vital and health statistics; series 11, no. 146 (12/74) girls remained essentially stationary. in...
TRANSCRIPT
Data from the
NATIONAL HEALTH SURVEY
HematocritValuesOfYouths12=17YearsUnited States
Hematocrit values are presented and discussed by age, sex, race, and socio-
economic level of youths 12-17 years of age in the UnitedStates, 1966-70.
DHEW Publication No. (HRA) 75-1628
U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
Public Health Service -
Health Resources AdministrationNational Center for Health Statistics
Series 11
Number 146
Rockville, Md. December 1974
Series 11 reports present findings from the National Health Examination Survey, whichobtains data through direct examination, tests, and measurements of samples of the U.S.population. Reports 1 through 38 relate to the adult program, Cycle I of the HealthExamination Survey. The present report is one of a number of reports of findings from thechildren and youth programs, Cycles II and III of the HeaIth Examination Survey. Theselatter reports from Cycles II and HI are being published in Series 11 but are numberedconsecutively beginning with 101. It is hoped this will guide users to the data in which theyare interested.
Library of Congress Cataloging in Publication Data
Heald, Felix.
Hematocrit values of youths 12-17 years, United States.
(National Center for Health Statistics. Data from the National Health Survey, series 11,no. 146) (DHEW publication No. (HRA) 75-1628)
Supt. of Dots. no.: HE 20.6209:11/146.1. Hematocrit. 2. Youth–Health and hygiene. 3. United States–Statistics, Medicul. I.
Levy, Paul S., joint author. II. Hamill, Peter V. V., joint author. 111, Rowland, hlichael, jointauthor. IV. Title. V. Series: United States. National Center for Health Statistics. Vital andhealth statistics. Series 11: Data from the National Health Survey, Data from the healthexamination survey, no. 146. VI. Series: United States, Dept. of Health, Education, andWelfare. DHEW publication no. (HRA) 75-1628. [DNLhl: 1. Hematocrit–In adolescence.W2A N14vk ser. 11 no. 146]RA407.3.A347 no. 146 [RB45] 312’.0973sISBN 0-8406-0025-9 [312’.31’500973] 74-12434
For sale by the Superintendent of Documents, U.S. Qovernmsnt Prfnting Office, Waehfngton, D.C. Z134GZ- Pric8 $1
NATIONAL CENTER FOR HEALTH STATISTICS
EDWARD B, PERRIN, Ph. D., Director
PHILIP S. LAWRENCE, SC.D., .Depttjv DirectorJACOB J. FELDMAN, Ph. D., Acting Associate Director for Analysis
GAIL F. FISHER, Associate Director for the Cooperatzlq Health Statistics SystemELIJAH L. WHITE, Associate Director-for Data Systems
IWAO M. MORIYAMA, Ph. D., Associate Direct&-for International StatisticsEDWARD E. MINTY, Associate Director for Management
ROBERT A. ISRAEL, Associate Director for OperationsQUENTIN R. REMEIN, Associate Director for Program lleve~oprnent
PAUL E. LEAVERTON, Ph. D., Acting Associa;e Director for Research
ALICE HAYWOOD, .Inforrnation Officer
DIVISION OF HEALTH EXAMINATION STATISTICS
ARTHUR J. McDOWELL, Director
JEAN-PIERRE HABICHT, M.D., SC.D., Specicd Assistant to the Director
PETER V. V. HAMILL, M.D., Medical Advisor
JEAN ROBERTS, Chie~ Medical Statistics Branch
ROBERT MURPHY, Chief Survey Pkming and Devei’opment Branch
COOPERATION OF THE BUREAU OF THE CENSUS
In accordance with specifications established by the National HealthSurvey, the Bureau of the Census, under a contractual agreement, partici-pated in the design and selection of the sample, and carried out the first stageof the field interviewing and certain parts of the statistical processing.
Vital and Health Statistics-Series 1l-No. 146
DH EW Publication No. (HRA) 75-1628
Library of Congress Catalog Card Number 74-12434
r
CONTENTS
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Blood Specimen and Microhematocrit . . . . . . . . . . . . . . . . . . . .Race . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .SexandAge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .GeographicRegion,Sex,andAge . . . . . . . . . . . . . . . . . . . . . .Race,Sex,andAge . . . . . . . . . . . . . . . . . . . . . . . . . . . . .AnnualFamilyIncome,Sex,andAge . . . . . . . . . . . . . . . . . . . .EducationofParent,Sex,andAge . . . . . . . . . . . . . . . . . . . . .Annual Family Income, Education of Parent, Race, Sex, and Age . . . . .
.-.
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .AgeandSex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Race . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .IncomeandEducation . . . . . . . . . . . . . . . . . . . . . . . . . . .Interaction Between Race and Socioeconomic Status . . . . . . . . . . . .GeneraI Comparability With Other Studies . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ListofDetailedTables . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
,4ppendixI. Statistical Notes . . . . . . . . . . . . . . . . . . . . . . . . .TheSurveyDesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .SomeNotesonResponseRates . . . . . . . . . . . . . . . . . . . . . . .ParameterandVarianceEstimates . . . . . . . . . . . . . . . . . . . . . .Standards of Reliability and Precision . . . . . . . . . . . . . . . . . . . .Imputation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AppendixII. Demographic Variables . . . . . . . . . . . . . . . . . . . . .
Appendix III. Techniques ofMeasurement and Quality Control . . . . . . .Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .MonitoringSystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .hflethods of Taking Replicate Measurements . . . . . . . . . . . . . . . . .Results of the Replicate Studies for Hematocrit . . . . . . . . . . . . . . .
1
;3
3344677
91011121213
15
17
323233343535
36
3737383839
.
...Ill
SYMBOLS
Datanot available --------------------------------------- ---
Category not applicable ------------------------------ . . .
Quantity zero --------------------------------------------- -
Quantity morethan Obutless than0.05---- 0.0
HEMATOCRIT VALUES OF
Felix Heald, M. D., Paul S. Levy, SC.D., Peter V. V.
YOUTHS 12-17 YEARS
Hamill, M.D., M.P.H., and Michael Rowlanda
INTRODUCTION
This report of hematocrit values as deter-mined from blood samples of youths 12-17years of age in the United States is one of aseries of reports presenting findings from CycleIII of the Health Examination Survey. Themeans and selected percentiles of the hematocritvalues are examined here by sex, age, race,family income, education, and geographic regionof the United States. As described in a detailedreport of its general plan and operation, 1 theHealth Examination Survey (HES) is conductedin a succession of cycles.
Cycle I of the HES, conducted from 1959 to1962, obtained information on the prevalence ofcertain chronic diseases and on the distributionof a number of anthropometric and sensorycharacteristics in the civilian, noninstitution-alized population of the continental UnitedStates, aged 18-79 years. The detailed plan ofCycle I is described in a previous report,2 and
aProfcssor of Pediatrics, Director, Division of Adolescent
Medicine, University of Maryland, School of Medicine; AssociateProfessor of Biometry, University of Illinois, School of Public
Healtti Medical Adviser, Chfldren and Youth Programs, Division
of Heahh Examination Statistics; Analytical Statistician, Divi-sion of Health Examination Statistics, respectively.
most of the results are published in other reportsof Series 11 in Vital and Health Statistics. -
Cycle II of the HES, conducted from July1963 to December 1965, involved selection andexamination of a probability sample of noninsti-tutionalized U.S. children, 6-11 years. Thisprogram succeeded in examining 96 percent ofthe 7,417 children selected for the sample. Theexamination had two focuses: (1) factors relatedto healthy growth and development as deter-mined by a physician, a nurse, a dentist, and apsychologist and (2) a variety of somatic andphysiologic measurements performed by spe-cially trained technicians. The detaiIed plan andoperation of Cycle II and the response resultsare described in Vital and Health Statistics,Series l-Number 5.3
HES Cycle III, conducted from March 1966to March 1970, was essentially an agewiseextension of Cycle II into adolescence. Asdescribed in detail in “Plan and Operation of aHealth Examination Survey of U.S. Youths,12-17 Years of Age,”4 Cycle III was moresimilar to Cycle II than to Cycle I not only inform, content, and style but also in having itsmajor emphasis on factors of “normal” growthand development rather than on chronic dis-eases. These analyses on “normal” growth and
1
development of adolescents have been wellunderway since 1970, and some of the resultsfrom the battery of body measurements havealready been published.5 -7
This report, which presents the first of thefindings from the sample of blood drawn fromeach youth, is closer in content to Cycle I thanto Cycle II, because no blood specimen was ob-tained from the children in Cycle II. One wasobtained from the adults in Cycle I and amongthe determinations made was the hematocrit. Onthe other hand, the present report, like those ofCycle II, primarily deals with factors relating togrowth and development. As will be shown, therewas a marked “growth” in the erythrocytevolume for adolescent boys, while that for ado-lescent girls remained essentially stationary.
In this report, the means and selected percen-tiles of the hematocrit values are examined bysex, age, race, family income, education of
parent, and geographic region of the UnitedStates.
METHOD
At each of 40 preselected locations (seeappendix I for sample design) throughout theUnited States, the youths were brought to thecentrally located mobile examination center foran examination that lasted about 3% hours. Sixyouths were examined in the morning and six inthe afternoon. Except during vacations, theywere transported to and from school and/orhome.
When the youths entered the examinationcenter, their oral temperatures were taken, and acursory screening for acute illness was made; ifillness was detected, the youth was sent homeand reexamined later. The examinees changedinto gymnasium-type shorts; cotton sweat socks;a terry-cloth robe; and, for the girls, a light,sleeveless topper. All six then proceeded todifferent stages of the examination, each onefollowing a different route. The 3Yz-hour exami-nation was divided into six 35-minute timeperiods, each consisting of one or more detailedexaminations at a designated station. At the endof each period, the youths rotated to anotherstation so that at the end of 31% hours eachyouth had been given essentially the sameexaminations by the same examiners but in a
different sequence. Four of these examinationtime periods were allocated to examinations bya pediatrician, a dentist, and a psychologist ,band the other two were allocated to a group ofexaminations performed by highly trained tech-nicians. This last group of examinations con-sisted of X-rays of the chest and hand-wrist,hearing and vision tests, measures of respiratoryfunction, a 12-lead electrocardiogram, a sub-maximal exercise tolerance test on a treadmillwith chest leads to a continuous electro-cardiogram, a battery of body measurements,grip strength, examination of blood and (on girlsonly) urine cultures for bacteria, and a privatelyadministered health behavior and attitude ques-tionnaire.
Blood Specimen and Microhematocrit
The development of the technique used byHES for obtaining and processing the bloodspecimen is described in appendix HI. When thevenipuncture needle was withdrawn from thearm, the tubing was partially drained, and twomicro capillary tubes were two-thirds filled withthe residual blood in the tubing. The ends werefjlled with Sealease and the tubes placed into aspecial autocrit carrying tray. The tray wascentrifuged at 12,500 r.p.m. for 5 minutes, anda reading was taken on each tube directly on thetray. Care was taken to read immediately aftercentrifugation—the tubes were not allowed to sitfor long periods of time. The reading wasexpressed as the percentage that the volume ofthe packed red blood cells comprised of theinitial whole blood specimen in the microcapil-lary tube.
The microhematocrit was the only laboratorydetermination performed in the examinationtrailers on the blood specimen. The bulk of theblood samples, after preliminary preparation inthe trailer’s laboratory, were properly separatedinto their various sub samples and packed intospecially devised Styrofoam containers for ship-ment via air freight to either the Immunogene-tics Laboratory of The Johns Hopkins Univer-
bThe entire examination by the psychologists consisted of
two consecutive time periods (70 minutes). Two psychologists
performed identical examinations simultaneously at separate
stations.
2
sity, Baltimore, Md., or the Center for DiseaseControl (PHS), Atlanta, Ga.
Race
Race was recorded as “white,” “Negro,” and“other races”c (see appendix II). In Cycle III,
the white youths constituted 84.74 percent ofthe tokd, the Negro youths 14.76 percent, andyouths of other races only 0.50 percent. InCycle II, white chikiren constituted 85.69 per-cent of the examined subjects and Negro chil-dren, 13.86 percent. (The differential responserate by age, sex, and race is analyzed anddiscussed in appendix L The increased propor-tion of Negro subjects in Cycle III was due totheir better response rate–the overall Negroresponse rate was 96.6 percent and the overallwhite response rate was 89.1 percent. ) As inCycle II, because so few youths of “other races”were part of the sample, data for them have notbeen analyzed as a separate category. Wheneverdtita are analyzed independent of a classificationby race, however, data for these youths areincluded.
tions. Percentage distribution of hematocrits isgiven for these youths by age and sex in table 2.
In male youths, mean hematocrit levels in-creased consistently with age, from a low of40.5 ml percent for 1Z-year-olds to a high of45.8 ml percent for 17-year-olds (table 1 andfi~re 1). This represents an estimated increaseof 5.3 ml percent or 1.06 ml percent per yearover the 6 years of adolescence. In females,however, no such increase with age was seen; themean hematocrit levek remained within the40.3-40.7 ml percent range throughout adoles-cence (table 1 and figure 1). For each of the sixage groups, mean hematocrit levels were higherin males than in females. Differences betweenthe mean hematocrit levels of male and femaleyouths increased consistently with age, from 0.2ml percent for 12-year-olds to 5.2 ml percent for17-year-oIds.
The increase in mean hematocrit levels withage for male youths is also seen when the dataare examined by percentiles (table 1). The 5thpercentile for male youths increased consistentlywith age, from a low of 36.0 percent for
RESULTS
Sex and Age
The estimated number of youths falling intomch of 16 groups according to hematocrit levelsis shown by age and sex in table 1 along withsckcted percentiles, mean hematocrit levels, andsttindard deviations of the hematocrit distribu-
cThe same classification scheme as used in the 1960 censuswas employed here. As deseribed in the reviously mentioned
1’rcpcwt on the operation of HES Cycle III, this information wasobtained at the initiaI household interview by the Bureau of the
Census ficldworker. The accuracy of the information waschcckcd at the subsequent home visit by the highly experiencedrcprmxmtativc from HES and again at the examination in the
tr.silcr. A final record check by birth certificate turned up onlyscvtm inconsistencies, and these were mostly pertaining to thecatc~ory “other races.” Hence, the possible extent of misclassifi-
c~tion of the variable race, as described, is so minimal that itcould have no effect on the data anatyzed in tbk report.
Howcw,r, when comparing the present HES findings with those ofother variously defined racial groupings in the world, the degrees
of genetic admixture, as first discussed by Herskowitz8 in 1928
“ q by Roberts,10~11 and by Reed, 12and Iatcr by Glass and LI,-should bc t~en into consideration.
12 13 14 15 16 17
AGE IN YEARS
Figure 1. Mean hematocrit levels for youths aged 12-17 years,by age and sex: United States, 1966-70.
3
●
12-year-olds to a high of 41.2 ml percent for17-year-olds. No such increase with age in the5th percentile was seen for females. Values ofthis percentile for females remained within the35.8 -37.0 ml percent range throughout the 6years of adolescence. The median, or 50thpercentile, of the hematocrit distribution inmale and female youths at each age showed thesame variations with age and sex as the mean. Atthe upper end of the distribution, the 95thpercentile for male youths increased with agefrom 45.2 ml percent in 12-year-olds to 50.3 mlpercent in 17-year-olds. No such increase withage was seen in females.
In an attempt to gain further insight into thenature of the increase in hematocrit levels formale youths, data were amdyzed for 6-monthage groups from 12 to 18 years (table 3). Thoseyouths who were between 12 years and 12!4years were included in the 12-year age group;those between 12% and 12% years of age, in the12%-year age group; those between 12% and13?4, the 13-year age group; and so forth. Thosemale youths 173A years and over were includedin the 18-year age group. There was an irregulari-ty in the classification at the two extreme agegroups because only youths over 12 years andunder 18 years of age were eligible to participatein the examination.
Mean hematocrit levels in male youths for6-month age groups increased consistently andrather uniformly with age, from a low of 40.-~ml percent for 12-year-old males to a high of46.0 ml percent in 18-year-old males (table 3).An examination of successive differences be-tween these means shows that the largest differ-ence (0.8 ml percent) occurred between 14 and14Y2 years of age. Differences of 0.6 ml percentor more, however, occurred between 13 and13% years (0.6 ml percent), 14% and 15 years(0.6 ml percent), and 15 and 15% years (0.7 mlpercent). The smallest differences betweenmeans occurred at the two extreme age groups(which have irregular age compositions), wherethe difference in mean hematocrit between 12-and 12?4-year-olds was 0.1 ml percent, and thatbetween 17?4+ and 18-year-olds was 0.2 mlpercent.
In describing the shape of the distribution ofhematocrits, we have used three summary statis-tics–namely, the distance between the mean and
the median, the distance between the 5thpercentile and the median, and the distancebetween the 95th percentile and the median(table 4). In male youths, the median at each agewas slightly higher than the mean, but thedistance from the 5th percentile to the medianwas approximately the same as that from themedian to the 95th percentile. Thus, the distri-bution for males at each age group was fairlysymmetrical.
For females aged 12, 16, and 17 years, themean and median ht -natocrits were alwayswithin 0.1 ml percent of each other; and thedistance between the 5th percentile and medianwas nearly the same as that between the medianand 95th percentiles, indicating symmetricaldistributions. On the other hand, although themeans and medians for females aged 13, 14, and15 years were also virtually identical, the dis-tance between the 95th and 50th percentiles wasgreater than the dktance between the 5th and50th percentiles by 1.0 ml percent or more.Since the means and medians coincided, thisindicates that the lower 5-percent tail of thedistribution is longer than the upper 5-percenttail, which, in turn, implies the presence of somevery low hematocrit levels for 13-, 14-, and15-year-old females.
Geographic Region, Sax, and Age
Mean hematocrits and standardmean hematocrits are shown in
errors of thetable 5 by
geographic region, sex, and age; and the numberof sampIe persons are shown in tabte 6 by thesame variables. Within each geographic region,mean hematocrit increased with age in males butnot in females, a finding consistent with thepattern discussed previously for the UnitedStates as a whole. The findings show no evidenceof differences among regions with respect tomean hematocrit levels.
Race, Sex, and Age
Findings by race, sex, and age are shown intables 7 through 11, and in figures 2 and 3.
The increase with age in mean hematocritfound for male youths in the United States as awhole was also found for Negro and white males
z
50
45
40IA.
/,/ 95th percentile white females
-” 95th percentile Negro females
4 50th percentile white females.--- +.. -b.--- - 50th percentile Negro females-----
● ● 5th percentile white females*% \ .6.
-“ .-. -- 5th percentile Negro females--0-
35 ““
30- -
012 13 14 15 16 1“7
AGE IN YEARS
Flgure2. Median, 5th, and95th percentiles of thehematocrit distribution ofwhite and Negro females by age: United States, 1966-70.
50-
45- -
40- “
35 “-
*-
*
----
95th percentile white males
95th percentile Negro males
50th percentile white males
50th percentile Negro males
5th percentile white males
5th percentile Negro males
01 1 I I t m
12 13 14 15 16 17
AGE IN YEARS
Figure3, Median, 5th, and 95th percentiles of thehematocrit distribution of white and Negro males byage:United States, 1966-70.
5
separately (table 7). Mean hematocrit levelsincreased in white males, from a low of 40.6 mlpercent for 1Z-year-oIds to a high of 45.8 mlpercent for 17-year-olds. In Negro males, meanhematocrit levels increased consistently with agefrom 39.6 ml percent for 1 Z-year-oIds to 45.7ml percent for 17-year-olds. At each age, themean hematocrit for Negro males was lowerthan that for white males of the same age.
In white female youths, there was no increasewith age in mean hematocrit, with mean levelsfluctuating between 40.4 and 40.9 ml percent.In Negro females, mean hematocrit levels werelower in the 16- and 17-year-olds than in the13-, 14-, and 15-year-old youths. At each age,Negro females had lower mean hematocrit IeveIsthan white females, with the differences be-tween races being very large (2.4 ml percent) for16- and 17-year-old girls (table 7).
Selected percentiles of hematocrit levels areshown for white and Negro youths in tables 10and 11 and figures 2 and 3. The differences inhematocrit levels between white and Negroyouths at the 5th, 50th, and 95th percentiles areshown in table 8.
Differences between white and Negro maleswere primarily at the lower end of the distribu-tion (fi~re 3 and table 8). The 5th percentilefor white males was higher at each age than thecorresponding percentile for Negro males, withthese differences averaging 1.25 ml percent overall ages. At the middle of the distribution,differences between Negro and white males weremuch smaller, with the median for white malesaveraging 0.5 ml percent higher than that forNegro males. In sharp contrast, at the upper endof the distribution, there were no consistentdifferences between white and Negro males atthe 95th percentile.
The hematocrit distribution for Negro maleshad more variability associated with it than thatfor white males (tables 10 and 11). The standarddeviation of the hematocrit distribution wasgreater at each age for Negro males than forwhite males, as was the distance between the95th and 5th percentiles. If one distribution hasa lower median and greater variability y thananother, one would expect the differences be-tween the two to be most striking at the lowerend of the distribution, w!~ich is exactly whatwas observed in the hematocrit distributions ofwhite and Negro males.
Similar patterns appear when the differencesbetween the hematocrit distributions of whiteand Negro females are analyzed (tables 8, 10,and 11 and figure 2), except that the differenceswere more apparent for females and they weremost striking for the older females (15-, 16-, and17-year-olds). The 5th percentile for whitefemales averaged 3.0 ml percent higher than thatfor Negro femzdes, with 16-year-oId white fe-males having a 5th percentile that was 4.4 mlpercent higher than their Negro counterparts(figure 2).
The median of white females was higher ateach age than that of Negro females, and thedifference was highest at the two oldest ages(2.4 ml percent for 16-year-olds and 2.0 mlpercent for 17-year-oIds). At the upper end, the95th percentile for white females was higherthan that for Negro females at all ages except 13years. Again, as in males, differences betweenNegro and white females were largest at thelower end of the distribution, and smallest at theupper end.
In all but 1Z-year-oIds, the standard deviationof the hematocrit distribution was higher forNegro females than for white females (tables 10and 11).
Another way of lo~king at differences inhematocrit levels by race is shown in table 12, inwhich the percentage of persons with hematocritlevels below three cutoff point s–i.e., 35.0 mlpercent, 37.0 ml percent, and 39.0 ml percent–is shown by race, age, and sex. For both malesand females at each age group, the percentage ofNegroes having hematocrit levels below each ofthese cutoff points was much greater than thecorresponding percentage of whites.
Annual Family Income, Sex, and Age
Mean hematocrit levels are given for maIe andfemale youths by annuzd family income in table13. A small but clear increase in mean hemat o-crit occurred with increase in annual familyincome for both male and female youths at eachage group. Hematocrit levels of males in familieswith annual incomes of $10,000 or more aver-aged (over all ages) 0.67 ml percent higher thanthose of males in families with annual incomesof less than $3,000, while the level of females infamilies with annual incomes of $10,000 ormore averaged 1.08 ml percent higher than those
6
of females whose families had annual incomes of increase in mean hematocrit was foundless than $3,000 (figure 4). Table 14 shows the10th, 50th, and 90th percentiles of the hemato-crit distribution by age and sex for those infarnilics with annual incomes less than $4,000and for those in families with annual incomes of$10,000 or more. Differences between youths~,h(osc fanl~ies had ~nu~ incomes of $10,000”or more and those whose families had annualincomes less than $4,000 were generally greaterat the 10th percentile than at the median and90th percentiles.
Education of Parent, Sex, and Age
Mean hematocrit levels by education of par-ent, SCX, and age are shown in table 15. An
L
withincrease in education of parent. Over all ages,those male youths whose parents had 13 yearsof education or more had mean hematocritlevels approximately 0.82 ml percent higherthan those males whose parents had less than 8years of education, while those females whoseparents had 13 years of education or more hadmean hematocrit levels approximately 1.13 n-dpercent higher than those females whose parentshad less than 8 years of education (figure 5).
Annual Family Incomer Education of Parent,Race, Sex, and Age
Mean hematocrit levels are given by annualfamily income, sex, and age (tables 16 and 17)
Males, annual family income less than $3,000
------ Males, annual family income $10,000 or more
--------- Females, annual family income less than $3,000
47 T-------- Females, annual family income $10,000 or more
A//
.
1- 45- -z ~- &“
w /-0 /“”aLu 43- - // 4“”n-1 Rti”
z 41- - w-Z““””””””.......................”””-”.”...........
..................
0---- .-,-,--,-------.----------- .-”-”------~ -.-.+. _._._.-----
K0
39- -
01-a2 37- -
$z$ 35
z
10 I I 1 I 1 ,12 13 14 15 16 17
AGE IN YEARS
Figure 4, Mean hematocrit levels for youths aged 12-17 years, by annual family income, age, and sex: United States, 1966-70.
7
— Males, education of parent lessthan 8 years
------ Males, education of parent 13 years or more
isz
47
45
43
41
39[
---- ----- Females, education of parent lessthan 8 years
.. . . . . . . . . .. Females, education of parent 13 years or more.-4
**./”
//
.“O/...... . . . . ... . . . ... . .....~.. . .... . . . . . . . . . . . .. . . . . . . . . . . . . . . . . ...* -. ..... . ..-
.- .-.=.-.-------- -------- ---- --------- -------- .-. **.-
37- -
35- “
1 I I t I t
12 13 14 15 16 17
AGE IN YEARS
Figure 5. Mean hematocrit levelsfor youths aged 12-17 years, by education of parent, age, and sex: United States, 1966-70.
and by education of parent, sex, and age (tables Negroes in the sample, especially at the higher
18 and 19) separately for white and Negroyouths.
The relationships between mean hematocritand annual family income and between meanhematocrit and education of parent, discussedpreviously for the population as a whole, heldalso for white youths when examined separately.Among white males and white females in all agegroups, a consistent increase in mean hematocritwas found both with increase in annual familyincome and with increase in education of parent.
When data were examined separately for eachage group (tables 17 and 19), there dld notappear to be a consistent increase in meanhematocrit of Negro youths with increase inannual family income or with increase in educa-tion of parent. These estimates of mean hemato-crit for Negroes, however, were quite unstablebecause of the relatively small numbers of
income and education groups. To compensatefor this, the six age groups for females in tables16 through 19 were collapsed into one table,and the combined data were examined for Negroand white female youths by education of parentand by annual family income (table A andfigures 6 and 7). Even with all ages combined,the Negro female youths did not show aconsistent increase in mean hematocrit withincrease in annual family income or with in-crease in education of parent. At all levels ofincome and education, white female youths had,mean hematocrits approximately 1.0-2.0 mlpercent higher than Negro females of compara-ble family income or parental education.
A similar analysis was not performed formales because of the increase with age in meanhematocrit observed in both Negro and whitemales.
Table A. Sample size, mean hematocrits, and standard errors of the means for Negro and white female youths 12-17years of age, by annual family income and education of parent
1ncome and education of parent
Annual family income
Lessthan $3,000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
$3,000-$4,999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .$5,000-$9,999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
$10,000 ormore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Education of parent
Lessthan 8years . . . . . . . . . . . . . ., . . . . . . . . . . . . . . . . . . .8-n years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13yearsormore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1-
White females _-------- -----
~~ I Negro females!-
: o~Lessthan $3,000- $5,000-$10,000
zf $3,000 $4,999 $9,999 ormore
s. ANNUAL FAMILY INCOME
Hematocrit in ml percent
495400407
117
302679363127
39.338.9
39.7
39.4
38.939.439.838.5
.32
.49
.19
.72
.28
.22
.32
.90
9281,102
3,760
3,258
7842,3373,6002,752
40.240.340.6
41.0
40.140.440.8
41.0
.39
.24
.14
.15
.62
.19
.10
.17
45
40
35 1
White females* ---------- ti----~
.Negro females
o~Less 8-11 12 130r
than 8 more
EDUCATION OF PARENT IN YEARS
Flaure6. Mean hematocrit for white and Nesrrofemalevouths Figure7. Mean hematocrit for white and Negro female youths
aged 12-17 years by annual family income: United States,
1966-70.
DISCUSSION
The packed cell volume, or hematocrit value,is determined by centrifuging a standard amountof whole blood at 12,500 r.p.m. for 5 minutes.
aged 12-17 years by education of parent: United States,
1966-70.
As a result of this procedure, the blood’s cellsare separated from the plasma and packed in thelower portion of the tube. The resultant hemato-crit packed red cell volume can be measuredvery accurately.
9
The hematocrit value gives similar, but notidentical, information on the hematological sta-tus of a subject, as does a hemoglobin or redblood cell count. Currently, the hematocrit isthe simplest and single most accurate measure-ment of anemia or polycythemia available to thehealth profession. If the hematocrit is abnormal,then the hemoglobin and the number of red cellscan be measured for more definitive hematologi-cal studies. Unfortunately, standard values foradolescents are difficult to construct due to thescarcity of studi& that controlled the numerousvariables influencing the normal hematocritvalue in adolescence.
There are many variables that influencehematocrit values in healthy individuals, so thathealthy persons vary widely with respect to theirblood values.l 3 Small differences may be associ-ated with such factors as height, weight, andbody types. Under ordinary circumstances, asmall diurnal variation of about 4 percent fromthe mean hemoglobin has been observed. 14Acute or chronic changes in hydration willinfluence the hematocrit concentration. It hasbeen suggested that muscular activity can alterhematopoietic values. For example, strenuousexercise may lead to accelerated destruction ofred cells; 15 however, sustained exercise byathletes can stimulate the erythropoietic system.Olympic athletes have mean hemoglobins thatvary from 13.7 to 18.6 grams percent.l 6 Be-cause of this, blood values obtained after ex-treme exercise are of dubious significance. Ex-treme excitement resulting in erythrocytosis hasbeen described. 17 On the other hand, climate,temperature, and season have relatively littleinfluence on blood values. 18 Studies reportingon racial differences in erythropoietic values arefew, inconclusive, and deal primarily withadults. 18 ~19 Low barometric pressure, whichincreases anoxemia, is accompanied by a com-pensatory increase in erythrocyte count andthus in the quantity of hemoglobin. The magni-tude of the response depends on the degree,duration, and constancy of exposure to the lowbarometric pressure associated with high alti-tude.d Sustained exposure to high altitudes overmonths or years will lead to an increased
‘A preliminary analysis of unweighed data showed theexpected altitude gradient with a range of approximately 8 mlpercent.
concentration of red cells. 13 The significantphysiologic factors (i.e., excluding diseases) af-fecting the erythrocyte count are (1) extremephysical exercise or excitement, (2) severe dehy-dration, (3) age, (4) sex, and (5) high altitude. zoAge and sex, two of the most sibmificantvariables in this type of study, were carefullycontrolled in this survey. Extreme physicalexercise, excitement, and high altitudes play acasual role in the results in this study. Severedehydration, an uncommon event in an ambula-tory population, is also a most unlikely factorthat could possibly affect the results of theinvestigation.
Finally, most of the data reported in theliterature differ from those of the HES becauseof the omission of those subjects who wereanemic or suffered from recent blood loss, poornutrition, and diseases of the blood. Therefore,those results are not strictly comparable to HESdata, which were obtained from a population atlarge with no specific medical exclusions.
Anemia in some of our sample subjects maysignificantly affect the outcome of these HESdata. (This important variable will be discussedfully under racial differences.) It has beendemonstrated in a recent study that a normaljuvenile population given therapeutic iron for aperiod of 3 months will show an increase inhemoglobin and mean corpuscular hemoglobinconcentration.z 1 However, there was no changein mean hematocrit values, indicating that a mildiron deficiency will not affect mean populationhematocrit values. When using and interpretingHES values, it is important to keep in mind thatthese values represent the hematocrit status forthe entire U.S. population aged 12-17 years andmay or may not represent the optimal ormaximal hematocrit values for this age group.
Age and Sax
Striking variations in erythropoiesis areknown to be associated with age and sex. Inthese HES data, adjustment for age and sex iscritical because the major differences in eryth-rocyte formation between males and femaIesfirst occur during adolescence. In order todemonstrate this, studies need to be carefullydesigned to control for age and sex; most of thestudies in the literature have failed to do this.
10
The effect of age and sex on hematocritduring childhood is not striking. Reportedstudies indicate a slight rise in hematocrit duringchildhood with no difference between boys andgirls. 13 ~~1~Zz During adolescence, girls in thepresent study have a stable hematocrit, withlittle if any change with age (tables 1 and 3).This is also true in the few comparable studiesthat are available in the literature. Z1‘z 3 In sharpcontrast, boys who enter adolescence with ahematocrit similar to that of girls show a steadyincrease with age during adolescence, almostreaching their adult values by the 17th year(tables 1 and 3). The increase in hematocrit inboys appears to reach its maximum between theages of 14 and 14% years. These ages are theperiod of life when many of the boys would beexperiencing the biological expression of pubes-cence. Therefore, more than a casual relation-ship between the events of pubescence andchanges in hematocrit in males is suggested. Ithas already been shown that increment inhemoglobin in adolescent males is closely relatedto the events of pubescence. z A The postulationthat this increase in erythropoiesis in the male isinduced by androgens is compatible with ourobservations on males of same age.zs Thus, thesex difference in hematocrit develops in the13th year in our population and widens per-ceptibly by the 17th year. In a similar andcomparable study of Swedish children and ado-lescents, the significant sex difference inhematocrit also occurred during the 13thye:ir. z 1
By age 15, most teenage girls are menstruat-ing. The daily loss of iron through menstruationranges from 0.1 to 4.0 mg/day, the average being0.7 mg.~ G Under normal living conditions andwith an adequate diet, the teenager is able toreplace these additional iron losses. If the diet isinadequate or if pregnancy with its higher ironrequirement occurs, the teenage female maydevelop anemia.
Race
In one earlier study of healthy adult men, noracial differences in hemoglobin, hematocrit, orerythrocyte count were detected. z 7 Subsequent-ly, Cycle I HES data showed that with theexception of males aged 18-24, white adults had
consistently higher mean hematocrits thanNegro adults.
Differences in blood values between varioussamples of Negro and white children and adoles-cents have been reported recently. Owenz 8 hasreported a significantly lower hemoglobin inyoung Negro children compared with that inwhite children, even though an adequate amountof circulating iron in the serum was demon-strated. Pearson reported a prevalence of anemia(which he defined by a hematocrit value lessthan 34 ml percent) in Negro girls aged 14-17years of 8.0 percent and a prevalence of anemia(as he defined by a hematocrit value less than 36ml percent) in Negro males aged 14-17 years of1.2 percent.z 9 Unfortunately, Pearson’s studyhad no white subjects to be used as a compara-tive reference. Data from the recent l.O-StateNutrition Survey suggest that hematocrit valuesof Negroes are lower than those of white personsat most ages.30 In addition, the study foundthat Negro children 6-12 years of age in Stateswith both low and high income ratios have lowerhematocrits than white children of the same agegroup. Further, it was found that only in Stateswith low income ratios were the hematocritvalues for Negro children 2-5 years of age lowerthan those for white children. Most recently, thepreliminary findings from the Health and Nutri-tion Examination Survey (HANES). of theNational Center for Health Statistics, which havejust been published,a 1 estimate mean hemato-crit values for white youths 12-17 at 42.2 mlpercent and those of Negro youths at 40.2 mlpercent.
In the currently reported HES survey ofyouths 12-17 years of age, racial differencesbetween Negro and white adolescents are veryapparent. Because of the size and nature of thesample, these currently reported HES estimatesare much more reliable than those referred toabove. Both male and female white youths haveconsistently higher mean hematocrit values thanNegro male and female adolescents, with differ-ences between Negro and white youths beingmore apparent in females than in males andmost apparent among older female adolescents(table 7). Descriptively, our findings furthershow th?t a substantially greater percentage ofNegro female youths have hematocrit valuesbelow 35, 37, or 39 ml percent at all ages than
.11
white girls; the same is true, although to a lesserextent, of Negro male youths when comparedwith white males of similar age.
It is unlikely that the racial differences inhematocrit found in our survey can be explainedby anemias resulting from abnormal hemoglo-bin. The prevalence of sickle cell disease, sicklecell trait, and thalassemia, all known to causeanemia, was found to be only 0.24 percent in agroup of Job Corps enrollees. I g This low preva-lence of anemia due to abnormal hemoglobincould not explain the racial differences foundin our population.
One determinant of low hematocrit in adoles-cent girls is the occurrence of pregnancy with itshigher demands for dietary iron at a time whengrowth and onset of menstruation also requireincreased iron. When this demand for increasediron is superimposed on chronically marginalnutritional intakesj excessively low hematocritsmight result. This might be particularly true ofNegro girls, since the rate of pregnancy forfemales of races other than white has beenreported to be 11 times that of white girlsbetween the ages of 10 and 14 years and 2.4times that of white girls between the ages of 15and 19 years.32 This may partially explain theexcessive number of Negro girls with low hemat-ocrits, particularly as age increases.
If we apply Pearson’s criterion for anemia infemales (hematocrit less than 34 ml percent) toHES data, we would estimate that 6.8 percent ofNegro adolescents 14-17 years of age are anemic,which closely compares with a prevalence of b.Opercent found by Pearson. Likewise, applyingPearson’s criterion for anemia in males (hemato-crit less than 36 ml percent), we would extrapo-late from HES data that 1.9 percent of Negroyouths 14-17 years of age are anemic, a preva-lence very close to the 1.2 percent found byPearson for Negro males 14-17 years old. z 9
It is quite apparent from the recent reportsmentioned previously as well as from HES datathat a significant difference exists betweenhematocrit values for Negro and white teenagers.Furthermore, this difference is greater betweenfemales than between males. It is unlikely thatall of the difference is due to an excessiveproportion of cases of anemia among Negroes.There remains an apparent residual racial differ-ence in the comparative percentile distributions
(even at the 50th and 95th percentiles) amongfemales that cannot be satisfactorily explainedby these HES data.
income and Education
There is clear evidence that level of familyincome is related to the level of hematocrit inboth boys and girls (table 13 and figure 4). Thiseffect is much more pronounced at the 10thpercentile of the distribution than at the 90thpercentile (table 8).
In a similar fashion, the level of education isclearly correlated with the level of hematocrit inthe population. This i. apparent at all ages forboth boys and girls (table 15 and figure5). This observation is not unexpected, sinceeducation and family income are known to behighly correlated.
Presumably, the major part of this socioeco-nomic effect lies in the greater amount ofanemia found in lower socioeconomic groups.Perhaps, a rise in income results in betternutrition or medical care and, hence, a decreasein the likelihood of developing anemia. Alterna-tively, there may be a tendency fcir “anemic”families to be less successful economically. Wheninterpreting associations between income andconditions of ill health, there is always thequestion of whether ill health is a result of lowincome or whether it is a cause. The presentdata, however, cannot ascertain which, if either,of these hypotheses is the important determi-nant of the observed socioeconomic differences.
Interaction Between Race and SocioeconomicStatus
An important finding in these data is the lackof a clearcut increase in hematocrit with increasein socioeconomic status among Negro adoles-cents (tables 17 and 19). Even when all the agegroups for females were combined (table A),there appeared to be no consistent socioeco-nomic effect for Negro females. The explanationfor this may be nutritional. That is, even asNegro families rise up the socioeconomic ladder,they may stall maintain their dietary habits ofless affluent periods. Alternatively, there may begenetic or physiological factors that affecthematocrit and are resistant to changes insocioeconomic status. Again, it cannot be deter-
12
mined from the present data which, if any, ofthese speculations is the important determinantof the lack of a strong socioeconomic effectobserved in Negro adolescents.
There was, however, in white adolescentmales and females, a consistent increase in meanhematocrit with increase in socioeconomic sta-tus as measured by either family income (table16) or education of parent (table 18). Sincethere was no age effect in female adolescents, itwas possible to combine age groups for Negroand white females and, hence, to examineNegro-white differences at each level of incomeor education (table A). From this, it appearsthat the differences in mean hematocrit betweenwhite females at the highest socioeconomiclevels (annual family income of $10,000 or moreor education of parent of 13 years or more) andthose at the lowest end (annual family incomeless than $3,000 or education of parent less than8 years) are generally smaller than the differ-ences observed between white and Negro femaleyouths of comparable socioeconomic status.
General Comparability With Other Studies
There are only a few studies in the literaturewith which HES data can be somewhat com-pared. Unfortunately, most of the studies eithercombine boys and girls during adolescence andcalculate a mean value or otherwise constructtables or arrange the data by age in such a waythat makes a comparison with these HES datadifficult.
Dibble,z 3 however, reported a study in whichthe hematocrit values of boys and girls 12-15years of age can be compared with those of HESyouths. The hematocrit values among 12-year-old boys in three different school populationsranged from 38.7 ml percent to 43.0 ml percent;and for 15-year-olds, the range was from 43.3 mlpercent to 45.0. The mean for HES boys at age12?4 was 40.5 ml percent and at age 15%, 43.9ml percent, For 15-year-old girls, the range wasfrom 37.1 ml ~ercent to 40.7. Similarly, themean for HES girls at age 12Y2 was 40.3 mlpercent and at 15%?, 40.7. Thus, the databetween the two studies are quite comparable.
Wlntrobe,l 3normal values,years and then
.& .
a frequently cited reference forgrouped data by age up to 15reported normal values for adult
males and females. At ages 11-15 years, theoverall mean hematocrit is 39.0 ml percent. Hispublished mean values for adult females andadult males are 42.0 and 47.0 ml percent,respectively. The most comparable HES femalesand males (i.e., 17-year-olds) have hematocritvalues of 40.6 and 45.8 ml percent, respectively.Both values are slightly under the values foundby Wintrobe. It appears that HES values wouldbe similar to Wintrobe’s if the age limits of ourpopulation had been extended into theirtwenties. Further comparisons with Wintrobe’snormal data are not valid because of thegrouping of his data.
Albritton’sz z standard values for 12-year-oldboys and girls combined are slightly lower (39.6ml percent) when compared with HES boys(40.5 ml percent) and girls (40.3 ml percent) ofthe same age. For boys and girls 14 years andover, Albritton’s standard values are 47.0 and42.0 ml percent, respectively. HES boys andgirls aged 17 years have a mean hematocrit of45.8 and 40.6 ml percent, respectively; bothvalues are slightly lower than Albritton’s.
Pearsonz g reported the hematocrit values fora large number of Negro children and youths.Again, his grouping of data during adolescencemakes comparisons difficult. For Negro malesaged 14-17 years in the Pearson study, meanhematocrit was 43.2 ml percent, whereas forthose similarly grouped by age in the HES, theoverall mean was 43.6 ml percent. For females1417 years, the mean hematocrit was 38.8 mlpercent compared with that of 38.9 ml percentfor comparable females in the HES. Thus thedata from Pearson’s study and the nationalestimates from the HES for Negroes aged 14-17years are, on comparison, almost identical. Arecent Norwegian study reports hematocrits foreach year of age from 7 through 20 years. z 1 Themean hematocrit for bo ys at age 12 years is 40.7ml percent (40.5 for those in the HES) and at 17years, 45.2 ml percent (45.8 for those in theHES). For girls at age 12, the mean hematocritwas 41.0 ml percent (40.3 for those in the HES)and at 17 years, 41.1 ml percent (40.6 for thosein the HES). The data are remarkably compara-ble, most especially that for boys. Finally, themost recently available comparison of these HESdata is that with the preliminary findings of theHANES data referred to earlier,31 in which the
13
overall mean values are almost identical: In boththe HES and the HANES, mean hematocrits ofmales aged 12-17 years are estimated to be 43.1ml percent, while those for females in theHANES are estimated to be 40.7 ml percent andin the HES, 40.5 ml percent. Thus, the meanvalues of HES hematocrit data compared withsimilar data in the literature on adolescence arein close agreement.
The HES data can also be compared with twostudies of young adults. First Greendyke’s 27
study of 950 healthy enlisted men aged 17-29years (with exclusions for any factor known toaffect normal blood values) found the meanhematocrit to be 45.8 f 2.5 ml percent, againquite close to HES values. Second, current dataon teenagers can be compared with Cycle I datafor adults (tabIe 20).18 It appears from thecombined Cycle I and Cycle III data that theincrease with age in mean hematocrit observedin Cycle III adolescents continues in males 18-24years oId (as observed from Cycle I data), but ata much sIower rate, until “mature” levels ofabout 47.0 ml percent are reached. There thenappears to be a decrease in mean hematocritwith age beginning with the age group 45-54years. In Cycle III females, there was little or noapparent increase in mean hematocrit observedat ages 12-17. The mean hematocrit Ievel for17-year-old adolescent females was 40.6 mlpercent and 41.4 ml percent for females aged18-24. This would be compatible with an annualincrease of 0.2 ml percent beginning with age17. Female hematocrits seem to increase duringadulthood to levels over 43 ml percent amongpostmenopausal women.
It appears from Cycle I data “that meanhematocrit increases slowly throughout theadult life of women to reach levels of 43.7 mlpercent around age 60. The data from Cycle III,
however, indicate little if any evidence ofincrease during adolescence. Therefore, #o ac-count for the difference of 0.8 ml percentbetween 17-year-oId females in Cycle III andthose aged 18-24 years in Cycle I, one wouldhave to hypothesize that the increase observedin Cycle I adult women begins shortly after age18 or that there is a systematic artifact that givesartificially high levels to the Cycle I data. Onewould tend to reject the latter hypothesis,because the laboratory artifact would also haveto be operating on the men, and the observedcurve for men appears completely reasonable.
Finally, the question can be raised concerningthe use of these data. Can the data be used asoptimal physiological standards, or is their userestricted to being descriptive of the adolescentpopulation in the United States during thesampling period? It is obvious that this is arepresentative sample for the United States thatalso includes all the abnormally low and highvalues secondary to disease. It is of interest,however, that the HES results for adolescentsare almost identical to those reported byNatvig.21 His study had one characteristic thatis important to note. After initial measurementsof blood indexes, a group of Norwegian childrenand adolescents were given elementaI iron for 3months. During this period, there was a signifi-cant rise in hemoglobin and mean corpuscularhemoglobin concentration, indicating mild irondeficiency in his “normal” population. At thesame time, there was no change in their hemato-crit. Therefore, the hematocrit may be normal inmild iron deficiency. The suggestion that HESdata for adolescents may be quite close to oreven representative of the optimal hematocritvalues for adolescents could be supported by thesimilarity between the Norwegian and HES dataon hematocrits during adolescence.
14
lNational Center for Health Statistics: Origin, program, and
Opcraticm of the U.S. National Health Survey. Vital and HealthStatistics. PHS Pub. No. 1000-Series I-No. 1. Public HealthScmicc, Washington. U.S. Government Printing Office, Sept.1969.
—
2Natirmaf Center for Health Statistics: Plan and initiafprob~arn of the Health Examination Smvey. Vital and HealthStatistics. !kks 1, No. 4. DHEW Pub. No. (HRA) 741038.H~dth Resources Administration. Washington. U.S. GovernmentPrinting Office, Nov. 1973.
3Nationaf Center for Health Statistics: Plan, operation, andresponse results of a program of chlklren’s examinations. Vitaland Hculth Statistics. Series 1, No. 5. DHEW Pub. No. (HSM)73-1251. Health Services and Mental HeaMs Administration.Washington. U.S. Government Printing Office, Oct. 1967.
‘$Wtional Center for Health Statistics: Plan and operationof .r Hcdth Examination Survey of U.S. youths 12-17 years of
We. ~rltaland Health Statistics. PHS Pub. No. 10oo.senes I.No.
8. Public Hwdth Service. Washington. U.S. Government PrintingOffice, Sept. 1969.
‘N,ition.d Center for Heafth Statistics: Height and weight ofyouth$ 12-17 ywrm, United States. Vital and Health Statistics.Scriw 11, No. 124. DHEW Pub. No. (HSM) 73-1606. HealthScn ices :md Mental Heafth .4dministration. Washington. U.S.Govcmmmrt Printing Office, Jan. 1973.
6Nution~i Center for Heafth Statistics: Body Weight, stat-
ure, tmd sitting height: white and NeLgroyouths 12-17 years,Unitmi St~tes. Vital and Health Statisti-cs. Series 11, No. 126.DHEW Pub. No. (HRA) 74-1608. Heafth Resources Administra-tion. Wilshi.ngton. U.S. Government Printing Office, Aug. 1973.
7N~tiunuf Center for Health Statistics: Blood pressure levelsof chMrcm 6-11 years: Relationship to age, sex, race, andsaciocconomic status, United States. V7taland Health StatRtics.Series 11, NCI. 135. DHEW Pub. No. (HRA) 74-1617. Health
Rcsourws Administration. Washington. U.S. Government print.in~ ONicc, fk. 1973.
8Hmshowitz, M. J.: The American Negro, A Study in RacialCkmin.q. New Y,ork. iilfred A. Knopf, 1928.
%lJSS, B., und Li, C. C.: The dynamics of mcird mtermix-tnrc –.m .mtfysis based on the American Negro. Am.J. Hum.-Gcoftlt. 5(1 ): 1-20, hlarch 1953.
1%ubcrts D. F.: “rhe dymunics of racial intermi..ture in theAmcric:m Negro-some anthropcdogicat considerations. Am.J.-Ifunf. (kwct. 7(4): 361-367, Dec. 1955.
11Robt.rts, D. F., and Hioms, R. W.: The dynamics of raciafintwmixturc. Am. J.Hum. Genet. 14(3):261-277, Sept. 1962.
12Rccd, T. E.: Caucasian genes in American Negroes. Science165(3895):762-768, Aug. 22, 1969.
13Wiitrobe, M. M.: Clinical Henratologv, ed. 6. Philadelphia.Lea and Febiger, 1967, p. 91.
14Brown, A., and Goodafl, A. L.: Normrd variations in bloodhemoglobin concentration. J.Physiol. 104(4):404407, Apr.1946.
15Kmonen, M. J., and Kunnas, M.: Erythrocyte and
hemoglobin changes during protracted heavy muscular work.Ann. Med.Exper.etBiol.Fennr2te 30:180-185, 1952.
16Berry, W. T. C. et af.: The diet hemoglobin vrdues, andblood pressures of Olympic athletes. Br.Med.J. 1:300-304, Feb.19, 1949.
17Wintrobe, M. M.: The erythrocyte in man. Medicine9(2):195-255, May 1930.
18Nation~ Center for Health Statistics: Mean blood hemato-
cnt of adults, United States, 1960-1962. Vital and HealthStatistics. Series 11, No. 24. DHEW Pub. No. (HShl) 73-1080.Health Services and Mental Herdth Adrninktmtion. Washington.U.S. Government Printing Office, Apr. 1967.
19Field@, J., Batilden, P., Tolbert, G., Bennett, R.> ad
Nelson, S. H.: A coordinated sickle celf program for economical-ly disadvantaged adolescents. Am.J.Publ 64(5):427-432, hlay1974.
%fiale, J. B.: Laborato~ Medicine –Henratology. St. Louis.C. V. Mosby CO., 1958, p. 296.
21Natvig, H. et al.: Studies on hemoglobin vafues in Norway.ActaMed.Stand. 182:183, Aug. 1967.
22A1bntton, E. C. (cd.): Standard Values in Blood. Phila-delphia. W. B. Saunders Co., 1952.
23Dibble, M. V. et af.: Some preliminary biochemicalfindings in junior high school children in Syracuse and OnondagaCounty, New York. Anr.J. CZin.Nutr. 17(4): 218-239, Oct. 1965.
2’%oung, H. B.: Aging and adolescence. Dev.Med. ChildNeurol. 5:451, 1963.
25 Kennedy, B. J., and Gdbertsen, A. S.: Increased erythro-poiesis induced by androgenic-hormone therapy. N.EngLJ.Med.
4256 16):719-726, Apr. 18, 1957.
6World Health Organization: Nutritional Anaemias. WHOTech.Reposer., No. 405, Genev~ 1968.
27Greendyke, R. M. et al.: A suggested revision of normafvalues for hemoglobin, hematocrit, and erythrocyte count inhealthy adult men. Techn. Bull. Regist. Med. Techn.
k32(3 .41-48, Mar. 1962.
2 Owen, G., Lubin, A. H., and Garry, Philip J.: Preschoolchildren in the United States: Who has iron deficiency?J.Pediatr. 79(4):563-568, Oct. 1971.
29Pearson, H. A., McLean, F. W., and Brigety, R. E.: Anemiarelated to age: Study of a community of young black Americans.JAMA 215(12):1982-1984, Mar. 22, 1971.
15
30Center for Disease Control: Ten-State Nutrition Survey inthe United States, 1968-1970. IV. Biochemical. DHEW Pub. No.(HSM) 72-8132. Atlanta. Health Services and Mental HealthAdministration. 1972.
31National Center for Health Statistics: Preliminary Findingsof the First Health and Nutrition Examination Survey, UnitedStates, 1971-1972: Dietary Intake and Biochemical Findings.DHEW Pub. No. (HRA) 74-1219-1. Health Resources Adminis-
tration. Washington. U.S. Government Printing Office, Jan.1974.
32National Center for Health Statistics: Teenagers: Mar-riages, divorces, parenthood, and mortality. Vital and HealthStatistics. Series 21, No. 23. DHEW Pub. No. (HRA) 741901.Health Resources Administration. Wash@on. U.S. GovernmentPrinting Office, Aug. 1973.
33Nationaf Center for Health Statistics: Sample design andestimation procedures for a national health examination survcyof children. Vital and Health Statish”cs. Series 2, No. 43. DHEWPub. No. (HRA) 741005. Health Resources Administration.Washington. U.S. Government Printing Office, Aug. 1973.
3%ational Center for Health Statistics: Cycle I of the HealthExamination Survey: Sample and response, United States,1960-62. Vital and Health Statistics. PHS Pub. No. 1000-Series1l-No. 1. Public Health Service. Washington. U.S. GovernmentPrinting Office, Apr. 1964.
35Nationaf Center for Health Statistics: Replicatiorx An
approach to the analysis of data from complex surveys. Vital andHealth Statistics. PHS Pub. No. 1000-Series 2-No. 14. PublicHealth Service. Washington. U.S. Government Printing Office,Apr. 1966.
LIST OF DETAILED TABLES
Table 1,
2.
3.
4.
5.
6.
7.
8.
9.
10,
11.
12.
13.
14.
15.
16.
17.
Estimated number of youths aged 12-17 years in the population by hematocrit group and selected characteristics of
hematocrit distribution, by sex and age: United States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . .
Percent distributions of youths aged 12-17 years by hematocrit intervals, according to sex and age: United States,
1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mean hematocrit, standard deviation, standard error, and percentiles for youths aged 12-17 years, by sex and by age
in%year intervals: United States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Means, medians, differences between means and medians, differences between medians and 5th percentiles,
differences between 95th percentiles and medians of hematocrits, and a ratio for youths aged 12-17 years by sex and
age: United States, 1966-70.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mean hematocrits and standard errors of themeans for youths aged 12-17 years, bygaographic region, sex, and age:
United States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Number of youths aged 12-17years inthesample of the Health Examination Survey, bygeographic region, sex, and
age: United States, 1966-70.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mean hematocrits and standard errors of the means for youths aged 12-17 years, by race, sex, and age: United
States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ditierences between white and Negro youths aged 12-17 years with respect to selected percentiles of hematocrit
levels: UnitedStates,1966-70. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Number of youths aged 12-17 years in sample of the Health Examination Surveyrby race, sex, and age: United
States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Estimated number of white youths aged 12-17 years in the population by hematocrit group and selected
characteristics of hematocrit distribution, bysexand age: United States, l966-7O . . . . . . . . . . . . . . . . . .
Estimated number of Negro youths aged 12-17 years in the population by hematocrit group and selected
characteristics of hematocrit distribution, bysex and age: United States, 1966-70 . . . . . . . . . . . . . . . . . .
Percent of youths with hematocrit determinations below certain levels, byage, sex, and race . . . . . . . . . . . .
Weighted sample size, mean hematocrits, and standard errors of the means for youths by age, sex, and normal family
income: UnitedStates, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10th, 50th, and 90th percentiles of hematocrit distributions for youths by age at last birthday and sex for the
following income groups: lessthan $4,000 and$lO,OOOOr more: United States, 1966-70 . . . . . . . . . . . . . .
Weighted sample size, mean hematocrits, andstandard errors of themeans foryouths, byage, sex, and education of
parent: UnitedStates, 1966-70.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Weighted sample size, mean hematocrits, andstandard errors of themeans forwhite youths, byage, sex, and annual
familyincome: UnitedStates, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Weighted sample size, mean hematocrits, and standard errors of the means for Negro youths, by age, sex, and annual
family income: UnitedStates, 1866-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Page
19
20
21
21
22
22
23
23
24
25
26
27
28
28
29
29
30
17
LIST OF DETAILED TABLES–Con.
Page
Table 18. Weighted sample size, mean hematocrits, and standard errors of the means for white youths, by age, sex, and
education ofparent: United States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
19. Weighted sample size, mean hematocrits, and standard errors of the means for Nagro youths, by age, sex, and
adulation ofparent: United States, 1966-70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
20. Mean hematocrit byageand sex for HESCycle llland Cycle ldata . . . . . . . . . . . . . . . . . . . . . . . . . 31
18
Table 1. Estimated number of vouths aaed 12-17 years in the population by hematocrit group and selected characteristics of hematocrit distribution,by sex and age: United States, 1886-70
IMale Female
12 13 14 15 16 17 12 13 14 15 16years years years years years years years years years Vears years
Selected variable
.17
years
Estimated number of youths in POPUlation in thousandsHematocrit group
1.970 1,946
32765
352588466286101
4811
YwE!All groups . . . . . . . . . . . . . . .
Undw29.0ml percent . . . . . . . . . . .
2,032 2,008 E!-l-E! 1,836 1.764
3
1073
248
335
132322465
489305173
243
2
1681
208364
454436
3
73171
24255543432211844
9
335
2669
433608410251
5630
5
39
1878
343561427
257112
296
6
33
122091
343537434
32
102193
278521410
3036536
4
35.837.339.240.7
42.7
44.245.3
17.5040.6
2.84.16
81057
135
3864794501734915
1894
391
22
15130328652487
2978924
3
36.037.339.140.642,4
44.145.2
12.5440.5
2.87.17
29.0-30,9 ml percent31.0-32.9 ml percent
33.0-34.9 ml percent35.0.36.9 ml percent37.0-38.9 ml percent39.0-40.9 ml percent41. C-42.9 ml percent43.044,9 ml percent45.0-46.9 ml percent47.0.48.9 ml percent49.0.50.9 ml percent51.0.62.9 ml percent53.0-54.9 ml percent55.0-56.9 ml percent
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
496 608554
376159
7116
37.138.240.041.843.3
45.248.3
13.4941.5
2.86.20
506
305 26839
63
58
193
36.237.338.640.4
42.4
44.144.7
16.4940.4
2.73.18
2009927
33657. Omlpercentandwer . . . . . . . . . .
Percentile
6th . . . . . . . . . . . . . . . . . . . . . . . .10th . . . . . . . . . . . . . . . . . . . . . . .26th, . . . . . . . . . . . . . . . . . . . . . .both, . . . . . . . . . . . . . . . . . . . . . .76th, . . . . . . . . . . . . . . . . . . . . . .
90th . . . . . . . . . . . . . . . . . . . . . . .95th . . . . . . . . . . . . . . . . . . . . . . .
Mean age at time of examination(inyears) . . . . . . . . . . . . . . . . . .
Mean hematocrit inmlpercenf . . . . . .Standard deviation inml percent . . . . .Standard errorofmean . . . . . . . . . . .
Hematocrit in ml percent
38.1 40.341.743.245.0
47.1
48.849.9
16.4744.9
2.99.21
41.242.444.246.1
48.0
49.650.3
17.5145.8
2.82.16
36.837.638.940.4
42.3
43.444.3
12.5340.3
2.37.17
37.037.839.240.7
42.6
44.345.8
13.4840.7
2.80.16
36.737.638.740.3
42.2
44.044.9
14.4940.3
2.77.18
38.840.142.144.2
46.1
48.349.7
15.4943.9
3.36.21
36.737.739.140.6
42.7
44.545.7
15.5140.7
3.00.19
39.240.943.0
45.1
47.047.8
14.5142.7
2.92.17
19
Table2, Percent distributions of youths aged 12-17 years byhematocrit intervals, according tosex and age: United Statas, 1966-70
Hematocrit group
All groups , . . . .
Under 29.0 ml
percent . . . . . . . . .
28.0-30.9 ml percent . .
31.0-32.9 ml percent . .
33.0-34.9 ml percent .,
35.0-36.9 ml percent . .
37.0-38.9 ml percent . .
39.0-40.9 ml percant . .
41.0-42.9 ml percent . .
43.0-44.9 ml percent . .
45.0-46.9 ml percent . .
47.0-48.9 ml percent . .
49.0-50.9 ml parcant . .
51.0-52.9 ml percent . .
53.0-54.9 ml p¢ . .
55.0-56.9 ml percant . .
57.0 ml percent
And over . . . . . . . .
Male Female
12 13 14 15 16 17 12 13 14 15 16 17
years years years years years years years years years years years years
Percent distribution
100.0 I 100.0 I 100.0
.1
.1
.7
6.4
16.2
32.1
24.0
14.6
4.4
1.2
.1
.1
.5
3.6
12.4
24.7
27.6
18.7
7.9
3.5
.8
.2
1.8
6.8
16.5
23.8
25.1
15.6
8.9
1.2
.2
100.0
.1
.8
4.3
10.9
19.2
23.9
22.9
10.5
5.2
1.4
.2
.2
.3
100.0
.2
.4
1.7
3.9
13.2
30.2
23.6
17.5
6.4
2.4
.5
100.0
.5
.6
3.2
7.7
21.9
27.2
25.5
9.8
2.8
.9
100.0
.9
4.8
19.8
30.9
25.7
15.5
2.0
.3
.2
100.0
.2
1.4
3.3
18.1
30.2
23.9
14.7
5.2
2.5.
.6
100.0
.2
.2
.3
1.4
3.6
22.8
32.0
21.6
13.2
2.9
1.6
.3
100.0
.2
.5
1.0
4.2
18.5
30.3
23.1
13.9
6.1
1.6
.3
.3
100.0
.2
.2
.7
1.1
5.1
19.2
30.0
24.3
15.0
3.1
1.1
.2
100.0
.2
.1
.6
1.2
5.3
15.9
29.8
23.5
17.4
3.7
2.1
.2
Note. –Percents may not add to 100.0 due to rounding.
20
Table 3, Mean hematocrit, standard deviation, standard error, and percentiles for youths aged 12-17 years, by sex and by age in %-year
intervals: United States, 1966-70
F
40.4
40.5
40.9
41.5
41.9
42.7
43.3
44.0
44.4
44.9
45.3
45.8
46.0
39.9
40.3
40.7
40.8
40.6
40.2
40.2
40.8
40.5
40.4
40.6
40.5
40.6
Percentile
25
Sex 5X ST
3.11 .47
2.52 .15
2.85 .24
2.85 .19
2.71 .20
2.87 .22
3.19 .15
3.24 .23
3.41 .27
2.79 .24
2.83 .?0
3.14 .22
2.28 .25
2.16 .22
2.45 .21
2.45 .18
2.82 .21
2.85 .20
2.77 .27
3.03 .19
3.03 .28
2.77 .17
2.64 .2a2.67 .24
2.90 .20
3.33 .37
5 10 50 75 90 95
Male
12yews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12!4 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13!: years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14ytxrrs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14)L years, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15Xycars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16% years, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17ycars, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17!4 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35.2
36.7
36.6
37.0
38.0
38.1
38.4
39.0
39.4
40.4
41.4
40.7
42.8
36.4
36.2
37.2
37.0
36.2
37.1
35.9
37.2
36.2
36.4
36.3
35.8
35.7
35.7
37.6
37.9
38.1
38.7
39.2
39.4
40.2
41.0
41.7
42.3
41.9
43.2
37.5
37.5
38.0
37.8
37.7
37.7
37.2
38.1
37.2
37.3
37.7
37.2
37.2
38.9
39.1
39.3
40.0
40.2
41.1
41.3
42.2
42.7
43.2
43.4
44.1
44.7
38.6
38.8
39.3
39.1
38.9
38.6
38.6
39.1
38.8
38.9
39.2
38.9
39.5
40.5
40.4
41.1
42.0
42.2
43.1
43.2
44.6
44.8
45.1
45.3
46.1
46.3
40.2
40.4
40.8
40.8
40.3
40.2
40.3
40.6
40.4
40.3
40.8
40.4
41.1
42.7
42.3
43.0
43.4
43.8
45.0
45.3
46.0
47.1
47.1
47.5
48.1
48.0
41.6
42.4
42.4
42.6
42.6
42.2
41.9
42.7
42.6
42.4
42.6
42.6
42.7
44.4
44.1
44.7
45.1
45.8
46.7
48.1
48.2
48.3
48.7
49.1
50.2
48.9
42.8
43.4
43.8
44.3
44.6
43.8
44.1
44.8
44.2
44.1
44.1
44.3
44.3
45.7
44.8
45.7
47.0
47.0
47.6
49.1
49.9
49.3
49.7
50.5
50.8
49.7
43.5
44.2
45.1
45.7
45.7
44.4
45.6
45.8
45.1
44.9
44.4
45.7
45.6
Female
12yedrs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12!4 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13ytx.ws. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13?4years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14yoars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14% years . . . . . . . . . . . . . . . . . . . .’ . . . . . . . . . . .
15yews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15!4 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16’A years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17%years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 4. Means, medians, differences between means and medians, differences between medians and 5th percentiles. differences between 95th
percentiles and medians of hematocrits, and a ratio for youths agad 12-17 years by sex and age: “Unitad States, 1966-70
Male Female
12
years
40.5
40.6
-0.1
4.6
4.6
1.00
13
years
41.5
41.8
-0.34.7
4.5
0.86
14
years
42.7
43.0
-0.34.9
4.8
0.98
15
years
43.9
44.2
-0.35.4
5.5
1.02
16
years
44.9
45.0
-0.1
4.7
4.9
1.04
17
years
45.8
46.1
-0.3
4.9
4.2
0.86
12
years
40.3
40.4
-0.1
3.8
3.9
1.03
13
yeers
40.7
40.7
0.0
3.7
5.1
1.38
14
years
40.3
40.3
0.0
3.6
4.6
1.28
15years
40.7
40.6
+0. 1
3.9
5.1
1.31
Selected variable16
years17
years
Mcan(ml percent) . . . . . . . . . . . . . .
Median (ml percent) . . . . . . . . . . . . .
Mean-median (mlpercent) . . . . . . . . .
50th.5th percentile (ml percent) . . . . .
95th-50th percentile (ml percent) . . . .
40.440.4
0.04.2
4.3
1.02
40.6
40.7
-0.1
4.9
4.6
0.94
~atiO 95th.50th
50th-5th’”’”’”””””””””””
21
Table 5. Mean hemetocrits and standard errors of the means for youths aged 12-17 years, by geographic region, sex, and age:United States, 1966-70
Total Northeast Midwest South West1 1
Sex and age Total I I Northeast Midwest I South\ West
Male Mean hematocrit in ml percent Standard error in ml percent
40.0
41.1
42.7
43.6
44.6
45.6
40.4
40.4
39.9
40.5
39.7
39.9
40.7
41.8
42.9
44.4
45.8
46.2
40.1
41.2
40.4
40.9
40.2
40.6
.17
.20
.17
.21
.21
.16
.17
.16
.18
.19
.18
.16
.18
.30
.26
.44
.26
.28
.38
.27
.29
.45
.25
.23
.37
.55
.27
.31
.22
.40
.28
.19
.43
.24
.22
.33
.22
.19
.47
.38
.33
.32
.27
.37
.31
.22
.57
.42
.65
.52
.48
.65
.75
.44
.31
.43
.61
.57
.53
.46
12 years . . . . . . . ...<.....
13years . . . . . ..i . . . . . . . .
14 years . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . .
16 years . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . .
40.5
41.5
42.7
43.9
44.9
45.8
40.3
40.8
40.3
40.7
40.4
40.6
40.7
41.4
42.7
43.9
44.9
45.9
40.5
40.8
40.3
40.5
40.6
40.5
40.5
41.5
42.6
43.9
44.3
45.5
40.3
40.6
40.5
40.8
40.941.2
Female
12years . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . .
Table6. Number of youths aged 12-17 yearsin the sample of the Health Examination Survey, by geographic region, sex, and age:
United States, 1866-70
11 1 1 1
Total Northeast ‘ Midwest South westSex and age
Male— Number of youths
922 890
m’
3,!%15
643
626
618
613
556
489
3,223
547
582
586
503
536
469
870
173
162
156
137
124
118
771
Allages, 12-17years . . . . . . . . . . . . . . . . . . . . . . . . .
12years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17years . . . . . . . . . . ..I. . . . . . . . . . . . . . . . . . . . . . . . 116 130 125
835 814 803
Female
All ages, 12-17years . . . . . . . . . . . . . . . . . . . . . . . . .
135
145
139
121
125
106
150
157
148
108
140
132
128
142
161
139
126
118
134
138
138
135
145
113
12years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
22
Table7. Mmn hematocrits and standard errors of the means for youths aged 12-17 years, by race, sex, and age: United States,
1966-70
II I
Sex and age
Male
12ycwrs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13ymtrs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14yecirs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15ytmrs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16 years . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17ycmrs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Female
12yews . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14ytx3rs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . .. . . . . . . . . . . . . . . . . .
17y0ws . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Total White Negro
Mean hematocrit in ml percent
40.5
41.5
42.7
43.9
44.9
45.8
40.3
40.7
40.3
40.7
40.4
40.6
40.6
41.5
42.9
44.1
44.9
45.8
40.4
40.8
40.4
40.9
40.7
40.9
39.6
41.4
41.7
42.9
44.4
45.7
39.740.2
39.4
39.3
38.3
38.5
Total White Negro
Standard error in ml percent
.17
.20
.17
.21
.21
.16
.17
.16
.18
.19
.18
.16
.19
.21
.19
.25
.24
.16
.19
.18
.19
.18
.17
.17
.35
.41
.27
.32
.43
.47
.21
.29
.46
.39
.47
.51
TotIIe8. Differences between white and Negro youths aged 12-17 years with respect to selacted percentiles of hematocrit
levels: United States, 1966-70
Percentile in white youths minus percentile in Negro youths
Percentile IMale I Female
12 13 14 15 16 17 12 13 14 15 16 17
years years years years years years years years years years years years
5th . . . . . . . . . . . . . . . . . . . 1.2 1.7 1.5 1.7 0.6 0.8 0.6 3.1 2.1 3.8 4.4 3.9
both . . . . . . . . . . . . . . . . . . 0.9 -0.3 1.0 0.6 0.4 0.2 0.5 0.6 0.8 1.3 2.4 2.0
95ih . . . . . . . . . . . . . . . . . . -0.1 -1.1 0.4 -0.2 0.1 -0.1 0.7 -1.6 0.3 0.5 1.1 0.6
23
Table9. Number of youths aged 12-17 years in sample of the Health Examination Survey, by race, sex, and age: United States,
1866-70
Sex and age
Male
Allages, 12-17 years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Female
All ages, 12-17years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
II I
‘eta’ II ‘hi’e I ‘egro
3,545
643
626
618
613
566
489
3,223
547
582
586
603
536
469
Number of youths
3,047
540
542
527
525
496
417
2,688
455
490
484
425
441
393
479
10180
88
84
57
69
520
88
91
101
73
93
74
24
Table IO. Estimated number of white youths aged 12-17 years in the population by hematocrit group and selected characteristics of hematocrit
distribution, bysex andage: United States, 1866-70
White male White female
12 13 14 15 16 17 12 13 14 15 16 17years years years years years years years years years years years years
Selected variable
Hemamcrit group Estimated number of youths in population in thousands
1,542 I 1,502All groups . . . . . . . . . . . . . . . 1,747 1,729 1,686 1,646 1,594 1,528 1,685 1,667 1,633 1,584
Under 29. Oml percent . . . . . . . . . . .
29.0-30.9 ml percent . . . . . . . . . . . .31. O-32.9 ml percent . . . . . . . . . . . .33. O-34.9 ml percent . . . . . . . . . . . .35.0-36.9ml percent . . . . . . . . . . . .37.0-38.9ml percent . . . . . . . . . . . .39.0.40.9ml percent . . . . . . . . . . . .41.0-42.9ml percent . . . . . . . . . . . .43.0.44,9ml percent . . . . . . . . . . .46.0-46,9ml percent . . . . . . . . . . . .47.0-48,9ml percent . . . . . . . . . . . .49.0.60.9mi percent . . . . . . . . . . . .61.0.52.9ml percent . . . . . . . . . . . .63.0-54.9ml percent . . . . . . . . . . . .56.0.66.9ml percent . . . . . . . . . . . .57.0mlpercentand ovar . . . . . . . . . .
10101252575445264
7324
3
36.437.6
39.3
40.742.544.145.2
12.5440.6
2.59
0.198.84.54.3
67200434513319133
5111
37.238.2
40.0
41.843.245.146.2
13.4941.5
2.68
0.219.04.44.6
2388
289410409296146
223
38.339.5
41.1
43.145.147.048.0
14.5142.9
2.s4
0.199.74.94.8
1049
1813264093771768424
336
39.340.3
42.2
44.246.248.349.7
15.4944.1
3.20
0.2610.4
5.54.9
32461
198497394272
9937
9
41046
1123344273881494215
1276
319512439282
366
3
1148
298511425239
9339
3
35
1744
3555483462334830
2
37.137.7
38.8
40.342.244.144.9
14.4940.4
2.60
0.197.84.63.2
56
55274495373245101
29
6
4
37.138.0
39.3
40.742.844.645.7
15.5140.9
2.86
0.188.65.03.6
55
58266477407
41253
222446381
2535614
3
37.137.9
39.3
40.742.644.244.8
16.4940.7
2.47
0.177.74.13.6
2905536
4
37.037.7
39.7
41.143.044.245.4
17.5040.9
2.64
0.178.44.34.1
Percentile Hamatocrit in ml percent
5th . . . . . . . . . . . . . . . . . . . . . . . .10th . . . . . . . . . . . . . . . . . . . . . . .
2Mh . . . . . . . . . . . . . . . . . . . . . . .
50th . . . . . . . . . . . . . . . . . . . . . . .76th . . . . . . . . . . . . . . . . . . . . . . .90th . . . . . . . . . . . . . . . . . . . . . . .95th . . . . . . . . . . . . . . . . . . . . . . .
40.641.9
43.2
45.047.148.749.9
16.4844.9
2.79
0.249.34.94.4
41.242.4
44.2
46.147.949.650.3
17.5045.8
2.79
0.169.14.24.9
36.837.7
39.1
40.542.443.544.3
12.5440.4
2.37
0.197.53.83.7
37.238.0
39.3
40.842.644.245.7
13.4940.8
2.62
0.188.54.93.6
Mean age at time of examination
(inyears) . . . . . . . . . . . . . . . . . .Mean hematocrit inml percent . . . . . .
Standerd deviation in ml percent . . . . .
Stendard errorofmeen . . . . . . . . . . .95th-5th percentile . . . . . . . . . . . . . .96th-50th percentile . . . . . . . . . . . . .50th.5th percentile.. . . . . . . . . . . . .
Table If. Estimated number of Negro youths aged 12-17 years in the population by hematocrit group andsalected characteristics of hematocrit
Selected variable
Hematocrit group
All groups . . . . . . . . . . . . . . .
Under 29. Oml percent . . . . . . . . . . .29. O-30.9ml percent31.0-32.9 ml percent33.0-34.9 ml percent
35.036.9 ml [email protected] [email protected] ml percent41.0-42.9 ml percent43.0-44.9 ml percent45.0-46.9 ml percent47.&46.9 ml percent49.0-50.9 ml percent51.0-52.9 ml percent
53.0-54.9 ml percent55. C-EJ5.9ml percent57.0 ml percent and over . . . . . . . . . .
Percentile
5th . . . . . . . . . . . . . . . . . . . . . . . .10th . . . . . . . . . . . . . . . . . . . . . . .25th . . . . . . . . . . . . . . . . . . . . . . .50th . . . . . . . . . . . . . . . . . . . . . . .75th . . . . . . . . . . . . . . . . . . . . . . .
both . . . . . . . . . . . . . . . . . . . . . . .95th . . . . . . . . . . . . . . . . . . . . . . .
Mean age at time of examination
(inyears) . . . . . . . . . . . . . . . . . .Mean hematocrit inml percent . . . . . .Standard deviation inml percent . . . . .Standard errorofmean . . . . . . . . . . .95th-5th percentile . . . . . . . . . . . . . .
95th-50th percentile . . . . . . . . . . . . .5&h-5th percentile . . . . . . . . . . . . . .
distribution, bysex and age: United States, 1886-70
Negro male Negro female
12 13 14 15 16 17 12 13 14 15 16 17years years years years years years years years years years years years
Estimated number of youths in population in thousands
280
226
287772433316
35.235.938.139.842.043.945.3
12.5539.6
2.980.35
10.16.5
4.6
262
3
10
6485142572616
4
35.5
37.338.942.144.245.4
47.3
13.4641.4
3.79
0.4111.8
5.2
6.6
256 ] 241
=i-
3 -
12 644 3233 2852 3875 46
9 5527 18
2 123
36.838.039.542.143.9
47.147.6
14.52
41.73.260.27
10.8
5.5
5.3
37.638.540.343.645.448.149.9
15.49
42.94.140.32
12.36.3
6.0
231
3
44
114358404219
8
225
3
122345535925
7
Hematocrit
40.0
41.042.744.647.249.149.8
16.4544.4
4.000.439.85.2
4.6
40.4
42.443.445.948.250.0
50.4
17.57
45.73.0:0.47
10.0
4.55.5
272
6
1969936023
3
275
ml percent
316
1754774247
597
36.2
37.138.340.041.542.843.6
12.48
39.72.320.217.4
3.6
3.8
34.1
35.738.340.242.444.447.3
13.4340.2
3.620.29
13.2
7.1
6.1
266
3
9
2675606418
8
3
35.0
36.138.139.541.443.644.6
14.46
39.43.490.469.65.1
4.5
33
12
23
6458461211
2
33.3
35.237.739.441.343.346.2
15.5039.3
3.620.39
11.95.8
6.1
243
337
15
3276612713
5
32.7
34.737.038.340.242.3
43.7
16.4738.3
3.280.47
11.0
5.4
5.6
237-
3269
40
5772251310
33.1
35.136.739.140.642.8
44.8
17.50
38.53.820.50
11.7
5.7
6.0
26
Table 12. Percent ofyouths,with hematocrit determinations below certain levels, byage, sex, and race
Age and race
Total
12ytmrs . . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . . . . .
17 years . . . . . . . . . . . . . . . . . . . .
White
12ycars . . . . . . . . . . . . . . . . . . . .
13yews. . . . .. m. . . . . . . . . . . . .
14 years . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . .
16yuars . . . . . . . . . . . . . . . . . . . .
17 years, . . . . . . . . . . . . . . . . . . .
Nearo
12years, . . . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . . . . .
14 years . . . . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . . . . .
16 years, . . . . . . . . . . . . . . . . . . .
17 years . . . . . . . . . . . . . . . . . . . .
Percentage of males with hematocrit
determinations below:
35.0 percent
0.9
0.6
0.2
0.1
0.2
0.6
3.6
5.0
1.2
0.8
1.3
37.0 percent
7.3
4.3
1.9
0.9
0.5
0.4
6.4
3.9
1.4
0.6
0.2
0.3
13.6
7.3
5.9
3.3
3.0
1.3
39.0 percent
23.5
16.7
8.7
5.2
2.2
1.0
20.8
15.4
6.6
3.6
1.7
0.9
41.1
25.6
23.0
16.6
4.8
1.3
Percentage of females with hematocrit
determinatiorrs below:
35.0 percent
0.9
1.5
1.9
1.6
2.3
2.1
0.7
0.7
1.5
0.7
0.6
1.1
2.2
6.9
4.5
7.7
11.5
8.4
37.0 percent
5.7
4.9
5.6
5.8
7.2
7.5
5,2
3.5
4.2
4.1
4.3
4.6
9.2
13.1
14.3
17.4
24.7
25.3
39.0 percent
25.5
23.0
28.4
24.4
26.4
23.3
24.1
21.4
26.0
21.3
21.5
19.4
34.6
32.7
42.5
44.7
56.0
49.4
r
27
Table 13. Weighted samp[esize, mean hematocrits, andstandard errors of themeans foryouths byaF, sex, andnormal fami[y income: LJnited States, 1966.70
12 years 13 years 14 years 15 years 16 years 17 years
N z *X N z Sx N F Sx N z ST N F % N z Sz.%x and annual family income
Total male
All incomes . . . . . . . .
Lessthan $3,000 . . . . . . . . .
$3,000$4,999 . . . . . . . . . . .$5,000’$9,999 . . . . . . . . . . .$l0,0000rmOre . . . . . . . . .Don<tknmw . . . . . . . . . . . .Blank mrefused . . . . . . . . .
Total female
All incomes . . . . . . . .
Lewthan $3,000 . . . . . . . . .
$3,00M4,999 . . . . . . . . . . .
$5,(XXI.$9,999 . . . . . . . . . . .
$lO,~OrmOre . . . . . . . . .Don’t know . . . . . . . . . . . .Blink orrefused . . . . . . . . .
Hematocrtt in ml percent
1,951 [ 42.7 0.17
0.67
0.48
0.220.220.810.72
0.18
0.310.46
0.25
0.220.420.65
1,900 I 43,9 44.9 0.21 1,764
_ll_
44.6 0.W5 22645.1 0.34 245
44.8 0.35 69245.0 0.25 47944.8 0.85 49
!v!-L!x
I45.3 0.6045.3 0.49
46,0 0,1846.1 0.1744.5 08245.6 0.84
40.6 016
T
39.6 0.4739.6 0.45
40.7 0.27
41,1 0.2840.8 0.5240.7 1.25
2,o32 40.51 0.17 2,0C6 41.5 0.20 !zL.L.E
I0.53 2010.49 265
0.30 7190.24 6160.47 62OS’S 55
0,19 1,789
T0.50 2460.31 166
0.26 650
0.27 5940.68 633,16 60
180308
869575
6546
1,970
257274
722
6007937
139.8 0.38
40.2 0.38
40.6 0.2440.6 0.2140.9 0.5741.2 0.89
40.3 0.17
T
40.2 0.4540.0 0.3540.3 0.25
40.7 0.2139.3 0.6340.4 9.15
212
i64
767619
6580
1,846
268301
802
4944636
40.941.0
41.841.541.341.1
40.7—
40.240.6
40.741.340.341.0
0.44
0.46
0.260.300.531.12
0.16.
0.360.3B
0.15
0.240.501.07
1200 42.2
240 42.5
746 42.7587 42.9100 43.5
67 42.5
1,90140.3
I
213 39.7300 39.9
720 40.4
554 40.772 39.542 39.6
1241 43.4219 44.2
719 43.9598 44.1
76 44.246 43.6
1,851 40.7
T
258 40,2264 40.0
700 40.7
533 41.263 40.032 42.4
44.C 1.07 74
WL?2!hK
IT39.6 0.46 18239.2 0.44 [email protected] 0,17 5s6
40.9 0.27 62841.5 0.53 9240.8 0.62 52
T6ble 14. 10th, 50th, and 90th percentiles of hemtmcritd istributionsfor youths byage8t Iastbirthday andsexfor the following
income groups: [essthan $4,0006 nd$lO,OOOOr more: United States, 1966-70
I I
Distribution at the I Distribution at the I Distribution at the
10th percentile 50th percentile 90th percentile
Age and sex
$10,000 ~ota, Less than $10,000or more $4,000 or more
Less than
$4,000Total
Male
12years . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . . .
Female
12years . . . . . . . . . . . . . . . . .
13years . . . . . . . . . . . . . . . . .
14years . . . . . . . . . . . . . . . . .
15years . . . . . . . . . . . . . . . . .
16years . . . . . . . . . . . . . . . . .
17years . . . . . . . . . . . . . . . . .
Hematocrit in ml percent
40.6 40.041.8 41.043.0 43.044.2 44.245.0 45.046.1 45.1
40.4 40.340.7 40.340.3 39.840.6 40.140.4 40.040.7 40.1
37.338.239.240.141.742.4
37.637.837.637.737.337.3
36.138.238.638.841.241.2
36.436.336.237.236.335.7
40.442.143.144.245.246.3
40.641.240.441.141.141.5
44.145.247.048.348.849.6
43.444.344.044.544.144.2
44.344.647.248.249.749.7
43.444.343,743.844.144.0
43.945.147.149.048.249.8
43.645.244.245.344.244.2
38.138.339.339.742.543.1
38.138.638.138.138.038.2
28
Table 15. Weighted sample size, mean hematocrits, andstindard errors of themeans foryouths, byage, sex, andeducation of parent: United States, lS66-70
12 years 13years 14 years 15 years 16 years 17 years
N x SE N x Sz N x Sz N x s% N z Sx N x SxSIN and education of parent
Hematocrit in ml percentTotal male
All education groups
Lessthm 8 years . . . . . . . . .&ll years . . . . . . . . . . . . . .lzyl!ars . . . . . . . . . . . . . . .13 years or mom . . . . . . . . .Unknown . . . . . . . . . . . . . .
Total female
Alleducmlon groups . .
Lrss than .9 years . . . . . . . . .S.llycms . . . . . . . . . . . . . .
12years . . . . . . . . . . . . . . .
13ymrs0rm0re . . . . . . . . .Unknown . . . . . . . . . . . . . .
42.7 I 0.17 I 1,900 43.9 0.16=
0.570.340.220.1s0.90
0.16=
0.360.190.210.351.28
2,0% 41.51 0.20 I 1,961 0.21 E!?_lJH
T
217 44.5474 44.863o 44.8480 45.1
36 45.7
1,789 40.4
0.21—
0.600.250.240.370.58
0.18—
0.560.28
0.18
0.361.5S3
X2i_!2!
T20S 45.2446 45.8668 45.7388 46.2
54 45.8
1,746 40.6
2,032
176543736%2
16
1,970
211460
729509
41
40.51 0.17—
139.7 0.4840.0 0.2640.8 0.2240.9 0.3139.6 6.96
40.3 0.17
T
39.4 0.4640.2 0.26
40.5 0.21
40.7 0.1839.4 0.56
175502780627
23
1,946
228507727
46322
41.1 0.75 17241.2 0.28 47441.7 0.24 73041.4 0.38 549
41.6 0.71 196
42.8 0.2s 48742.7 0.17 72743.1 0.17 472
43.843.644.144.145.1
40.7—
40.140.1
41.0
41.240.9—
0.480.290.270.352.01
0.19
42.3 0.89 25 41.9 1.69 16
MINl_.W
lT40.1 0.69 16040.4 0.24 53740.9 0.17 71941.3 0.24 465
,. 21
~
II39.6 0.56 176
40.1 0.26 596
40.5 0.23 640
40.5 0.34 411
39.1 1.38 22
—
1167 39.7463 39.7623 40.6
512 40.934 40.7
146 39.7465 40.6525 40.8542 40.8
67 39.6
0.690.23
0.22
0.329.29
—
Table 16. We[ghted sample size, mean hematocrits, andstandard errors of themeans forwhite youths, byage, sex, andannual family income: United States, 1966-70
Stw and annual f~mily income12 years 13 years 14 years 15years 16 years 17 years
N x s% N x Sz N x SK N x Sx N z Sx N x Sz
White male
Al18nc0mes . . . . . . . .
Lcssthrm $3,000 . . . . . . . . .83,000+4,999 . . . . . . . . . . .$5,00C!-$9,999 . . . . . . . . . . .$i0,0000rmOrc . . . . . . . . .Don’t know . . . . . . . . . . . .L31imkorrefuscd . . . . . . .
White female
Allincomcs . . . . . . . .
Ltssstht?n$3,000 . . . . . . . . .
$3,0QM4,999 . . . . . . . . . . .$6,000$9,999 . . . . . . . . . . .
$l0,000ormOre . . . . . . . . .Don’t know . . . . . . . . . . . .
81ank m refused . . . . . . . . .
Hematocrit in ml percent
42.9.
42.342.7
42.843.043.742.8
40.4
40.040.240.4
40.7
39.7
39.7
0.19 I 1.546 44.1 0.25 1,594
T
43.6 0.54 12944.6 0.60 211
44.0 0.32 63244.1 0.24 50144.4 0.38 6743.6 0.94 55
!4J!2z
I0.70 1100.47 1900.26 6620.30 5790.5+ 671.14 57
0.18 1,633
I
0.55 1360.45 2110.97 661
0.24 535
0.59 51
0.99 36
44.9 I 0.24 1,628.
142161
540459
3768
1,502
111146529
591
86
40
45.8=
44.845.745.946.144.145.8
40.9=
40.040.641.0
41.1
41.1
41.6
0.16.
1.220.600.160.170.950.94
0.17
1,747.
111209
7s5557
4143
1,665
1s0195641
571
63
34
40.6.
39.940.740.640.641.541.1
40.4—
40.3
40.140.3
40.8
39.7
40.4
J’221JE
I0.76 1270.46 160
0.25 6810.21 6060.49 580.92 76
0.19 1,667
I0,61 1600.42 2290.27 729
0.23 482
0.63 36
9.17 31
41.5—
41.041.0
41.741.540.941.2
40.6 L0.81 1530.62 151
0.23 6530.23 5780.79 640.82 46
0.19 1,594
I
0,32 183
0.56 2070.26 611
0.23 508
0.50 59
0.7.1 27
44.9 0.6945.4 0.3644.9 0.3645.0 0.2644.7 1.0244.0 1.07
40.7 0.17K!L?LlT40.1 0.62 15540.5 0.41 11340.9 0.27 569
41.3 0.25 571
40.0 0.72 69
40.540.640.7
41.3
40.6
40.6— 1
40.5 0.44
39.7 0.6340.5 0.19
41.0 0.22
41.8 0.52
41.2 0.57
0.6S0.330.27
0.27
0.50
9.3642.9 3.93 44
29
Table 17. Weighted sample size, mean hematocrits, and standard errors of the means for Negro youths, by age, sex, and annual family income: United States, 1‘%6-70
12 years 13 years 14 years 15 years 16 years 17 years
N x SF N x s~ N x SF N ~ SE N ~ SE N ~ Sz
Sex and annual family income
Hematocrit in ml percentNegro male
Al[ incomes . . . . . . . . .
Less than S3,ClXJ . . . . . . . . . .
$3,000$4,999 . . . . . . . . . . . .
$5,000-$9.999 . . . . . . . . . . .$10,0CY30rmcie . . . . . . . . . .Don’t know . . . . . . . . . . . . . .
Blank orrefused . . . . . . . . . . .
Negro female
All incomes . . . . . . . .
Less than $3,000 . . . . . . . . . .
$3,000$4,899 . . . . . . . . . . . .
$5,W$9,899 . . . . . . . . . . . .
$l0,0000rmOre . . . . . . . . . .Don’t know . . . . . . . . . . . . . .
Blank orrefuwd . . . . . . . . . . .
41.7
42.1
41.7
41.8
39.642.0
.
39.4
39.1
39.1
40.540.0
39.0●
—
0.27 [ 241 q43.1
43.1
42.542.6
,
39.3
231 I 44.4262—
85
80
78
9
6
3
275—
108
72
74
9
9
4
41.4—
40.8
41.0
42.3.
●
●
40.2=
39.7
40.4
41.038.8
●
●
0.41—
0.58
1.25
0.73..
●
0.2s=
0.57
1.00
0.652.37
●
●
—
256 0.32=
0.9B
0.68
0.752.10
●
0.39=
0.70
0.69
0.712.12
●
●
0.43—
1.05
0.77
1.18..
0.47=
0.90
0.75
0.492.32
1.51,
226=
85
58
4817
12
6
237=
71
6C
58
3C
6
12
45.7=
46,0
44.3
46,746.7
45.7.
38.5—
39037,2
379
41.7.
37.7
047=
0.53
12!3
0.920.57
1029●
050=
0.60128
0.671.76
●
1.23
280—
69
96
8116
14
2
272=
77
75
8119
16
4
39.6
39.6
39.0
40.440.3
39.2●
39.7
40.0
39.5
40.139.0
●
●
0.35
0.39
0.53
0.76
0.739.24
●
0.21
0.62
0.48
0.270.56
.
●
0.66
0.72
0.920.64
1.84.
88
65
5917
12
89
50
78
15
13
11
266 172 44.0
54 43.9
80 44.710 ‘
15 “. .
243 38.3
T
69 37.852 38.1
57 38.7
23 38.0
14 40.0
7’
0.47 235
7589
5719
21
6
0.63 751.13 52
0.90 811.76 1?0.62
● E
40.4
38.4
39.3
37.1.
●
Table 18. Weighted sample size, mean hematocrits, and standard errors of the means for white youths, by age. sex, and education of parent: United States, 1966-70
12 years 13 years 14 years 15 years 16 years 17 years
N x 5X N z .x N x s~ N x s~ N z s~ N z s~
Sex and education of parent
White male
All educat,on groups .
Lessthan 8years . . . . . . . .
f$ll years . . . . . . . . . . . . . .12 years . . . . . . . . . . . . . . .13years0rm0re . . .Unknown . . . . . . . . . . . . .
Hematocrit in ml pwcent
0.21 I 1,685 I 42,9
T
1.20 145 41.80.30 341 42.90.23 662 42,90.40 623 43.10.95 14 43.6
0.19=
0.830.350.170.169.98
0.19
1,646
139380652461
14
1,5s4
127479578
39120
3!-I-E3
T43.8 0.6243.9 0.3544.1 0.3244.2 0.3346.5 1.64
40.9 0.18
*
T45.1 0.6344.7 0.2844.8 0.2445.1 0.4045.6 0.67
40.7 0.17
E_13?
I119 40.1412 40.1675 40.8526 41.0
13 “
1,685 40.4
I165 39.7379 40.3
650 40.6464 40.6
26 39.5
0.19=
0.76
0.280.230.32
.
0.19=
0.850.320.24
0.190.55
—
1,729
10439t713501
19
1,667
182356
561452
16
41.5—
41.141.44t.641.442.2
40.8=
40.340.341.0
41.3.
1.584
1693715s0446
28
1,642
1,528
143342632363
49
1,502
45.8—
45.1458
45.846.245.7
40.9=
40.541.141.041.139.3
0.16=
0.790.370.220.180.99
0.17
White female
All education groups 0.18 I 1,633 I 40.4
Lessthan 8 years . . . . . . .Ell years . . . . . . . . . . . .
12years . . . . . . . . . . . . .13years0rm0re . . . . . . .Unknown, . . . . . . . . . . .
1.01 107 39.50.33 429 40.20.17 648 40.60.24 440 40.7
● 9 “
0.830.320.24
0.28.
40.0 0.9340.4 0.2141.2 0.2541.3 0.3341.1 9.38
106
349672490
24
40.8 0.65
40.2 0.3+740.6 0.1841.1 0.2841.3 1.63
96
344481516
55
0.750.24
0.210.341.67
30
Table 19. Weighted sample stze, mean hematocrits, and standard errors of the means for Negro youths, by age, sex, and education of parent: United States, 1966-70
SW and cducaticm of parent
12 years 1 13 years I 14 years I 15 years I 16 vears 17 vears
Hematocrit in ml percentNegro knde
All t?ducatinn groups . . . .
Lmsthan8yaws . . . . . . . . . . .
,%llyuars . . . . . . . . . . . . . . . .
12ycws . . . . . . . . . . . . . . . . .
13 vticwsw more . . . . . . . . . . .Unhrmwn. . . . . . . . . . . . . . . .
Neqru female
All educatitm groups . . . .
Lwsthan8vears . . . . . . . . . . .S. Ilvmrs . . . . . . . . . . . . . . . .
12yews . . . . . . . . . . . . . . . . .
13yimrs0rm0re . . . . . . . . . . .
(.hknown . . . . . . . . . . . . . . . .
41.8
40.6
42.5
40.9
42.4*
39.4
39.6
39.6
40.1
37.4●
0.27
9.28
0.45
0.48
1.13,
0.46
0.82
0.59
0.64
3.06●
N-135
I
57 43.5
103 42.4
69 43.8
8“
3“
235 39.3
I
47 40.6109 39.0
63 39.0
14 39.3
3’
28o I 39.7 0.35
II
57 38.9 0.44
130 39.7 0.60
68 40.6 0.44
32 39.1 1.18
3*”
272 39.7 0.21
*
I41.2 1.32
40.7 0.72
41.9 1.00
43.5 0.98.*
40.3 0.29
I
39.4 0.92
40.4 0.54
40.6 0.56**
● .
*
T42.4 1.29
45.2 0.64
44.4 0.81
44.6 0.62,.
38.3 0.47
fz-u!cE
T46.5 0.62
45.7 0.84
44.9 1.17
47.0 0.78. .
38.5 0.50
262
—
71
102
64
23
3
275—
46
147
65
11
5
256
27
130
68
20
11
266
531C6
71
25
12—
&-1-2
T0.67 48
0.66 103
0.73 46. 27. 7
0.39 243
225—
62
104
31
23
6
237—
50
117
35
24
12
38.o 0.77
39.2 0.63
38.9 1.13
35.2 1.32
40.5 9.08
46 38.5
96 39.7
79 39.8
35 41.3
15 39.1
0.64
0.31
0.29
0.39
1,01
0.96 610.64 104
0.78 50
0.40 18● 10 1
37.6 0.9538.0 0.43
40.2 0.56● *
. .
Table 20. Mean hematocrit by age and sex for HES Cycle I I I and Cycle I data
Midpoint of
age interval
12.5
13.5
14.5
15.5
16.5
17.5
21.5
30
40
50
60
70
77.5
Male Female
Mean hamatocrit
in ml percent
40.5
41.5
42.7
43.9
44.9
45.8
46.7
47.0
46.6
46.5
46.1
45.8
45.1
Estimated
increase per year
in ml percent
1.0
1.2
1.2
1.0
0.9
0.2
0.04
-0.04
-0.01
-0.04
-0.03
-0.07
Mean hematocrit
in ml percent
40.3
40.8
40.3
40.7
40.4
40.6
41.4
41.8
42.0
42.5
43.7
43.2
43.1
Estimated
ncrease per year
in ml percent
0.5
-0.5
0.4
-0.3
0.2
0.2
0.05
0.02
0.05
0.12
-0.05
-0.01
Cycle
Cycle Ill
Cycle I
Age group
12 . . ! . . . . . . . . . . . . .
13. .’ . . . . . . . . ...4..
14 . . . . . . . . . . . . . . . .
15 . . . . . . .. . . . . . . . . .
16 . . . . . . . . . . . . . . . .
17 . . . . . . . . . . . . . . . .
18-24 .. . . . . . . . . . . . .
26-34 . . . . . . . . . . . . . .
35-44 . . . . . . . . . . . . . .
45-5 $. . . . . . . . . . . . . .
55-64 . . . . . . . . . . . . . .
65-74 .. . . . . . . . . . . . .
75-79’ . . . . . . . . . . . . . .
31
APPENDIX I
STATISTICAL NOTES
The Survey Design
The sampling plan of Cycle III of the HealthExamination Survey followed a multistage,stratified probability sample of clusters ofhouseholds in Iand-based segments in which asample of the U.S. population (including Alaskaand Hawaii ) aged 12 through 17 years wasselected. Excluded were those youths confinedto institutions and those residing on any of thereservation lands set aside for use by AmericanIndians.
The sample design of Cycle III is similar tothat of Cycle II in that it uses the same 40sample areas and the same segments. The deci-sion to incorporate this feature into Cycle IIIwas not made prior to the selection of the CycleII sample, although it is consistent with the earlyconcept of a single program for persons 6-17years old. The final decision to use this identicalsampling frame was made during the operationof the Cycle II program.
The successive elements for this sample designare primary sampling unit; census enumerationdistrict; segment (a cluster of househoMs);household; all eligible youths; and finally, thesample youth. Every eligible youth within thedefined population has a known and approxi-mately equaI chance for selection into thesample.
The steps of drawing the sample were carriedout jointly with the U.S. Bureau of the Census;the starting points were the 1960 decennialcensus lists of addresses and the nearly 1,900primary sampling units (PSU’S) into which theentire United States was divided. Each PSU is astandard metropolitan statistical area (SMSA), acounty, or a group of two or three contiguouscounties. These PSU’S were grouped into 40strata so that each stratum had an average size of
about 4.5 million persons. This grouping wasdone in a manner which maximized the degreeof homogeneity within strata with regard to thepopulation size of the PSU’S, degree of urbaniza-tion, geographic proximity, and degree of in-dustriaIization. The 40 strata were then classi-fied into four broad geo~aphic regions of 10strata each and, within each region, cross-classi-fied by four population density classes and bythe rates of population change from 1950 to1960. Using a modified Goodman-Kish con-trolled-selection technique, one PSU was drawnfrom each of the 40 strata.
The sampling within PSU’S was carried out inseveral steps. The first was the selection ofcensus enumeration districts (ED’s). These ED’sare small well-defined areas of about 250 hous-ing units into which the entire Nation wasdivided for the 1960 population census. EachED was assigned a “measure of size” equal tothe rounded whole number resulting from a“division by nine” of the number of childrenaged 5-9 in the ED at the time of the 1960census. A sample of 20 ED’s in the sample PSUwas selected according to a systematic samplingtechnique with each ED having a probability ofselection proportional to the population ofchildren 5-9 years at the time of the 1960 censusdate. From each ED a random selection of onemeasure of size (segment) was taken.
Minor changes required in the Cycle HI designwere that it be supplemented for new construc-tion to a greater extent than had been necessaryin Cycle II and that reserve segments be added.Although it was the plan for Cycle III to use theCycle II segments, it was recognized that withinseveral PSU’S additiomd reserve segments wouldbe needed to avoid the risk of having aninsufficient number of examinees. This wasprompted by the fact that four of the PSU’S in
32
Cycle H had yields of less than 165 eligiblechildren and several others were marginal intheir yield. In addition, there was a 3-Yearinkrval between Cycle II and Cycle III, so thatit was quite possible for some segments to havebeen completely demolished to make room forhighway construction or urban redevelopment.
The time available for examinations at aparticular location or stand, as they have beendesignated, is necessarily set far in advance ofany preliminary field work at the stand. There-fore, the number of examinations that can beperformed at a particular location is dependenton the number of examining days available. Atthe majority of locations, the number of daysmnilable, excluding Saturdays, is 17. At the rateof 12 examinations each day, this provides for~0~ examination slots. Exam~ations are con.
ducted on Saturdays if, for some reason, it isnecessary. Because of rescheduling for cancella-tions or no-shows, the maximum number ofyouths that is considered for inclusion in themmple is 200. When the number of eligibleyouths exceeds this number, subsampling isperformed to reduce the number to manageablelimits. This is accomplished through the use of,amaster list, which is a listing of all eligibleyouths in order by segment, serial number(household order within segment), and columnnumber (order in the household by age). Afterthe subsampling rate has been determined, everynth name on the list is deleted, starting with theyth name, y being a randomly selected numberbetween 1 and n. Youths who are deleted fromthe Cycle III sampIe bbt who were examined inCycle 11 as well as any twin who may have beendeleted are, if time permits, scheduled for anexamination for inclusion only in the longitudi-nal study portign or twin study portion of thesurvey. Their data are not included in the reportas part of the regular sample.
Since the strata are roughly equal in popula-tion size and a nearly equal number of sampleyouths were examined in each of the samplePSU’S, the sample design is essentially, self-weighting with respect to the target population;that is, each youth 12 through 17 years old hadabout the same probability of being drawn intothe sample.
The adjustment upward for nonresponse isintended to minimize the impact of nonresponse
on final estimates by imputing to nonrespon-dents the characteristics of “similar” respon-dents. Here “similar” respondents were judgedto be examined youths in a sample PSU havingthe same age (in years) and sex as those notexamined in that sample PSU.
The poststratified ratio adjustment used inCycle HI achieved most of the gains in precisionthat would have been attained if the sample hadbeen drawn from a population stratified by age,color, and sex, and it made the final sampleestimates of population a~ee exactly with in-
dependent controls prepared by the U.S. Bureauof the Census for the noninstitutional popula-tion of the United States as of March 9, 1968(approximaate midsurvey point), by color andsex for each single year of age 12 through 17.The weight of every responding sample child ineach of the 24 age, race, and sex classes isadjusted upward or downward so that theweighted total within the class equals the in-dependent population control.
A more detailed description of the samplingplan and estimation procedures is included inVital and Health Statistics, Series 2, Number43,3s and in Series 1, Numbers 1,1 5,s and 8,4which describe the plan and operation of thefirst three cycles of the Health ExaminationSurvey (HES).
Some Notes on Response Rates
As mentioned previously, the sample designsof the second and third cycles of the HES weresimilar. Differences did occur, however, in re-sponse rates of various subgroups of thesesamples, and these differences deserve someconsideration here.
Most importantly, the number of youthsselected for examination increased from 7,417in Cycle II to 7,514 in Cycle III. The responserate, that is, the number of youths selected whowere actually examined, decreased horn 96percent in Cycle II to 90 percent in Cycle III.Ofthe examined youths of Cycle II, 13.9 percentwere Negro compared with 14.8 percent ofthose examined in Cycle III. This differencedoes not reflect a difference in the percentage ofNegro youths selected for examination, butinstead, a smaller decrease in response rate for
33
Negro youths between the two cycles than was
the case for the white youths. In actuality, 13.8percent of the sample selected for examinationwas Negro in Cycle III corresponding to 13.5percent for Cycle II. However, whereas theresponse rate for white youths dropped from95.6 percent in Cycle II to 89.1 percent in CycleIII, the response rate for Negro youths droppeda far lesser degree from 98.4 percent to 96.6percent. Thus, better relative response from theNegro portion of the sample yielded a greaterpercentage of these youths actually examinedduring Cycle 111 than was the case during theprevious sample.
Examination of sample sizes in this reportclearly shows that at every age group there werefewer females actually examined than there weremales of the same age. This again is notattributed to differences in numbers of youthsselected in the sampling design, but rather to thefollowing differential response rates betweenmales and females:
Age Male Female
Total . . . . . . . . . . . . . . . . . . . . . . . 91.4 88.7
12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93.5 91.3
13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93.2 91.914 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.7 90.715 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.6 87.916 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89.8 87.717 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87.6 81.8
Note that at each age group the response rate formales exceeded that for females.
A similar analysis of response rates can bedone by age, race, and sex as foIlows:
AgeWhite Negro White Negro
male male female female
Total . . . . . . . . . 90.5 87.6 87.4 95.8
12 . . . . . . . . . . . . . . . 92.6 99.0 90.1 98.9
13 . . . . . . . . . . . . . . . 92.5 98.8 91.1 96.8
14 . . . . . . . . . . . . . . . 91.0 87.8 89.6 96.215 . . . . . . . . . . . . . . . 90.7 97.7 86.4 98.616 . . . . . . . . . . . . . . . 89.2 95.0 86.6 93.0
17 . . . . . . . . . . . . . . . 86.5 95.8 80.2 91.4
The above clearly indicates that for all agesunder consideration in Cycle 111 of the HES theresponse rate for Negro youths exceeded thatfor white youths of the same sex and age.
Reasons for differences in response rates aremany, but may range from the incentive to getexamined in order to miss a day of school, tofear of the examination itself, to inhibitionswith respect to being examined. The worstresponse rate was recorded for the oldest fe-males, that is, those aged 17 years.
Parameter and Variance Estimates
Because each of the 6,768 sample children hasan assigned statistical weight, all estimates ofpopulation parameters presented in HES publi-cations are computed taking this weight intoconsideration. Thus, ~, the estimate of a popula-tion mean, “P,” is computed as follows:
where Xi is the observation or measurementtaken on the ith person and Wi is the statisticalweight assigned to that person.
The HES has an extremely complex samplingplan, and obviously the estimation procedure is,by the very nature of the sample, complex aswell. A method is required for estimating thereliability of findings that “reflects both thelosses from clustering sample cases at two stagesand the gains from stratification, ratio estima-tion, and poststratification. “s A
The method for estimating variances in theHES is the half-sample replication technique.The method was developed at the U.S. Bureauof the Census prior to 1957 and has at timesbeen given limited use in the estimation of thereliability of results from the Current PopulationSurvey. This half-sample replication technique isparticularly well suited to the HES because thesample, although complex in design, is relativelysmall (6,768 cases) and is based on but 40 strata.This feature permitted the development of avariance estimation computer program that pro-duces tables containing desired estimates ofaggregates, means, or distributions, togetherwith a table identical in format but that containsthe estimated variance of the estimated statis-tics. The computations required by the methodare simple, and the internal storage requirementsare well within the limitation of the IBM 360-50computer system used at the National Center forHealth Statistics.
34
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. Tocompute the variance of any statistic, thisstatistic is computed for each of the 20 halfsamples. Using the mean as an example, this isdenoted .~i. Then, the wefihted mean of theentire, undivided sample (~) is computed. Thevariance of the mean is the mean square devia-tion of each of the% half-sample means aboutthe overall mean. Symbolically,
.fl(Xi- X)2Var(X) = 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 in Vital and HealthStatistics, Series 2, Number 14.35
Standards of Reliability and Precision
All means, variances, and percentages appear-ing in this report met defined standards beforethey were considered acceptably precise andreliable.
The rule for reporting means and percentilesconsisted of two basic criteria. The first criterionWJS that a sample size of at least five wasrequired. If this first criterion was met, then thesecond criterion, that the estimated coefficientof variation [i.e., the standard error of the meandivided by the mean (sx/X)] was to be less than~J percent, must have been demonstrated. Thus>
if either the sample size was too small, or the
variation with respect to the mean was too large,the estimate was considered neither precise norreliable enough to meet the standards estab-lished for publication.
To illustrate these criteria, in table 17 valuesof the mean and standard error for 16-year-oldNegro males from families with a total yearlyincome of $10,000 or more were replaced byasterisks (*) since the standard error withrespect to the mean exceeded the criterionpreviously stated.
Imputation
In addition to the subject nonresponse dis-cussed previously, the problem of item non-response merits consideration here. In this situa-tion, information about a respondent is com-plete with the exception of a missing hematocritvalue.
Missing data and values that fell outside anormal range and were deemed unlikely afterexamination of physician records were replacedby the value recorded for a randomly selectedrespondent of similar age, sex, and race. How-ever, when only one of the two hematocritvalues was missing, the recorded measurementwas substituted for the unknown value. Imputa-tion, where there was no value recorded forexaminees, was necessary in 190 cases forhematocrits (2.9 percent of total respondents).
In effect, sample cases were sorted intocategories within each of which the cases wereexpected to have high intraclass correlation (i.e.,being relatively homogeneous). Those with miss-ing values were then completed with a valuerandomly selected from within the category.This method of imputation preserves both theexpected values and the distribution of values ineach category to those of the respondent cases.
35
APPENDIX II
DEMOGRAPHIC VARIABLES
(
Regional and demo.qaphic characteristics bywhic~ the population ~as-been classified for tti-sreport are defined as follows.
Age and sex. –Population was classified into12 age-sex groups-the six ages 12-17 years bysex. Birth certificates were obtainable for verifi-cation of age for 92 percent of the youths. Agestated by the parents was accepted as the trueage for the other 8 percent. Age is expressed asyears attained at last birthday.
Race. –Hematocrit was reported by race forwhite and Negro youths. Youths of other raceswere not sampled sufficiently for comparisonpurposes; these youths represented only 0.55percent of the sample.
Region. –Regional data are presented for fourregions of the continental United States.
Region States Included
Northeast . . . Maine, Vermont, New Hamp-shire, Massachusetts, RhodeIsland, Connecticut, NewYork, Pennsylvania, New Jer-sey
Midwest . . . . Minnesota, Wisconsin, Michi-gan, Iowa, Missouri, Illinois,Indiana, Ohio
South . . . . . . Delaware, Maryland, Virginia,District of Columbia, WestV@inia, Kentucky, Tennes-see, North Carolina, SouthCarolina, Georgia, Florida,
Alabama, Mississippi, Arkan-sas, Louisiana
West . . . . . . . Washington, Oregon, Idaho,Montana, North Dakota,South Dakota, Wyoming,Nebraska, Kansas, Colorado,Utah, Nevada, California, Ari-zona, New Mexico, Texas,Oklahoma, Alaska, Hawaii
Family income. –The income recorded wasthe total income received during the past 12months by the head of the household and allother household members related to the head byblood, marriage, or adoption. This income wasthe gross cash income (excluding pay in kind)except in the case of a family with its own farmor business, in which case net income wasrecorded.
Education of paren t or guardian. –This itemwas recorded as the highest grade that had beencompleted in school. The only grades countedwere those that had been completed in a regularschool in which persons were given formaleducation in graded or private schools, eitherday or night schools , with either full-time orpart-time attendance. A “regular” school is onethat advances a person toward an elementary orhigh school diploma, or a college, university, orprofessional school degree. Education in voca-tional, trade, or business schools outside theregular school system was not counted in deter-mining the highest grade of school completed.
36
APPENDIX Ill
TECHNIQUES OF MEASUREMENT AND QUALITY CONTROL
Measurement
The HES blood-drawing technique and itshistorical development. –For a variety of rea-sons, it was decided not to attempt to draw ablood specimen on children 6-11 years of age inCycle II. The children included in the nationalprobability sample came from all regions of thecountry and all cultural and socioeconomicgroupings and ways of life; some would neverhave been to a physician before; some wouldhave had very bad memories and associations. Inaddition, by sampling the entire spectrum ofbehavioral and physical development extant inthe United States, some of the 6-year-old chil-dren and even some of the 7-year-old childrenwould be so immature as to present severetechnical and behavioral problems. It was be-lieved that fear, or extreme distaste, of having ablood sample drawn felt by many potentialsubjects might severely affeet the response rate,which is so crucial to a survey like this. When itis remembered that the overall response rate forCycle II was a remarkable 96 percent, it isdifficult to argue with that line of reasoning. Ofcourse, now it can never be known how great adiminution of the response rate the addition of ablood sample would have caused.
In the early planning stages of Cycle III, itwas decided to obtain a blood specimen if at allfeasible, that is, primarily if the price in dimin-ished response rate and cooperation of thewmminees would not be too high. Accordingly,in three separate pretests, investigations wereconducted regarding the problem of developinga satisfactory blood-collection technique for thisage ~group and examination setting. The opti-mum amount of blood drawn that would nothave an emotional impact on the examinee andwould not affect his performance in any of the
procedures to follow was desired. The amountof usable blood that could be drawn posed alimiting factor on the number of blood chemis-try tests that could be performed and made adifference in accepting or rejecting an entirepossible area of the examination such as thenutritional assessment. Logistical problems alsohad to be resolved involving the handling,separating, and packaging of drawn blood sothat there would be a minimum of blood lossand packaging error. For the refrigerated butunfrozen bloods, time from shipment to deliverywas critical; therefore, arrangements had to bemade with postal authorities to assure promptdelivery to the laboratories in order to avoidspoi~age.
There was a trial-and-error process, and therewas good advice and help from many sources indeveloping a satisfactory blood-drawing tech-nique. The chief sources of help, outside of theimmediate HES technical staff, were Dr. Wi~aBias and Dr. Bernice Cohen of The JohnsHopkins University; Dr. Gerald Cooper, Chief ofLaboratories, Communicable Disease Center,Public Health Service, Atlanta, Georgia; themany teenage subjects during our pretest whogave valuable suggestions and who pointed out,either as overt advice or by their immediatereactions, specific points to be avoided; and,finally, the professional and technical division ofthe Becton-Dickinson Company, Rutherford,New Jersey. The latter, through several personalvisits by a representative of their technicaldivision, not only gave excellent technical adviceon blood-collection techniques and the use ofalternative equipment but also devised a specialfitting that made the transfer from one vacutain-er tube to another much smoother.
During the pretesting, it was learned thatmany subjects did not like to see any part of the
37
blood-drawing procedure, including their ownblood in tubes. Therefore, a technique wasemployed that minimized the subject’s attentionto the operation. With the subject lying supineand by draping and by keeping the arm andtubes well below the level of the examinationtable, effective screening was achieved. After theskin area was cleansed with alcohol, the bloodwas then drawn from the antecubital fossa bythe physician-nurse team. At the discretion ofthe physician, a tourniquet was used to fill thevein; however, once the needle was inserted intothe vein, the tourniquet was taken off the armso that the blood flowed freely.
A B-D blood culture needle and tube wereused to draw blood. Using the specially preparedlink fitting, the nurse inserted the short needleinto a vacutainer tube holder. The tube wasclamped with a hemostat until the vein waspunctured and the vacutainer inserted into theholder.
From the one free-flowing venipuncture, atotal of only 55 cms of blood was collected infoLIr separate vacutainer tubes from all male andalmost half of the female subjects. (The differ-ence being that all males had separate plasmadrawn to be frozen “and stored for futuretestosterone determination and almost half ofthe females provided a replicate blood specimenfor quality control of the laboratory determina-tions; the remaining females had 40 cm3 drawn.)
All test tubes were labeled with the exami-nee’s number and left in the test-tube holdingrack at room temperature for 1 hour. The nursethen placed the tubes in the laboratoryrefrigerator, along with 10 extra examinee iden-tification labels, for the technicians.
An analysis that attempts to estimate theimpact of the addition of a blood sample on theCycle III sample response rate is in progress.
Hematocrit determinations. –After the B-Dblood culture needle set was withdrawn fromthe arm, the tube was drained; and the twomicrocapillary tubes were two-thirds filled withthe residual blood in the tubing. They weresealed with Sealease and centrifuged (Interna-tional Model MB) for 5 minutes. Both specimenswere read to the nearest percent directly froman International Micri-capillary Reader, ModelCR.
The mean of the two readings expressed asthe volume (in milliliters) occupied by the red
cells per 100 milliliters of whole blood wasrecorded as the subject’s hematocrit value.
Monitoring Systems
In addition to the sampling considerationsalready discussed, the quality of data collected isalso a special concern. One of the main purposesof the monitoring system employed in thesurvey was to indicate whether the measure-ments produced by our measurement processattained the desired quality. A second majorpurpose was to make possible quantitative sum-mary descriptions of residual measurementerrors to aid in the interpretation of survey data.
The monitoring system as applied to thetaking of blood samples consisted of a formalsystem of replicate examinations (described laterin this appendix). Replicate measurements areuseful for a variety of reasons, for example, as ameans of increasing precision of estimates ofindividual measurements, as a training tech-nique, and as a monitoring system that includesthe objective of overaIl evaluation of measure-ment errors. These objectives are not incom-patible, and replicate data collected primarilyfor one of these objectives often indirectly, ifnot directly, accomplish one or both of theremaining two. For this reason, replicate dataare most often collected with a combination ofthese objectives in mind.
Methods of Taking Replicate Measurements
A major source of uncertainty in estimatesderived from replicate measurements is in theinability to make the replicate measurementunder precisely the’ same conditions and in thesame manner as the original measurement. Thisuncertain y is difficult to evaluate, and mostattempts are restricted to subjective statementsconcerning the direction and/or size of the biasand the need for concern in the analysis of data.
Each sample examinee had blood drawn intotwo microhematocrit tubes, and readings wereobtained by the same technician according tothe procedure described previously. Since thetwo blood samples were taken at the same time,diurnal variation in the hematocrit of the ex-aminee would not be expected to be a source of
3a
variation bet ween the two readings. It is possiblethat differences may have occurred between thetwo specimens in duration of time centrifugedand in the time interval between the drawingand centrifuging of the blood sample. However,these differences were held within strict limits,and it is unlikely that they are a major source ofvariation between the two readings. The mostimportant source of variation between the tworeadings by the same technician is likely to be areading error. Part of the reading error is due toa “rounding error “ in the last significant figure,since readings were taken to the nearest percent.Since the hematocrit value reported in thesurvey is an average of the duplicate readings bythe same technician, the effect of rounding errorwould tend to be ameliorated. In the followinganalysis, the differences between the two read-ings by the same technician will be referred to as“intraobserver” differences.
A study was undertaken to evaluate theextent of variation between two observers read-ing the same microhematocrit tube. This wasdone by having a supervisory technician onMonday mornings read the hematocrits for themorning and enter the findings beside theoriginals. In the ensuing discussions, such differ-ences will be referred to as “interobserver”differences.
Results of the Replicate Studies forHematocrit
Two hematocrit readings were obtained bythe same technician on 6,375 examinees (94.2percent of the total sample). Replicate readingsby another technician were obtained on 640examinees (9.5 percent of the total sample) for atotal of 1,280 replicate measures. All together,during the 4 years, 12 technicians participated inobtaining replicate measurements for this phaseof the quality control progam.
Since there were 12 technicians employedduring Cycle III, it is of interest to ascertainwhether each of the examiners had a representa-tive number of replicate measurement sessionswith respect to the number of examinationsperformed during the survey. It should becarefully noted that it was not possible to insurethat each technician had equal chances toperform replicate measures, since the length oftime various technicians associated with thesurvey team varied.
Table I presents the percentage of totalexaminations done in the survey, the percentageof intraexaminer replicates, and the percentageof interexaminer replicates participated in byeach of the 12 technicians. Table I indicatessome possible sources of bias that may affect the
Table 1. Percent of regular Cycle III hematocrit examinations and replicate hematocrit examinations participated inby each technician
Technician number
1 . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 .,, .,... . . . . . . . . . . . . . . . . . . .
3 .,, ..,,. . . . . . . . . . . . . . . . . . . .
4 . . . . . . . . . . . . . . . . . . . . . . . . . . .5 . . . . . . . . . . . . . . . . . . . . . . . . . . .
6 . . . . . . . . . . . . . . . . . . . . . . . . . . .
7 ,, .,,,., . . . . . . . . . . . . . . . . . . .
8 . . . . . . . . . . . . . . . . . . . . . . . . . . .9 . . . . . . . . . . . . . . . . . . . . . . . . . . .
lo . . . . . . . . . . . . . . . . . . . . . . . . . .11, . . . . . . . . . . . . . . . . . . . . . . . . .12 . . . . . . . . . . . . . . . . . . . . . . . . . .
Percent of regular
Cycle I I I examinations
0.113.7
25.8
7.213.94.3
3.7
12.6
13.5
0.04.2
1.0
Percent of replicate examinations
Intraobserver
0.013.9
25.8
7.114.04.2
3.6
12.813.4
0.04.21.0
Interobsetver
15.8
9.4
35.6
1.710.74.1
2.2
7.5
9.0
2.31.8
39
analysis of replicate data. The number of inter-examiner replicate examinations is distributedbetween technicians disproportionately withrespect to regular and intraexaminer replicatereadings. For example, assume technician 3 wasin very close agreement on his own readings, buthis readings were very different from the otherexaminers’. Because of this technician’s over-representation, the distribution of interexaminerdifferences would be considerably more skewedthan it should have, since the technician doesnot agree well with the other technicians’measurements. Thus, the various combinationsof observers for the interexaminer replicates andthe proportions of intraexaminer replicates werenot controlled so as to be balanced among theobservers. In the survey proper, the examina-tions were similarly not (of necessity, sincelength of time the various technicians wereassociated with the survey varied) proportion-ately distributed among the observers.
The foregoing indicates that the distributionof numbers of replicate examinations done byeach technician is not the same as the distribu-tion of the total number of survey examinationsdone by each in Cycle III. This represents one ofthe inherent problems of the present replicatedata, and, limits to some extent implications tothe survey as a whole. Nevertheless, the readershould be aware of the many problems confront-ing those who conduct large-scale health sur-veys,a 6 and in this context, the present systema-tic approach to the collection of replicate data isadequate.
For both the intraobserver study and theinterobserver study, frequency distributions ofthe absolute differences between replicatehematocrit readings are presented in table II.
As a summary statistic of the distribution ofdifferences between replicate hematocrit read-ings displayed in table H, we have computedV–the percentage technical error of m&sure-ment—which is given by
I
where
n is the number of pairs of measurements inthe study,
Table 11. Frequency and percent distribution of absolute differ-
ences between replicates for the interobserver and intra-observer studies of hematocrit measurement error
Absolute
difference
(in milliliters)
Total . .
0 . . . . . . . . .
1 . . . . . . . . .
2 . . . . . . . . .
3 . . . . . . . . .
4 . . . . . . . . .
5 . . . . . . . . .
6 . . . . . . . . .
7 . . . . . . . . .
8 . . . . . . . . .
9 . . . . . . . . .
lo . . . . . . . .
11 . . . . . . . .
1nterobserver study
Frequency
1,280
732
544
4
0
0
0
0
0
0
0
0
0
Percent
of total
100.0
57.2
42.5
0.3
Intraobserver study
Frequency
6,375
3,858
2,488
16
6
2
3
0
0
1
0
0
1
‘ercen t
>f total
100.0
60.5
39.0
0.20.1
0.1
df is the square of the difference betweenmembers of the ith pair of measurements(i= 1,.. .,n), and
~ is the arithmetic mean of the 2n measure-ments in the study.
The percentage technical error, V, can be inter-preted as a “coefficient of variation” and is adimensionless constant. It essentially describesthe size of measurement error relative to themean value of a measurement. For replicatehematocrit determinations obtained in the inter-observer and intraobserver studies, the values ofV are the following:
Study
InterobserverIntraobserver
The interpretationtwo observers reading
v
1.12 percent1.14 percent
of these statistics is thatthe same microhematocrit
tube would, on the _-average, record values thatdiffer from each other by about 1 percent.Likewise, a single technician reading two micro-hematocrit tubes obtained from a single veni-puncture processed under identical laboratoryconditions would record values that differ fromeach other by an average of approximately 1percent.
40 .
Sm”es 1.
Sm”es 2.
Sm’es 3.
Series 4.
Series 10.
Sera”es 11.
Series 12.
Sm”es 13.
Series 14.
Sm”es 20.
Sw”es 21.
Series 22.
*
VITAL AND HEALTH STATISTICS PUBLICATION SERIES
Formerly public Health SetvicepublicationNo. 1000
Pngmms and collection Procedures. - Reports which describe the general programs of the National
Center for Health Statistics and its offices and dfvisions, data collection methods used, defioftions,and other material necessary for understanding the data.
Lkzta eualwxtion and methods research. —Studies of new statistical methcxfology inclu~ experi-mental testa of new survey methods, studfes of vital statistics collqdon methods, new snaiytfcaltechniques”, objective evaluations of reliability of collected data, contributions to statistical theory.
/
AWvticaZ stwdies. —Reporta presenting analytical or interpretive studfes basedon vital and healthstatistics, carrying the analysis further than the expository types of reports in the other series.
Documents and committee rs@rts.— Final reports of major committees concerned with vital andhealth statistics, and documents such as recommended model vital registration laws and revisedbirth and death certificates.
Data from the Health Intm”ew Survev. —Statistics on illness, accidental injuries, disability, useof hospital, medical, dental, and other services, and other heal&related topics, based on datacollected in a continuing national household interview survey.
Data fio?p the Health Examination Survey. —Data from direct examination, testing, and measure-ment of natimal samples of the civilian, noninstitutional population provide the basis for two typesof reports: (1) estimates of the medically defined prevalence of specific diseases in the UnitedStates and the distrilmtions of the population with respect to physical, physiological, and psycho- “logical characteristics; and (2) analysis of relationships among the various measurements withoutreference to an explicit finite universe of persons.
D@ from the Institutional Pojndation Surveys .- Statistics relating to the health characteristics ofpersons in institutions, and their medical, nursing, and personal care received, based on nationalsamples of establishments providing these services and samples of the residents or patients.
Data from the Hospitil Discharge Survey. —Statistics relating to discharged patienta in short-stayhospitals, based on a sample of patient records in a national sample of hospitals.
Data on health resources: manpower and facilities. —Statistics on the numbers, geographic dfstri-Ixttion, and characteristics of health resources including physicians, dentists, nurses, other healthoccupations, hospitals, nursing homes, and outpatient facilities.
Daka on mortality. —Various statistics on mortality other than as included in regular annual ormonthly reports-special analyses by cause of death, age, “and other demographic variables, alsogeographic and time series analyses.
Data on natality, mawiage, and divorce. —Various statistics on natality, marriage, and divorceother than as included in regular animal or monthly reports-special analyses by demographicvariables, also geographic and time series analyses, studies of fertility.
Data from the National Natality and Mortality surveys. — Statistics on characteristics of birthsand deaths not available from the vital records, based on sample surveys stemming from theserecords, including such topics as mortality by socioeconomic class, hospital experience in thelast year of life, medical care during pregnancy, health insurance coverage, etc.
For a list of titles of reports published in these series, write to: Office of InformationNational Center for Health StatisticsPublic Health Service, HRA
Rockville, Md. 20852
