Data from theNATIONAL HEALTH SURVEY

Forced Vital Capacity

of Children 6-11 YearsUnitedStates

Forced vital capacity measurements of children 6-11 years, UnitedStates, 1963-65, by chronologic age, sex, race, stature, and sittingheight are presented and discussed.

DHEW Publication No. (PHS) 78-1651

U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFAREPublic Health Service

National Center for Health StatisticsHyattsville, Md. February 1978

Series 11Number 164

—

Library of Congress Cataloging in Publication Data

Hdmill, Peter V. V.Forced vital capacity of children 6-11 years, United States.

(Vital and health statistics: Series 11, Data from the National Health Survey; no. 164)(DHEW publication; No. (PHS) 78-1651).

Bibliography: p. 191. Spirometry-United States–Statistics. 2. Children in the United States–

.hthropometry. 3. Health surveys-United States. I. Palmer, Alan, joint author. II. Drizd,Terence, joint author. III. Title. IV. Series: United States. National Center for HealthStatistics. Vital and health statistics: Series 11, Data from the National Health Survey, Datafrom the health examination survey; no. 164. V. Series: United States. Dept. of Health,Education, and Welfare. DHEW publication; no. (PHS) 78-1651.RA407.3.A347 no. 164 [RJ433.5.S6] 312’.0973sISBN 0-8406 -0102-6 [312’.6] 77-608143

NATIONAL CENTER FOR HEALTH STATISTICS

DOROTHY P. RICE, Director

ROBERT A. ISRAEL, Deputy Director

JACOB J. FELDMAN, Ph.D., Associate Direciorfor Analysis

GAIL F. FISHER, Associate Director for the Cooperative Health S tatistics System

ELIJAH L. WHITE, Associate Director for Data Systems

JAMES T. BAIRD, JR., Ph.D., Associate Director for International Statistics

ROBERT C. HUBER, Associate Director for Management

MONROE G. SIRKEN, Ph. D., Associate Director for Mathematical Statistics

PETER L. HURLEY, Associate Director for Operations

JAMES M. ROBEY, Ph.D., Associate Director for Program Development

PAUL E. LEAVERTON, Ph.D., Associate Director for Research

ALICE HAYWOOD, Information Officer

DIVISION OF HEALTH EXAMINATION STATISTICS

MICHAEL A. W. HATTWICK, M.D., Director

PETER V. V. HAMILL, M.D., M.P.H., Medical Adviser

JEAN ROBERTS, Chief Medical Statistics Branch

ROBERT S. MURPHY, Chiej Survey P&znning and Development EWznch

COOPERATION OF THE U.S. BUREAU OF THE CENSUS

Under the legislation establishing the NatiOn~ He~th Smey, the public He~th service isauthorized to use, insofar as possible, the services or facilities of other Federal, State, or privateagencies.

In accordance with specifications established by the National Center for Heafth Statistics,the Bureau of the Census, under a contractual arrangement, participated in Plmning the sumeYand collecting the data.

Vital and Health Statistics-Series 11-No. 164

DHEW Publication No. (PHS) 78-1651Library of Congress Catalog Card Number 77-608143

CONTENTS

\

Introduction ............................................................................................................................................

Method ....................................................................................................................................................Spirometry Memuement ...................................................................................................................

Results ......... .......................................................................................................................... .................Sex and Age ................................................................................................................................ .......Race ...................................................................................................................................................FVC by Stature and by Sitting Height ...............................................................................................

Discussion ...............................................................................................................................................Growth and Development ..................................................................................................................Ratio of Sitting Height to Stature ......................................................................................................Epidemiologic Implications ................................................................................................................Testing Immaturity ............................................................................................................................Data Adjustments ..............................................................................................................................Final Note .................................................................................................................................. ........

References ...............................................................................................................................................

List of Detailed Tables ............................................................................................................................

Appendix: Statistical Notes .....................................................................................................................

1.

2.

3.

4.

50

6.

7.

8.

9.

LISTOF FIGURES

Schematic drawing of a wet system spirometer ................................................................................

Forced expiatory spirogram illustrating measurements of FVC, FEV1 .. . and FEF25.75% andidentification of zero time ............................................................................................................

Mean forced vital capacity for chkfren aged 6-11 years, by age and sex: Health ExaminationSurvey, Cycle II, weighted results .................................................................................................

Mean forced vital capacity for children aged 6-11 years, by age and race: Health ExaminationSmvey, Cycle II, weighted results .................................................................................................

Mean forced vital capacity for children aged 6-11 years, by age, race, and sex: HealthExamination Survey, Cycle II, weighted results ............................................................................

Mean forced vital capacity for children aged 6-11 years, by sex or race and stature: HealthExamination Survey, Cycle II, weighted results ............................................................................

Mean forced vitaf capacity for children aged 6-11 years, by sex and race and stature: HealthExamination Survey, Cycle II, weighted restits ............................................................................

Mean forced vitaf capacity for children aged 6-11 years, by sex or race and sitting height: HealthExamination Survey, Cycle II, weighted resdts ............................................................................

hfean forced vital capacity for children aged 6-11 years, by sex and race and sitting height: HealthExamination Survey, Cycle II, weighted results ............................................................................

1

33

5556

99

1014141618

19

20

26

3

4

5

5

6

6

7

7

8

...Ill

10. Stature of children and youths aged 6-17 years, by age, race, and sex: Health ExaminationSurvey, Cycles II and ~ ... .. ... .. .. ... .. ... .. .. .... .. .. .. .. .... .. .. .. .... .. . ... .. .. .... . ... ... .. .. ... .. ... .. .. ... .. .. .. .. .... .. ... ... . 10

11. Shting height of children and youths aged 6-17 years, by age at last birthday, race, and sex:Health Examination Survey, Cycles II and III .. .. ... ... .. .. .. ... .... .. .. .. ... .. ... . .... ... .. .. .. ... .. ... .. ... .. .. .. ... ... .. . 10

12. Mean ratio of sitting height to stature of chiidren aged 6-17 years, by age, race, and sex: HealthExamination Survey, Cycles fI and III ... ... ... .. .. .. ... .. .. ... ... ... . .. ... ... . .... .. ... .. .. .. .. . .. ... ... .. ... . ... ... .. .. ... .. .. 11

13. Increase in sitting height of white children and youths aged 6-17 years, by age, race, and sex:Health Examination Survey, Cycles II and HI.. ... ... .. .. ... ... .. ... .. .. .. ... ... .. ... .. .. ... .. .. ... .. ... . .... .. .. .. ... ... .. . 12

14, Percent of examinees with unsatisfactory spirograms, by whole year for ages 6-11 years and byquarter year for ages 6-8 years: Health Examination Survey, Cycle II .. ... . .... .. ... .. . .... .. . ... .. ... . ... ... .. 15

TEXT TABLE

A. Zero-order and partial correlation coefficients for forced vital capacity (FVC) versus stature orsitting height, by race and sex: United States, 1963-65 .. .. .. ... .. ... . .. .... .. .. .. ... ... . ... .. .. .... .. .. .. .. ... .. ... .. . 12

SYMBOLS

Data not available--–--––--–----—--—-–– ---

Category not applicable----------—----—--- . . .

Quantity zero–----—------------–—------– -

Quantity more than Obut less than 0.05--–- 0.0

Figure does not meet standards ofreliabilityy or precision --------------------------- *

iv

FORCED VITAL CAPACITY OF CHILDREN 6-11 YEARS

Peter V. V. Hamill, M.D., M.P.H., Dimkiorz of Health Examination Statistics; Alan Palmer, Ph.D.,Center for Occupational Environmental Safetv and Health. Stanford Research Institute; and

Terence A. Drizd, M. S.P.H., Di~isi&

INTRODUCTION

This report presents data on forced vital ca-pacity (FVC) of children ages 6-11 years in theUnited States, by age, sex, race, stature, and sit-ting height obtained in Cycle II of the HealthExamination Survey (HES) as the most reliableof the ventilator parameters measured on a 6-liter Collins Vitalometer. This publication, oneof a series reporting the findings from the HealthExamination Survey, assesses the heaIthy growthand development of children. There is discussionand appraisal of these data considered as clinicalstandards; the discussion stresses their epidemio-logic value and their consideration as one of themany physical and physiologic parameters of thegrowing child. At the end of the report is a listof recommended adjusting factors to these re-ported values, and the critical unanswered ques-tions are identified.

The Health Examination Survey, conductedby the National Center for Health Statisticsfrom 1959 to 1970 to collect and analyzehealth-related data on the American peoplethrough direct examination of selected subjects,was a succession of separate programs (or“cycles”); each cycle lasted from 2 to 4 years.Since 1971 the Survey has been called theHeal th and Nutrition Examination Survey(HANES) because it assumed the additiomd re-sponsibility of assessing nutritional status.

Cycle I of HEfj, conducted from 1959 to

1962, obtained information on the prevalence ofcertain chronic diseases and the distribution of anumber of anthropometric and sensory charac-teristics in 18-70-year-old civilian noninstitution-alized people living in the conterminous United

of Health Examination Statistics

States. The general plan and operation of thesurvey and Cycle I have been described in twoprevious reports,l ~z and most of the results werepublished in other Series 11 reports of Vital andHealth Statistics.

Cycle II, conducted from July 1963 to De-cember 1965, involved selection and examina-tion of a probability sample of noninstitution-alized children in the United States aged 6-11years. In this program, 96 percent of the 7,417children selected for the sample were examined.The examination had two emphases. The firstconcerned factors related to healthy growth anddevelopment as determined by a physician, anurse, a dentist, and a psychologist; the secondconcerned a variety of somatic and physiologicmeasurements performed by specially trainedtechnicians. The detailed plan and operation ofCycle 11 and the response results are described inVital and Health Statistics, Series 1, Number 5.3A comparable program of examinations and datacollection for youths aged 12-17 years, CycleIII, was completed in 1970, and the plan andoperation are described in Series 1, Number 8.4Many of the findings on growth and develop-ment have also been published.

The present report is the first in a series ofHES spirometry reports that will present anddiscuss findings from the U.S. population aged6-75 years. Several, more technically specialized,methodologic reports will also be included inthis series. Although the present report considersonly FVC, subsequent reports will considermany additional spirometric parameters andtheir interrelationships among themselves andwith other physical, physiologic, and behavioralvariables that are obtained in the HES examina-

1

tions. Spirometric data using increasingly sophis-ticated spirometry equipment and techniquesand collected in cycles subsequent to these cur-rently reported Cycle II data will have been re-corded on magnetic tape.

Besides being the first in a series of HESspirometry publications on all ages, the presentreport also is part of the HES series on parame-ters of human growth and development, espe-cially the physiologic ones (namely, physician’sexamination of heart and lungs, medical and ex-ercise history, spirometric testing, chest X-ray,EKG, measurements of blood pressure, physicalexercise capacity, hand-grip strength, and hema-tocrit determinations).

Analysis of these present data was first be-gun in 1966, but an almost endless succession ofmethodologic and interpretive problems delayedcompletion for publication from 1969, as origi-nally estimated, to the present time. Besides re-quiring several major methodologic substudies,the multiplicity of problems also required shelv-ing two of the original three spirometric vari-ables that were measured—forced expiatory vol-ume at 1 second (FEV1.0 ) and forced expiat-ory flow rate between 25 percent and 75 per-cent of the FVC (FEF2 5.7s % ). They have beenshelved temporarily or permanently; only FVCis reported at this time.

The first problem encountered in 1966–andstill the most serious—was that the spirometricrecordings of many of the children containedobvious technical flaws, in most cases proceduralerrors by the subjects. A preliminary assessmentestimated that between 10 and 20 percent of thesubjects’ spirograms so obviously did not repre-sent a properly performed maximum expiatoryeffort that they would have to be discarded;those of the remainder appeared “probably ac-ceptable.” This naturally forced the followingquestion (which has been asked again andagain): “Are these data worth trying to clean upfor publication; that is, will the resultant qualityand uniqueness of the data, together with thegreat quantity and quality of other data col-lected on the subjects to which the spirometryfindings can be related, justify both the enor-mous additional effort (and possible failure) andeventual formal publication?”

The cost/benefit considerations can be sum-marized: (1) the extent and severity of the tech-

nical imperfections in the data; (2) the feasibil-ity of culling the technically flawed data andadjusting the remaining data; (3) the time, ef-fort, and money required for this effort; (4) thequality of the final product and the spinoffmethodologic side benefits; (5) the meagernessof other published spirometry data from theseage groups; (6) the probable value of population-wide estimates of “developing” ventilator func-tion as one of the parameters of physiologic andsomatic growth and development of children ob-tained by HES.

The complicated sequence of procedures bywhich the data from all of the original record-ings were technically reassessed and edited in theattempt to adjust them and to arrive at the “bestcurrent U.S. estimates of FVC values for chil-dren 6-11 years of age in the United States,1963-65” is described in detail in a companionmethodologic report.5 An adjustment to pro-duce “best-estimated values” could never becompleted because several essential adjustingfactors are still missing; instead, these factors arediscussed, listed, and summarized. Separate pub-lication in the National Center for Health Statis-tics methodologic series was chosen for theother report because of the great amount oftechnical detail (both spirometric and epidemio-logic-statistical), which would appeal to alimited audience with specialized interests andskills, and also to preserve the narrative continu-ity and rationale of the entire editing-adjustingprocess. However, every effort has been made tokeep this report (which presents the actual find-ings of FVC, stature, and sitting height) as self-contained as possible with the caveat that thepopulation estimates reported here have beenedited by deleting 1,129 subjects’ spirometric re-cordings (15.9 percent) because of technicalflaws and imputing by stepwise multiple regres-sion their values together with those of 187other subjects (2.6 percent) whose spirogramswere either lost or never obtained.

This long undertaking would have been afoolhardy one without the benefit of encourage-ment, advice, and criticism of many colleaguesover the years. Three colleagues require specialthanks: David Discher and Giles Filley for theirhelp with matters of spirometry, pulmonaryphysiology, and chest diseases; and Robert Reedwho assisted with biostatistical perspective,

2

At each of 40

METHOD

locations preselected random-ly throughout the United Slates (see appendixfor survey design), the children were brought tothe centrally located mobile center for an exami-nation which lasted about 2?42 hours. Six chil-dren were examined in the morning and six inthe afternoon. They were transported to andfrom school or home.

When the children entered the center, theiroral temperature was taken and a cursory screen-ing for acute illness was made; if illness was de-tected in a child, he was sent home and ex-amined at a later date. The examinees changedinto shorts, cotton sweat socks, and a light,sleeveless top and proceeded to different stagesof the examination, each child following a dif-ferent route. There were six different stationswhere examinations were conducted simultane-ously, and the stations were exchanged so thatby the end of 2% hours each child had under-gone essentially the same procedures by thesame examiners but in a different sequence. Atthree of these stations a pediatrician, a dentist,and a psychologist did the testing, and at theother three stations well-trained technicians per-formed a number of other examinations (chestand hand-wrist X-rays, hearing and vision tests,respiratory function tests and electrocardio-

graphy, a bicycle exercise test, a battery of bodymeasurements, and a grip-strength test).

Spirometry Measurement

Respiratory function was assessed on eachchild by means of a set of forced expiatoryspirograms (FES) obtained from a 6-liter CollinsVitalometer. Each child was instructed in thetest procedure’ by the technician, who actuallydemonstrated the maneuver.

aThe set of instructions used WZ%based on the stand-ard instruction booklet (alternativeprocedure) suppliedby the equipment manufacturer. Although the authorssuspected (retrospectively) that this closed procedure(as distinguished from the more commonly used openprocedure whereby the subject reaches maximal inspira-tion using ambient air) might have been a significantcause of underestimation of FVC due to the inhibitionof maximal inhalation in a large proportion in children,a subsequent substudy, which is described in the meth-odologic report, demonstrated otherwise. We concludethat the choice of the open or closed system has nodiscernible effect on FVC in young children.

The procedure required that the subject,standing comfortably erect, be fitted with a noseclip (considered necessary to prevent expiatoryair leakage among young children) and a dispos-able cardboard mouthpiece (with. Iips sealedtightly to prevent air leakage) that was insertedinto the rubber hose leading to the spirometer;when reasonably comfortable with this, the sub-ject was to maximally inhzde (from a state of“resting respiration”) and then forcefully expireall his air as rapidly and completely as possible.During the maneuver the attending technicianverbally exhorted the subject to blow out all hisair as hard as possible. One or two practice trialswere usually performed; and then, at least two,but seldom more than three, trials were re-corded, with special care taken to insure that thespirometer’s breathing tube was not looped orkinked. (A kinked tube could cause air turbu-lence and result in an increased airway resistancewhich, in turn, could artificially diminish flow-rate measurements. )

Prior to each day’s testing, the barometricpressure was noted and recorded. In addition,the room temperature was recorded before eachexamination. The water level in the spirometerwas carefully maintained within 1Y2 inches ofthe top of the bell cannister to prevent excesswespirometer dead space, which would cause anunderreading of the recorded signal.

The spirometer (Collins Vitzilometer) (figure1) is a water-sealed counterbalanced, lightweightcylindrical bell that is filled or emptied duringexhalation and inhalation, respectively. Its dis-

Pullw--nBead chain

fl)

Mouthpiece

Recording pen

Kymcgraph

Figure 1. Schematic drawing of a wet system spirometer

3

placements are recorded on a moving chart(kymograph). The writing device is connected tothe bell by a pulley system. The kymograph isdriven by a motor at a constant speed of 32millimeters per second.

The resultant kymographic tracing is a time-volume coordinate providing a monotonically in-creasing signal which ultimately reaches a pla-teau and should be free from inhalation artifacts(figure 2). From this waveform, calculation ofvolume and flow-rate measurements are made bycareful measurement of specified time and vol-ume increments.

The best trial of each child’s recorded set offorced expiatory spirograms was chosen to bemeasured.

The three measurements made on the besttrial were:

1. Forced vital capacity (FVC), the largestvolume measured on complete forced ex-piration after the deepest inspiration.

2. Forced expiatory volume at 1 second(FEVI.0 ), the volume of air expired dur-ing the first second of the FES.

3. Forced expiatory flow rate between 25percent and 75 percent of the FVC

(FEF2 5.7s % ), the average rate of flowduring the middle 50 percent of FES.

Measurements were made by determiningthe number of millimeters traveled by the kymo-graph pen from the baseline to the intersect of

—FEF25,5%

5-g

?3 4-g

: 3-

;FV c

2-

-/25%FVC

1-

0 1 2 3 4

TIME IN SECONDS

Figure 2. Forced expiatory spirogram illustrating measurements

of FVC, FEV1.O, and FEF 25-7570 and identification ofzero time

the volume line at defined points bydividing caliper, and converting these

use of adistances

to vol~me ;n milliliters by m~tiplying by thespirometer bell factor. For example, the FEV1.0is the distance in millimeters from the 1-secondpoint on the baseline to the intersect of the vol-ume curve (figure 2). The point from which thel-second point is measured (zero time) is deter-mined by extrapolating from the steepest part ofthe curve to a juncture of intersection on thebaseline. This technique is used to approximatethe point at which the spirogram would havebegun if the expiatory effort had been devoidof hesitation and equipment inertia. The meth-od, in fact, served to lend precision to theFEV1.0 measurement, since it controls for thosevariables that cause the initiation of a spiromet-ric signal to be slow in onset; that is, poor pa-tient cooperation or poor comprehension of thetest instructions. The FVC is measured from thebaseline to the lowest point reached by the vol-ume curve. Measurement of the FEF2s.7 5 % is

determined by the volume and time incrementbetween 25 percent and 75 percent of FVC andis expressed as milliliters per second.

To eliminate errors in conversion and com-putation, the original direct measurements inmillimeters were recorded and punched ontocards. These values then were converted auto-matically to volumes and flow rates, correctedto body temperature and fn-essure and saturatedwith water vapor (BTPS), by a programed digitalcomputer.

The findings are reported as FVC valuesfrom all of the 7,119 children examined in CycleII weighted by their individual sampling weightsto obtain the estimated national values (de-scribed in appendix). No recordings were ob-tained from 187 children; and, as described, thespirograms “of 1,129 additional children were dis-carded as technically unsatisfactory. The missing

FVC value for each of these children was im-puted by multiple stepwise regression to pre-serve the individual sampling weights for alldata. The ratiomile and method for imputing themissing data is discussed in the appendix.

As in all HES reports, age is basically definedas age attained at la& birthday (verified from acopy of the birth certificate in 95 percent ofCycle II examiners). The mean age of each cate-

gory, therefore, approximates the midpoint of

4

II the whole year; for instance, the 8-year-old male

group consists of a l-year cohort” whose meanage is 8.51 years, while the corresponding femalesample averages 8.49 years.

Race was recorded as “white,” “Negro,” and“other races. ” White children comprised 85.69percent of the total; Negro children, 13.87 per-cent; and children of other races, only 0.45 per-cent. Because so few children of “other races”are included in a national probability sample,data from them cannot be analyzed separately.These data are included, however, when “totzd”is used, but are dropped when a white/Negrodichotomy is used. The differential response rateby age, sex, and race is discussed in the appen-dix.

RESULTS

Sex and Age

Mean FVC’S by sex and by single year of ageat last birthday show a monotonic increase withchronologic age which is almost straight throughage 10.5 years, at which time both lines deflectupward (figures 3 and 4 and table 1). The twoIincs connecting the mean FVC’S are generallyparallel (bm = 206.3, bf = 202.1). The boys-

1.0

0.8 [

,b~12

AGE IN YEARS

versus-girls difference averages 146 milliliters (ml)over the 6 years. At age 6.5 years, it is 127 (ml),while at 11.5 years, it is 157 ml. Although thereis a slight unexplainable dip in the boys’ line at10.5 years, it shows a commensurately greaterincrease to 11.5 years to regain the overall paral-lelism to the girls’ line.

Race

The marked differences in mean FVC’S be-tween Negro and white children seen in figures4 and 5 and table 1 are quite consistent. As withmost HES data, when values for Negro childrenare examined separately, they fluctuate morethan those of the white children because of themuch smaller size of the sample (approximately6.2:1).

The Negro girls’ line is closer to the regres-sion line for all girls than is that of the Negroboys as compared with that for all boys. Thevalues for Negro girls also have the same upwarddeflection at 10?42years that is seen in the datafor all boys and all girls.

The Negro boys’ line, on the other hand,follows the slopes of the other lines for the first3 years but then progressively falls away so that

‘or

1.0

t

AGE IN YEARS

Figure 3, Mean forced vital c8pacity for children aged 6-11

years, by age and sex: Health Examination Survey, Cycle 11,

weighted results

Figure d. Mean forced vital capacity for chi[dren aqed 6-II

Years, by age and race: Health Examination- Survey,

Cycle 11, weighted results

5

3.0 —

2.8 —

2,6

2.4gu!k 2.22g

g 2.0G<~ 1.8

J<k 1.6>a: 1.4

0u.

1.2

1.0

_ White boys

/ . . . . . . . White girls

. . . . . . . Negroboys

--- Negrogirls

AGE IN YEARS

by age 11?4 years the mean FVC for Negro boysis only 21 ml greater than is that of the Negrogirls, whereas for the first 3 years it averages 129

ml greater.From 6 years to 9 years of age, at least, the

sex and race differences are both consistent andadditive. All males during the 6-year age span(between 6 and 12 years) average 146 ml greaterFVC than do all females; all white children dur-ing the 6-year age span average 226 ml greaterFVC than do all Negro children; all white malesaverage 263 ml greater than do Negro males andall white females average 186 ml greater than doNegro females.

In the age group 6-11 years there is a greateraverage difference when the data are grouped byrace than when grouped by sex (i.e., 226 mlbetween white and Negro children, and 146 mlbetween all boys and girls).

FVC by Stature and by Sitting Height

, When the mean FVC’S of the children areFigure 5. Mean forced vital capacity for children aged 6-11

years, by age, race, and sex: Health Examination SurveY, examined, first by stature (figures 6 and 7 andQcle 11, weighted results table 2) and then by sitting height. (figures 8 and

3.0.—

2.8 —

2.6 —

2.4 —g

g 2.2. —

g

~ 2.0

G<~ 1.6u

<h 1.6>08 1.4 —K0&

1.2 .—

1.0 —

0.6 —

—....- ............. ---....- ........

---

e0y5While

Girls

Nqro

o%%’:05 I I I I I I I i I I I

110 116 120 125 130 135 140 145 150 155 160 165

STATURE IN CENTIMETERS

Figure 6. Mean forced vital capacity for children aged 6-11 years, by sex or race and stature: Health Examination Suwey, Cycle 11,

weighted results

6

3,0

2.8

2.6

2.4

2,2

2,0

1,8

1.6

1.4

1.2

1.0

0.8

0

— white boys

. . . . . . . White girls

. . . . . . . . Negro boys

--- Negro girls

hL 1 I I I i I I I I I I I I

105 110 115 120 125 130 135 140 145 150 155 160 165

STATURE IN CENTIMETERS

Figure 7. Mean forced vital capacity for children aged 6-11 years, by sex and race and stature: Health Examination Survey, Cycle 11,waighted results

‘0r2,8

2.6 —

2.4 —i?.

Z 2.2z

~ 2.0 —3$?“5 1.8Ja=> 1,6

35z 1,43.

1.2 —

1.0 —

0,sFoG/ ;6 :s :0 :2 :4 ;6 ;e ;0 I I I I I I I I

72 74 76 78 80 82 84 66

SITTING HEIGHT IN CENTIMETERS

Figure 8. Mean forcad vital capacity for childran aged 6-11 years, by sex or race and sitting height: Health Examination Survey,Cycle 11, weighted results

7

9 and table 3), several relationships become ap-parent.

When examined by stature, the lines con-necting the mean FVC’S show a general similar-ity to those data just examined by chronologicage. However, there are some small but notabledifferences.

The most striking change is that the fourlines in figure 7 have a more consistent and abetter balanced relationship to each other thandid the four lines connecting mean FVC’S bychronologic age in figure 5. The erratic upwardhump plotted at 9% years for Negro girls hasbeen “replaced” by a smaller deflection at 132ml; but more pronounced is the “correction” ofthe unexplainable general relative downward de-flection of the Negro boys’ line seen after 8Y2years in figure 5. When grouped by increasingheight rather than by chronologic age, the line ofthe Negro boys’ mean FVC’S maintains its rela-tive position with the other three lines.

AI-I important similar note, however, be-tween figure 5 and figure 7, is the upward de-flection of the lines at the far right of the graphsfor most of the taller and older children (exceptthe Negro boys).

When sitting height replaces stature as theclassifying variable, even though the lines be-come more erratic (which will be considered inthe section, “Discussion”), there is a more dra-matic change than that seen between the figuresfor age and stature; the four lines move muchcloser together. Both girls’ lines move a littlecloser to the lines of the two groups of boys, butthe lines representing the FVC’S of white andNegro children, which signify the mean racialdifference in FVC, narrow the previously ob-served gap by approximately half. In fact, themean FVC’S of the Negro boys in figure 9 arenow essentially equal to those of the white girls(when classified by sitting height, the overallmeans for all ages are: Negro boys, 1,557 ml;white girk, 1,542 ml).

Figures 6 and 8 contrasted with figures 3and 5 summarize the changing relationship ofFVC by sex and race when first classified bychronologic age, then by stature, and finally bysitting height. The original ordinal relationshipsstill obtain (the mean FVC of white children isgreater than that of Negro children and that ofboys is greater than that of girls); but while thedifferences both by sex and by race have been

nug 1.4 —

oL1.2 1-

1.0

}

ob:6 :8 : :2 :6 ; :8 ;0 ;2 ;4 ;6 :8 :0 :2 :4 /6SITTING HEIGHT IN CENTIMETER

Figure 9. Mean forced vital capacity for children aged 6-11 years, by sex and race and sitting height: Health Examination Survey,Cycle 11,weighted results

diminished by the consideration of sitting heightdifferences, the magnitude of the difference byrace has been reduced by a considerably greatermargin.

DISCUSSION

Growth and Development

One of the most apparent and visibly dra-matic changes from early childhood to adult-hood among all normal pe~ple is the enormousincrease in stature. Increase in stature was theparameter used in one of the earliest publishedaccounts of a systematic study on humangrowth: A nobleman naturalist, PhilibertGueneau de Montbeillard, from 1759 to 1777,made serial height measurements every 6 monthson his first-born son, which were later publishedby his encyclopedist friend, DeBuffon.6 Sincethat time, stature has probably been used as aparameter of human growth as frequently as allother parameters combined because of its manyconceptual and technical advantages: it lends it-self to easily obtainable yet relatively accuratemeasurement; it is readily understood by almostall people; it is useful; and it is highly correlatedboth with the chronologic age of the growingchild and most of his other physical, physiolog-ic, and behavioral parameters. Also since deMontbeillard’s time, the biologic mechanism ofthe elongation of the growing long bones hasbecome rather well described and documented,and many of the factors controlling and influ-encing the rate of elongation have also beenstudied.

Although not as visible, nor as easily under-stood, nor as easy of accurate measurement, theincrease in FVC, nevertheless, shares many ofthe characteristics of the increase in stature dur-ing these growing years of middle childhood.

From 6-10 years of age both mean FVC andmean stature display an almost linear increasewith chronologic age, with a slight upward de-flection from 11 to 12 years, especially amonggirls. The coefficient of correlation of chrono-Iogic age from 6 to 12 years with stature is0.824 while with FVC it is 0.699. Dramatic asthe increase in stature is, the proportionate in-crease in FVC is more than three times greaterfrom 6 to 12 years: there is an average 24-

percent increase in stature but an average83.8-percent increase in FVC during these same6 years. The proportionate increase in each ofthe parameters is approximately 10 percentgreater in girls than in boys because, as a group,girls reach the adolescent growth spurt almost 2years earlier than boys do. Even though the“growing,” or increasing, FVC is as valid a pa-rarnet er of human growth, as truly characteristicof the growing child as is the increase in stature,there does not seem to be much imminent dan-ger of it replacing stature as the usual yardstickby which most people understand adolescentgrowth spurt.

The many physical, physiologic, and behav-ioral variables that systematically increase duringthe growing years in either dimension or quan-tity (and also in functional capacity) and thatcan usefully be considered as parameters of hu-man growth and development are all, more orless, highly correlated statistically with chrono-Iogic age and, hence, by association with eachother. Some of those growth changes, like skele-tal maturation and developing intellectual capac-ity, are merely concomitants, or parallel devel-opments, and are only very remotely, if at all,functionally related, while others, like the in-crease in stature and the increase in bodyweight, are very intimately related, both struc-turally and functionally. FVC is essentially ameasure of physiologic or functional capacity,but it is much more than casually related tobody size and structure. Both because it ismeasured as a volume (i.e., in milliliters, cubiccentimeters, or liters) and because its upper limitat full inspiration (i.e., total lung capacity) andlower limit at full expiration (i.e., residual vol-ume) require physiologic behavior and cannot beachieved passively, it should be considered anactive or physiologic volume. As such, it has anintimate relationship to both the size and theshape of the growing musculoskeletal thorax.

It has long been known that FVC correlatesvery higl-dy with stature; in fact, sex-specificstature represents the chief measured element inmost prediction equations in adult’s FVC. Onewould intuitive y think that FVC would be mosthighly associated with the more anatomicallyspecific part of the elongating body (the trunksegment ) rather than simply overall stature. Twoother related and well-established facts could

9

give further promise to this line of reasoning. Asstated by Damon in his review in 1966,7 theliterature for almost a hundred years has demon-strated the fact that Negro men have, on theaverage, 10 to 15 percent lower vital capacitythan have white men of similar stature and age(and this is very well borne out in our even moreaccurate HES data on children). And anotherwell-established, but perhaps lesser known factis that Negroes, of both sexes and all ages, havelonger legs and shorter trunks, proportionately,and thus, a shorter sitting height, than have cor-responding white people.8 )9 Furthermore, withthese HES data, the comparison of figure 6 withfigure 8 (i.e., FVC grouped by sitting heightrather than by stature) demonstrates that themean differences in FVC between Negro andwhite children were reduced by approximately50 percent when classified by sitting heightrather than by stature.

Ratio of Sitting Height to Stature

To explore further the relationship of FVCwith stature, we can examine the relationship of

180 —

170 —

160 —

)

#

j 150

*)

; 140—)

,130—

120—

..

— White boys

. . . . . . . White girls

. . . . . . . Negroboy$

--- Negro girls

11OL

L I I I I I I I I I I I I Io 678910111213 1415161718

AGE IN YEARS

Figure 10. Stature of children and youths aged 6-17 years, byage, race, and sex: Health Examination Sutvey, Cycles I Iand Ill

stature itself with its two most frequentlymeasured component segments, sitting heightand leg length. The relationship between sittingheight and stature and age, race, a,ncl sex in

growing children has been examined in threeprevious HES reports. s-l 0 The data from allthree reports can be brought together here andexamined in a little more detail because the rela-tionship to FVC might be so crucial and alsobecause the relationships are complicated.

Figures 10-12 and table 4 present the dataseparately for Negro and white children, andover the whole age range from 6 to 18 years togive a better perspective. In comparing figures10 and 11 it can readily be seen that there is amuch greater difference in sitting height amongthese four race-and-sex groups than there is instature.

94 —

92 —

90 —

88 —

86

84 —

z 82

)

j *O

I 78

~ 76

! 74

i 72

70 —

68 —

66 —

64 —

62 —

/“”/

/

.“” x’” — White boys

...””...;/’” . . . . . . . Whlteglrk

..””..” / . . . . . . . Negro boys

.“””..”””/’ --- Negro girls

.“”:.”;/”

..””;.;/

..””..;/

...”>

/’

AGE IN YEARS

Figure 11. Sitting height of children and youths aged 6-17 years,

by age at last birthday, race, and sex: Health ExaminationSurvey, Cycles II and Ill

10

.... ....,.....,. .....”” ... ,.....-”””.........:...”

t

— White boys

.49 . . . . . . . White girls

. . . . . . . Negrobays

---- Negrogirls

AGE IN YEARs

Figure 12. Mean ratio of sitting height to stature of children

aged 6-17 years, by aga, race, and sex: Health ExaminationSurvey, Cycles I I and I I I

To summarize the highlights of stature first:

there is a more consistent racial difference be-tween the two groups of girls over this 12-yearage span than there is between the two groups ofboys. The mean heights of the Negro girls areconsistently greater than those of white girlsfrom 7 until 14 years of age; from 14 until 17years the mean heights for the white girls areslightly greater. The relationship between thetwo groups of boys is less clear (same height at 6years of age; Negro boys slightly taller at 7-9years; white boys slightly taller at 9-12 years;same height at 12-14 years. Then, from 14 to 17years the mean heights of the white boys areslightly, but consistently, greater than those oftheir Negro counterparts). The overall meanfrom 6 to 14 years is essentially the same in thetwo groups of boys. In other words, the differ-ences, though slight, are real, but they go in bothdirections so that they cancel each other outfrom 6 to 14 years of age. However, after age 14there is enough consistency from 14-17years tosay that the white boys average very slightlytaller over the entire 12-year span.

The difference in stature by sex is a littlemore consistent and a little more clear than by

race. The boys start out slightly taller; the girlsbecome taller at 9?4 years and remain so for thenext 4 years until the boys rather dramaticallypass the girls at 14% years and increasingly wid-en the gap for the remaining 4 years.

In contrast to stature, the data on sittingheight display clear and consistent distinctionsover the entire 12-year age span between thefour age-sex groups. For the entire age span 6-18years, white boys have distinctly and markedlygreater sitting heights than Negro boys have, andthe same is true of the relationship betweenwhite girls and Negro girls (the average differ-ence of the means is 2.4 centimeters betweenthe two groups of boys, and 2.0 centimeters be-tween the two groups of girls over the 12-yearspan). In addition, there is a very marked,though not as consistent, difference between thetwo sexes. The sitting height of the boys is great-er from 6 years to 9~z or 10 years of age, andbecomes slightly less from 10 to 14 years (dueto the earlier growth spurt in girls, as very vivid-ly demonstrated for white children and youthsin figure 13); but at age 14 the boys’ sittingheight becomes almost as distinctly greater thanthe girls’ as was just seen for stature, but neverby quite as much.

Figure 12 graphs the mean sitting height/stature ratios of the four age-sex groups from6-17 years. When interpreting this figure it mustbe kept in mind that the ratios are dynamic inthat at every age and for both sexes neither thenumerator nor denominator ever decreases or re-mains constant from one period to the next;that is, from 6-17 years of age both the numera-tor and denominator at every single point areincreasing, with the result that the ratios expressthe relatively greater increase of either the nu-merator or the denominator from one time tothe next (obviously when the ratio is decreasing,as it does consistently from age 6 to 11Y2 in thegirls and until about 13% years in boys, it meansthat the legs are growing faster than the upperhalf of the body–trunk, neck, and head–butwhen the ratios start increasing again, althoughthe legs are continuing to increase in length, theupper half of the body is increasing at an evenfaster rate). This figure clearly shows how muchmore discriminating, in growth differencesamong the four race-sex groups, sitting height is

than is stature. Except for a little bit of sampling

11

4.50 —

4.25 —.

4.00 —

3.75 –

3,50 –

3.25 –,

:3,00 —

! 2.75 –>

! 2.50 –

j

i 2.25 —j

: 2.W —j

; 1.75 —

?

j 1.50:

1.25 —

1.00 —

0.75

0.50

0.251

........... .,.t:,

$.$,....

White girls t

I I I I I I I0.00 L I 1 I I

6.7 7.8 8.9 9.10 10-11 11.12 12.13 13.14 14.15 15.16 16.1

AGE IN YEARS

Figure 13. Increase in sitting height of white children and

youths aged 6-17 years, by age, race, and sex: Health

Examination Survey, Cycles II and III

“noise “ in the much smaller Negro sample, thedistinctions are extraordinarily consistent andquite marked. The sex difference, though rela-tively small, is completely consistent until 11years of age (the boys’ relative sitting height isgreater than the girls’, hence, the relative in-crease in their leg length is less), but the racialdifference from 6 through 17 years is quite con-stant, with a slightly increasing gap at the upperages. And also from 12 through 17 years of age,the differences between the two sexes becomesalmost as marked as the differences between thetwo races. From 6-11 years, the average differ-ence in sitting height/stature ratio between Ne-gro and white children being so much greaterthan the average difference of that between boysand girls immediately explains the much greaterrelative reduction of the average difference inFVC by race than by sex when the data aregrouped by sitting height rather than by stature.

Because of this distinct, and now under-standable, reduction in the mean racial differ-ences when FVC data are grouped by sittingheight, and because there is a much closer ana-tomical relationship with sitting height thanthere is with leg length, one could be tempted togo a .Iittle further and say that sitting height“correlates” better with FVC and is a better pre-dictor of FVC than is total stature. If one isexpecting that, then, after examining the coef-ficients of correlation (table A), one would bequite surprised, and possibly disappointed. One

Table A. Zero-order and ~artial correlation coefficients for forced vital cauacitv (FVC) versus stature or sittinq heiqht, bv race and sex:

-----U@ted States, 1963-65 “--- ,

Zero-order and partial correlation coefficients

Zero-order

FVC: stature . . ... ... .. .. .. ... ... .. .. ... .. .. .. ... ... .. .. ... .. .. ... ... . ... .. .. ... ... .. .. .. ... ... .. .. .. ..

FVC: sitting height .. .. ... ... .. .. ... .. .. .. .. .... .. .. ... .. .. ... . .. .... .. ... . ... ... .. .. .. .... . ... .. ..

Sitting height: stature ... . .... .. .. ... .. .. .. ... .. .. .. .... .. .. ... .. .. ... .. .. ... .. ... .. .. .. ... ... .. ..

Partiall

FVC: stature/sitting height .. .... .. .. .. .. .... ... . .. .. .. .. . ... . .... .. .. ... .. .. ... .. .. ... .. .. ... .FVC: sitting height/stature .. .... .. ... . ... .. ... .. .. ... ... . ... .. .. .. .. . .. .... .. .. . .. ... ... .. ... .

Nonimputed cases ... . ... ... .. .. .. .. ... ... .. .. ... .. .. .. ... ... ... . .. ... .. ... . ... ... ... . .. ... .. .. ... ..

lA: ~/c: A VS. B, C partiaued out.

White NegroTotal White Negro

Boys I Girls Boys I Girls

Correlation

0.806 0.8240.817 0.8170.923 0.935

0.231 0.2930.322 0.233

EE1l.EE

0.7940.8070.920

0.2240.319

778

0.8400.8280.938

0.3330.208

Number

2,611

0.831

I

0.793 0.8100.825 0.795 0.8250.934 0.907 0.930

0.301 I0.282 0.2020.246 0.294 0.335

2,414 370 I 408

12

probably expected that the correlation coeffi-cient of FVC would be higher not only by race-and sex-specific groups with sitting height thanwith stature but by an even greater magnitude inracially mixed groups, because sitting height hasa much greater race specificity than has stature.

That none of these conditions obtained wasthe occasion of much of this detailed considera-tion of the relationship of stature with its com-ponent parts (especially sitting height). The rela-tive contributions to differences in correlationsby differential relative variances among and be-tween the major body segments (with eachother and also with FVC) and, in turn, the dif-ferential contributions to these variances madeby true biologic variation versus that made bymeasurement error is discussed in detail in theappendix. There is a greater relative varianceamong the sitting heights than among the stat-ures, as measured in Cycle II, but that, in itself,begs the question of biologic variability versusmeasurement error. A definite answer is notavailable at this writing. In Cycle II the onlybody segments directly measured were statureand sitting height; but data will be availablefrom Cycle III on adolescents aged 12-17 yearsin which, in addition to stature and sittingheight, measures of cervicale height (vertical dis-tance from the floor up to the seventh cervicalvertebra, or the bottom of the neck) were alsoobtained. Cervicale height, when subtractedfrom stature, gives the height of the neck-headsegment of the body and trochanteric (lateralbony hip) height, which has a very high correla-tion with subischial height (leg length). Thesemeasures also were replicated on several occa-sions to give estimates of the technical error ofmeasurement (or that amount of variation dueto measurement error) so that when the FVCdata are also available on this age group, theanswer might become much clearer.

In the meantime, the best summary at thepresent time appears to be that although group-ing by sitting height reduced by almost half themean racial difference in FVC than when FVC isgrouped by stature (and, intuitively, seems to bea much “closix relationship”) in the presentHES data, sitting height does not correlate anybetter with FVC than stature does, by eitherrace or sex. Nor would sitting height predictFVC any better than would stature. In addition,

although the racial differences in mean sittingheight account for almost half of the observeddifferences in mean FVC, there are no additionalanatomical measurements which seem capable ofaccounting for the residual mean difference.b

Another way of formulating the authors’present quandry is to restate Damon’s questionsof 19667 and then try to answer them withthese data. In his study relating anthropometryto lung function, he states that, both from hisown findings (when matching Negro and whitesoldiers of comparable age, sex, body size, smok-ing habits, and other attributes) and also fromhis summary of pulmonary function literaturecovering 100 years (in which the relationship hasbeen found “without exception”), Negroes (chil-dren and adults) have 10 to 15 percent loweraverage FVC than comparable white personshave. As an eminently skilled anthropologist, heobtained muItiple body measurements, includinga grip-strength test (partly as an index of relativeskeletal muscle mass and athletic ability andpartly as an index of motivation), on these sol-diers and tried partial correlations searching forhis answers. He found that stature had a slightlyhigher partial correlation with FVC than sittingheight had when he was trying to account forthese differences. At the present time, HES dataconfirm the magnitude of the observed racialdifference in FVC when matched by age, sex,and stature, and although we cannot “explain”Damon’s dilemma, we can reduce his observeddifference by more than 50 percent; namely, hisobserved mean difference in FVC between thetwo groups could have been reduced by approxi-mately 50 percent if grouped by sitting heightinstead of by stature. However, the reduction islimited to the overall mean difference betweenthe two groups. When trying to apply it on anindividual level and ask, “Does sitting height cor-relate with, or predict, FVC much better thanstature?”, Damon concluded from his data andhis analysis that it did not. In the present stageof the analysis of HES data, we have to reach asimilar conclusion. In a future paper we hope tobe able to distinguish whether the slightly great-er relative variance in sitting height than in stat-

bCubing the differences in sitting height, to approxi-mate more closely a volumetric relationship, makesslmost no difference.

13

ure is an inherent biologic relationship or is pri-marily a function of the differences in technicaldifficulty in measuring body segments.

Epidemiologic Implications

This rather detailed description and analysisof FVC, as a significant and measurable physio-logic attribute of the growing child (with its sim-ilarities and differences between major age, sex,and race groupings and with its relationship tostature and to the components of stature) as-sumes greater significance when FVC is con-sidered in the context of the total human life-span. FVC, together with flow rates such asFEV1.0 and FEF25.7 ~%, are extremely impor-tant in estimating both the frequency of occur-rence and the magnitude of disability fromchronic obstructive pulmonary disease, whichhas become so prevalent among men after earlymiddle age in most of tie technologically devel-oped countries. When these population estimatesof FVC among young children are related to thismedically useful parameter in later life and thesedescriptive estimates are considered as the lead-ing edge, or the youngest chronologic ages, atwhich we can obtain population estimates onthis parameter because of behavioral testing limi-tations, and when these data are considered ascompleting the picture of the natural history ofthis parameter, the effort takes on added mean-ing.

There are two facets to the natural historyof a parameter such as FVC: (1) the descriptionand understanding of its behavior as a physio-logic variable in health and disease and (2) theconsideration of it as a measuring or testing phe-nomenon. Because of the scarcity of well-developed disease entities in this young agegroup which can significantly affect FVC (whichwould have made it valuable as a clinical case-finding tool), the chief uses, epidemiologically,for these accurate HES population estimates areto look for predisposing factors (either host orenvironmental) which might impair FVC, eithertemporarily or permanently. This kind of searchfor incipient change, either in an individual or inpopulation data, can sometimes yield valuableanswers. Some of the questions to be asked ofthe data are: Is there a critical age of occurrence

which would indicate greater age- or growth-related susceptibility? Among what specialgroups of children do any changes occur (and, ofcourse, what are the changes and what may besome of the causes)? Is there any suggestion of adose-response type of relationship? What are theeventual health consequences of these physio-logic changes? Can anything be done to preventthe changes from occurring?

Population data such as these can be usedepidemiologically in another way, too. Impair-ment in the rate of growth of children, espe-cially in stature, is being used increasingly as oneof the most sensitive indicators of adverse envi-ronmental factors, especially malnutrition. 1 I -18By analogy, it is probable that alterations in theexpected rate of growth of spirometric param-eters in children might also be more sensitiveindicators of environmental stress (such as airpollution) than is alteration in adult pulmonaryfunction values. The Environmental ProtectionAgency, through its air pollution program, theCommunity Health and Environmental Surveil-lance System, is studying this premise by usingchildren almost as miners used canary birds inthe old days to detect poisonous gases. They willfolIow panels of children over time, in a varietyof geographic sites, to correlate their changes inpulmonary function with data from its air-sampling network.

Testing Immaturity

Spirometric parameters do not arise sudden-ly and fully developed at about the age of 6. Infact, many of the ventilator movements havebeen present since the last few months in utero.But because the testing of most of the common-ly used spirometric parameters requires a highdegree of cooperation and understanding on thesubject’s part, there are severe age limitationswhen considered as a populationwide testingphenomenon. A fairly reasonable FES is probablyobtainable as early as age 2?/2to 3 years on a veryfew precociously bright, imitative, and co-operative children who also have an unusuallyhigh degree of control over their voluntaryphysiologic functions. In fact, Luft, at the Love-lace Clinic in Albuquerque, is getting good FESdata on children as young as 3 years of age and

14

“most all of them by 5 or 6 years of age”l 4 (thequality validated by comparable results from gasmixing and perfusion tests on the same sub-jects). However, these results are achieved in anoptimal clinic setting: Luft has enormous exper-ience, skill, and patience in spirometry with chil-dren; his staff has been with him for some 10 or15 years and is similarly skilled; and there are notime pressures–the children can return if theybecome confused, tired, irritable, or just don’tquite get the proper knack of it. In addition,they are self-selected by word-of-mouth contactsof friendly, interested, and cooperative families.

We will see, as this discussion of FVC con-siders “the natural history of FVC as a testingphenomenon,” that when attempting to get to-tal population estimates (which requires maxi-mal participation by as many people as possiblewithout bias among such major classifying vari-ables as age, race, sex, socioeconomic level, intel-ligence, environmental region, schooling, etc.),there are distinct lower age limitations to accur-ate and reliable data. On a total population level,and in a survey setting, the testing problems areso numerous that at age 6 years, and probablyeven at 7 years of age, too many children aresimply too immature behaviorally and sociallyto produce reliable and unbiased test data–eventhough many do produce good data. The verylengthy methodologic companion report5 tothis one describes several substudies and an enor-mous amount of data, argument, and analysisexamining this point, which had become promi-nent when attempting to clarify and, if possible,to “adjust or correct” some of the inconsisten-cies found several years ago during analysis ofthese data. “Spirometric immaturity of a largeproportion of 6- and 7-year-olds and an ever-decreasing proportion of 8-, 9-, 10- and 1 l-year-olds, in a large survey setting such as HES, ” isone of the chief conclusions of that long andtortuous methodologic analysis. Figure 14 illus-trates the age relationship to the proportion ofchildren whose spirograms had. to be discardedbecause of gross, observable errors. The almostasymptotic rise in those younger than 6Y2 years(compared with the llfi-year-olds) is readilyseen.

The major substudy, which is the heart ofthe methodologic report,5 was performed in Salt

5

Lo w,

AGE IN YEARS

50

45

40

35

I30

25

20

15

10

1

L

5

0 ~2

AGE IN YEARS

Figure 14. Percent of examinees with unsatisfactory spirograms,

by whole yaar for ages 6-11 years and by quarter year for

ages 6-8 years: Health Examination Suwey, Cycle I I

Lake City in May 1974, on a very select subpop-ulation under optimal testing conditions for chil-dren, which probably cannot be repeated andalmost certainly cannot be appIied in all regionsof the country and to all segments of the popu-lation. As described in detail in the report, it wasperformed by an unusually experienced team,using very efficient and sensitive equipment on aself-selected population of highly motivated 6-and 7-year-old children almost totally from up-per middle class , managerial, and professionalfamilies. The whole study was completed in afew short days before the enormous enthusiasmthat had been generated could dissipate. TheseSalt Lake City data are used in this report and inthe methodologic one as if they are optimalphysiologic estimates, free of contaminationfrom testing inadequacies. Even with these opti-.mal testing circumstances, several of the subjectswere finally judged to be too immature to per-form properly a forced expiatory spirogram;their test results had to be discarded.

In all spirometry testing administered in asurvey setting to broad populations of people ofall ages there are always some people who areeither incapable of or unwilling to perform aproper spirogram. Their data must be recognizedand flagged so that it may be either adjusted or

15

discarded. This general phenomenon is also cer-tainly true in this young age group, but the chiefepidemiologic problem encountered here is theadditional one of immaturity—not physiologicimmaturity but immaturity as a test subject.This testing immaturity is a behavioral-socialphenomenon like a lack of “school readiness”;that is, the subjects would be unable to followproperly instructions like “Please sit down andbe quiet, “ “Raise your hand when you want tospeak, “ “Pick up your pencil and copy the pic-ture and the words in your book,” all of whichrequire understanding and willingness andenough self-control to perform properly and ef-fectively. In the absence of adverse factors, thisimmaturity is “outgrown” in a year or two inmost children. We know that proficient testing,skillfully used motivational techniques, andplenty of patience can surmount the immaturityin some proportion of the youngsters (as isdone by skillful kindergarten and first gradeteachers). Apparently, this was accomplished inthe Salt Lake City and Annapolis substudies5performed by HES and, as has been recentlybeen reported,l G by another group of investiga-tors working independently, also in Salt LakeCity.

In a substudy of 222 children reported bythe pulmonary division of the Latter-day SaintsHospital (LDS), Salt Lake City, the find-ingsl 5 ~I G generally reinforce all HES findings,but the data are not sufficiently detailed to sup-ply essential missing information. Specifically, asnoted above, both they and the investigators inthe HES Annapolis substudy achieved an ap-proximate 10-percent improvement in FVC bya spirited group demonstration and instructiongiven prior to individual instructions. (The LDSgroup had to discard the data of 10 subjects astechnically inadequate.)

Data Adjustments

The methodologic companion piece containsdetailed rationale and discussion of the recom-mended adjustments. The resolution of many ofthe problems discussed in this report is beingimplemented in a current cycle of national datacollection (19 76 through 1980) on these sameage groups: many refinements in testing proce-

dures, hardware, and data handling have beeninstituted by HES over the past 8 years. Whenavailable, these new data willthe validity, or lack thereof,recommended adjustments:

. Processing and refinement

help substantiateof the following

of original dataset

1.

2.

No observed HES values have been ad-justed or changed; however, the impu-tation of FVC values for 18 percentof the 7,119 examined children to ob-tain weighted national population esti-mates slightly changed the overall meandata and the percentile distributions inthe tables (as discussed in the methodo-Iogic report). The imputation procedure(described in detail in the appendix) em-ployed multiple stepwise regression usinga wealth of other variables. If it is as-sumed that most of the correlated attn-but es (sitting height, stature, chestbreadth, race, sex, etc.) would be moreclosely associated with the true physio-logic FVC capacity than with the testingbehavior of the subject, then it wouldfollow that the imputed values are rea-sonably fair and unbiased physiologicestimates.

Except for reclassifying by quality (i.e.,by c~lling and regroup~ng- and then- im-

puting), no other modifications of theoriginal data set have been made. At onetime it had been fully intended to adjustthese observed HES values and appendthe adjusted values as the final section ofthis publication. However, repeated at-tempts failedc to obtain other popula-tion data on children to supplementthose from the two methodologic studieswhich were undertaken (i.e., Salt Lake

cThe only two exceptions have already been men-tioned: (1) Dr. Luft, Lovelace Clinic, Albuquerque, NewMexico, who was very generous with advice and generalconclusions, but, unfortunately, whose data have neverbeen systematically analyzed and (2) Dr. C. DuWayneSchmidt, Chief of Pulmonar Division, LDS Hospital,

1 ~6 and personal commun-Salt Lake City (published 5~icationsl 7,.

16

City, 1974, and Annapolis, 1976); thisambitious project had to be abandonedbecause several pieces of key informa-tion are still missing.

. Recommended adjusting factors for clinical orepidemiologic comparisons: The follow-ing adjusting factors have been explained ex-tensively both in this report and in the meth-odologic one, but they have not actuallybeen applied to these data because they areincomplete.

10

2.

The sex difference in median FVC re-ported in these HES data (table 1) is 9.8percent at 6..5 years. of age. AlthoughSchmidt 17 observed even greater differ-ences than these, he and we believe thatthese magnitudes overestimate the truephysiologic difference and that someportion of the difference is due to atti-tudes and expectations on the part ofboth subjects and examiner. If this test-ing difference is overcome, we believethe residual physiologic difference willrange from 3 to 6 percent at age 6.5years.

Because almost all errors in performanceof the FES result in underestimates andthe effects are additive, it follows thatalmost all imperfect data are underesti-mates to varying degrees and that opti-mal data will be generally “maximal”data. The HES Salt Lake City substudydata are maximal (and optimal) withspecial attention to the problems of testinstructions as they relate to testing im-maturity, so a general adjustment couldbe made for the most immature children(i.e., the 6- and 7-year-olds). No compar-able data are available for the older agegroups, but it is assumed that among the1 l-year-olds, the special age-relatedproblem of immaturity (although therewill always be a residual group of inade-quate subjects in any population of peo-ple, as described earlier) is no longeroperative so that, as the general popula-tion of children “outgrows “ its testingimmaturity (i.e., testing maturity wouldgrow in a manner similar to the increase

3.

4.

of FVC and stature and other “growthvariables” in young children), the impactof its prevalence and severity decreaseswith age. Therefore, the mean FVCshould be increased by approximately20-25 percent at ages 6 and 7 years andby somewhere between 5 and 20 percentat ages 10 and 11 years (ages 8 and 9 tobe interpolated linearly or curvilin-early).d

The shape of the distributions willchange slightly when testing conditionsmore closely approach the optimal; thatis, although the whole distributionwould move up, by approximately 10-20percent, the lower percentiles wouldmove up proportionately more, thussomewhat narrowing the distance be-tween the 5th and 95th percentile as re-ported in table 1.

The mean difference can be reduced byabout 50 percent by standardizing forthe racial differences in sitting height;8 Ygpossibly a greater mean reduction wouldbe obtained if thoracic height (i.e., fromsitting position at table top up to theseventh cervical vertebra) can be mar-tialled out. Can the residual differencebe reduced somewhat more by over-coming attitude/expectation differences(as is predicted for girls in footnote d)?Because of the similarity of the esti-mated magnitude of difference withmuch more varied and extensive adultdata, it is predicted that, even in an idealtesting situation, there will remain a re-sidual mean FVC black-white difference,by stature, of 5-10 percent.

Five residual questions. –The following ques-tions still remain:

1. Although classifying FVC by sittingheight rather than by stature is intuitive-ly more appealing and reduces the ob-

‘By integrating the sex difference into the adjust-ments, it follows that HES girls’ FVC is more under-estimated than the boys’; therefore, whatever upwardadjustments axe made, the girls’ data would require rela-tively greater accommodation.

17

2.

3.

4.

served mean differences by sex and race(than when classified by stature orchronologic age) by as much as 50 per-cent, why doesn’t FVC correlate betterwith sitting height than with stature? Isthe slightly greater variance in HESCycle II sitting-height data than that inthe stature data primarily due to greaterbiological variability (especially becausesitting height includes both thoracicheight and head-neck height) or is it dueto greater measurement error? (Becauseof more detailed and better replicatedanthropometric data this probably canbe answered using data from HES CycleIII.)

What is the earliest age at which a reli-able FES can be performed in a surveysetting by enough children to get validpopulation estimates by race, sex, socio-economic, culturzd, and regional classifi-cations? There are many interesting sub-sidiary questions that could be studiedby behaviorists; for example, What fac-tors comprise testing immaturity (howsimilar is this phenomenon to school“readiness”) and how are they best over-come? How much extra skill, time, pa-tience, sensitive equipment, and sophisti-cated milieu are required to obtain rea-sonably good population estimates onthe youngest possible age group? (a prac-tical cost/benefit survey question).

Looking at FVC as a physiologic param-eter of the growing child, and assumingbetter testing conditions and qualitycontrol than obtained in HES Cycle II,what regression line would most accur-ately describe FVC and chronologic ageand/or stature (or sitting height)?

What are the true physiologic differencesin FVC between boys and giris-and be-tween Negro and white children?e

5. What are the best estimated test differ-ences obtained under optimal condi-tions? And what factors contribute tothese differences?c

In other words, how much of the5-20-percent differences in these population sub-grouping (i.e., sex and race), which have beenalmost universally found in the past, are due totrue physiologic capacity and how much are dueto attitudes and expectations on the part ofboth the examinee and the subject?

Final Note

Because of the technical limitations of thesespirometry data in children, they are not recomm-ended to be used as clinical standards. In mostlaboratories, with skill and patience and time,and opportunity to have the subject returnanother day (when necessary) researchers canexpect to get mean FVC values approximately150 ml higher than these HES values among6-8-year-olds and perhaps 100 ml higher among9-1 I-year-oIds.

These data have been presented as popula-tion distributions with subgroup comparisonsfor their epidemiologic value.

This is epidemiologicaUy useful both as ameasurable physiologic parameter of the grow-ing child and by filling in as the youngest ages inthe natural history of a parameter that assumesclinical importance in adults, especially for thedetection and measurement of chronic obstruc-tive lung disease. Although the absolute valuesaxe underestimates by approximately 10-20 per-cent, it is considered that the comparative valuesby age, race, and sex are reasonably accurate(although the observed sex difference requiresmore study), and that the population distribu-tion of values by percentiles has some merit as ameasure of population variabilityy, especially be-cause of their pioneering nature.

‘Questions 4 and 5 are related and are perhaps themost interesting because of their extraphysiologic im-plications..

18

REFERENCES

1National Center for Health Statistics: Plan and initialprogram of the Health Examination Survey. Vital andHealth Statistics. PHS Pub. No. 1000-Series l-No. 4.Public Health Service. Washington. U.S. GovernmentPrinting Office, July 1965.

~National Center for Health Statistics: Cycle I ofthe Health Examination Survey: Sample and response,United States, 1960-1962. Vital and Health Statistics.PHS Pub. No. 1000-Series 1l-No. 1. Public HealthService. Washington. U.S. Government Printing Office,Apr. 1964.

3National Center for Health Statistics: Plan, opera-tion, and response results of a program of children’sexaminations. Vital and Health Statistics. PHS Pub.No. 1000-Series l-No. 5. Public Health Service. Washing-ton. U.S. Government Printing Office, Oct. 1967.

4National Center for Health Statistics: Plan and oper-ation of a Health Examination Survey of U.S. youths12-17 years of age. Vital and Health Statistics. PHSPub. No. 1000-Series l-No. 8. Public Health Service.Washington. U.S. Government Printing Office, Sept.1969.

5National Center for Health Statistics: Methodo-logic problems in children’s spirometry. Vital and HealthStatistics. Series 2-No. 72. DHEW Pub. No. (PHS)78-1346. Public Health Service. Washington. U.S. Gov-ernment Printing Office, Nov. 1977.

60euvres completes DeBuffon: Histoire NaturelleDe L’Homme. Brussels. Th. LeJeune, 1829. pp. 17-20.

7Damon, A.: Negro-white differences in p~mon~y

function, vital capacity, timed vital capacity, and expiat-ory flow rate. Hum. Biol. 38:380-393, 1966.

8National Center for Health Statistics: Selected bodymeasurements of children 6-11 years, United States.Vital and Health Statistics. Series 1l-No. 123. DHEWPub. No. (HSM) 73-1605. Health Services and MentalHealth Administration. Washington. U.S. GovernmentPrinting Office, Jan. 1973.

‘National Center for Health Statistics: Body weight,stature, and sitting height: white and Negro youths 12-17 years, United States. Vital and Health Statistics.Series 11-No. 126. DHEW Pub. No. (HRA) 74-1608.Health Resources Administration. Washington. U.S.Government Printing Office, Aug. 1973.

1‘National Center for Health Statistics: Body dimen-sions and proportions, white and Negro children 6-11years, United States. Vital and Health Statistics. Series1 l-No. 143. DHEW Pub. No. (HRA) 75-1625. HealthResources Administration. Washington. U.S. Gover-nmentPrinting Office, Dec. 1974.

11National Center for Health Statistics: NCHS growthcharts, 1976, by P. V. V. Hamill et al. Monthly VitalStatktics Report, Vol. 25, No. 3, Supp. (HRA) 76-1120,Public Health Service. Washington, D.C.

12Hamill, P. V. V. et al.: Physical growth: NCHSpercentiles. Am. J. Clin. Nutr. To be published.

13National Center for Health Statistics: NCHS growthcurves for children birth-1 8 years, United States. Vitaland Health Statistics. Series 11-No. 165. DHEW Pub. .No. (PHS) 78-1650. Public Health Service. Washington.U.S. Government Printing Office, Nov. 1977.

14Luft, U.C.: Personal communication.

] 5Dickman, M. L., Schmidt, C. D., and Gardner, R.M.: Spirometric standards for normal children andadolescents (ages 5 through 18 years). Am. Rev. Respir.Dis. 104:680-687, 1971.

16 Brough, F. K., Schmidt, C. D., and Dickman, M.L.:Effect of two instructional procedures on the per-formance of the spirometry test in children<ive throughseven years of age. Am Rev. Respir. Dis. 106: 604-606,1972.

17 Schmidt, C. D.: Personal communications.

18Nation~ Center for Health Statistics: Replication:an approach to the analysis of data from complexsurveys. Vital and Health Statistics. PHS Pub. No.1000-Series 2-No. 14. Public Health Service. Washington.U.S. Government Printing Office, Apr. 1966.

19

LIST OF DETAILED TABLES

1. Forced vital capacity of children 6-11 years, by age, sex, and race–sample sizas, standard deviation, standard error of themean, maan, and salacted percentiles: United States, 1963-65 .. .... .. .. .. ... .. .. ... ... .. ... .. .. ... .. .. ... ... ... .. . .... .. .. .. .. ..... .. .. .. .. ... .. ... ... ... . ... ... 21

2. Mean forcad vital capacity of children 6-11 yaars, by sex, race, and stature-sample siza: United States, 1963-65 . ... .. .. ... ... ... .. 23

3. Maan forced vital capacity of children 6-11 years, by sex, race, and sitting height–sample size: United States, 1963 -65 . .. .. ... ... 23

4. Stature, sitting height, and sitting height/stature ratio for children 6-11 years, by race, sex, and age-sampla siza, mean, andstandard deviation: United States, 1963-65 ... ... .. .. .... .. .. .. ... .. ... .. .. ... ... .. .. .. .. ... ... . .... .. .. .. .. .. ... ... .. .. ... .. .. .. .... .. .. . ... .. .... .. .. .. .... .. ... .....O.. 24

20

Table 1. Forced vital capacity of children 6-11 years, by age, sex, and race–sample sizes. standard deviation, standard error of the mean,mean, and selected percentiles: Uni@d States, 1963-65

I 1 IPercentile

SD SE~ x5th

I10th 25th I 50th 75th 90th 95th

Age, sex, and race

Totall

6-11 years .........

6 years ..............................7 years .m......m..o..o.o.o...........8 years ..............................9 years ..............................10 years ............................11 years ............................

BOYS

Nn

7,119

Milliliters

2.035 2,391

1,5841,7952,0222,3012,5152,803

2,464

1,6281,8402,0822,3812,5682,894

2,298

2,595

1,7111,9012,1382,4352,6412,992

2,653

23,784 488

279286307354379430

491

11.3 1,726 1,009 1,127

9071,0741,2281,3851,5431,745

1,202

1,371

1,0481,2281,4261,5881,7421,982

1,441

1,684

1,2321,4241,6211,8121,9882,254

1,758

-

4,0984,0843,9863,9573,8673,792

12,081

12.411.911.015.619.116.3

14.5

1,2401,4291,6281,8292,0122,279

1,798

807974

1,1201,2941,4041,613

1,068

1,4121,6241,8292,0492,2662,552

2,113

1,1111,2411,2311,1841,1601,192

3,632

575632618603576628

3,487

536609613581584564

6,100

6-11 years .........

6 years 0.............................7 yaars ..............................8 years ...............0..............9 years ..............................10 years .................... ........11 years ............................

Girls—,

6-11 years .........

6 years ..............................7 years ..............................8 yaars ..............................9 years .... ..........................10 years .. ..........................11 years ............................

White

6-11 years .........

6 years ..............................7 years ..............................8 years ..............................9 years ..............................10 years ............................11 years ............................

1,3131,4941,7011,8982,0662,337

1,611

1,4741,6891,9082,1362,3372,619

1,948

1,7451,9352,1942,5552,7083,078

2,505

2,0822,0742,0262,0121,9631,924

11,703

2,0162,0101,9601,9451,9041,868

20,403

3,5093,4973,4133,3933,3243,267

3,272

570570560534530507

271287299367377433

474

272267295322368412

488

272283301351368421

440

276264283319362391

14.316.513.825.922.322.9

9.8

15.610.614.715.721.713.9

12.0

13.613.612.116.218.617.6

11.0

17.520.221.918.433.729.5

1,3021,4981,7031,9082,0802,356

1,652

1,1751,3571,5501,7471,9422,199

1#757

1,2611,4541,6571,8572,0482,320

1,531

1,0981,2731,4481,6561#7792,010

8781,0581,2181,3401,4751,709

961

9551,1431,3301,4621,6141,846

1,078

1,1121,3001,5051,6641,8342,056

1,302

8561,0231,1581,336I ,507I ,681

1,155

9331,1041,2731,4241,5851,792

988

747941

1,0921,303I ,3071,523

1,0021,1731,3401,5341,6901,898

1,401

1,1591,3471,5531,7341,9272,179

1,712

1,3401,5211,7541,9462,1902,466

2,072

1,4951,7071,8982,1612,4272,717

2,419

1,6411,8172,0442,2882,5672,901

2,618

744917

1,0781,2171,3651,548

1,039

8621,0181,1581,3161,4701,673

860

642876

1,0281,1581,1461,348

1,5931,8122,0392,3232,5302,646

2,093

1,7201,9162,1612,4582,6553,028

2,280

9501,0631,0351,0191,0141,019

987

1,0731,2591,4651,6241,7932,025

1,207

9051,0901,2421,4301,5711,785

1,2541,4471,6471,8332,0352,301

1,505

1,4251,6511,8512,0792,2982,585

1,822

Negro

6-11 years .........

6 years .... ..........................7 years ... .... .. .. .. ... .. ... .. .. .... .8 years ..............................9 years ..............................10 years ............................11 years ............................

*InAsdes other races.

1,2691,4571,6271,8661,9442,236

1,4361,6211,8332,0652,2012,523

1,5881,7181,9562,1832,4032,681

156172192158142167

1,1001,2481,4381,6381,7801,989

21

Table 1. Forced vital capacity of children 6-11 yaars, by age, sex, and race–sample sizes, standard deviation, standard error of the mean,mean, and selected percentiles: United States. 1963-65—Con.

Age, sex, and race

White boys

6-11 years .........

6years ..............................7 yeare ..............................8 years ..............................9years ..............................10 years ............................11 years ............................

White girls

6-11 years .........

6 years ..............................7 years ..............................8 years ..............................9 years i, ............................10 years ............................11 years ............................

Negro boys

6-11 years .........

6 years ..............................7 years ..............................8 years .................... ..........9 years ..............................10 years ............................11 years ............................

Negro girls

6-11 years .........

6 years ..............................7 years ..............................8 years ..............................9 years ..............................10 years ....... .....................11 years ............................

-

N

10,391

1,7871,7811,739 I1,7301,6921,662

10,012

1,7221,7161,6741,6631,6321,605

1,642

289286279269264255

1,629

n

3,153

489551537 I525 I509542

2,947

461512498494505477

464

847979746583

523

7293

113847784

PercentileSD SEX j?

5th 10th 25th 50thI

75thI

90thI

95th

Milliliters

1,793

1,3291,5191,7361,9272,1192,388

1,637

2,155 2,496 2,690492

263282296358364419

472

15.6

17.219.214.326.921.624.1

9.6

1,833 1,104 1,244 1,475

1,1321,3321,5361,7001,8742,119

1,330

1,0251,2021,4031,5561,7161,845

1,279

9731,1241,3471,4301,6541,805

1,163

9121,0881,2631,3851,5191,767

992

790967

1,1041,2361,4141,615

924

841921

1,0751,1581,1301,483

828

642860982

1,0991,2211,323

1,6401,8582,1052,3972,5802,930

2,328

1,5101,7331,9102,1692,4482,738

2,110

1,7641,9462,2162,5862,7243,119

2,527

1,6401,8462,0592,2902,5752,915

2,237

1,6171,7912,0292,1462,3462,701

2,348

1,3261,5241,7321,9472,1222,406

1,678

9941,1821,3591,5181,6461,903

1,101

9001,0581,1891,3561,5431,716

1,053

8031,0061,2141,3091,3461,573

939

1,4971,7071,9322,1702,3732,645

1,974

1,1781,3771,5771,7471,9542,206

1,559

1,1601,3211,5271,6321,7991,991

1,442

1,3541,5401,7681,9532,2122,487

1,845

266265286317356404

419

275262265309347366

456

16,212,117.415,621.815.9

15.2

1,1941,3621,5791,7631,9722,230

1,570

1,3671,5191,7041,6291,9482,219

1,799

1,1971,3701,5551,9171,9402,282

1,4581,6851,8902,0632,1872,492

2,080

1,3441,5341,7712,0682,2132,550

38.624.934.435.250.544.8

15.4

1,1531,3311,5281,6511,8122,021

1,492

1,0401,2151,3701,6611,7462,000

1,4821,6071,9132,2432,4042,667

726912

1,0661,2861,2861,473

8361,0411,1651,4291,4911,760

1,0211,2021,3181,6481,7501,984

281284281265266253

266232278328373415

24.426.822.326.839.037.7

NOTE: n=s&mple size; N=estimated number ofchildren inpopulation inthousands; SD=standard deviation; SE~= standarderror of the mean; X = mean. . .

22

Table2. Mean forced vital cap

Sex and race

lIndudes other races.2Smnplcsizelessthm 14.

NOTE:rI ‘sllIllP!e SiZ?.

Table3. Mean

~

White ................... .... .... ........... ... 6,100Negro ........ .. ....... ....................... . 987

White bnys ............... .............. .. .. 3,153Whito girls ........... .. ... ...... ............ 2,947

Negro btws ........ .................... ... .. 454Negro girls............. ... ................. . 523

ty.afchildren 6-11 years, bysex, race, ands?ature–sample size: United States, 1963.85

Stature in centimeters

n1o5- 11o- 115- 120- 125- 130- 135. 140- 145. 150- 155- 160-110 115 120 125 130 135 140 145 150 155 160 165

7,119=

3,6323,487

6,100987

3,1532,949

484523

1,026

1,148956

1,020(1)

1,137953

(1)(1)I

1,106 1,230

1,147 1,2811,088 1,183

1,124 1,256994 1,068

1,167 1,3141,082 1,204

1,024 1,090967 1,045

Mean forced vital capacity in milliliters

1,361

1,4331,326

1,4071,207

1,4551,355

1,2651.156

1,549

1,6181,474

1,5681,318

1,6511,514

1.3611,266

T1,711 1,885

1,762 1,9611,633 1,796

1,745 1,9121,516 1,656

1,825 1,9921,659 1,818

*

2,090

2,1801,985

2,1311,827

2,2172,026

1,9061,747

2,247

2,3462,151

2,2951,970

2,3912,201

2,0721,665

2,470

2,6142,323

2,5222,164

2,6592,375

2,2772,082

2,714-

2,6672,594

2,7832,357

2,9312,662

F)2,336

2.627-

[2)2,544

2,805[2)

(2)(2)

F)(2)

ct!dvit”l capacity ofchildren 6-11 years, by sex, race, andsitting height–sample size: United States, 1963.85

Sitting height in centimeters

56-56 56-60 60.82 62-64 64.86 66-88 66-70 70-72 72-74 74-78 76-78 76-80 80.82 62-84 94-66

Mean forced vital capacity in milliliters

77942 976 1,110

(1) 1,000 1,157977 964 1,077 7

1,204 1,318

1,235 1,3681,183 1,265

911 1,023 1,108 1,226 1,332(1) 909 1,118 1,125 1,249

(1) 1,112 1,158 1,284 1,380(1) 975 1,072 1,200 1,280

;:; 864 1,156 1,144 1,302949 1,0901,109 1,203

1,483 1,638 1,803 1,979 2,1601,426 1,566 1,702 1,919 1,994

1,531 1,692 1,867 2,056 2,2361,431 1,582 1,727 1,677 2,070

1,479 1,623 1,752 1,958 2,0731,365 1,521 1,656 1,872 1,920

2,307

2,4032,195

2,3192,129

2,4102,209

2,2232,064 T

2,509 2,673

2,651 2,8282,372 2,559

2,525 2,7202,875 {2)

2,664 2,8562,374 2,606

(2) (%2,355 (2)

2,745 I 2,861

,2) (2)2,666 2,752

F) (’42,704 2.762

llncludes other races.2Sample size lessthLwI 14.

NOTE: II = sample size.

23

Table4. Steture, sitting height, and sitting height/stature ratio for children 6-11 years, by race, sex, and age–sampla size, mean, andstandard deviation: United States, 1963-65

Race, sex, and age

White boys

6 years .. ... .. .. ... ... .. ... .. .. ... .. .. .. .... .. ... .. .. ... .. ... ... .. .. ... . .... .. ... .. .. .. ... ... . .... .. .. .. .... .. .. .. .. ..7 years .. ... .. .... . ... .. ... .. .. ... ... .. . .... .. ... .. .. ... ... .. .. ... ... .. .. ... .. ... .. .. ... ... .. ... .. . .... .. .. ... ... .. ...8 years ... ... .. ... .. ... .. ... .. ... . .... .. .. ... . ... .. .... . .... . .... . ... .. ... ... . .... . ... .. ... .. ... .. .. ... ... .. .. ... ... ..9 years . .... . ... .. ... ... .. ... .. .. .. .... .. .. .. .... .. .. .. .. ... ... .. .. .... . ... ...o... ... .. .. ... .. ... .. .... . .. ... .... . .. ..10 years ... .. ... .. .. ... ... .. .. ... .. ... . ... .. ... ... .. ... ... .. .. .. .. .. .. .. ... . ... .. ... .. .... . .... . .. ... ... .. .. .... .. .. .11 years .... .. ... .. ... .. ... . ... ... . .... . .... .. .. ... .. ... .. .. ... .. ... .. .. ... ... .. .. .... . ... .. .. ... ... .. .. ... .. ... .. ...

White girls

6 years .. .. .. .. .. .... . ... ... . .... . ... .. .. ... .. ... .. ... .. ... .. ... .. ... .. .. .. . ... .. .. .. ... ... .. .... . .. ... . .... . ... ... ...7 years . .... . ... .. .. ... .. ... .. ... .. ... .. ... .. .. .. ... ... .. ... .. ... .. ... .. ... .. .. ... .. ... . .... .. ... .. ... .. .. .. .. ... ... ..8 years .... . . ... ... .. .. ... ... .. .. .. .... . ... .. .... . ... .. .. .... .. ... . ... .. ... .. .. ... ... .. .. ... .. ... .. ... . .... .. ... . ... ..9 years .. ... ... .. .. ... . ... ... .. ... .. ... . .... . ... .. .. .. .. .. ... ... .. . .... . .... .. ... .. .. ... .. ... ... .. .. .. .. .... .. .. .. ... .10 years .. .. .. ... .. ... . ... ... .. .. .. .. . .... . .. ... .. ... ... .. ... .. .. ... .. ... .. ... ... .. .. .. ... . .... .. ... .. .. ... .. ... .. ...

11 years ... . .. .. ... .. ... .... . .. . . .... .. ... .. .. ... .. ... . .... . .... .. ... .. ... . ... ... .. ... .. .. ... .. .. ... ... . .. .. .. .. . ....

Negro boys

6 years .. ... . .... ... .. .. ... .. .. .. ... .. ... .. .. .... . ... ... . ... .. ... .. ... . ..... . .... .. .. .. ... ... ... .. .. ... .. ... . .... .. ...7 years .. ... .. ... .. .. ... ... .. ... .. .. ... . .... . .... . ... ... ... .. .. .. ... .. . .. ... .. ... .. ... ... ... . ... .. ... ... .. ... .. .. .. ...8 years . .. ... .. ... .. .. ... ... .. .. .. ... ... . .... .. ... . ... .. ... .. ... .. ... .. .. .... ... .. ... .. ... .. ... .. ... .. ... ... .. .. ... .9 years . ... .. ... .. .. ... .. ... . .... .. ... .. .. ... .. .. .... .. .. ... .. ... .. ... .. ... .. ... .. ... .. ... .. ... .. ... ... .. .. ... .. ... .. .10 years ... . .... .. .. .. ... ... .. .. .. .... . ... .. ... ... ... .. .. ... .. ... ... . ... ... .. ... . .... .. ... .. ... .. .... .. .. ... ... . .. .. .11 years ... .. ... .. ... . .... .. .. ... ... .. .. ... ... .. .. ... ... .. .. .. ... ... .. .. .... .. .. .. ... ... ... .. ... ... .. .. ... .. .. ... .. ..

Negro girls

6 years ... ... .. .. .. ... .. .... . ... .. ... ... ... .. .. ... ... .. .. .. .. .. .. .. .... . ... .. .. .... .. .. ... .. ... ... . ... .. ... .. ... .. ....

7 years . ... .. .. .. . . .. .. .. ... .. ... .. ... .. ... .. ... .. ... ... .. ... . ... .. .... .. ... .. ... . .... . .... .. ... .. ... .. .. ... .. ... .. ...8 years . .. ... .. ... ... ... .. .. .. ... .. .... . .... . .... . .... .. .. ... ... .. ... .. .... .. .. .. ... .. ... .. .. .... . .... .. .. ... . .... .. ..9 years .. .. . .. ... .. ... .. ... .. ... ... . ... .. ... .. .... . .... .. ... .. ... .. ... .. ... . ... .. .... .. .. ... . .. .. ... .. .. ... ... .. .. ... .10 years .. . ... ... .. ... .. ... . .... .. ... .. ... .. .. ... ... ... ... .. .. .. .. ... .. .... .. ... .. ... .. .. ... ... .. .. ... ... .. .. ... .. ...11 years .. .. .. ... ... .. .. .. .. .... .. ... ... ... .. ... .. ... .. ... .. ... .. ... .. ... .. .. ... .. ... .. ... .. ... ... ... . ... .. ... .. ... ..

N

1,7871,7811,7391,730

1,6921,662

1,7221,7161,6741,6631,632

1,605

289286279269264255

281284281

265266253

Stature

(centimeters)

118.5124.5129.8135.5140.3145.7

117.7123.4129.4135.1140.8147.3

119.1125.2131.3135.0139.6145.7

118.5124.6129.4

137.5141.8149.2

SD

5.155.525.706.77

6.626.69

5.475.866.196.727.00

7.89

5.115.505.336.467.928.08

5.755.556.697.809.857.42

Sitting haight

(centimeters)

65.067.269.571.6

73.375.6

64.26E.468.871.173.5

76.5

63.265.667.869.370.973.4

62.364.966.6

69.671.975.1

NOTES: N = estimated number of children in population in thousands; ~ = mean: SD = standard deviation.

SD

2.682.742.943.153.073.10

3.002.992.893.193.39

3,96

2.482.582.813.693.543.36

2.783.053.503.433.774.08

;itting height/

stature(percant)

54.854.063.652.952.351.9

54.653.853.252.752.252.0

53.052.451.651.450.850.5

52.652.161.650,750.850.4

SD

1.331.341,201.661.411.36

1.641.511.791,461.281.92

1.391.191.241.512.262.35

1.231.181.521.492.771.23

[ :Omparatjle data for ~hildlen 12.17 years can be found in Series 11, NO. 126 of Vital and Health Statistics.

24

APPENDIXCONTENTS

Statistical Notes .................................................................................................................................. .... 26The Survey Design ............................................................................................................................. 26Notes on Response fites ................................................................................................................... 27Parameter and Variance Estimation ................................................................................................... 28Imputation .......................................................................................................................................... 28Correlations of Sitting Height Versus WC and Stature Versus FVC .................................................. 30

LIST OF APPENDIX TABLES

L Response rates for children 6-1 I years, by sex and age: United States, 1963-65 ............................. 27

11. Response rates for children 6-11 years, by sex, race, and age: United States, 1963-65 .................... 28

25

APPENDIX

STATISTICAL NOTES

The Survey Design

The sampling plan of the second cycle of theHealth Examination Survey (HES) followed ahighly stratified, multistage probability design inwhich a sample of the U.S. population (includ-ing Alaska and Hawaii) from the ages of 6through 11 years, inclusive, was selected. Ex-cluded were those children confined to an insti-tution or residing upon any of the reservationlands set up for the American Indians.

In the first stage of this design, the nearly2,000 primary sampling units (PSU’S), geograph-ic units into which the United States was di-vided, were grouped into 357 strata for use inthe Health Interview Survey and the CurrentPopulation Survey of the U.S. Bureau of theCensus and were then further grouped into 40superstrata for use in Cycle II of HES.

The average size of each Cycle II stratumwas 4.5 million persons, and all strata fell be-tween the limits of 3.5 and 5.5 million persons.Grouping into 40 strata was done in a way thatmaximized homogeneity of the PSU’S includedin each stratum, particularly with regard to thedegree of urbanization, geographic proximity,and degree of industrialization. The 40 stratawere classified into 4 broad geographic regions(each with 10 strata) of approximately equalpopulation and cross-classified into 4 broad pop-ulation density groups (each having 10 strata).Each of the resultant 16 cells contained either 2or 3 strata. A single stratum might include onlyone PSU (or only part of a PSU as, for example,New York City, which represented two strata)or several score PSU’SO

To take account of the possible effect thatthe rate of population change between the 1950

and the 1960 census might have had on health,the 10 strata within each region were furtherclassified into 4 classes ranging from those withno increase to those with the greatest relativeincrease. Each such class contained 2 or 3 strata.

One PSU was then selected from each of the40 strata. A controlled selection technique wasused in which the probability of selection of aparticular PSU was proportional to its 1960 pop-ulation. In the controlled selection an attemptwas also made to maximize the spread of thePSU’S among the States. While not every one ofthe 64 cells in the 4X4X4 grid contributed a PSUto the sample of 40 PSU’S, the controlled selec-tion technique ensured the sample’s matchingthe marginal distributions in all 3 dimensionsand being closely representative of all cross-classifications.

Generally, within a particular PSU, 20 cen-sus enumeration districts (ED’s) were selectedwith the probability of selection of a particularED proportional to its population in the agegroup 5-9 years in the 1960 census, which by1963 roughly approximated the population inthe target age group for Cycle IL A similar meth-od was used for selecting one segment (cluster ofhouseholds) in each ED. Each of the resultant20 segments was either a bounded area or a clus-ter of households (or addresses). All of the chil-dren in the age range properly resident at theaddress visited were eligible children (EC’s).Operational considerations made it necessary toreduce the number of prospective examinees atany one location to a maximum of 200. TheEC’s to be excluded for this reason from thesample child (SC) group were determined bysystematic subsampling. If one of the samplechildren had a twin who was not a sample child,

26

this otherand while

twin was brought in for examination,the results were recorded for use in a

special substudy of twins, this twin was not in-cluded in the 7,119 children under the presentanalysis.

The total sample included 7,417 children6-11 years of age of whom 96 percent were fi-nally examined. These 7,119 examined childrenrepresented the roughly 24 million children inthe United States who met the general criteriafor inclusion in the sampling universe as of mid-1964.

All data presented in this publication arebased on “weighted” observations; that is, datarecorded for each sample child are inflated inthe estimation process to characterize the largeruniverse of which the sample child is representa-tive. The weights used in this inflation processare a product of the reciprocal of the probabilityof sckcting the child, an adjustment for non-rcxponsc cases, and a poststratified ratio adjust-ment which increases precision by bringing sur-\JeJrresults into closer alinement with ~om

U.S. population figures by race and sex forsingle years of age group 6-11.

In the second cycle of HES, the sample wasthe result of three stages of selection—the single

PSLT from each stratum, the 20 segments fromeach sample PSU, and the sample children fromthu eligible children. The probability of selectingan individual child is the product of the proba-bilities of selection at each stage.

Since the strata are roughly equal in popula-tion size and a nearly equal number of samplechildren were examined in each of the sampleI%U’S, the sample design is essentially self-weighting with respect to the target population;that is, each child 6-11 years old had about thesamc probability of being drawn into thesample.

The adjustment upward for nonresponse isintended to minimize the impact of nonresponseon final estimates by imputing to nonrespond-ents the characteristics of “similar” respondents.Here, “similar” respondents were judged to beex;imincd children in a sample PSU having thesame age (in years) and sex as children not ex-amined in that sample PSU.

The poststratified ratio adjustment used inthe second cycle achieved most of the gains in

precision which would have been attained if thesample had been drawn from a population strati-fied by age, race, and sex and made the finalsample estimates of population agree exactlywith independent controls prepared by the U.S.Bureau of the Census for the noninstitutionalpopulation of the United States as of August 1,1964 (approximate midsurvey point), by raceand sex for each single year of age 6-11. Theweights of every responding sample child in eachof the 24 age, race, and sex classes are adjustedupward or downward so that the weighted totalwithin the class equak the independent popula-tion control.

A more detailed description of the samplingplan and estimation procedures is included inearlier reports of the Vital and Health Statisticsseries. Series 11, No. 1z described the tech-niques used in Cycle I, which are similar to thoseof Cycle II.

Notes on Response Rates

There were 7,417 children aged 6-11 yearsselected for examination. Of these, 7,119 wereactually examined, which made an overall re-sponse rate of just under 96 percent. The re-sponse rate by sex and l-year age groups isshown in table I. It can be seen that only at age8 years was the response rate for girls betterthan that for boys.

A similar analysis can be done by sex, race,and age as shown in table II where a strikingdifference in response is readily seen. Negro chil-dren responded better than their white counter-parts at every age ‘group, and 9-year-old Negrogirls had an extraordinary 100-percent responserate.

Table 1. Response rates for children 6-11 years, by sex and age:United States, 1963-65

Age Boys Girls

Percent

Total .. .. ... . .. .... .. ... .. .. .. ... .. .... . .... .. ... . ... 96.5 95.5

6 years . ... .. ... .. .. ... .. ... . ..... . ... .. .... . .. ... .. ... ... .. .. ... .. .. 96.5 94.9

7 years . ... ... .. ... .. ... . ... .. ... .. ... .. ... .. .. ... .. ... .. .... .. .. .. .. 96.5 95.5

8 years . .... .. .. .. ... ... .. .. .. .. .. . ... .. .... . .. ... .. .... .. ... . ... ... . 95.2 97.0

9 years . ... .. .... .. .. ... .. .. .. ... .. .... .. ... . .... .. .. .. .. .... . ... ... . 97.6 94.810 years ... .. .. .. .... . .. ... ... ... .. .. .. .... . .... . ... ... ... . .. ... .. .. 97.0 95.111 years ... .. ... .. ... .. ... . ... .. ... ... .. ... . .... .. ... .. ... .. .. ... . .. 96.2 95.6

27

Table 11. Response rates for children 6-11 years, by sex, race,and age: United States, 1963-65

Boys GirlsAge

Negro White Negro I White

I Percent

Total .. . ... ... . ... .. . 98.1 I 96.2 I 98.7 1 94.9

I I I6 years .. ... .. ... . .. . .. .... .. .. .. .. .7 years .. ... .. . .. . ... .. .... . ... .. .. .8 years .. ... .. ... . .... .. ... . .. .... . .9 years .. ... . .... . .... . ... ... . ... .. .10 years ... .. ... . .... .. .. .. .. .... . .11 years ... .. ... .. ... .. .. .. ... .. .. .

97.797.597.598.798.598.8

96.396.3

95.097.496.895.8

97.398.999.1

100.097.598.8

94.794.896.593.994.795.0

Parameter and Variance Estimation

As each of the 7,119 sample children has anassigned statistical weight, all estimates of popu-lation parameters presented in HES publicationsare computed taking this weight into considera-tion. Thus, ~, the estimate of a population meanp is computed as follows:

z’

where Xi is the observation or measurement tak-en on the ith person and Wi is the statisticalweight assigned to that person.

The Health Examination Survey has an ex-tremely complex sampling plan, and obviouslythe estimation procedure is, by the very natureof the sample, complex as well. A method isrequired for estimating the reliabilityy of findingsthat “reflects both the losses from clusteringsample cases at two stages and the gains fromstratification, ratio estimation, and poststratifi-cation. ”z

The method for estimating variances in HESis the half-sample replication technique. Themethod was developed at the U.S. Bureau of theCensus prior to 1957 and has at times been givenlimited use in the estimation of the reliability ofresults from the Current Population Survey. Thishalf-sample replication technique is particularlywell suited to HES because the sample, althoughcomplex in design, is relatively small (7,1 19

cases) and is based on but 40 strata. This featurepermitted the development of a variance estima-tion computer program that pmcluces tables

containing desired estimates of aggregates,means, or distributions, together with a tableidentical in format but with the estimated vari-ances instead of the estimated statistics. Thecomputations required by the method aresimple, and the internal storage requirements arewell within the limitation of the IBM 360-50computer system used at the National Center forHealth Statistics.

Variance estimates computed for this reportwere based on 20 balanced half-sample replica-tions. A half sample was formed by choosingone sample PSU from each of 20 pairs of samplePSU’S. The composition of the 20 half sampleswas determined by an orthogonal plan. To com-pute the variance of any statistic, this statistic iscomputed for each of the 20 half samples. Usingthe mean as an example, this is denoted xi.Then, the wei~ted mean of the entire, undi-vided sample (~ is computed. The variance ofthe mean is the mean square deviation of each ofthe 20 half-sample means about the overallmean. Symbolically,

20

Z(xi - =)2

Var(x) = ‘=1 Z.

and the standard error of the mean is the squareroot of this, In a similar manner, the standarderror of any statistic may be computed.

A detailed description of this replicationprocess has been published.18

Imputation

As described earlier, the 7,119 spirogramscollected by means of the HES varied greatly inquality. Since the HES is a weighted survey, in-tended to represent the 6-1 l-year-old populationof the United States, spirometric values are nec-essary for every examined subject. It is impor-tant, then, to determine (1) for which subjectsvalues should be imputed and (2) which methodof imputation is most satisfactory.

Distributions of FVC’S and FEVI’s werecomputed by age, race, and sex for each quality

28

group (reference 5, tables 1 and 2), along withdistributions of a variety of other physical anddemographic characteristics (reference 5, tables4-7). Inspection of these tables led the authorsto collapse groups IA-IE into a group called “ac-ceptable” and the other groups into a groupcalled “unacceptable.” Groups IA-IE (n =5,803) comprised approximately 82 percent ofthe examined subjects, and their distributions ofspiro metric parameters and physical and demo-graphic variables were sufficiently similar totreat them as one group.

Several methods of imputation were con-sidered, but multiple stepwise regression waschosen as the most appropriate for the followingreasons: (1) the percent of “unacceptable” wastoo high to permit simple substitution of age-race-sex specific means or random assignment ofthe values listed for a similar respondent case;(2) the percent of “unacceptable” was too highto permit simply dropping the unknown casesand distributing the accompanying samplingweights; and (3) relevant information was avail-able for each unacceptable case, making possiblea more accurate prediction than could be ob-tained by any of the other methods, particularlyfor those cases evidencing some unusual physio-logical or demographic characteristics. Thus, the5,803 cases of groups IA-IE became the data setfrom which equations to predict spirometric pa-rameters for the 1,316 “unacceptable” weregenerated.

The authors determined that, among themany physical and demographic variables avail-able as regressors, 37 of these (see list that fol-lows) might manifest significant correlationswith the spirometric variables. The 37 independ-ent variables submitted for multiple stepwise re-gression equations were:

Race Elbow-elbow breadthSex Sitting heightAge StatureNorth Region Subischial lengthMidwest Region Biacromial breadthSouth Region Chest breadthWest Region Chest depthSize of place Bicristal breadthFamily income Chest girthParental education Waist girthSeat breadth Hip girth

Muscle circumferenceShoulder/hip ratioWeightRight-grip strengthLeft-grip strengthSystolic blood

pressureDiastolic blood

pressure

Ponderal indexWISC I.Q.Oral health indexNumber of decayed,

missing, filIed teethAsthmaRestricted activityHeart abnormalityBirthweight

The computer was fed the values for these 37possible regressors and the two dependent vari-ables (FVC and FEVl ) for the 5,803 “accept-able” cases, and the following equations to pre-dict FVC and FEVI for the 1,316 “unaccept-able” were the result.

FVC = - 2982.5- 141.7 X race - 69.3 X sex+ 22.4 X age + 24.7 X sitting height+ 11.8 X stature + 33.7 X chest breadth+ 11.6 X chest girth - 7.5 X hip girth+ 15.0 x left grip strength+ 2.0X WISC I.Q.

r = 0.8689 ~2 = 0.7550 aY = 238.3

FEVI =- 2624.0- 95.8 X race - 35.8 X sex

+ 19.7 X age + 20.3 X sitting height+ 10.6 X stature + 27.7 X chest breadth+ 15.7 X chest depth + 8.6 X chest girth- 7.4 X hip girth+ 10.7X left grip strength+ 1.5 X WISC I.Q. + 167.5 X restrictedactivity

r = 0.8279 ~2 = 0.6854 OY = 236.3

The reader should be forewarned about mak-ing unwarranted conclusions from these equa-tions. The absence of a particular independentvariable (e.g., chest girth) does not imply that itis not related to the dependent variable; it sim-ply means that some other variable or combina-tion of variables that is in the equation has ac-counted for the effect of the absent variable.Also, the resultant equation may not necessarilybe the best predictor–as the number of depend-ent variables goes up, the number of possiblemultiple regression equations increases at a ratethat makes the computation of every possibleequation a prohibitive effort. The stepwise tech-nique is a compromise in which the loss of pre-

dictive power is usually negligible.The values for the independent variables in

the two equations were then collected for the1,316 “unacceptable” cases. The nominal vari-

29

ables, such as region, were recoded into “O-l”dichotomies, and all the data were edited. Thespirometric values for each case were predicted,and to each value was added a random errorcalculated from the standard error of estimate ofthe appropriate regression equation and theIBM-written, random deviate generator. This laststep is necessary to provide a parametrically nor-mal dispersion of predicted values for any givenset of regressor values; that is, to prevent an im-proper lumping of values at or near the popula-tion mean that would speciously reduce thestandard error of that mean.

The validity of the prediction scheme wasverified several ways. Distributions of the pre-dicted values by age, race, and sex were com-puted and compared with the same distributionsfor the “acceptable” population. No significantdifference was shown by the Sign Test at the5-percent level. When these distributions werefurther crossed by sitting height (the first vari-able to be included in both the FVC and FEV1equations), again the Sign Test revealed no sig-nificant difference. Visual inspection showedthat the relationship by age, race, and sex ofFVC and FEV1 was very similar in the “accept-able” and “unacceptable” classes. Finally,combining the groups resulted in values thatwere very near those of the original respondentgroup. Generally, the values for summary age,race, or sex groups were reduced somewhat,while there was no consistent increase or de-crease of age-race-sex specific values. This wasexpected, since younger children, females, andNegroes (all of whom tend to have lower valuesthan their respective complements) were over-represented in the “unacceptable” class.

Correlations of Sitting Height Versus FVC

and Stature Versus FVC

The problem of determining the exact rela-tionships of FVC with stature or sitting heightwhen race is a qualifier has stubbornly eludedsolution. Graphs of these relationships (see fig-ures 6 and 8) show clearly that there is a raceeffect and that nearly half of the FVC differencebetween Negro and white people can be ac-counted for by the shorter trunk lengths of Ne-

groes. This suggests that race-specific product-moment correlations between FVC and sittingheight should be greater than correlations thatare not race specific. Also, since only the upperpart of the body is significantly involved in theperformance of an FES, it seems reasonable thatcorrelations between FVC and sitting heightshould be greater than those of FVC and stature.Subsequent analysis was unable to substantiateany of this conjecture (see table A in text).

Not only are none of the race-specific corre-lations significantly better than the total corre-lations, but they also vary erratically. Also, thecorrelation between FVC and sitting height isonly slightly greater than that between FVC andstature, where a large difference was anticipated.

One explanation for these unexpected re-sults may lie in the relatively larger technicalerror (t.e. ) associated with sitting height. Thisstatistic, calculated from the formula

@ /

n

t.e. = d: 2ni=l

where di is the difference between repeated meas-urements on an examinee, reflects the reproduci-bility of a given measurement and is nearly twiceas 1arge (relatively) for sitting height as forstature.8 It is mathematically and empiricallydemonstrable that increasing the technical errorof a variable decreases its correlation with othervariables. However, no replicate studies weredone on Cycle II measurements. So it was notpossible to calculate t.e.’s for these data or toquantify their relative effects, except to say thatif the t.e. for sitting height had been as low (rela-tively) as that for stature, then the correlationbetween sitting height and FVC would havebeen greater. Also, due to the lack of t .e.’s onsitting height and stature, estimating what pro-portion of the total variability of each is due toreal biologic variation and what proportion isdue to examiner error is impossible. If these esti-mates had been available, the variances and co-variances could have been adjusted to reflectequal t.e. ‘s, and the pseudocorrelations calcu-lated from these estimates would have been abetter description of the relationships betweenthe three variables.

* U. S.GOVERNMENT PRINT~G OFFICE :1978 260-937/4

30

VITAL AND HEALTH STATISTICS PUBLICATIONS SERIES

Formerly public Health Service tiblication No. 1000

Series 1. Programs and Collection Procedures. -Reports which describe the general programs of the NationalCenter for Health Statistics and its offices and divisions, data collection methods used, definitions, andother material necessary for understanding the data.

Series 2. Data Evaluation and Methods Research. –Studies of new statistical methodology including experimentaltests of new survey methods, studies of vital statistics collection methods, new analytical techniques,objective evaluations of reliability of collected data, contributions to statistical theory.

Series 3. Analytical Studies. –Reports presenting analytical or interpretive studies based on vital and healthstatistics, cssrying the analysis further than the expository types of reports in the other series.

Series 4. Documents and Committee Reports. –Final reports of major committees concerned with vital andhealth statistics, and documents such as recommended model vital registration laws and revised birthand death certificates.

Series 10. Data j+om the Health Interview Survey. –Statistics on illness; accidental injuries; disability; use ofhospital, medical, dental, and other services; and other health-related topics, based on &ta collected ina continuing national household ititerview survey.

Series 11. Data porn the Health Examination Survey .–Data from direct examination, testing, and measurementof national samples of the civilian, noninstitutionalized population provide the basis for two types ofreports: (1) estimates of the medically defined prevalence of specific diseases in the United States andthe distributions of the population with respect to physical, physiological, and psychological charac-teristics; and (2) analysis of relationships among the various measurements without reference to anexplicit finite universe of persons.

Sert”es12. Data from the Institutt”onalized Population Surveys. –Discontinued effective 1975. Future reports fromthese surveys will be in Series 13.

Series 13. Data on Health Resources Utilization. –Statistics on the utilization of health manpower and facilitiesproviding long-term care, ambulatory care, hospital care, and family planning services.

Series 14. Data on Health Resources: Manpower and Fizcilities. –Statistics on the numbers, geographic distrib-ution, and characteristics of health resources including physicians, dentists, nurses, other health occu-pations, hospitals, nursing homes, and outpatient facilities.

Series 20. Data on Mortality .-Various statistics on mortality other than as included in regular annual or monthlyreports. Special analyses by cause of death, age, and other demographic variables; geographic and timeseries analyses; and statistics on characteristics of deaths not available from the vital records, based onsample surveys of those records.

Series 21. Data on Natality, Marriage, and Divorce. –Various statistics on natality, marriage, and divorce otherthan as included in regular annual or monthly reports. Special analyses by demographic variables;geographic and time series analyses; studies of fertility; and statistics on characteristics of births notavailable from the vital records, based on sample surveys of those records.

Sert”es22. Data from the National Mortality and Natality Surveys. –Discontinued effective 1975. Future reportsfrom these sample surveys based on vital records will be included in Series 20 and 21, respectively.

Scrt”es23, Datu j%om the National Survey of Family Growth. –Statistics on fertility, family formation and disso-lution, family planning, and related maternal and infant health topics derived from a biennial survey ofa nationwide probability sample of ever-mamied women 15-44 years of age.

For a list of titles of reports published in these series, write to: Scientific and Technical Information BranchNational Center for Health StatisticsPublic Health Servicei+yattsville, Md. 20782