

PersonYears

Number

Rate

PersonYears

Number

Lower Socioeconomic Class

New Cases

New Cases

New Cases

New Cases
PersonYears

Middle Socioeconomic Class

Rate

PersonYears
Number

Rate

Number Rate

Total
Total

18,624.50

32

1.7

1,261.00

0

6,544.75

16

2.4 10,818.75

16

<5
5-14
15-24
25-34
35-54
55+

1,151.00
4,271.75
3,906.75
1,685.50
3,894.25
3,715.25

0
1
13
2
7
9

0.2
3.3
1.2
1.8
2.4

51.00
128.00
150.75
147.75
358.50
425.00

0
0
0
0
0
0

225.75
839.00
1,113.50
687.75
1,646.25
2,032.50

0
0
7
2
1
6

6.3
2.9
0.6
3.0

874.25
3,304.75
2,642.50
850.00
1,889.50
1,257.75

0
1
6
0
6
3

2.2

5,925.00

8
0
0
3
0
4
1

1.5

0.3
2.3
3.2
2.4

Male
Total

9,760.50

15

<5
5-14
15-24
25-34
35-54
55+

523.25
2,257.75
2,164.00
816.25
2,065.75
1,933.50

0
0
4
1
5
5

1.5

598.00

0

3,237.50

7

1.8
1.2
2.4
2.6

25.00
82.50
84.50
53.50
188.50
164.00

0
0
0
0
0
0

96.50
410.25
547.50
325.75
843.75
1,013.75

0
0
1
1
1
4

1.8
3.1
1.2
3.9

401.75
1,765.00
1,532.00
437.00
1,033.50
755.75

2.7

4,893.75

8

10.6
2.8

472.50
1,539.75
1,110.50
413.00
856.00
502.00

0
1
3
0
2
2

1.4

2.0
3.9
1.3

Female

Total

8,864.00

17

<5
5-14
15-24
25-34
35-54
55+

627.75
2,014.00
1,742.75
869.25
1,828.50
1,781.75

0
1
9
1
2
4

1.9

0.5
5.2
1.2
1.1
2.2

663.00

0

3,307.25

9

26.00
45.50
66.25

0
0
0
0
0
0

129.25
428.75
566.00
362.00
802.50
1,018.75

0
0
6
1
0
2

94.25
170.00
261.00

2.0

1.6

0.6
2.7
2.3
4.0

75

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XVL4
Tuberculosis Attack Rates per 1,000 Person-Years among Nonwhite Associates of “Sputum­
positive” Index Cases, by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Upper and Middle
Socioeconomic Classes

Total

Age Groupjn Years

New Cases

Person-Years

Number |

Lower Socioeconomic Class

New Cases

Person-Years
Rate

Number |

New Cases

Person-Years
Rate

Number

Rate

Total
Total

3,780.00

31

8.2

908.00

2

<5
5-14
15-24
25-34
35-54
55 +

128.25
802.25
851.00
518.75
801.75
678.00

1
11
6
1
6
6

7.8
13.7
7.1
1.9
7.5
8.8

22.00
98.00
135.50
119.00
245.00
288.50

0
0
0
0
1
1

2.2

2,872.00

29

10.1

1
11
6

4.1
3.5

106.25
704.25
715.50
399.75
556.75
389.50

5
5

9.4
15.6
8.4
2.5
9.0
12.8

1

Male
Total

1,821.50

12

G.6

397.00

0

1,424.50

12

8.4

<5
5-14
15-24
25-34
35-54
55 +

54.75
354.25
481.50
254.00
365.50
311.50

1
1
4
0
2
4

18.3
2.8
8, 3

10.50
52.50
79.00
57.50
84.50
113.00

0
0
0
0
0
0

44.25
301.75
402.50
196.50
281.00
198.50

1

1
4
0
2
4

22.6
3.3
9.9
7.1
20.2

3.9

1,447.50

17

11.7

6.2
5.7

62.00
402.50
313.00
203.25
275.75
191.00

0
10
2
1
3
1

24.8
6.4
4.9
10.9
5.2

5.5
12.8

Female
Total

1,958.50

19

9.7

511.00

2

<5
5-14
15-24
25-34
35-54
55+

73.50
448.00
369.50
264.75
436.25
366.50

0
10
2
1
4
2

22.3
5.4
3.8
9.2
5.5

11.50
45.50
56.50
61.50
160.50
175.50

0
0
0
0
1
1

76

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XV W
Tuberculosis Attack Rates per 1,000 Person-Years among Nonwhite Associates of “Otherthan-Sputum-positive” Index Cases, by Age, Sex, and Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Upper and Middle
Socioeconomic Classes

Total

Age Group in Years

Person-Years

Person-Years

Number |

New Cases

New Cases

New Cases
Person-Years

Lower Socioeconomic Class

Rate

Number

Rate

4.2

Number

Rate

1,588.75

4

2.5

143.75
448.00
330.50
191.75
258.50
216.25

1
2
0
0
0
1

7.0
4.5

956.25

3

3.1

82.25
278.00
204.25
106.75
136.00
149.00

1
2
0
0
0
0

12.2
7.2

632.50

1

1.6

61.50
170.00
126.25
85.00
122.50
67.25

0
0
0
0
0
1

14.9

Total
Total

2,066.75

6

2.9

478.00

2

<5
5-14
15-24
25-34
35-54
55+

160.75
483.75
380.50
233.75
445.25
362.75

1
2
0
0
2
1

6.2
4.1

17.00
35.75
50.00
42.00
186.75
146.50

0
0
0
0
2
0

4.5
2.8

10.7

4

4.6

Male

Total

1,144.50

4

3.5

188.25

1

<5
5-14
15-24
25-34
35-54
55+

88.75
298.75
238.75
116.25
197.50
204.50

1
2
0
0
1
0

11.3
6.7

6.50
20.75
34.50
9.50
61.50
55.50

0
0
0
0
1
0

5.1

5.3

16.3

Female

Total

922.25

2

<5
5-14
15-24
25-34
35-54
55+

72.00
185.00
141.75
117.50
247.75
158.25

0
0
0
0
1
1

2.2

289.75

1

4.0
6.3

10.50
15.00
15.50
32.50
125.25
91.00

0
0
0
0
1
0

3.5

8.0

WILLIAMSON COUNTY TUBERCULOSIS STUDY

77

TABLE XVIL4
Number of Deaths and Death Rates per 1,000 Person-Years for All Causes, for Tuberculosis,
AND FOR NONTUBERCULOUS CAUSES AMONG WlIITE HOUSEHOLD ASSOCIATES OF “SPUTUM-POSITIVE”

Index Cases, by Age, Sex, and Relationship to the Index Case
Williamson County, Tennessee, 1931-1955
Total

Age Group in
Years

v

Male

All
Tuber­
Causes culosis

Female

All
Tuber­
Causes culosis

Other

PersonYears

PersonYears

E
3

z

ad
cd

Pi

E
3

oj
cd

Pi

E

3

E

3

cd

z

Pi

OJ
cd

Pi

i
z

PersonYears

id

cd

Pi

All
Tuber­
Causes culosis

Other

i
z

<D

t%

o
cd

Pi

Other

CD

cd

Pi

S3

o
cd

Pi

1
%

cd

Pi

Total

Total
Adjusted
<5
5-14
15-24
25-34
35-54
55+

7,474.00 94 12.6 13 1.7 81 10.9 3,801.75 55 14.5 6 1.649 12.9 3,672.25 39 10.6 7 1.9 32 8.7
12.3
15.4
9.0
333.75 4 12.0 1 3.0 3 9.0
1,351.50 5 3.7 0 — 5 3.7
1,489.00 7 4.7 2 1.3 5 3.4
903.50 7 4.7 3 3.3 4 4.4
1,729.00 13 7.5 21.211 6.3
1,669.25 58 34.7 5 3.0 53 31.7

196.00 3 15.3
772.00 4 5.2
850.25 3 3.5
429.75 4 9.3
814.50 9 11.0
741.25 32 43.2

15.1 2 10.2
0 — •1 5.2
1 1.2 2 2.3
1 2.3 3 7.0
22.5 7 8.5
1 1.3 31 41.9

137.75 1 7.3
579.50 1 1.7
638.75 4 6.3
473.75 3 6.3
914.50 4 4.4
928.00 26 28.0

0 — 1 7.3
0 — 1 1.7
1 1.6 3 4.7
24.2 1 2.1
0 — 4 4.4
44.322 23.7

Close Relatives

Total
Adjusted
<5
5-14
15-24
25-34
35-54
55+

4,018.7547 11 .79*2.238 9.5 2,175.00 29 13.3 5 2.324 11.0 1,843.75 18 9.8 42.214 7.6
12.2
176.50 4 22.7
861.75 4 4.6
1,004.75 5 5.0
533.75 4 7.5
690.00 5 7.2
752.00 25 33.2

1 5.7 3 17.0
0 — 4 4.6
22.0 3 3.0
23.7 2 3.7
1 1.4 4 5.8
34.022 29.2

111.50 3 26.9
519.75 3 5.8
621.50 1 1.6
278.75 2 7.2
304.75 3 9.8
338.75 17 50.2

19.0 217.9
0 — 3 5.8
1 1.6 0
1 3.6 1 3.6
1 3.3 2 6.5
1 3.0 16 47.2

65.00
342.00
383.25
255.00
385.25
413.25

1 15.4
1 2.9
4 10.4
2 7.8
2 5.2
8 19.4

0 — 1 15.4
0 — 1 2.9
1 2.6 3 7.8
1 3.9 1 3.9
0 — 2 5.2
24.8 6 14.6

Other Members
Total
Adjusted
<5
5-14
15-24
25-34
35-54
55+

3,455.25 47 13.6 4 1.243 12.4 1,626.75 26 15.9 1 0.6 25 15.3 1,828.50 21 11.5 3 1.618 9.9
11.2
157.25 0
489.75 1 2.0
484.25 2 4.1
369.75 3 8.1
1,039.00 8 7.7
917.25 33 36.0

0 — 0
0 — 1 2.0
0 — 2 4.1
1 2.7 2 5.4
1 1.0
6.7
22.231 33.8

84.50 0
252.25 1 4.0
228.75 2 8.7
151.00 2 13.2
509.75 6 11.8
402.50 15 37.3

0 — 0
0 —, 1 4.0
0 — 2 8.7
0 — 2 13.2
1 2.0 5 9.8
0 — 15 37.3

72.75 0
237.50 0
255.50 0
218.75 1 4.6
529.25 2 3.8
514.75 18 35.0

0 — 0
0 — 0
0 — 0
1 4.6 0
0 — 2 3.8
23.9 1631.1

* JIn a previous report (20) the total deaths were shown as 10. Since that report it was learned that a
child of two years who had been included in that total had actually died some six months after re­
moval from his household.

78

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XVIIB
Number of Deaths and Death Rates per 1,000 Person-Years for All Causes, for Tuberculosis
AND FOR NoNTUBERCULOUS CAUSES AMONG WlHTE HOUSEHOLD ASSOCIATES OF “OtHER-THANSputum-positive” Index Cases by Age, Sex, and Relationship to the Index Case

Williamson County, Tennessee, 1931-1955

Age Group
in Years

All
Causes

PersonYears

Female

Male

Total

Tuber­
culosis

All
Causes

Other

Tuberculosis

o
aJ

E
Z

JU
cd

Pi

£
Z

Other

PersonYears

PersonYears

1z

Tuber­
All
Causes culosis

Other

£

3

cd

P4

£

ju
cd

pi

£

3

z

ju
cd

Pi

<
E3

z

JU
cd

Pi

£3

z

JU
oj

E3

z

<

JU
cd

Pi

Total
18,835.50 185 9.8100.5175 9.39,870.75 108 10.9 50.5 103 10.48,964.75 77 8.6 50.672 8.0
Total
9.3
11.8
10.6
Adjusted

<5
5-14
15-24
25-34
35-54
55+

1,158.50 5 4.3
4,315.25 3 0.7
3,955.75 4 1.0
1,716.50 5 2.9
3,939.75 27 6.9
3,749.75 141 37.6

0 — 5 4.3 523.75 3 5.7
0 — 3 0.7 2,283.75 3 1.3
0 — 4 1.02,185.00 2 0.9
0 — 5 2.9 828.25 3 3.6
20.5 25 6.4 2,095.75 15 7.2
82.1 133 35.5 1,954.25 82 42.0

2 3.2
0 — 3 5.7 634.75 2 3.2 0
0
0
0 — 3 1.3 2,031.50 0
2 1.1
0 — 2 0.9 1,770.75 2 1.1 0
2 2.3
2
2.3
0
888.25
0 — 3 3.6
10.5 14 6.7 1,844.00 12 6.5 10.5 11 6.0
42.0 78 40.0 1,795.50 59 32.9 4 2.2 55 30.7

Close Relatives

9,558.25 60 6.3 40.4 56 5.9 5,071.25 31 6.1 20.4 29 5.7 4,487.00 29 6.5 2 0.4 27 6.1
Total
11.2
Adjusted
<5
5-14
15-24
25-34
35-54
55+

651.75 2 3.1
2,979.75 3 1.0
2,831.00 1 0.4
910.75 2 2.2
1,168.00 10 8.6
1,017.00 42 41.3

0 — 2 3.1 302.75 1 3.3
0 — 3 1.0 1,580.75 3 1.9
0 — 1 0.4 1,594.00 0
0 — 2 2.2 497.25 1 2.0
1 0.9 9 7.7 633.75 4 6.3
3 2.9 39 38.4 462.75 22 47.5

1 2.9
0 — 1 3.3 349.00 1 2.9 0
0
0
0
1.9
1,399.00
3
0 —
1 0.8
0 — 0 — 1,237.00 1 0.8 0
1 2.4
0 — 1 2.0 413.50 1 2.4 0
0 — 4 6.3 534.25 611.2 1 1.9 5 9.3
24.3 20 43.2 554.25 20 36.1 1 1.8 19 34.3

Other Members

9,277.25 125 13.5 6 0.6 119 12.9 4,799.50 77 16.0 30.6 74 15.44,477.75 48 10.7 30.745 10.0
Total
10.0
Adjusted
<5
5-14
15-24
25-34
35-54
55+

1 3.5
506.75 3 5.9 0 — 3 5.9 221.00 2 9.0 0 — 2 9.0 285.75 1 3.5 0
0
0
0
632.50
0
0

703.00 0
0 — 0
1,335.50 0
1 1.9
1,124.75 3 2.7 0 — 3 2.7 591.00 2 3.4 0 — 2 3.4 533.75 1 1.9 0
1 2.1
805.75 3 3.7 0 — 3 3.7 331.00 2 6.0 0 — 2 6.0 474.75 1 2.1 0
6
4.6
0
6
4.6
6.8
1,309.75
1
0.7
10
2,771.75 17 6.1 1 0.4 16 5.7 1,462.00 11 7.5
3
58
38.9
1,241.25
39
31.4
2.4
36
29.0
2
1.3
60
40.2
94
34.4
1,491.50
2,732.75 99 36.2 5 1 .8

o

79

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XVIIIA
Number of Deaths and Death Rates per 1,000 Person-Years for All Causes, for Tuberculosis,
AND FOR NoNTUBERCULOUS CAUSES AMONG NONWHITE HOUSEHOLD ASSOCIATES OF “SPUTUMpositive” Index Cases, by Age, Sex, and Relationship to the Index Case

Williamson County, Tennessee, 1931-1955
Total

Age Group
in Years

All
Causes

PersonYears

1z

o
cd

Male

Tuber­
culosis

1
z

PersonYears

<u
cd

P4

All
Causes

Other

B

4)

S

3

3

z

z

a;
cd

Female

Tuber­
culosis

Other

aj

CD

z1

a;
cd

s3

All
Causes

Tuber­
culosis

PersonYears
-Q

e3
z

cd

P4

Other

J;
JU
cd

1

3

z

cd

B

z3

JU
cd

Pi

Total
Total
3,780.00 84 22.2 25 6.6 59 15.6 1,821.50 39 21.4 11 6.028 15.4 1,958.50 45 23.0 14 7.1 31 15.9
Adjusted
24-9
24.6
25.1

•v

<5
5-14
15-24
25-34
35-54
55 +

128.25 646.8
802.25 4 5.0
851.00 11 12.9
518.75 7 13.5
801.75 15 18.7
678.00 41 60.5

323.4 3 23.4
4 5.0 0
9 10.6 2 2.4
1 1.9 6 11.6
2 2.5 13 16.2
6 8.8 35 51.7

54.75 354.8
354.25 0 —
481.50 5 10.4
254.00 519.7
365.50 6 16.4
311 .50 20 64.2

2 36.5 1 18.3
0
0 —
4 8.3 1 2.1
0
5 19.7
1 2.7 5 13.7
4 12.8 1651.4

73.50 3 40.8
448.00 4 8.9
369.50 6 16.2
264.75 2 7.6
436.25 9 20.6
366.50 21 57.3

1 13.6 227.2
4 8.9 0 —
5 13.5 1 2.7
1 3.8 1 3.8
1 2.3 8 18.3
2 5.5 1951.8

Close Relatives
Total
2,294.25 58 25.3 20 8.7 38 16.6 1,151.00 26 22.6 9 7.8 17 14.8 1,143.25 32 28.0 11 9.6 21 18.4
Adjusted
26.8

<5
5-14
15-24
25-34
35-54
55+

75.50 5 66.2
480.50 4 8.3
545.50 11 20.2
292.25 3 10.3
440.00 7 15.9
460.50 28 60.8

2 26.5 3 39.7
4 8.3 0
9 16.6 2 3.6
0 — 3 10.3
0 — 7 15.9
5 10.9 23 49.9

34.50 2 58.0
224.75 0 —
349.00 5 14.3
160.00 3 18.8
138.25 2 14.5
244.50 14 57.3

1 29.0 1 29.0
0 — 0 —
4 11.5 1 2.8
0 — 3 18.8
0 — 214.5
4 16.4 10 40.9

41.00 3 73.2
255.75 4 15.6
196.50 6 30.5
132.25 0 —
301.75 5 16.6
216.00 14 64.8

1 24.4 248.8
4 15.6 0
525.5 1 5.0
0
0
0
516.6
1 4.6 13 60.2

Other Members
Total
1,485.75 26 17.5 5 3.421 14.1
Adjusted
21.9

<5
5-14
15-24
25-34
35-54
55 +

52.75 1 19.0
321.75 0 —
305.50 0 —
226.50 4 17.7
361.75 8 22.1
217.50 13 59.8

1 19.0 0 —
0
0 —
0
0 —
1 4.4 3 13.3
2 5.5 6 16.5
1 4.6 1255.2

670.50 13 19.4 2 3.0 11 16.4
20.25
129.50
132.50
94.00
227.25
67.00

1 49.4
0 —
0 —
221.3
4 17.6
6 89.6

149.4
0
0
0
1 4.4
0

0 —
0 —
0 —
2 21.3
3 13.2
6 89.6

815.25 13 15.9 3 3.7 10 12.2
32.50
192.25
173.00
132.50
134.50
150.50

0 —
0 —
0 —
2 15.1
4 29.7
7 46.5

0
0
0
0
0
0
1 7.5 1 7.5
1 7.4 3 22.3
1 6.6 6 39.9

80

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XVIIIB
Number of Deaths and Death Rates per 1,000 Person-Years for All Causes, for Tuberculosis,
AND FOR NoNTUBERCULOUS CAUSES AMONG NONWHITE HOUSEHOLD ASSOCIATES OF “OtHER-THANSputum-positive” Index Cases, by Age, Sex, and Relationship to the Index Case

Williamson County, Tennessee, 1931-1955
Male

Total

Age Group
in Years

All
Tuber­
Causes culosis

PersonYears

Female

All
Tuber­
Causes culosis

Other

PersonYears

u
v

z

o
rt

E

3

C4

E3
Z

PersonYears

<D

-D

E

3

Other

E

3

z

E

JU

3

%

E

3

rt

Pi

All
Causes

Tuberculosis

<D

<u

Other
1

1

z

Pi

§

Oj

E

0J

z

Pi

3

z

rt

Total
Total
Adjusted
<5
5-14
15-24
25-34
35-54
55+

2,076.75 30 14.4 4 1.9 26 12.5 1,153.50 15 13.0 2 1.7 13 11.3923.25 15 16.2 2 2.213 14.0
18.5
15.3
16.7
161.25 2 12.4
487.25 3 6.2
381.00 4 10.5
233.75 1 4.3
450.75 7 15.5
362.75 13 35.8

0 — 2 12.4
1 2.1 2 4.1
25.2 2 5.2
0 — 1 4.3
1 2.2 6 13.3
0 — 13 35.8

88.75
302.25
238.75
116.25
203.00
204.50

2 22.5
2 6.6
2 8.4
1 8.6
2 9.9
6 29.3

0 —
1 3.3
0 —
0 —
1 4.9
0 —

2 22.5 72.50
1 3.3 185.00
2 8.4 142.25
1 8.6 117.50
1 4.9 247.75
6 29.3 158.25

0
1 5.4
2 14.1
0 —
520.2
744.2

0 —
0 —
2 14.1
0 —
0 —
0 —

0
1 5.4
0
0
5 20.2
7 44.2

Close Relatives
Total
Adjusted
<5
5-14
15-24
25-34
35-54
55+

894.25 11 12.3
15.2

1 1 .1 10 11.2

454.25 5 11.0

0

5 11.0 440.00 (i 13.(>

112.50
242.00
178.00
82.50
148.25
131.00

0 — 2 17.8
0 — 0 —
1 5.6 0 —
0 — 0 —
0 — 3 20.2
0 — 5 38.2

66.00
166.00
107.25
44.50
31.00
39.50

2 30.3
0 —
0 —
0 —
0 —
3 75.9

0
0
0
0
0
0

2 30.3 46.50
0 — 76.00
0 — 70.75
0 — 38.00
0 — 117.25
3 75.9 91.50

2 17.8
0
1 5.6
0
3 20.2
5 38.2

1

2.3 5 11.3

o —
0
o —
0
1 14.1 1 14.1 o —
0 —
0
0
3 25.6 0
3 25.6
221.9
2 21.9 0
0
0

Other Members

Total
Adjusted
<5
5-14
15-24
25-34
35-54
55 +

1,182.50 19 16.1 3 2.5 16 13.6
17.9

48.75
245.25
203.00
151.25
302.50
231.75

0 —
3 12.2
3 14.8
1 6.6
4 13.2
8 34.5

0 —
1 4.1
14.9
0
1 3.3
0

0
2 8.1
2 9.9
1 6.6
3 9.9
8 34.5

699.25 10 14.3 2 2.9 8 11.4 483.25 9 18.6 1 2.1 8 16.5
22.75
136.25
131.50
71.75
172.00
165.00

0 —

2 14.7
2 15.2
1 13.9
2 11.6
3 18.2

0 —
1 7.3
0 —
0 —
1 5.8
0 —

0 — 26.00
1 7.3 109.00
2 15.2 71.50
1 13.9 79.50
1 5.8 130.50
3 18.2 66.75

0
0
1 9.2 0
1 14.0 1 14.0
0
0
2 15.3 0
5 74.9 0

0
1 9.2
0
0
2 15.3
574.9

4

WILLIAMSON COUNTY TUBERCULOSIS STUDY

81

TABLE XIXA
Death Rates per 1,000 Person-Years for All Causes and for Tuberculosis among White.
Household Associates of “Sputum-positive” Index Cases, by Age, Sex, and
Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Total

Age Group
in Years

All Causes
PersonYears

Male
Tuberculosis
Deaths

a>
cd

E

Tuberculosis
Deaths

PersonYears

Jd
E

All Causes

Female

1

o
cd

PersonYears

CD

0J

42

E

C4

All Causes

5

cd

Tuberculo­
sis Deaths

CD

1

CD

cd

1

■S'
cd

Total
Total
7,375.50 94
Adjusted
<5
5-14
15-24
25-34
35-54
55+

317.75 4
1,310.00 5
1,473.50 7
900.00 7
1,707.50 13
1,666.75 58

12.7
11.9

13

1.8 3,757.25 55

14.6
18.Jf.

6

1.6 3,618.25 39

10.8
10.3

7

1.9

12.6
3.8
4.8
7.8
7.6
34.8

1
0
2
3
2
5

3.1

16.1
5.3
3.6
9.4
11.2
43.2

1
0
1
1
2
1

5.4

131.75 1
556.00 1
628.50 4
474.75 3
901.50 4
925.75 26

7.6
1.8
6.4
6.3
4.4
28.1

0
0
1
2
0
4

1.6
4.2
4.3

10.7

1

2.1

5.0

1.4
3.3
1.2
3.0

186.00 3
754.00 4
845.00 3
425.25 4
806.00 9
741.00 32

1.2
2.4
2.5
1.3

Upper Socioeconomic Class

Total
Adjusted
<5
5-14
15-24
25-34
35-54
55+

696.00 11

16.75
35.75
104.50
89.00
164.00
286.00

0
0
0
1
3
7

15.8
10.8

11.2
18.3
24.5

1
0
0
0
0
0
1

1.4

226.75

6

3.5

11.75
18.25
46.00
20.75
46.00
84.00

0
0
0
1
1
4

26.5

0

469.25

5

48.2
21.7
47.6

0
0
0
0
0
0

5.00
17.50
58.50
68.25
118.00
202.00

0
0
0
0
2
3

16.9
14.9

0
0
0
0
0
1

10.4

3

2.1

4.7

1.8

Ik

Middle Socioeconomic Class

Total
2,898.75 46
Adjusted
<5
5-14
15-24
25-34
35-54
55+

121.00 0
362.25 0
474.00 2
398.00 1
700.25 7
843.25 36

15.9
12.8

4.2
2.5
10.0
42.7

6

2.1 1,462.00 31

0
0
1
1
1
3

76.50 0
236.00 0
259.25 0
179.50 I
357.00 7
353.75 23

2.1
2.5
1.4
3.6

21.2

3

2.1 1,436.75 15

5.6
19.6
65.0

0
0
0
1
1
1

44.50 0
126.25 0
214.75 2
218.50 0
343.25 0
489.50 13

26.6

0
0
1
0
0
2

5.6
2.8
2.8

9.3

4.1

Lower Socioeconomic Class
Total
3,780.75 37
A djusted

9.8
114

6

1.6 2,068.50 18

8.7

3

1 .5 1,712.25 19

11.1

3

4
5
5
5
3
15

22.2
5.5
5.6
12.1
3.6
27.9

1
0
1
2
1
1

5.6

30.7
8.0
5.6
8.9
2.5
16.5

1
0
1
0
1
0

10.2

12.2
2.4
5.6
16.0
4.5
42.7

0
0
0
2
0
1

<5
5-14
15-24
25-34
35-54
55 +

180.00
912.00

895.00
413.00
843.25
537.50

1.1
4.8
1.2
1.9

97.75
499.75
539.75
225.00
403.00
303.25

3
4
3
2
1
5

1.9
2.5

82.25 1
412.25 1
355.25 2
188.00 3
440.25 2
234.25 10

10.6
4.3

WILLIAMSON COUNTY TUBERCULOSIS STUDY

82

TABLE XIXB
Death Rates per 1,000 Person-Years for All Causes and for Tuberculosis among White
Household Associates of “Other-than-Sputum-positive” Index Cases, by Age, Sex,
and Socioeconomic Status
Williamson County, Tennessee, 1931-1955

Age Group in
Years

All Causes

Tuberculosis
Deaths

PersonYears
Num­
ber

Female

Male

Total

All Causes

PersonYears
Num­
ber

Rate

Rate

Tuberculosis
Deaths

Rate

gs

z

All Causes

PersonYears
g OJ

Rate

Z

Rate

Tuber­
culosis
Deaths

go

z

Rate
I

Total
Total
Adjusted

18,624.50 182

<5
5-14
15- 24
25-34
35-54
55+

5
1,151.00
3
4,271.75
4
3,906.75
1,685.50
5
3,894.25 27
3,715.25 138

9.8 10
104

0.5 9,760.50 106

4.3
0.7
1.0
3.0
6.9
37.1

523.25
2,257.75
2,164.00
816.25
0.5 2,065.75
2.2 1,933.50

10.9

5

0.5 8,864.00 76

8.6 5
9.3

0
0
0
0
1
4

627.75 2
2,014.00 0
1,742.75 2
869.25 2
0.5 1,828.50 12
2.1 1,781.75 58

3.2 0
0
1.1 0
2.3 0
6.6 1
32.6 4

114

0
0
0
0
2
8

3
3
2
3
15
80

5.7
1.3
0.9
3.7
7.3
41.4

0.6

0.5
2.2

Upper Socioeconomic Class
Total
Adjusted

1,261.00

13

<5
5-14
15-24
25-34
35- 54
55+

51 .00
128.00
150.75
147.75
358.50
425.00

0
0
0
0
2
11

10.3
6.7

5.6
25.9

0

598.00

6

0
0
0
0
0
0

25.00
82.50
84.50
53.50
188.50
164.00

0
0
0
0
1
5

10.0

0

663.00

5.3
30.5

0
0
0
0
0
0

26.00
45.50
66.25
94.25
170.00
261.00

10.6 0

0
0
0
0
1
6

0
0
0
0
5.9 0
23.0 0

Middle Socioeconomic Class

I Total
Adjusted

6,544.75

93

14.2
10.2

4

0.6 3,237.50

54

16.7

2

0.6 3,307.25 39

11.8 2

0.6

<5
5-14
15-24
25-34
35-54
55+

225.75
839.00
1,113.50
687.75
1,646.25
2,032.50

1
0
1
1
17
73

4.4

0
0
0
0
1
3

96.50
410.25
547.50
325.75
843.75
0.6
1.5 1,013.75

1
0
1

10.4

0
0
0
0
0
2

129.25 0
428.75 0
566.00 0
362.00 1
802.50 6
2.0 1,018.75 32

0
0
0
2.8 0
7.5 1
31.4 1

1.2
1.0

0.9
1.5
10.3
35.9

1.8

0
11
41

13.0
40.4

Lower Socioeconomic Class
Total
Adjusted

10,818.75

76

7.0
10.9

6

0.6 5,925.00

46

7.8

3

0.5 4,893.75 30

6.1 3

0.6

<5
5-14
15-24
25-34
35-54
55+

874.25
3,304.75
2,642.50
850.00
1,889.50
1,257.75

4
3
3
4
8
54

4.6
0.9
1.1
4.7
4.2
42.9

0
0
0
0
1
5

401.75
1,765.00
1,532.00
437.00
0.5 1,033.50
755.75
4.0

2
3
1
3
3
34

5.0
1.7
0.7
6.9
2.9
45.0

0
0
0
0
1
2

472.50 2
1,539.75 0
1,110.50 2
413.00 1
1.0
856.00 5
502.00 20
2.6

4.2 0
0
1.8 0
2.4 0
5.8 0
39.8 3

6.0

83

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XXA
Death Rates per 1,000 Person-Years for All Causes and for Tuberculosis among Non white
Household Associates of “Sputum-positive” Index Cases, by Age, Sex, and
Socioeconomic Status
Williamson County, Tennessee, 1931-1955
Male

Total

Age Group in
Years

All Causes

PersonY ears

Tuberculosis
Deaths

0 Rate

PersonYears

Rate

Female

All Causes Tuberculosis
Deaths


5^
Z

PersonYears
Rate

E
£
5-0

z

Rate

All Causes Tuberculosis
Deaths

0 Rate 2;g £ Rate

Total

Total
Adjusted

3,780.00 84

<5
5-14
15-24
25-34
35-54
55+

128.25 6
802.25 4
851.00 11
518.75 7
801.75 15
678.00 41

22.2 25
2^.2

6.6 1,821.50 39

46.8
5.0
12.9
13.5
18.7
60.5

23.4
5.0
10.6
1.9
2.5
8.8

3
4
9
1
2
6

54.75 3
354.25 0
481.50 5
254.00 5
365.50 6
311.50 20

21 .4 11
23.7

54.8
10.4
19.7
16.4
64.2

2
0
4
0
1
4

6.0 1,958.50 45

23.0 14

7.1

24.7
73.50 3
448.00 4
369.50 6
264.75 2
436.25 9
366.50 21

40.8
8.9
16.2
7.6
20.6
57.3

1
4
5
1
1
2

13.6
8.9
13.6
3.8
2.3
5.5

21.5

1

2.0

16.3
12.5
45.6

0
0
0
0
0
1

5.7

7.7 1,447.50 34

23.5 13

9.0

45.2

48.4
10.0
19.2
4.9
25.4
68.1

16.1
10.0
16.0
4.9
3.6
5.2

36.5
8.3
2.7
12.8

Upper and Middle Socioeconomic Classes

u

Total
Adjusted

908.00 23

<5
5-14
15-24
25-34
35-54
55+

22.00 0
98.00 0
135.50 0
119.00 3
245.00 4
288.50 16

25.3
17.4

25.2
16.3
55.4

1

0
0
0
0
0
1

1.1

397.00 12

3.5

10.50
52.50
79.00
57.50
84.50
113.00

0
0
0
2
2
8

30.2

0

511.00 11

34.8
23.7
70.8

0
0
0
0
0
0

11.50
45.50
56.50
61.50
160.50
175.50

0
0
0
1
2
8

Lower Socioeconomic Class

Total
Adjusted

2,872.00 61

<5
5-14
15-24
25-34
35-54
55+

106.25 6
704.25 4
715.50 11
399.75 4
556.75 11
389.50 25

21.2 24
26.1
56.5
5.7
15.4
10.0
19.8
64.2

3
4
9
1
2
5

8.4 1,424.50 27

19.0 11

28.2
5.7
12.6
2.5
3.6
12.8

67.8

44.25 3
301.75 0
402.50 5
196.50 3
281.00 4
198.50 12

12.4
15.3
14.2
60.5

2
0
4
0
1
4

9.9
3.6
20.2

62.00 3
402.50 4
313.00 6
203.25
1
275.75 7
191.00 13

1
4
5
1
1
1

84

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XXB
Death Rates per 1,000 Person-Years for All Causes and for Tuberculosis among Nonwhite
Household Associates of “Other-than-Sputum-positive” Index Cases, by Age, Sex, and
Socioeconomic Status
Williamson County, Tennessee, 1931-1955

Age Group in
Years

PersonYears

Female

Male

Total
All Causes

Tuberculosis
Deaths

0 Rate


%

All Causes

Tuberculosis
Deaths

PersonYears

PersonYears

I

Rate

Rate

Rate

All Causes Tuberculosis
Deaths

Rate 3x>
Z

Rate

2

2.2

Total

Total
Adjusted

2,066.75 30

14.5
154

4

<5
5-14
15-24
25-34
35-54
55+

160.75 2
483.75 3
380.50 4
233.75 1
445.25 7
362.75 13

12.4
6.2
10.5
4.3
15.7
35.8

0
1
2
0
1
0

1.9 1,144.50

15

13.1
13.5

2

1.7 922.25 15

88.75
298.75
238.75
116.25
197.50
204.50

2
2
2
1
2
6

22.5
6.7
8.4
8.6
10.1
29.3

0
1
0
0
1
0


72.00
3.3 185.00
— 141.75
— 117.50
5.1 247.75
— 158.25

0
1
2
0
5
7

20.2
44.2

0
0
2
0
0
0

1

5.3 289.75

9

31.1

1

0
0
0
0
1
0


10.50

15.00

15.50

32.50
16.3 125.25

91.00

0
0
1
0
3
5

2.1
5.3
2.2

16.3
17.5

5.4
14.1

14.1
1

Upper and Middle Socioeconomic Classes

Total
Adjusted

<5
5-14
15-24
25-34
35-54
55+

478.00 15

31.4
2^.5

2

1
0
1
0
5
8

58.8

0
0
1
0
1
0

17.00
35.75
50.00
42.00
186.75
146.50

20.0
26.8
54.6

4.2

188.25

6

20.0

6.50
20.75
34.50
9.50
61.50
55.50

1
0
0
0
2
3

5.4

31.9
*

32.5
54.1

24.0
54.9

0
0
1
0
0
0

9.5

1

64.5

3.5

64.5

Lower Socioeconomic Class
Total
Adjusted

1,588.75 15

9.4
10.7

2

<5
5-14
15-24
25-34
35-54
55+

143.75
448.00
330.50
191.75
258.50
216.25

1
3
3
1
2
5

7.0
6.7
9.1
5.2
7.7
23.1

0
1
1
0
0
0

1.3

2.2
3.0

956.25

9

9.4

1

1.0 632.50

6

82.25
278.00
204.25
106.75
136.00
149.00

1
2
2
1
0
3

12.2
7.2
9.8
9.4

0
1
0
0
0
0


61.50
3.6 170.00
— 126.25

85.00
— 122.50

67.25

0
1
1
0
2
2

* Less than 10 person-years; no rate calculated

20.1

5.9
7.9
16.3
29.7

0
0
1
0
0
0

1.6

7.9

85

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XXI
New Cases of Tuberculosis with Cumulative Probability per 100 of Children of “Sputum­
positive” Index Cases Developing Tuberculosis, according to Age at First Exposure
Williamson County, Tennessee
Age at First Exposure
Age Group in Years

Less than 1 Year

New Cases

t

<1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

36
37
38
39

8
1
1
1
1
1

1

1

1

Cumulative
Probability
per 100

1 to 4 Years

New Cases

Cumulative
Probability
per 100

5 to 14 Years

New Cases

Cumulative
Probability
per 100

15 Years and Older

New Cases

Cumulative
Probability
per 100

6.6
7.4
8.3
9.2
10.2

10.2
11.3
11.3
11.3
11.3
11.3
12.5
12.5
12.5
12.5
12.5
12.5
12.5
12.5
14.2
14.2
14.2
14.2
14.2
14.2
16.8
16.8

16.8
16.8

16.8
16.8
16.8
16.8
16.8
16.8

16.8
16.8
16.8
16.8
16.8

1
1

1
2
2
1
1

2.0
6.2
10.7
13.1
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6
15.6

15.6
15.6
15.6
15.6

1
3
2
2
1

1
1

1

0.7
0.7
1.3
1.9
3.8
3.8
5.1
6.5
7.2
7.2
7.2
7.2
7.2
7.2
8.3
8.3
9.6
9.6
9.6
9.6
9.6
9.6
11.8
11.8
11.8
11.8
11.8
11.8

1
2
1

1
1
1
1

1

1

2.9
7.6
9.6
11.4
11.4
13.2
14.8
16.4
16.4
18.1
18.1
18.1
18.1
18.1
19.7
19.7
19.7
19.7
19.7
19.7
19.7
19.7

86

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XXI1
Number of Persons Tested for the First Time with Tuberculin (0.1 mg. of OT—2+ on More
Reaction) and Number and Percentage Positive by Age Group, Sex, and Race
Williamson County, Tennessee, 1951-1955

Age Group in Years

Positive
Number
Tested

Female

Male

Total

Number | Per Cent

Positive

Number
Tested
Number

Per Cent

15.8
26.4

Positive

Number
Tested
Number

Per Cent

4,122

635

15.4
23.3

8
28
44
33

2.2
2.8
5.5
9.1
19.0
30.8
39.3
45.2
47.0

Total
*

Total
Adjusted.*

7,765

1,210

15.6
24.8

3,643

575

<5
5-9
10-14
15-19
20-29
30-39
40-49
50-59
60+

743
1,971
1,627
702
679
769
615
351
308

11
56
84
72
148
269
261
168
141

1.5
2.8
5.2
10.3
21.8
35.0
42.4
47.9
45.8

377
963
830
340
252
311
261
152
157

3
28
40
39
67
128
122
78
70

0.8
2.9
4.8
11.5
26.6
41.2
46.7
51.3
44.6

366
1,008
797
362
427
458
354
199
151

141
139
90
71

81

1

White
Total
A djusled*

6,205

901

14.5
23.2

2,938

439

14.9
24-6

3,267

462

14.1
21.7

<5
5-9
10-14
15-19
20-29
30-39
40-49
50-59
60+

559
1,571
1,319
544
560
617
492
294
249

9
41
60
51
109
199
192
133
107

1.6
2.6
4.5
9.4
19.5
32.3
39.0
45.2
43.0

285
762
682
269
215
259
217

3
23
29
27
54
98
93
62
50

1.1
3.0
4.3
10.0
25.1
37.8
42.9
49.2
40.7

274
809
637
275
345
358
275
168
126

6
18
31
24
55
101
99
71
57

2.2
2.2
4.9
8.7
15.9
28.2
36.0
42.3
45.2

19.3
33.1

855

173

20.2
29.1

92
199
160
87
82
100
79
31
25

2
10
13
9
26
40
40

2.2
5.0
8.1
10.3
31.7
40.0
50.6
61.3
56.0

126

123
Non white

Total
Adjusted*

1,560

309

19.8
31.1

705

136

<5
5-9
10-14
15-19
20-29
30-39
40-49
50-59
60+

184
400
308
158
119
152
123
57
59

2
15
24
21
39
70
69
35
34

1.1
3.8
7.8
13.3
32.8
46.1
56.1

92
201
148
71
37
52
44
26
34

0
5
11

61.4
57.6

12

13
30
29
16
20

2.5
7.4
16.9
35.1
57.7
65.9
61.5
58.8

19
14

* Population of Williamson County estimated as of July 1, 1953, used as standard population

4

87

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XXIII/1
Adjusted* Prevalence of Infection among School Children 5 to 19 Years of Age, by Age and
Race for Three Periods of Time
Williamson County, Tennessee, 1932-1955

Age Group
in Years

Number

Per
Cent

Number
Tested

Number

Per
Cent

Positive

Positive

Positive

Positive

Number
Tested

1948-1955

1940-1947

1932-1939

Total

Number
Tested

Number
Tested

Per
Number Cent

Number

Per
Cent

4,330

440

10.2
10.3

4,452

1,200

1,845
1,285

66
169
205

5.5
9.2
16.0

2,030
1,653
769

59
87
88

2.9
5.3
11.4

291

8.3
8.1

3,508

160

4.6

4.6
7.0
13.4

1,622
1,336
550

Total

*

Total
Adjusted*

11,354

5-9
10-14
15-19

4,107
4,732
2,515

9.0
9.2

2,572

210 5.1
444 9.4
364 14.5

877
1,234
461

1,018

344

13.4
12.7

85
188
71

9.7
15.2
15.4

234

5.3
6.2

White

Total
Adjusted*

8,761

645

7.4
7.8

1,736

194

11.2
11.3

3,517

5-9
10-14
15-19

3,211
3,621
1,929

144 4.5
265 7.3
236 12.2

632
792
312

56
98
40

8.9
12.4
12.8

957
1,493
1,067

44
104
143

5.5

44
63
53

2.7
4.7
9.6

I



Nonwhite
Total
Adjusted*

2,593

5-9
10-14
15-19

896
1,111
586

373 14.4

836

150

813

149

245
442
149

29
90
31

11.8
20.4
20.8

18.3

944

74

243
352
218

22
65
62

9.1
18.5
28.4

7.8
8.8

18.3

17.6

U4
66 7.4
179 16.1
128 21.8

17.9

408
317
219

15
24
35

3.7
7.6
16.0

* Using the percentage distribution of 5- to 19-year-old subjects in the Williamson County population,
1950 census, as standard

!

88

I

WILLIAMSON COUNTY TUBERCULOSIS STUDY

TABLE XXIIIB
Adjusted* Prevalence of Infection among 5- to 19-Year-Old Household Associates of Index
Cases, by Age Group and Race, for Three Periods of Time
Williamson County, Tennessee, 1932-1955
1948-1955

1940-1947

1932-1939

Total


Age Group
in Years

Positive

Positive
Number
Tested

Number

Per
Cent

Positive

Number
Tested

Per
Cent

Number

Number
Tested

Number

Positive

Per
Cent

Number
Tested

Per
Number Cent

Total

I

-i

Total
Adjusted*

635

303

47.8

495

260

52.5
50.8

86

37

43.0
47.6

54

16

29.6
27.9

5-9
10-14
15-19

236
264
135

91
133
79

38.6
50.5
58.5

175
210
110

73
111
66

41.7
52.9
60.0

40
32
14

15
13
9

37.5
40.6
64.3

21
22
11

3
9
4

14.3
40.9
36.4

*

White
*

Total
Adjusted*

485

219

45.1
46.8

364

178

49.0

71

26

36.6

50

15

30.0

5-9
10-14
15-19

179
201
105

63
94
62

35.2
46.8
59.0

125
156
83

48
79
51

38.4
50.6
61.4

35
25
11

12
7
7

34.3
28.0
63.6

19
20
11

3
8
4

15.8
40.0
36.4

Nonwhite

Total
Adjusted*

150

5-9
10-14
15-19

57
63
30

84

56.0

131

72

55.7

15

11

73.3

4

1

25.0

50
54
27

25
32
15

50.0
59.2
55.6

5
7
3

3
6
2

t
t
t

2
2
0

0
1
0

t

54.2

28
39
17

49.1
62.0
56.7

* Using the percentage distribution of 5- to 19-year-old subjects in the Williamson County population,
1950 census, as standard
t Less than 10 subjects; no percentage calculated

3

Media
10615.pdf

Position: 1209 (7 views)