TE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING A SAMPLING SURVEY STUDY OF ARTERIAL BLOOD PRESSURE LEVELS IN NASSAU, NEW PROVIDENCE, BAHAMAS, 1958 FOR DESCRIPTION OF LEVELS OF BLOOD PRESSURE IN A POPULATION IN RELATIONSHIP TO AGE, SEX, RACE AND OTHER FACTORS Benjamin Charles Johnson A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Public Health Department of Epidemiology School of Public Health The University of Michigan Ann Arbor, Michigan 1959 July, 1959 IP-375

ACKNOWLEDGEMENTS The author is greatly indebted to many people foy assistance and guidance in the course of this investigationo Financial assistance for the field work was provided by Dr, Marvin Moser of Grasslands Hospital., Valhalla, New York, from funds granted by the National Heart Institute and by Dr. Sibley Hoobler of the University of MichiganO Sincere appreciation is expressed to Dr Eo Ho Murcott, Chief Medical Officer of the Bahamas Medical Service and his staff for their splendid cooperation and assistanceo Special thanks are due the four senior medical students from the University of Michigan whose careful work in the field contributed greatly to the completion of the study. These students were Walter Briney, Herbert Kaufer, Edward. Pierce, and Donald Van der Poldero Dr Richard Remington. of the School of Public Health, University of Michigan. assisted greatly i:n planning of the study, in the design. of the sample, and in statistical. guidanceo Mrs. Louise Dapprich was of invaluable aid in processing,of the data. Preparation of the maniuscript was done by Mrso Glenda Buist and by the staff of the Industry Program of the College of Engineering, University of Michigan. Special indebtedness for direction and encouragement is due the members of the Doctoral Committee and Dro Thomas Francis, Jr. of the Department of Epid.emiology, School of Public Health, University of Michigan. ii

TABLE OF CONTENTS Page ACKNOWLEDGMENTS........ *,. i*........... ii LIST OF TABLES..... * 0........................... iv LIST OF ILLUSTRATIONS........ *, ~,.* 5 9, v 9 5 94 9 * 9 O. 9 * V CHAPTER I INTRODUCTION. o... * o. o.o*.,.. o.... o......o o, 1 II CHOICE OF SITE..............., *.....,,,........ o.*,. 5 III CHOICE OF SAMPLE DESIGN AND SAMPLE VALIDATION..,....,.., 8 IV THE INTERVIEW AND EXAMINATION PROCEDURE.... *..........* 21 V ANALYSIS OF POTENTIAL OBSERVER ERROR............,. 25 VI COMPARISON OF DIASTOLIC READING METHOD........,...,,,. 31 VII THE DISTRIBUTION OF BLOOD PRESSURE READINGS BY AGE, SEX AND RACE................34..9....... 9...... 5 34 A. Means and Standard Errors of the Distributions..,... 34 B. Graphs of the Distributions.......,0a*,,*..,o.,,iaa* 50 VIII ANALYSIS OF SPECIFIC BREAK POINTS OF BLOOD PRESSURE..... 71 A. Break Points Considered Singly....0 4o...........0*. 71 B. Study of Systolic and Diastolic Pressures in Combination......9...9 9U.,,. 9 * o 9,, 94*4,, 5,9,.9*- 82 IX RELATIONSHIP OF BLOOD PRESSURE TO BODY BUILD,...,...., 87 X RELATIONSHIP OF BLOOD PRESSURE TO ARM GIRTH......,,,... 94 XI PARITY IN RELATION TO BLOOD PRESSURE....,,..,,.... 100 XII -RELATIONSHIP OF PULSE RATE TO BLOOD PRESSURE........... 104 XIII SUMMARY OF THE STUDY............... o......... 106 XIV DISCUSSION.........9 9 49,.* 5.9*5*55 9 *94 a *-0* a, 111l APPENDIXa...~.S..................,,.......*a,....S...,.9.O...,.. 118 BIBLIOGRAPHY..........,.......,,.4. o..,.............*,,........,. 127 iii

LIST OF TABLES Table Page 31o Comparison of Age-Sex Proportions, 1953 Bahama Census, and Present Population Sample................ 20 501 Components of Variance Test for Variation Due to the Observero...............OOOO..o.......O............ 27 502 Terminal Digit Analysis~ Proportion of Readings by Terminal Digit per Examiner.....o........o...... 30 6o1 Mean Differences Between Diastolic Reading Method: Diastolic (Fifth Phase) Minus Diastolic (Fourth Phase) by Age, Sex and Race, Nassau, 1958 o... 00.oo...oo. 32 7o1 Age, Sex and Race Specific Mean Blood Pressures (Systolic, Diastolic and Pulse Pressures) with Standard Errors of Means, Standard Deviations of Distributions and Coefficients of Variation, Nassau, 1958....... a o o o 0 o o o o o o a 36 81l Proportions of Persons with Blood Pressures Greater Than Specific Break Points of Systolic and Diastolic Pressure by Age, Sex and Race Groups, Nassau, 1958,00 72 8,2 Race Specific Proportions of Persons Exceeding Blood Pressure Break Points with Significance of Differences Between Races, Nassau, 1958....000oo.... ooo... 78 803 Sex Specific Proportions of Persons Exceeding Blood Pressure Break Points (All Ages Included) with Signicance of Differences Between Sexes, Nassau, 1958..... 79 8~4 Sex Specific Proportions of Persons Exceeding Blood Pressure Break Points by Broad Age Grouping (PreMiddle Life and Post-Middle Life) with Significance of Differences Between Sexes, Nassau, 19580......... 80 8,5 Proportion of Persons in Race-Sex Groups with Both Systolic and Diastolic Pressure Elevations, in Relation to Either Pressure Elevated Singly, Nassau, 1958...... o o o 0 o o o o o o 0 o 00 o 0 0 0 o 000 0 0 o 0 0 o 0 0 84 9.1 Relationship of Ponderal Index to Mean Blood Pressure in Age-Sex Specific Groups, White Race, Nassau, 1958. 89 iv

LIST OF TABLES (CONT'D) Table Page 9o2 Relationship of Ponderal Index to Mean Blood Pressure in Age-Sex Specific Groups, Negro Race, Nassau, 1958 o o o..o.o oo o. o o o. o o o o o o o o a a 90 101o Relationship of Arm Girth to Mean Blood Pressure in Age-Sex Specific Groups, White Race, Nassau, 1958 o 0 0 0 0 0 0 O 0 0 0 0 0o o o a o..00 o 0 o o o o o o o. o o o 97 10o2 Relationship of Arm Girth to Mean Blood Pressure in Age-Sex Specific Groups, Negro Race, Nassau, 19580..000...0000 o0.o o o o o o o o o 0. o o o o o o o o o o o 98 1101 Relationship of Multiparity to Mean Blood Pressure in Age Groups, Negro Race, Nassau, 1958..00,, 00ooooo 102 1102 Mean Blood Pressures by Age Group for Pregnant Women Compared to Mean Blood Pressures for All Women of the Same Age (Including Pregnant) Negro, Race, Nassau, 1958o..oooQoo 0000000o 000000000 103 12ol Relationship of Pulse Rate to Mean Blood Pressure in Age and Sex Groups, Negro Race, Nassau, 1958.oo,,,,,, 105 APPENDIX DISTRIBUTION OF BLOOD PRESSURE READINGS CLASSIFIED BY SYSTOLIC PRESSURE AND DIASTOLIC PRESSURE ACCORDING TO AGE, SEX AND RACE GROUPS, NASSAU, 1958. o o o o.. 115

LIST OF ILLUSTRATIONS Figure Page 7o1 Frequency Distributions of Systolic and Diastolic Pressures by Sex, Race and Selected Age Groups, Nassau, 1958 (Eight Parts) o o..................... 40 7.2 Mean Systolic and Diastolic Blood Pressures by Age, Sex, Race, and Five Year Age Groupo White and Negro Races, Nassau, 1958..oo.....................oo.o 51 7.3 Mean Pulse Pressure by Age- and Sex, Negro Race, Nassau, 1958o.....o.o o.o....................o.. 53 7.4 Mean Pulse Pressure by Age and Sex, White Race, Nassau, 1958... 0 o o...........0..a. o... 54 705 Mean Systolic and Diastolic Blood Pressures by Age Group. Negro Race, Nassau, 1958 and Georgia (Comstock3), ooo..,,ooo...0..,.,O.,.O O-O-.o o 59 7.6 Mean Systolic and Diastolic Blood Pressures by Sex and Age Groupo Negro Race, Nassau, 1958, and Sto Kitts, Leeward Islands (Corcoran and Schneckloth 1958).....oo..o..o OO......oO........OO......O........ 61 7.7 Mean Systolic and Diastolic Blood Pressures by Age Group. White Race, Nassau, 1958 and Georgia (Comstock3).....................,,...,,o0o........... 63 7.8 Mean Systolic and Diastolic Pressures by Age and Sex, White Race, Nassau, 1958 and Wales (Miall and Oldham12) 7.9 Mean Systolic and Diastolic Blood Pressure by Age Groupo White Males, Nassau, 1958 and Bergen, Norway (Boe......... oo., o.."....o...... 67 7o10 Mean Systolic and Diastolic Blood Pressures by Age Group, White Females, Nassau, 1958 and Bergen, Norway (Boel )o o o o o.o o o o. o o o o o o o o o o o o o o o o o o o o o o 68 8o1 Proportion of Persons Within Age-Sex-Race Group with Blood Pressure Over Arbitrary Break Points, Negro and White Race, Nassau, 1958. 9.1 Mean of Height-Weight Index by Age. Negro and White Raceso Nassau, 1958..,,.,0.....0......0...0.0...,,,, 91 101ol Mean Arm Girth for Age, Sex and Race Specific Groups Negro and White Races, Nassau, 1958,.o o.,...o..,.o.. 95 vi

CHAPTER I INTRODUCTION Much has been written regarding the normal and abnormal blood pressure of man. In the more severe forms of hypertension there is no question that it constitutes a true disease state. Yet, many questions remain unanswered. First of these is the uncertainty of distinguishing the normal from the abnormal, except at the extremes. Considering the blood pressure reading alone, there is no point at which one can say that a specific person has hypertension or hypertensive diseaseo This reading, or preferably a series of readings over time, must be considered in relation to the patient s medical history, symptoms, and examination to determine the presence of hypertensive disease, Also to be considered are his race, age, sex, degree of obesity or type of body build, as well as emotional factors. The same applies to hypotension, if such a state can be considered a pathological entity. Numerous studies have been done which describe blood pressure in various groups and subdivisions of populationso Some have applied arbitrary division points delineating "normal" from "abnormal", others have expressed the blood pressures observed in terms of the arithmetic mean and variation about it- Many attempts at a definition of blood pressure values in a population have been made (Morsell(4) lists 51), but almost invariably studies are directed toward specific, easily obtainable segments of a population There has been considerable

-2disagreement, particularly in regard to the relative importance of the various factors mentioned, such as race, age and obesity. Since essential hypertension is still a disease of unknown etiology, such disagreement is inevitable, Much research activity has been directed toward the individua patient with the diseaseo From study directed toward the biochemical and physiological aspects of the condition, many important discoveries have been made, A few specific and sometimes remediable causes have been found, such as pheochromocytoma, primary aldosteronism and, of course, chronic renal disease. From this approach have come hypotheses which form the basis for much current researcho These include investigation into the broad categories of the neural, humoral and renal mechanisms in disease There is yet another approach to the study of chronic disease and one which has provoked much thought in recent years. This is the study of disease in the mass, or the relationship of persons with the disease in question to their environment, community, racial origin) sex and age, as well as past experience with disease, This is the epidemiological approach, as old as the affair Of the Broad Street Pump, yet new as long as its principles are intelligently applied. In relation. to hypertension, several studies in recent years have attempted to describe blood pressure levels in specific (1, 3, 1 3) populations o These have varied in scope from a relatively small community in Wales(') to the entire city of Bergen, Norway. 1) A goal of this method may be to establish more definitely the distribution of blood pressure levels in a population in reference to such

variables as age, sex, race, pregnancy, socioeconomic status, occupation, environmental conditions, etco Information may also be gained as to the possible role of genetic factorso Hopefully, the aggregate of such information may provide clues for more specific investigations. One aspect of such investigations deserves special mention. In any situation where conclusions are to be drawn beyond the individual case, special care must be taken that such results may withstand statistical scrutiny. The science of statistics has given us many invaluable toolso Since statistical techniques are an essential part of epidemiologic method, population studies of disease must be designed with full appreciation of the statistical basis of the analysis to be employed The study to be described was undertaken because of two separate stimulio The first was the findings of Moser, Hoobler, Remington and Dodge (unpublished) in the Bahamas of what appeared to be a population with an inordinately high proportion of people with elevated blood. pressure levelso Since the majority (three-fourths) of the population is Negro, this impression agreed with Comstockgs findings in Georgia 3) Since people of the Bahamas represent a geographic isolate, and perhaps a genetic isolate, these islands might be a useful location for the study of the effect of race on blood pressure. The second stimulus was the knowledge that the water supply in the Bahamas has.a relatively high salt content. If, in truth, excess sodium has a hypertensive effect, it should theoretically affect

-4both races equally and cause noticeable differences in both races from their counterparts in the United States and elsewhere. An additional incentive was the concern of the resident medical profession and the Colonial Medical Service over hypertension as a major public health problem, The specific aims of this study were then formulated. It was desired to study a representative sample of the Bahamian people with a sufficient degree of precision to ensure statistical validity and reproducibility of the results. The primary information desired was a knowledge of the distribution of blood pressure levels in relation to race, sex and ageo Secondarily, information regarding height, weight, build, arm girth, socioeconomic status, length of residence in the Bahamas and in Nassau, family history and parity would also be elicitedo If possible, information on relative intake of salt (in water supply) would be obtained for units within the sampleo In addition to these specific goals, it was felt that such a project could provide useful information in regard to the methodology itself. The two principal methodological problems were choosing a proper sample for a population study and standardization of procedure. This latter included the type and method of examination as well as an attempt to assess differences in results due to instrument error, error due to the observer and error due to variability within the person observedo

CHAPTER II CHOICE OF SITE Since it had been decided to study the people of the Bahamas, it was necessary to know something of the geography and history, as well as the population distribution of the Islands. The Bahama Islands constitute a British Colony separate from the federated British West Indies. The island chain contains hundreds of small islands and cays which extend almost 900 miles in a southeasterly direction from the northern tip of the chain which lies less than 100 miles off the southeastern Florida coast. Because of its separate political status, immigration is greatly restricted from islands outside the Bahamas and to a lesser degree from other parts of the world. This has enabled the Bahamians to maintain a considerable degree of racial purityo The original Indian inhabitants were either exterminated or sent off to other islands by the Spaniardso When the English took possession, all Spaniards were evacuated. The forebears of many of the population today descended from the white Tories of the American Revolution and their Negro slaveso Over the next fifty years more Negroes entered, being left on the islands as free men from slave ships intercepted by the Britisho These Negroes, then, are all of the same basic geographic origin as Negroes in the United States, This origin cannot be specifically identified other than West Africa, -5

-6The population of the islands remained relatively stable until the twentieth century when the universal population explosion had its effect, From available data(l9) it is seen that the relative proportions of Negro to white has been approximately unchanged since 1800. This suggests that most of the population increase is natural and not by immigration. In the 1953 census only 853% of the people were born outside of the Bahamaso Since 1953 the population has increased greatly, but good data is unavailable as to source of increase. Undoubtedly, the rate of immigration was increased somewhat. The total population in 1953 was listed at 84,000, and in 1956 it was grossly estimated at 118,000. If this rate of increase continued into 1958, total population would approach 150,000. No valid estimate of immigration or emigration is available. It is felt that the best estimate of stability of the population, under the circumstances, is the evaluation of the birthplace of the people in the sample finally chosen. It was found in the sample that 807% of the people were born outside of the Bahamas. This is remarkably close to the total of 8o.3 born outside of the colony in the 1953 censuso Without further examination, these figures indicate that our sample primarily represents native Bahamians without undue dilution by recent immigration. However, since it is known that the population of Nassau has increased significantly since 1953 beyond increase accountable for by births, one must postulate that most of the increase has resulted from movement of people to Nassau from the other islands of the colonyo In making this statement

-7it is necessary to state that the sample was drawn from essentially the "old city" and newer developed areas of the island may show greater proportions of immigrants. Since this total population is scattered over a great many widely separated islands, sampling of the total population was not considered feasible. A central population concentration was needed for studyo The island of New Providence has become not only the geographic cent of the colony, but also the cultural and population center, It is small (58 square miles) and quite densely populated (790 persons/square mile in 1953). The majority of the people are clustered on the northeast shore in the city of Nassau. The population of the city is unknown since the island is governed as a unit and no definite city boundaries existo It is likely that about 60,000 people live on the island. It was decided to perform a population study of New Providence, not only because it was more convenient geographically and contained a large enough population for sampling procedures, but also that this population probably represented a "melting pot" of the people of the entire colonyo In short, Nassau was selected as the site of the study because of its central location in the colony, its population concentration, and its "melting pot" quality. Because of the latter one feels that a sample studied here may represent not only the people of Nassau, but generally speaking, a cross-section of Bahamianso

CHAPTER III CHOICE OF SAMPLE DESIGN AND SAMPLE VALIDATION In order to describe a characteristic of a group of people, one of two alternatives may be used~ I) to measure this characteristic in the entire group, or 2) measure a sample of the people in such a way that results may be extended to the parent population with a calculable degree of sampling error. The sample size is chosen depending upon the expected variation of the results and the importance of precise accuracy. Various methods of sampling may be usedo The choice depends upon local conditions and problems of selection, the cost of obtaining the information in various sampling plans, the degree of accuracy desired, tiJhe number of variables to be studied and the varying complexity and cost of analysis. All of these factors must be weighed separately and in re lation to o ne another before the most practical. design may be selected In recen't years the applications of sample survey techniques have been extensively studied, The potential usefulness of these techniques in epidemiology is becoming increasingly apparent. In regard to studies on blood pressure, Comstock( 3) and Miall(13) have used types of sample surveyso Many others have selected people for study on a haphazard basis or have chosen special, restricted groupso The results of su.ch stuudies cannot be extended beyond these restricted groups. Boe, on the other hand, did not sample, but undertook the Herculean task of measuring blood pressure of the population of an entire cityo -8

-9Since we wish bette)r to understand the nat;,cual.level of the human blood pressure and how it varies wit.hin the population, we must sample from a defined populationr. One reason for the disparity between the findings of some blood pressure stAd.ies may be that t-he population was not truly represented in the sampleo Another reason is that of the variability which canr be accournted for by the process of sampling. The first cardinal rule is that the sample should be a probability sample, This implies t'hat the sampled units have a known chance of being included in the sampleo Calc-. ulation of' sampling error is dependent upon thiss known probabilityo If one wishes to oversample in a particular segment, he may do so, with proper weighting procedures in subsequent analysis. Quota sampling is the onl.y feasible alternative. This also is a population sampl.e but in this case a specified nurmber in each age, sex, or race group is interviewed. often o:n the basis of convenience It s mpossibl.e in this case to assign sarnpling error to results, but this is not to say that results are necessarily inaccurate In. the st udy at hand, more precision was desired so that probability sampling was choseno Within the realm of probability sampling lies a variety of sample designs. Primary types are simple random sampling, systematic sampling cluster sampling, stratified sampling and vario'us combinations of these, All have their place in specific situations and choice is made according to the factors mentioned. After the sample is drawn, the population may be post-stratified, e.go., by age, sex or race group for analysiso All of these possible designs will not be covered but

-L0o the discussion of specific problems in sampling in this area will describe why the design used was selectedo Goals of the study were: 1) To study distribution of blood pressure in a representative sample of Bahamianso This implies inclusion of all race groups, all age and sex groups and all socioeconomic groups 2) To study these variables as accurately. as possible, mean — while keeping cost at a mininrm. Accuracy implies careful work in all, phases, cognizance of potential sources of error, -with effort-exerted to minimize or eliminate these errors Cost also involves all phases from planning to data collection.-to analysiso In the planning phase, several facts were immediately apparent: 1) Budgetary restrictions were such that actual field work must be accomplished in about five weekso This meant that sample selection and operational planning facilitate rapid observations. 2) In order to include all types of persons in the population, the primary scheme for classification must be according to race and socioeconomic status. Information regarding specilac age-sex-race groups must be obtained by means of poststsratification. 3) Information about geographical location of people on the island was lackingo The following difficulties arose: a) People were too scattered to attempt a sampling study of the entire island.

-11-L b) City limits of Nassau were not demarcatedo c) In the thickly populated areas of the city, streets were reasonably marked, but house numbers were not usedo Houses were often scattered haphazardly in regard to street location, sometimes in clusters in the center of blocks. d) Races and socioeconomic classes were sometimes mixed on the same blocks, making such classification by block occasionally difficulto The perimeter of the area for sampling was then arbitrarily established at major arterial streets which enclosed the majority of the thickly populated areao Total population enclosed was estimated at 3035 thousand, by a rough extension of 1953 census data, Excluded were the mansion district of the white "winter resident" and the fringes of the colored areas, This did not appear to prejudice our goalso Also excluded with the sampling area were tourist hotels, hospital and institutions. Any selection scheme for sampling individuals was made impossible because of lack of any listing by person, Therefore, the household itself must be the selection unit. Simple random. selection of households was not feasible because of cost of mapping all. household locations within the areao Systematic sampling, though possible, was not desirable because of potential bias, but primarily because of time and cost of travel in reaching households in this mannero In order to keep cost at a minimum, the primary sampling unit was designed as a cluster of householdso "Sampling unit" by definition is that which is chosen to represent the whole of like unitso Therefore, the first step of outl.ing these clusters was to drive around every block

within the area chosen, counting the number of dwelling'units in each block. At, the sam..e t ai.m.:e a racial. composJition estimate was made and a socioeconomic score assigned for each: block, This yielded 261 blocks of greatly varying size o Before further...e subdivision, it is necessary to consider how large a samp.e is neededo Accrdin-g too the initial goal.s2 bl. ood pressure readings are'to be compared in three separate racial classes (white, Negro, mixed), the btwo sexes and twelve separate five-year age groupso This means an ult;mate sepa:ration of the data into seventy-two mutulally exc;i.sive groups, each of which is specified as to its race, sex and age composition. Variability witUhin..t;hese groups was one of the aspects to be studied and abo.t which'there was little:forek.owl.edgeo Therefore, it was estimated t;h.at an. average of 40 persorns per group would produce a reasonably reliable mean and standard error for the blood press-ure distribution.o These requirements su.ggested a woriking goal. of about 3000 perso.ns. It was then necessary to estimat.e the n:umber of persons per household in order to se lect an, adequate num.mber of households in the sample. From 1953 census figures'the average household contained 3o8 persons. Since Ich-ild:ren under six years were to be excluded, this average would be reduced to abo.ut three persons per ho.-seholdo Thus a yield of at least 35000 persons requ.ired that 1000 ho-useholds be included in. the sample, Since it was known that; the white and mixed races were minority groups, steps to ensure adequate numbers of these groups were necessaryo Since all blocks had been, assigned racial and socioeconomic scores, they were used as the basis for stratificationo Since the blocks

-13were of varying size, the concept of "measure of size" was usedo It was found`that 40 houses per block was a very rough estimate of average size, All of t+e blocks were then segregated by race and socioeconomic status and reapportioned into measures of size of approximately 40 houses (act+ually 35-45), This involved combining two or more blocks of like race and socioeconomic score where houses were sparse and dividing large blocks into two or more measures, By this sytem 13 measures of size were of predominantly white race, 16 measures were mixed white and Negro races and 1.74 were predominantly Negroo In sampling one must consider the ultimate cost of analysis as a feature of planningo Keyfitz (10) simplified computational scheme for means and standard errors stipulates the use of a stratified design such t;hat strata are constructed approximately equal in size and from these strata) two primary sampling units are chosen at random, The primary sampling unit is the measure of size, or 40 houses, This design proved th.e most convenient for the purpose, Since tlhe goal was 3000 persons, this was equivalent to 26 measures of size at 40 houses each in the final sample This would make the overall sampling ratio one house in eight of the sampled area, Size of stratwrm was dictated by these fact;s 26 measures were needed; oversam.pling in the white areas was needed to guarantee sufficient numberso (Oversampling is permissible in this scheme if weighting procedures are used,) If the entire white group (13 measures) is considered as a stratum, four of these would constitute a doubled sampling ratio The mixed stratum (16 measures) would be sampled with two unitso This leaves 20 units left to be drawn from 174 colored

-14measures, If these are divided into strata of approximately 17 measures from which two are sampled, our requirement of 26 measures or 1000 households would be met, These stratum sizes were felt to approximate equality sufficiently for this computation. The colored measures were grouped into strata on the basis of the socioeconomic score insofar as possible. The alternative method of allocation of portions of the sample to strata was that of optimum allocation, Such a method gives a greater precision of the estimate in relation to cost if sampling within a stratum is performed in a manner directly proportional to estimated variance and inversely proportional to cost per element of data collectiono In this case we had no reason to believe that either variance or cost per element was significantly different over the strata, Therefore, this method offered no advantage. The value of stratification may be viewed as two-foldo It tends to ensure a uniform coverage of all segments of the population and provides a measure of differences between these strata, while minimizing variance within strata, A further generalization is that efficacy of stratification is directly proportional to degree of homogeneity within strata, heterogeneity between strata and the extent to which the quality compared for homogeneity affects the variables under study. In the present study it becomes apparent that the principal value of stratification would be that of more complete coverage of the population, Since the measurements of blood pressure on people are to be the ultimate comparison items, the people must be grouped within

-15strata by age and sex (since these are variables known to be associated with variation in blood pressure) Even if race were completely homogeneous within strata, grouping by age and sex of 240 persons (average yield per stratum) makes such minute comparisons of doubtful validityo Overall stratum means and variances would probably be meaningless, It is important in any study to be aware of the potential sources of erroro These should be known in the planning phase so that steps may be taken to minimize or eliminate them. Kish 1) categorizes two main types of error, those of sampling and those classed as nonsampling. The nonsampling error includes "variable response error" and "biases," "Variable response" includes errors in data collection, punching, coding, as well as nonresponseo Errors of this type tend to cancel one another to some extent. "Bias" implies discrepancies between observations and the true values of the measurements takenO These may be of a systematic and noncanceling nature, Because of variable response errors, repeated observations of the same sample would yield a theoretical distribution of population valueso The mean of this distribution, is the "expected value." The difference between this and the "true value" is the bias0 The size of the bias is unknown in practice, The second category, sampling error, is the standard error of the sampling distribution over all possible samples. The total error is not the sum of these two types (nonsampling error and sampling error) since the former is of a partially self-canceling or compensating typeo This total error is, of course, larger than either of its two constituentsO

-16The errors in the mechanics of data collection and processing are minimized by a foreknowledge of these sources and the u.se of proper care. "Bias" ion the sense of the true value as opposed to measured value of blood pressure is impossible to study in this settingo However, the aim here is to study dist~+Sributi:ons of blood pressure using the measuremenr; tools available and in general. use Potential differences in the use of this meas- uring instrument is discussed in the chapter on observer variation (Chapter V)o. The abili.ty to calc.ulate sampling error is inherent in tZhe samp.le design and is one of the principal justifications for the sampling procedureo The material to follow in t.his chapter is properly considered to be a part of the result of the study but will h-e given ou.t of context for convenience in consider;ing the adequacy of the planning of the sampling method, At the conclusion. estimates of validity of the sample must be made, This is accomplished in four ways~ a) a cou.nt of the total households in sample and the yield in persons examined; b) an estimate of response rate; c) a comrparison of racial. proportions in the sample with census data; and d) a comparison of age arnd sex proportions within these races with cen....sus datao The total n.umber of households in the sample was 1008o The total number of persons with complete data for study was 3057. The average number of persons over age 5 per household was 35.3~ The response rate was 91.5%.o Approximately 2-3% of the sample were male migrant workers i:rn the Unrited States or seamen at home infrequently0

-17If these were to be excliuded, the proportion stundied of those selected would approach 94-95/o The refusal rate was o.8%o These were a mixed group of a few wealthy white persons, uncooperative children and persons with an insurmountable language or intelligence barrier The remainder (7o7%) were not at home after at least three call-backso Adequacy of coverage must be considered sufficient in view of the results to be expected from sample surveys. There is no convenient means by which to quantitate adequacy in this sense. However, the potential bias of the nonrespondents is felt to be minimal in this study because of the high proportion of completeness of the sample chosen., The age-sex proportions for nonrespondents does Ahow that more males than. females were lost to the sample at aboiut; a 2 to 1 ratio. The majority of both sexes followed approximately -the same relative age proportions of the population except that very few children under 10 were not locatedo Comparing racial proportions in the sample with census informatrion. some.discrepancies are seen~ 1953 Census Sample White 1.4 62% 8! 46% Negro 69 069 77 73% Mixed 14.71 3o 9.1 Other 0 o99 The major difference is that "mixed" racial stocks were noticeably fewer in'the sample. Census enumerators in 1,953 did not

ask specifically abou.t race, but estimated s'ubjectivelyo These enumerators were local people, much:more accustomed to detecting racial mixtures, It was also possible that they were hypersensitive to suc.h d.ifferences, Th.e examiners in the present study were not familiar with making such subjective ju.dgments in, regard to race and listed as "nmixed" or..y +those wit-h obvious racial mixing. Whichever was the case, the figures would. indicate that if approximately3 10% of t-he total in 1955 were subtracted from the 14% allotted to "mixed" and divided equally (5%) among the Negro and white totals, the percentages would be remarkably alike. One might speculate that of this 14% called "mixed" in 19553, 5 were principally white, 5% principally Negro, and 4% truly mixed, If this were accepted, it would tend. to strengthen the probability that the sample is valid~ For the purposes of,the study here reported, it would seem wiser not to be overly sensitive to minimal degrees of racial mixture, It seems logical to suppose that a person predominantly Negro or predominantly whkite wo-uld tend to be more 11ike his primary racial stock in regard to his blood pressvure, Sample validation was also attemp'ted by comparing age and sex proportions of the sample with the census data, Comparison was made with extracted data showing proportions by sex for similar fiveyear age groupings,.he census did not provide an age distribution by race, In Table 3l'this comparison shows that there is a reasonably good. agreement throughouto Most of the differences do not exceed ~%, A Chi square test for significance of differences between

-19these age-sex proportions was done showing X =- o68 (df = 23). This would indicate with a high degree of probability that no significant difference exists between the sample proportions and the census proportionso Further substantiation is given by the approximate identity of total sex proportions in the sample as compared to censuso The close approximation here must be considered in great measure fortuitous, since one would expect greater deviations by chance alone, However, these facts further support the claim to sample validityo In sumnmary, the problems in choice of sample design and its implementation have been presentedo The analysis necessary for validation of the sample has been reviewed, with the conclusion that the sample drawn has a strong claim to validityo

-20Table 3 1 COMPARISON OF AGE-SEX PROPORTIONS (PERCENT) OF THE POPULATION AS SHOWN BY 1955 CENSUS AND THE SAMPLE STUDIED AS TO BLOOD PRESSURE NASSAU, 1958 Numbers Proportion (Percent) of Total Age Males Females Male s Females Group in Years 1953 1958 1953 1958 1953 1958 1953 1958 ( Inclusive) Census Sample Census Sample Census Sample Census Sample 5-9 2524 188* 2506 203* 6.53 6.83* 6,48 7537* 10-14 2194 164 2191 169 5,68 5o96 5~67 6.13 15-19 2104 168 2362 177 5 44 611 6oll 643 20-24 1941 148 2401 201 5.00 5~38 6o21 7.30 25-29 1746 119 2148 154 4,52 4k32 5.56 5.59 30-34 1.458 85 1914 111 3577 3.09 4,95 4.03 55-59 1579 71 1811 128 4.09 2.58 4,69 4,65 40-44 1261 84 1583 105 3526 3.05 5.00 3581 45-49 935 80 1135 89 2o42 a291 2.94 3.23 50-54 625 48 876 59 1.62 1,74 2,27 2.14 55-59 427 25 574 43 1,10 o 91 l,49 1.56 6o+ 904 52 1462 82 2.34 1,89 3.78 2,98 Total 17688 1232 20963 1521 45.77 44.76 55,15 55o22 * Actually represents ages 6 through 9 years inclusive. Sample proportions corrected for oversampling of white stratum. "Mixed" omitted from sample calculations because of difficulty of this type of correction on small numbers

CHAPTER IV THE INTERVIEW AND EXAMINATION PROCEDURE One of the principal difficulties in comparing studies on blood pressure is that procedures vary between studies and often vary within individual. studies, There was no single, previously reported, study wh.ich was felt to be appropriate as a model in all essential respects o Blood pressure is known to vary under many circumstances; the magnitude and consistenlcy of these variations are largely unknowno In addition to age, sex, body build and arm girth, all of which can be measured or classified with a fair degree of accuracy, there are other factors such as emotion and state of rest, which cannot be classified with any hope of accuracyo Also, posture in which blood pressure is measured, e,g,, sitting, standing or recumbent, is known to affect blood pressure. This effect was not stu.died here, but was rather held constant by using the same posture throughout. This study was designed for examinations t;o be done in the home. Home examinations gave promise'to more complete coverage of families. It was also felt that a casual blood pressure taken at home would provide a more nearly accurate measure of the blood pressure the person "lives with" when not subjected to stress. This was undoubtedly counterbalanced to some degree by the presence of a strange person. in the home. Examinations were, of necessity, performed when the individual subject could be found at home, This frequently involved several callbacks before an entire family could be completed, For this reason, -21

-22variation in. blood pressure due to time of day, elapsed time since last meal, fatigue, etc., could not be compensated for by standardization. Most of the people were seen between 2-9 PoM,, although many were seen in the morningo Examinations were done in almost equal numbers by one physician (study director) and four senior medical students from the University of Michigan. The interview itself was standardized so that information was gathered in a systematic mannerO This helped to ensure completeness of information in addition to a standard order to the entire Qxamination, See Page 1, Appendix (Data Collection Sheet), All identifying information was obtained in the following mannero 1) arm girth measurement; 2) first casual blood pressure reading; 3) first pulse rate; 4) height measurement; 5) weight measurement; 6) second blood pressure; 7) second pulse rate, Blood pressures were taken from the left arm with the subject in the sitting position, Recorded were systolic, fourth phase diastolic (muffling) and fifth phase diastolic (disappearance). Studies have been done in all positions (sittin.g, standing, recumbent), but sitting appears to be more commonl.y used and was much more convenient in home examinations. Four of the five examiners were fourth year medical students and relatively unaccustomed to recording fourth phase diastolic pressures. However, standardization tests demonstrated that all were recording similar readings (see Chapter V)o Pulse rates were taken immediately after blood pressure readings, using a 15 second timing procedure,

-253Routinely, height and weight measurements were recorded after the first blood pressure and pulse The cuff was left on the ar.m. since i;t would speed up the procedure and more accurate weights were not necessary for the purposeo Scales used were of the simple bathroom type, portable in a briefcaseo Since they were spring scales, careful adjusting of zero point was necessary. All five scales agreed within 2 or 3 pounds when checked It was found, also, that considerable variabili:Cty resulted from placement of the scale on a rug.or an uneven surfaceo In some homes a level surface was difficult to findo Height was measured to the nearest half inch with a firm cloth t+ape after marking height on a wall from the subject standing without shoes. Arm. girth was measured to the nearest quarter inch at the point of maximum. circumference of the upper left arm. This firm cloth tape was felt to be sufficientily comparable in accuracy to a metal. flexible tape. After'these measurements, the second blood pressure and p-:_lse were recorded, Total. elapsed time between measu.rements was 1-2 minutes. Sphygmoma:nometers used were of the aneroid t'ype, compared daily for accuracy using a mercury-type instrument as a standard (see Chapter V) After thorough familiarization with the methods and procedures of examination, all examiners engaged in a field test in order to ensure uniformity of information elicited and the routine of examination. An area of the city not included in the sample was used in the pre-test.

-24The examiners worked in pairs, observing each other's techniques until it was felt that all were comparable. The progress of the work was facilitated greatly by the almost universal cooperation and interest of the people. Many were quite concerned about "high blood" in themselves and their families. Not a small problem were those people in the neighborhood (not included in the sample) who requested examination. Persons were told either that their blood pressure was normal, that it was borderline high, or high, The criterion used for the most part was diastolic pressure. All those over 110 were referred to their physician or to the Colonial Medical Serviceo Diastolic pressures 100110 were called borderline and advised to consult their physician at a later date. Others were called "normal", Some exceptions were made in young people, some older persons with extremely wide pulse pressures, and other with suggestive symptoms0 These also were referred to their physician. In summary, the location, content and method of the interview and examination have been discussed, as well as the referral system of persons with elevated blood pressures.

CHAPTER V ANALYSIS OF POTENTIAL OBSERVER ERROR Prior to actual observations, it was necessary to obtain assurance that the five observers, using similar techniques and similar instruments, did not introduce significant systematic differences into their readings of arterial blood pressureo Aneroid sphygmomanometers were used because of their greater practicality for examinations in the home. All of the instruments used were standardized daily against a mercury type instrument. None of the instruments varied by more than 2 mmo at any level. At the time of the experiment on observer error, one aneroid instrument was defective and the standard instrument was used by one observer. Subsequently, a new aneroid was used in the actual field work, The experiment to test observer error was done prior to any collection of datao Five different persons were to be observed by the five field examinerso The examinations were arranged so that the effects of the three potential variables could be measuredo These were 1) varition due to the subject himself; 2) variation due to the examiner; and 3) variation due to the order in which the persons were examined, These variations may be measured by a Latin Square analysis of variance design, Each subject and each observer were assigned a number, A Latin Square involving these numbers was then randomly selected *to determine order of observation, Each observer examined a different subject at the same time. When all had finished, each moved to the next subject in tl~e prearranged ordero The examination -.25

-26of each subject by each observer included three successive blood pressures and pulses. Both fourth and fifth phase diastolic pressures were recorded. Pulses were taken with fifteen second readings4 In the analysis of the above, three components of variances were compared~ It was found that: 1) Variation between subjects was great This also included a component of variability within the.subject himself. 2) Variation due to order of observation was slight or minimally significant, 3) Variation due to observers was insignificant for all blood pressure readings However, significant observer variation was present in pulse readings, This variation in regard to pulse was probably due to taking of fifteen second pulses where an error in counting one or two beats could produce such an error as seen here, This finding would not appear to be of such magnitude that its use for general comparisons is losto The level of significance was taken at the 5% level, utilizing the "F" testo It must be remembered that this determines a range of "F" ratios which enable one to accept or reject a null hypothesiso If these "F" ratios fall outside this acceptable zone, the null hypothesis is rejected For example, the first hypothesis was that of no difference between subjects. The "F" ratio fell outside the acceptable zone and, therefore, we state that there is a significant difference between subjects. However, "'F" ratios as such cannot be treated in a strictly quantitative fashion in order to compare accuracy of observing different

-27variables. These ratios tell us how strongly a belief should be held that no difference existso For example, (Table 5l1) the "F" ratio for systolic mean observer variation is ol935 This is equivalent to saying that if no difference between observers exists, an "F" ratio would be obtained as large as this between 90-95 times in a hundred. (.95 > p > 90) Table 5o1 LATIN SQUARE COMPON{ENTS OF VARIANCE TEST FOR VAR.IATION DUE TO OBSERVER Reading Effect "F" Ratio "p" Value Systolic (Mean) Subject Variation 85o215 p < o0005 Order Variation 4,147 o025 > p > oO10 Observer Variation ol93 o95 > p > o90 Systolic (First Subject Variation 39o133 p <.0005 Reading) Order Variation 2~989 olO > p > o05 Observer Variation ol143 o975 > p > o950 Diastolic (Fourth Subject Variation 52o908 p <.0005 phase) Mean Order Variation 3538 o050 > p > o025 Observer Variation 3015 0lO > p > o05 Diastolic (Fourth Subject Variation 37o594 p <.0005 phase) First Order Variation 25331 ~25 > p > lO0 Reading Observer Variation 2 597 o10 > p > o05 Diastolic (Fifth Sucbject Variation 38085 p < o0005 phase) Mean Order Variation 0278 ~ 90 > p > o75 Observer Variation lo08 o50 > p > ~25 Diastolic (Fifth Subject Variation 29068 p < o0005 phase) First Order Variation 0o440 o90 > p > o75 Reading Observer Variation o962.50 > p > ~25 Pulse (Mean) Subject Variation 59~856 p <.0005 Order Variation oO763 o75 > p > 050 Observer Variation 7~792.005 > p > oOOl At 5% level (two-tailed test) an "F" value of greater than 4.12 indicates significant difference,

-28Further study of potential observer variation is possible through analysis of the frequency of terminal digits in blood pressure readings. All examiners were instructed to read to the nearest 2mm. marking, i.e., 0,2,4,6.,8. Theoretically, the(distributions will be equal on these five numbersO It is unfortunately true that habit formation usually tends to cause a lumping at the usual 5 mm, breaks, e.go, 0 and 5o This is probably more true at the very high and very low levels of the measurement scale, since in these ranges one is usually not concerned clinically with precise readings, Table 5,2 shows the varying proportions of readings of terminal digit by the various examiners, The principal deviation from an even distribution is the excess proportion at the 0, at the expense of the 2 and 8. There were also some inadvertent readings at the 5 mark, at the expense of 4 and 6. However, considering 4, 5 and 6 together, there is no lumping apparent in this area, One might expect a degree of random deviation from an expected 20% in each of the five read pointso One examiner, No. 5, however, was remarkably close to this distribution, All of the others shared the lumping at 0. This disproportion at 0 does not cause great concern, although it was higher than hoped, Many investigators have found this true to an even greater extent. In this case it means, in effect, an error of plus or minus 2 mmo in reading which is within the scope of expected reading erroro Nor does it invalidate the attempt at greater accuracy by reading to the nearest 2 mm, rather than to the nearest 5 mm, Actually, this distribution should indicate that greater care in reading will provide a much more accurate description of blood pressure distributions than is possible with fewer reading points,

-2.9When the read points are grouped 0-4 and 5-9, it is seen that the proportions fall close to the expected 60%-40% distribution. This distribution is expected since three reading positions (0,2,4) are contained in the 0-5 portion of the dial, and only two positions (6,8) in the 5-9 portiono Systolic 0-4 65o6% 5-9 34.o4 Diastolic 0-4 - 64o8% 5-9 -55 2% Expected 0-4 60% (both) 5-9 40 It is of interest that proportions for systolic and diastolic are very similarO Although some observers showed varying patterns in reading positions for systolic as opposed to diastolic, the general distribution remained much the same. In summary, the analysis of terminal digits demonstrates that even wi;h care in reading, some individual variation is seen. between observers on reading the same variable, eogo, systolic pressure, and that the pattern may vary somewhat with a different variable (diastolic pressure). However,'these differences are seen to be not incompatible with expectations and do not prejudice the distribution. It is also seen that a quite good approximation of a continuous distribution may be measured by more precise reading of the instrumento Previously noted was the finding in the components of variance test tihat no significant systematic difference was apparent between observers in the reading of blood pressureo

-30Table 5.2 TERMINAL DIGIT ANALYSIS PROPORTION OF READINGS BY TERMINAL DIGIT PER EXAMINER NASSAU, 1958 Systolic (first reading) I Number l of Terminal Digit Obser- 1 2 4 5 6 7 8 9 X vations 1 602 30.2 20.0 20.9 0.3 10.0 18.8 2 714 44.7 10.5 0.2 21.3 1.1 12.6 9.5 3 534 27.7 0.2 13.3 0.2 16.1 3.4 20.6 o.4 18.0 0.2 4 550 36.3 13.2 0.2 16.0 18.4 6.9 o.6 8.9 5 677 17.9 17.2 19.8 23.2 0.1 21.32 3057 30.85% 0.035 14.528% 0.135 18.656% 4.03% 14.50% 0.19% 14.88% 0.03% Diastolic (fifth phase, first reading) k Number.g of Terminal Digit Obser- 0 1 2 3 4 5 6 7 8 9 vations 1 602 35.5 11.3 16.45 1.8 15.9 18.9 2 714 55.3 3.8 17.2 0.9 15.5 7.1 3 534 27.7 14.8 0.2 14.4 3.4 21.0 0.4 18.2 4 530 36.0 14.5 0.2 13.8 9.1 10.9 15.5 5 677 20.1 17.4 22.9 20.5 19.1 3057 34.69% 11.81% o.6% 16.87% 2.69% 16.51% 0.06% 15.14%

CHAPTER VI COMPARISON OF DIASTOLIC READING METHOD It; was previously noted in the chapter on observer variation that observers were able to read the fourth of diastolic pressure (muffling of sounds) equally as accurately as t;he fifth phase (disappearance of sounds). The questions as to which reading method approximates more closely the true diastolic pressure and which method is the most reliable and reproducible by different observers constitute a minor international disagreemento From the present data it appears that there is ample justification for each sideo Since, in this study, both fourth and fifth phases were recorded and age-sex-race specific means and standard errors were computed, further knowledge on this question was soughto Table 601 shows the age-sex-race specific mean differences between the two readings. It is seen that the greater differences are in the young, with a gradual diminution with; ageo However, the mean of the mean differences is only 306 mm. The difference between the races and sexes appears negligible. In regard to the question of which reading may be more variable (as evidenced by the standard deviations of the diastolic pressure distributions of both fourth and fifth phases for age-sex-race specific groups), a sign test was performedo This test showed that the variability of diastolic (fourth phase) was the greater in 24 groups, diastolic (fifth phase) was the greater in. 22 groups, and a tie resulted in 2 groups, 3.L

-32Table 61o DIFFERENCES BETWEEN DIASTOLIC READING METHOD DIASTOLIC (Fifth Phase)- DIASTOLIC (Fourth Phase) BY AGE, SEX, AND RACE NASSAU, 1958 Age White White Negro Negro Male Female Male Female;6-9 4 62 3.1-3 327 4,04 10-14 7.50 4oo00 4.55 4.54 15-19 5o54 4.45 5.84 5.30 20-24 5.94 3.91 5o06 4.77 25-29 4.20 1.76 4.43 3.05 30-34 3089 4.45 4.33 3o03 35-39 4o09 1078 4.39 2.79 40-44 3.09 1o56 3.58 3.05 45-49 2.45 1o90 2o35 2,83 50-54 3o00 0o40 3592 5.43 55-59 1o99 1o87 3500 2,66 60 + 2.14 2.60 3~45 4,15 Mean Difference 40o4 2o56 4.01 3 80 Mean of Mean Differences = 3o60mmo Mean Difference (Negro) = 3591mmo Mean Difference (White) =3 529mm,

_33and a tie resulted in 2 groupso Thus, no significant difference in variability between the two reading methods was apparent. In most of the sex-race categories, there was no definite relationship to age in regard to greater variability of one diastolic reading method as compared with the othero However, and for reasons unknown. the Negro male group showed a regular pattern with the fifth phase more variable up to the 25-29 age group, and the fourth phase more variable thereaftero No such pattern existed in the other race-sex groups. It must be noted again that four of the five examiners were medical students who by training are more accustomed to listening for and recording the fifth phaseo In summary, the findings of this study in regard to diastolic reading method are. 1) the difference between fourth and fifth phase is somewhat smaller than expected from previous data, 2) these differences tend to be greatest in the young, and 3) there appears to be ittle to coose in regard to vaabi y and reproducibility of either method

CHAPTER VII THE DISTRIBUTION OF BLOOD PRESSURE READINGS BY AGE, SEX AND RACE It is hoped that the material presented in this paper will be such that it may be compared with any existing or future population studies on blood pressureo For this reason, the recommendations of the American Heart Association in regard to the use of data from casual first readings and the use of the fifth phase of diastolic pressure in the analysis. Data describing distributions of the fourth phase of diastolic pressure are presented for potential comparison with studies using this method alone. A, Arithmetic Means and Standard Errors of the Distributions All participants in the study were classified by a poststratification procedure into age, sex and race specific groups, The age span within groups was five years, with the exceptions of the 6-9 year group, and the "60 and over" year groupo Race categories were pure Negro, pure white, and "mixed". A few Asians who fell in the sample were not included in the analysis. As previously noted, the "mixed" group was too small to permit detailed study. There were then 48 age-sex-race specific groups, categorized by two races, two sexes, and twelve age groups In order to compare different variables in these specific groups it was desirable to obtain measures of central value and variability within groups. To do this we must know something of the underlying distributions, since for full utilization of means and standard -34

-35errors we must be assured of an approximately normality of the distributions, It was noted by Boe( ) that systolic pressure showed a definite positive skewness in all age groups, becoming more skew as age progresseso However, Boe's distributions of systol.c pressure by age were shown to approximately the logarithmically normal distribution, The diastolic pressure distributions, on the other hand, agreed much more closely with the normal distribution, particularly in. the young age groups. The results of the present data corroborate these findings and are also shown graphically in Figures 71l (eight parts)o Systolic pressure distributions in several age groups were plotted on log normal probability paper; the results agreed generally with these conclusions It is known that the means of samples from non-normal distributions will be normally distributedo The use of a standard deviation for comparison is further strengthened by the fact that the distributitons compared across age groups are skewed in the same direction., varying somewhat in. degree. The findings for the age-sex-race specific groups in terms of the mean, standard error of the mean, standard deviation of the distribution and coefficient of variation are shown in Table 7olo It is seen that the standard errors of these means are generally small enough to give reasonable assurance of the probable true population valueso These standard errors were computed by the method of Keyfitz(10)o

MEANS OF BLOOD PRESSURE READINGS IN AGE, SEX AND RACE GROUPS, WITH STANDARD ERROR OF MEAN, DISTRIBUTION STANDARD DEVIATION AND COEFFICIENT OF VARIATION. NASSAU, 1958 Systolic Blood Pressure, White Race, by Sex and Age Males Females Systolic Standard - No. In Systolic Standard No. In Age Mean Deviation s/S Group Age Mean Deviation s/S Group 6-9 104.33 + 2.18 14.78.142 45 6-9 105.40 + 1.28 8.78.083 47 10-14 114.82 + 1.78 9.42.082 28 10-14 116.86 + 2.21 13.26.113 36 15-19 129.51 + 1.38 7.92.061 33 15-19 119.68 + 2.81 15.65.131 31 20-24 128.94 + 2.43 14.17.110 34 20-24 121.81 + 3.46 22.70.186 43 25-29 129.53 + 2.83 15.51.120 30 25-29 119.46 + 2.99 18.69.156 39 30-34 124.24 + 3.11 19.16.154 38 30-34 121.06 + 3.17 18.20.150 33 35-39 125.45 + 3.51 16.46.131 22 35-39 121.33 + 3.19 16.59.137 27 40-44 134.45 + 2.44 11.44.085 22 40-44 147.76 + 5.59 27.95.189 25 45-49 130.45 + 4.39 19.62.150 20 45-49 142.40 + 2.50 11.18.079 20 50-54 139.71 + 2.18 8.15.058 14 50-54 146.73 + 3.91 15.13.103 15 55-59 138.91 + 7.17 23.80.171 11 55-59 156.06 + 6.17 24.68.158 16 60+ 145.25 + 4.79 25.34.174 28 60+ 153.77 + 5.02 27.51.179 30 Systolic Blood Pressure, Negro Race, by Sex and Age Males Females Systolic Standard No. In Systolic Standard No.In Age Mean Deviation Group AgeMean Deviation s/ Group 6-9 101.59 + 1.02 12.66.125 154 6-9 102.60 + 0.50 6.44.063 166 10-14 109.31 + 0.63 7.53.068 143 10-14 111.70 + 1.01 11.91.107 139 15-19 125.01 + 1.41 16.62.133 139 15-19 118.84 + 1.20 14.84.125 153 20-24 128.00 + 1.22 13.53.106 123 20-24 119.67 + 1.45 18.91.158 170 25-29 132.90 + 1.83 18.12.136 98 25-29 126.84 + 2.01 22.73.179 128 30-34 135.84 + 2.68 20.42.150 58 30-34 132.29 + 1.49 13.81.104 86 35-39 142.89 + 5.03 37.32.261 55 35-39 140.31 + 2.15 22.45.160 109 40-44 142.22 + 3.19 26.13.184 67 4o-44 149.15 + 4.89 45.62.306 87 45-49 144.81 + 2.88 2.59.163 67 45-49 152.13 + 2.17 18.79.124 75 50-54 155.76 + 4.90 29.79.191 37 50-54 155.72 + 5.39 36.54.235 46 55-59 144.31 + 9.10 36.40.252 16 55-59 182.10 + 5.95 32.61.179 30 60+ 168.21 + 8.40 48.22.287 33 60+ 176.02 + 3.45 26.94.153 61

Table 7.1 Cont'd Diastolic Blood Pressure, Fourth Phase, White Race, by Sex and Age Males Females Diastolic Standard No. In Diastolic Standard No. In Age Mean Deviation s/D Group Age Mean Deviation s/D Group 6-9 66.11 + 2.41 16.17.245 45 6-9 69.94 + 0.98 6.72.096 47 10-14 68.79 + 1.75 9.26.135 28 10-14 71.11 + 1.23 7.38.104 36 15-19 76.55 + 2.39 13.72.179 33 15-19 70.97 + 1.27 7.07.100 31 20-24 78.82 + 3.17 18.48.234 34 20-24 77.12 + 1.57 10.30.134 43 25-29 78.53 + 2.11 11.56.147 30 25-29 77.79 + 0.81 5.06.065 39 30-34 81.00 + 1.43 8.81.109 38 30-34 80.79 + 2.22 12.74.158 33 35-39 82.45 + 3.29 15.43.187 22 35-39 79.67 + 0.85 4.42.055 27 40-44 87.18 + 1.63 7.64.088 22 40-44 90.04 + 2.36 11.8o.131 25 45-49 81.80 + 2.83 12.65.155 20 45-49 86.95 + 2.30 10.28.118 20 50-54 89.21 + 1.50 5.61.063 14 50-54 87.87 + 1.93 7.47.085 15 55-59 84.55 + 1.77 5.88.070 11 55-59 85.75 + 2.46 9.84.115 16 60+ 81.82 + 2.12 11.21.137 28 60+ 82.10 + 1.93 10.58.129 30 Diastolic Blood Pressure, Fourth Phase, Negro Race, by Sex and Age Males Females Diastolic Standard No. In Diastolic Standard No.In Age Mean Deviation s/ Group Age Mean Deviation Group 6-9 66.98 + 0.76 9.43.141 154 6-9 68.64 + 1.12 14.43.210 166 10-14 70-78 + 0.60 7.18.101 143 10-14.71.78 + 1.17 13.79.192 139 15-19 79.42 + 1.05 12.38.156 139 15-19 76.36 + o.88 10.89.143 153 20-24 83.78 + 1.13 12.53.150 123 20-24 78.61 + 1.18 15.39.196 170 25-29 87.34 + 1.46 14.45.165 98 25-29 83.53 + 1.13 12.78.153 128 30-34 91.90 + 2.25 17.14.187 58 30-34 87.58 + 1.07 9.92.113 86 35-39 95.80 +'2.86 21.22.222 55 35-39 90.78 + 1.70 17.75.196 109 40-44 95.80 + 1.94 15.89.167 67 40-44 93.68 + 2.65 24.72.264 87 45-49 96.54 + 2.27 18.59.193 67 45-49 94.63 + 1.30 11.26.119 75 50-54 97.11 + 2.71 16.48.170 37 50-54 96.91 + 2.61 17.70.183 46 55-59 89.94 + 4.82 19.28.214 16 55-59 100.63 + 1.55 8.49.084 30 60+ 93.03 + 4.89 28.07.502 33 60+ 95.56 + 2.96 23.12.247 61

Table 7.1 Cont'd Diastolic Blood Pressure, Fifth Phase, White Race, by Sex and Age Males Females c Diastolic Standard No. In Diastolic Standard No.In Age Mean Deviation s Group Age Mean Deviation s/~ Group 6-9 61.48 + 2.10 14.09.229 45 6-9 66.83.+ 1.84 12.62.189 47 10-14 61.29 + 1.96 10.37.169 28 10-14 67.11 + 1.59 9.54.142 36 15-19 71.00 + 1.87 10.73.151 33 15-19 66.52 + 1.29 7.18.108 31 20-24 72.88 + 1.99 11.60.159 34 20-24 73.21 + 1.23 8.07.110 43 25-29 74.33 + 2.71 14.85.200 30 25-29 76.02 + 0.67 4.19.055 39 30-34 77.10 + 1.43 8.81.114 38 30-34 77.45 + 3.11 17.85.230 33 35-39 78.36 + 3.69 17.31.221 22 35-39 77.89 + 1.13 5.88.075' 27 40-44 84.09 + 1.93 9.05.108 22 40-44 88.48 + 2.38 1l.90.134 25 45-49 79.35 + 2.08 9.30.117 20 45-49 85.05 + 1.86 8.31.098 20 50-54 86.21 + 2.13 7.97.092 14 50-54 87.47 + 2.14 8.28.095 15 55-59 82.54 + 2.10 6.97.084 11 55-59 83.88 + 7.25 29.00.346 16 60+ 79.68 + 1.83 9.68.121 28 60+ 79.50 + 0.95 5.21.066 30 k1 Oc Diastolic Blood Pressure, Fifth Phase, Negro Race, by Sex and Age Males Females Diastolic Standard No. In Diastolic Standard - No.In Age Mean Deviation s/D Group Age Mean Deviation /D Group 6-9 63.71 + 0.90 11.17.175 154 6-9 64.59 + 0.79 10.18.158 166 10-14 66.23 + 0.99 11.84.179 143 10-14 67.24 + 1.12 13.20.196 139 15-19 73.58 + 1.27 14.97.203 139 15-19 71.05 + 1.17 14.47.204 153 20-24 78.73 + 1.22 13.53.172 123 20-24 73.84 + 1.14 14.78.201 170 25-29 82.92 + 1.35 13.37.161 98 25-29 80.48 + 1.25 14.14.176 128 30-34 87.57 + 2.10 16.00.183,8 3Q-34 84.55 + 1.12 10.38.123 86 35-39 91.42 + 2.46 18.25.200 55 35-39 87.99 + 1.55 16.18.184 109 40-44 91.57 + 1.73 14.17.155 67 40-44 90.63 + 2.65 24.72.273 87 45-49 94.19 + 2.21 18.10.192 67 45-49 91.80 + 0o98 8.49.092 75 50-54 93.19 + 1.93 11.73.126 37 50-54 91.47 + 2.52 17.09.186 46 55-59 86.94 + 4.00 16.00.184 16 55-59 97.97 + 1.42 7.78.079 30 60+ 89.58 + 4.29 24.62.275 33 60+ 89.38 + 1.69 13.20.148 61

Table 7.1 Cont'd Pulse Pressure, White Race, by Sex and Age Males I Females Ag Mean Pulse Standard s/P No. In Mean Pulse Standard s/P No.In Age s/P Age Pressure Deviation Group Pressure Deviation Group 6-9 44.18 + 1.91 12.82.290 45 6-9 38.57 + 2.29 15.71.407 47 10-14 53.54 + 3.54 18.73.350 28 10-14 49.81 + 1.01 6.6.122 36 15-19 58.52 + 1.64 9.41.161 33 15-19 53.48 + 1.72 9.58.179 31 20-24 56.06 + 2.18 12.71.227 34 20-24 48.51 + 2.68 17.58.362 43 25-29 55.20 + 2.18 11.95.216 30 25-29 43.44 + 2.81 17.56.404 39 30-34 47.13 + 1.92 11.83.251 38 30-34 41.79 + 2.17 12.46.298 33 35-39 47.09 + 1.45 6.80.144 22 35-39 43.52 + 2.69 13.99.321 27 40-44 50.64 + 1.53 7.18.142 22 40-44 59.28 + 3.84 19.20.324 25 45-49 51.05 + 6.12 27.36.536 20 45-49 57.35 + 2.28 10.19.178 20 50-54 53.50 + 1.84 6.88.129 14 50-54 59.27 + 2.08 8.05.136 15 55-59 56.36 + 8.20 27.22.483 11 55-59 72.19 + 3.71 14.84.206 16 60+ 65.57 + 3.19 16.88.257 28 60+ 74.37 + 4.91 26.91.362 30 Pulse Pressure, Negro Race, by Sex and Age Males Females Mean Pulse Standard No. In Mean Pulse Standard No.In Age Pressure Deviation s/P Group Age Pressure Deviation s/P Group 6-9 38.57 + 0.97 12.03.312 154 6-9 37.89 + o.80 10.30.272 16 10-14 49.81 + 0.74 8.85.178 143 10-14 43.10 + 1.41 16.62.386 139 15-19 53.48 + 1.38 16.27.304 139 15-19 51.50 + 0.82 10.14.197 153 20-24 48.51 + 1.18 13.09.270 123 20-24 49.28 + 1.01 13.17.267 170 25-29 43.44 + 0.77 7.62.175 98 25-29 49.60 + 1.18 13.35.269 128 30-34 41.79 + 1.51 11.51.275 58 30-34 48.26 + 1.18 10.94.227 86 35-59 43.52 + 3.29 24.41.561 55 35-39 52.20 + 1.26 13.15.252 109 40-44 59.28 + 2.43 19.90.336 67 40-44 50.72 + 2.65 24.72.487 87 45-49 57.35 + 4.50 36.45.635 67 45-49 59.57 + 1.91 16.54.278 75 50-54 59.27 + 3.24 19.70.332 37 50-54 62.57 + 3.48 23.59.377 46 55-59 72.19 + 6.39 25.56.350 16 55-59 57.37 + 3.48 23.59.504 46 60+ 74.27 + 4.95 28.41.383 33 60+ 78.73 + 2.68 20.93.266 61

-4060 DIASTOLIC PRESSURE 50 _ A WHITE MALES..~~\'AGE ---- 20 - 24 35 - 39 40 50-54 ID~ j:: <'. - i........ 65 Over V o o o oio o w 0 e 0 0 0 C BLOOD PRESSURE, mm Figure 7.1 Frequency Distributions of Systolic and Diastolic Pressures by Sex, Race and Selected Age Groups, Nassau, 1958 (Eight Parts)

-4160 DIASTOLIC PRESSURE 50 - o WHITE FEMALES AGE ---- 20 - 24 35 - 39 40 - 40 50 - 54 0) — i- 65 - Over 0 30 I I 20 /. / / / 6 \ = / / 0 I I I I I I I I I I I v o o o t o o o0 - c?gBLOOD PRESSURE, mm BFigure 7.1 (Cont, mm Figure 7.1 (Cont'd)

-4260 DIASTOLIC PRESSURE o50 _A NEGRO MALES AGE 20 - 24 35 - 39 4 0............ 50 - 54 65 - Over z iu 30w / 0.. /.. / \.. e 20 i / Fi \'..o ~/ 1~\A A^'.. 0 o, o, o, o, / / ^ A o,..-.A. / I I I I I! I I I I I I I I 11 v oo 0 BLOOD PRESSURE, mm Figure 7.1 (Cont'd)

-4360 DIASTOLIC PRESSURE 50 - * NEGRO FEMALES AGE 20 - 24 35 - 39 40 50 - 54 - 65 - Over z w ~ 30 0. i \'.: O 0.../... Figure 7.1 (Cont'd) o- a a a....a a a.......' BLOOD PRESSURE, mm Figure 7.1 (Cont'd)

-44> - OL 692-09Z - 6SZ - 09Z ui cr O e- O Z. w w 0.. j Ilsa~~~~~~~~~~6.; - O 6-CZ a. tI I I I I6L - 0 iot 0 in o w u M ro t) to IJ 1 0 - 622-022 0 ") I'I m,) <- 61Z -00Z I i %__ d - 610-oO /.a < _601 -00z 1 _ ~661 -061 -I _ 681 -091 *^ ^\~6L OL1 0 II I I I -I I691 -091 691'091 d 6gl -01O ^^^.^ -~ 621 - 021 Jb~. <^... ^5^ ^-61 1 -Ol 601 -001,..'b 61 - 66 06 68 -09 I I I I I I I I.I I. > 0 0 0 0 0 0 Ut) ) Ct -

PERCENT o ro O. cO o 0 0.0 0 0 0 <- 79 80- 89 90- 99 100 -109 - C I 110-119 09 -...s 120-129 130-139 140-149 -,' 150- 159 4 6.... 160 -169 0.': 180- 189 0... 2202f- -t / 190- 199 210- 219 Q 1 220-229 / I a 230-239 6 \ 00 ClO ra1C m m 250-259 260 - 269 ~T — 2 3O 239 (:~_58

PERCENT 0 0 0 0 0 0 < -79 80- 89 90- 99 100-109 9 110-119 120-129 130-139 - 1 J 140-149 150-159- 160-169 CD 190-199 200-209 I: I 210-219 o.. E 0 -0-<../."' i' I "' 3> 0 220 -229 - Z O (j o ( o0 G.II: S 0 -1 230-239 o 0 270 - < 240 -2.49 m ^-9 250 - 259 - I -9+

50 SYSTOLIC PRESSURE 40 0 NEGRO FEMALES AGE 20 - 24 35 - 39 30 / \....... 50 - 54 I \ / / \ - - 65 - Over z w / U / / \ w O- 20 / / A/'-y' 0) M. I 0 -............. / /...o I I- - I I ^I1 I I —--- I I I I~. 0 0 _ N c')` - a 0- c. 0 V, 0,! 0 0, 0 0 0 0 0, 0, CM __ -- __.. tu it 0) 0) 0 - U N K) J) o O 0 00 0 0 0 0 BLOOD PRESSURE, mm Figure 7.1 (Cont'd)

-48The standard deviations of the distributions were calculated by multiplying the standard error of the mean by the square root of the number in the specific group. This calculated deviation has been based on the assumption of a low intra-class correlation, since the calculation contains components of variation due to differences between strata, units and households in addition to variation between individuals within age-sex-race specific groups. Boees exhaustive work in Norway provides the best basis for comparison. His findings were that the distribution standard deviations as well as the means show an orderly progression from the youngest age group to the oldest (with some diminution in the very old). To demonstrate this relationship a ratio of the distribution standard deviation to its corresponding mean (coefficient of variation) has been calculated. If this ratio remains constant;, the standard deviation of the distribution is increasing or decreasing in direct relat;ionship with the magnitude of the mean. If the coefficient of variation increases, the distribution standard deviation is increasing proportionately more rapidly with age than is the mean. In Boe's) study as well as Comstock's'3) the latter effect is noted for systolic pressures, i.e,, that the distribution standard deviation increases proportionately more with age than does the mean, The present study shows this same trend for systolic pressureso Diastolic pressures follow this pattern only for Negro malesa Boe's and Comstock's works also show a trend toward stability of coefficients of variation in regard to diastolic pressureo

-49In regard to the distribution standard deviations in the present study and the two studies mentioned, it is seen that the range of magnitude is roughly similar in spite of differences in numbers of persons studied. This similarity would be expected, since the dispersion of blood pressures at a given age might be expected to be generally similar even though values of the means might vary to some extento It; might be mentioned at this point that the above two studies risk a considerably greater potential bias as a result of nonrespondents in the population than does the present study. This does not imply that their results are in error, but* only that the risk of error is greater. In this section, then, it has been seen that the means and standard deviations of the distributions of systolic pressure all progress upward with age. A general similarity is noted with two similar studies. The increasing trend with age is seen to be proportionately greater for the standard deviations of the systol.ic distributions than for their corresponding meanso The sanle parameters for the diastolic distributions fail. to demonstrate a similar trend. since the means progress upward with age while the standard deviations are quite variableo However, it is felt that a slight trend similar to the systolic pressure trend is present. Means of the distributions for several studies will be compared in a later sectiono This trend of coefficients of variation enables one to predict that the curves of systolic pressure at a given age will become more platykurtic as age progresseso In addition to flattening, a positive skewness becomes evident with increasing age, as will. be seen in the following section. One might also predict that the diastolic curves will show a lesser trend toward flattening and skewnesso

-50diastolic means). It is also true that the true variability (between persons) of blood pressure increases with age as well as the means themselves, These seem to be the most important reasons for this observed phenomenono B. Graphs of the Distributions In studying the graphs of mean blood pressures by age (Figure 702), two aspects are considered~ 1) the magnitude of the respective means, and 2) the general configuration of the lineo 1 White Race Considering first the white race separately, the general configuration of the systolic pressure line with age shows a definite sigmoid curve in both sexes. These curves (reading to mid-poin.ts of age groups) cross in only two places, at: 1) about age 153 and 2) about age 38o These crossover points result from females having slightly higher systolic pressures prior to age 1.35 males then having considerably higher average pressures to age 38, and finally female pressures becoming markedly higher than the male after that age, The male curve rises sharply to an adult level by age 17. then remains relatively stable until about 40 years of age after which it rises slowly. The female curve shows that the adult level is reached at about the same time as the male, but the early adult plateau is lower, hovering at about 120 from age 17 to 37. The corresponding male plateau is closer to 130 during this period, Between the ages of 37-42

200 180- --- WHITE MALES 0 —- WHITE FEMALES -- - NEGRO MALES * —* NEGRO FEMALES 170 160 150100 - 140 E 30U, 120 a. 110 - 100 90 80 70 60 50 6-0 10-14 15-1 20-24 25-29 30-34 36-35 40-45 46-49 50-54 65-59 60 AGE Figure 7.2 Mean Systolic and Diastolic Blood Pressure by Sex, Race, and Year Age Group. White and Negro Races, Nassau, 1958.

-52a sharp rise in systolic mean occurred in the white female group. The means for the females after this age are not markedly different from the Negro male, but still lower than the Negro femaleo This sharpness has not been noted by other investigatorso It is possible that the magnitude of this increase may be made more apparent by sampling variation However, the standard errors of these means show that a true increase does occur at this timeo After the age of 42, the female curve progresses upward slowly, Substantiating evidence for these configurations is found in the analysis of break-points (proportion greater than 150 mm, systolic) as shown in Figure 81oo Here the crossover points are less distinct, but the configurations are substantially the same, The curve for diastolic pressure means shows little difference between the sexeso The range of the means is only about 20 mmo, which tends to mask any differences, However, the same crossover points are present, The crossover at puberty is now followed by a zone of no difference between sexes until the point of cross at age 38, after which the female is considerably higher, This is further illustrated in Figure 8o1 by the break-point graph (proportion greater than 100 diastolic)o Comparing sexes by mean pulse pressures in relation to age (Figures 7,3 and 7o4) shows much the same relationship

-5390 85 A NEGRO MALES 80- * NEGRO FEMALES,6) E 75 E 70 3 i w I )(i I I Nassau, 19658. 60 Q. z 55 50 45 40 35 6-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60 AGE GROUPS Figure 7.3 Mean Pulse Pressure (Systolic-Diastolic V) (First Reading) by Sex and Age, Negroes Nassau, 1958.

-5490 85 80- A WHITE MALES 0 WHITE FEMALES E 75- E 70 u) u) U 65 Q. w U 60 z 55 50 45 40 35 6-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60 AGE GROUPS Figure 7.4 Mean Pulse Pressure (Systolic-Diastolic V) (First Reading) by Sex and Age, White Race, Nassau, 1958.

-55except that the adolescent crossover is obliterated because of comparatively lower diastolic means in these young ageso The general configuration is retained, however, and shows an even deeper S-shaped curveo The point of cross in middle life remains the sameo 20 Negro Race It is of notable interest that the mean systolic pressure histograms for Negroes have few points of similarity to the white race except in childhood. In neither sex in the Negro is there any indication of a slowing down in the march of blood pressure elevation as age progresses. These curves indicate that blood pressure in the Negroes studied here progresses in an almost perfectly linear fashion from the youngest to the eldest. The difference in the curves between the sexes seems to be much smaller than for the white race, although the same temporal relationships existo Until puberty, the female is slightly higher, crossing the male curve at almost the identical point as'that for the white race, at about age 13. After this time the male is somewhat higher until the apparently identical point of cross at about age 38. The curve for diastolic means in the Negro shows, somewhat surprisingly, that after the pubertal crossover the male remains slightly but uniformly higher than the female until the upper ages when the curves coalesce.

-5630 Comparison of Negro with White Race It has been noted that the configuration and slope (magnitude) of the systolic curve is markedly different between the raceso These differences include: a) linear progression of the curves in the Negro as contrasted with the plateau effect seen in the white; b) slightly lower levels for Negroes than whites until late adolescence when the white plateau begins and the Negro curve continues to rise; c) diminution of sex difference as to blood pressure in the Negro as compared to the white, while the major points of cross in the sex curves remain almost identical, The comparison of races in the diastolic curves is somewhat differento The range of means for the Negro is only about 30 mm.., while in the white the range is about 20 mm, In this restricted range the race-sex curves do not show such a manifest difference in configuration as in the systolic curveso However, after adolescence the Negro is consistently higher than the white, Sex differences are less distinct, but still present in the white. The same basis coroasover points hold but there is diminution of sex differences during the early adult plateau. In the Negro significant crossover points are less evident and the male is slightly higher in almost all ages. Pulse pressure curves demonstrate again the smaller sex difference in the Negro and the tendency to linearity in the Negro as opposed to a sigmoid curve in the white, Pulse

-57pressure is seen to be slightly higher in the young whites (under 25-30 years) than the Negro, after which age the Negro has considerably higher mean pressureso 4. Comparison with Similar Studies Of the many attempts as description of blood pressure patterns, most are eliminated as potential sources of comparison by lack of similarity in choice of study population as well as differences in procedure and methods of measurement. Relatively few have attempted to define a complete population or to utilize adequate sampling methods. Comstockls(3 ) study in Georgia provides the points of comparison of 1.) sampling method, 2) biracial community in same type of climate as the present study. Miall and Oldham's ) work in Wales represents a sampling study in a community of the white race. Boe, et al.,( ) in Bergen, Norway attempted to study all of the persons in a major cityo These studies are felt to be the most nearly comparable of current studies for purposes of comparison with the study here reportedo The question of major concern in this study was "Is there a real difference between the patterns seen in the Negro and white races " The present data support an affirmative answer to this question. It must now be ascertained whether this is a finding restricted to a comparatively isolated group, geographically and genetically, or if there is a comparable pattern present in racial counterparts elsewhere in the world,

-58Figure 7 5 describes the pattern for Negroes in Comstock's study in Georgia, These persons would seem to be very similar in racial origin (West Africa) to the Bahamian Negro, although isolated geographically from each other for over 100 years. From the appearance of these mean blood pressures by age group curves, they indeed appear to come from the same population. In the systolic curves, the same unrelenting linearly upward progression is seen except for a slight wave in the young adult female. This may be due to sampling variation or small numbers in the groupo It is of interest to note that the major point of cross of the female over the male curve is at precisely the same point for both sets of datao However, Comstock's mean pressures apply to a ten year age group. The present data was converted to the same measure with the consequence that the crossover point is moved slightly to the left or to a younger age (about 36). Comparison of the Georgia Negroes' diastolic curve with the Bahamian Negroesv shows even less difference, The Bahamians seem to be slightly higher at the upper extreme of age, but this is of doubtful significance. Such a degree of similitude between two studies done at different times and places, using slightly different methods is truly surprising. Since, however, both studies were performed on seemingly valid population samples,one is encouraged to believe that this represents a fair picture of the blood pressure distribution by age and sex for Negroes of West

-59180 170 A —- NEGRO MALES (NASSAU) NEGRO FEMALES (NASSAU). A...A NEGRO MALES (COMSTOCK, GEORGIA) /........0 NEGRO FEMALES (COMSTOCK, GEORGIA) 160 150- 140 - 130 120 _ 0 0 *.' 100 - 90!." 10 20 30 40 50 6o AGE Figure 7.5 Mean Systolic and Diastolic Blood Pressures by Age Group. Negro Race, Nassau, 1958 and Georgia (Comstock5).

-6oAfrican descent. One must remember, unfortunately, that the fact of one study supporting another is similar to the house of cards where one card supports anothero One additional card (or study) may be disastrous. Two other investigations are currently being done on the Negro race, one in Liberia and another in the Caribbean. The results of these will. clarify the present knowledge regarding the Negro raceo Preliminary data of one of these, that of Page, Corcoran and Schneckloth (unpublished), done on the island of St, Kitts in the Leeward Islands is shown in comparison with Nassau Negroes in Figure 7,6 Although there were considerable differences in the method of study, the general. similarity of the curves for systolic and diastolic means is evidento Disregarding minor differences in degree, the linear configuration of the St. Kitts Negro systolic means by age group is similar to that shown for Bahamian and Georgia Negroes o The similarity in blood pressure findings in. the Negro in these three studies in. different areas suggests that there is no underlying cause related to geographical location or environmenat which would account for this elevated pattern in the Negro. It must now be seen if this is also true of the white race. The numbers of white persons seen in the present study are considerably smaller than the Negro However, it is seen that standard errors of the means for age-sex groups were sufficiently small to permit comparisono

-61170 -7 160 ~150~/ // / / 130 E I! 120 - 0. A -— A NEGRO MALES (ST. KITTS) o 110 - ---- NEGRO FEMALES (ST. KITTS) 0 m A NEGRO MALES (NASSAU) -* @* NEGRO FEMALES (NASSAU) 10090 80 70 10 20 30 40 50 60 AGE Figure 7.6 Mean Systolic andlic and Diastolic Blood Pressures by Sex and Age Group. Negro Race, Nassau, 1958, and St. Kitts, Leeward Islands (Corcoran and Schnecklothb 1958).

-62In Figure 7~7 the Bahamian white persons are compared with the white persons in Comstock's Georgia study. The curve for Bahamians for systolic pressure is seen to be of generally similar configuration (sigmoid in shape rather than linear), However, the crossover point is considerably delayed in the Georgia white persons (about age 43) Also, after the crossover the sex difference is greatly reducedo There also appears to be a significantly higher systolic curve for Bahamians than for the Georgianso This is apparent for the female sex where the lines run parallel for the entire age curve. In the male systolic curves this difference is less distinct with convergence of the two curves in the young and in the oldo In the diastolic curves no apparent difference is seen, The configuration is similar and the sex curves for Bahamians fall midway between their Georgia counterpartso The considerable difference between the sexes in the diastolic pressure curves seen in the Georgians is not found in the Bahamians o In viewing these curves one hesitates to attach significance to systematic differences of the magnitude seen in the systolic curveso It will be seen later that in the Norwegian study (Boe, et alo ()) that two large samples from the same population showed, inexplicably, a systematic difference0 It seems likely that slight differences in methodology may partially account for the differences between the Georgia

-63160 150 0 140 130 ) **,. e 0 n WHITE MALES (NASSAU) 0-0 — WHITE FEMALES (NASSAU) o 100 - A... WHITE MALES (COMSTOCK GEORGIA) 8 O~-~~~0 0 WHITE FEMALES (COMSTOCK, GEORGIA) a 0 90 ~~80,'~",.....~~~~~~~~.0' 70 - 60I.I 10 20 30 40 50 60 AGE Figure 7.7 Mean Systolic and Diastolic Blood Pressures by Age Group. White Race, Nassau, 1958 and Georgia (Comstock3).

-64Negroes and the Bahamian Negroes. Perhaps the most important feature to compare is configuration which is generally similar except for the points noted. Although the Georgia white community studied might be expected to be primarily of Anglo-Saxon origin (as are the Bahamians) it is of interest to compare the Bahamian white persons with those in other parts of the worldo Miall and Oldham's 3) work in Wales provides data which are shown in. Figure 7~08 There is a remarkable agreement in both magnitude and configuration in the curves for systolic and diastolic means by age groupo The.largest single group of white persons for comparison. is that of Boe, et al( ) in. Norwayo This study is more truly a population study than many otherso It encompasses a large segment (75-80%) of an entire large cityo In spite of large numbers, it is still hindered by the unmeasured portion of the people (nonrespondents) in regard to interpretation of results, just as in a sampling surveyo This fact is borne out by the difference between the first half of Boe s data and the second halfo These two groups in Boe s study are not actually different samples but are parts of the same population studied separately. The first group represents those studied in the first portion of the study and the second group is the remainder of the studyo Theoretically they represent the same population, separated onl.y by time of presentation for examinationo It is not known if

-65180 170 - L —-A WHITE MALE (WALES) P 0 —— o WHITE FEMALE (WALES) - WHITE MALE (NASSAU) o - 0 WHITE FEMALE (NASSAU) 160 / / / 150 - 140 130 ", 120 I X o -0 _100 90 _ 80 70 60 I 6- 10-14 I-19 20-4 25-S 30-34 3-0 40.44 45-40 50-54 500 60 AGE GROUPS Figure 7.8 Mean Systolic and Diastolic* Blood Pressure by Sex and Age Group. White Race, Nassau, 1958 and Wales (Miall and Oldhaml2) *(Fourth Phase Diastolic).

-66these two groups represent different types of people, or if a possible methodological difference was present, It seems that a more true picture of the distribution would have been shown by pooling the two groups. Conceivably, the unmeasured portion of the population could show a still different patterno The two groups are shown in Figures 7,9 and 7o10, and the Nassau white sample is contrasted, sex specifically. Whatever the reason for the systematic difference in Boevs study, evidenced in both male and female histograms for means of systolic and diastolic pressure by age group, the data for Nassau whites show the following points of comparison: 1) systolic curves show the same general configuration as the Norwegian data; 2) Nassau white male systolic curve seems to approximate quite closely in magnitude the second group of the Norwegian, while the Nassau white female curve strikes midway between Boe's two groupso In regard to diastolic pressure the Nassau males are slightly but uniformly lower than the Norwegian, while the Nassau. white females approximate Boe's except in the extremes of ageo Co Summary In this section has been presented the accumulated data regarding distributions of systolic, diastolic and pulse pressure in the Nassau study,. Distributions have been compared for Negro and white

-67160 / 150 - / 130 - 120 - E E w / A WHITE MALE 110- / BOE,GROUP I, NORWAY w BOE, GROUP II, NORWAY X~ ~ ---- ~ NASSAU 0 100 90 70 -9 10-14 15-19 20-24 25-29 30-34 3-39 40-44 45-49 50-54 55-59 60 AGE GROUPS Figure 7.9 Mean Systolic and Diastolic Blood Pressure by Age Group. White Males, Nassau, 1958 and Bergen, Norway (Boel).

-68160 150 140 - 130 - 120 - io cc 110 o0 WHITE FEMALE' - - BOE, GROUP I, NORWAY 100- BOE, GROUP II,NORWAY NASSAU 90 -.S 80 _ 60 ------ 6-9 10-14 15-19 20-24 25-29 30-34 38-39 40-44 45.49 50-54 55-59 O6 AGE GROUPS Figure 7.10 Mean Systolic and Diastolic Blood Pressures by Age Group. White Females, Nassau, 1958 and Bergen, Norway (Boel).

-69races and for the male and female sexes, Standard errors of means and of the distributions within age-sex-race specific groups have been shown, The principal. points of contrast were~ 1, Systolic means progress upward with age, almost linearly upward in the Negro from childhood to the oldest groups measuredo The white race shows a more sigmoid curve attaining an "adult" level of systolic pressure slightly earlier and resting in a plateau until middle life, then continuing the upward progression. 2o Diastolic means progress with age in a much more restricted range, but both races showing approximately the same configuration. The Negro is only slightly higher t;han the whit;e 3o Sex differences were somewhat similar between races, especially in systolic means, The curves cross in. adolescence and again in middle life, It is of interest that these crossover points are almost identical for the two races. Sex differences in. the diastolic are less distinct. The cross points are still present in the white diastolic curve but almost indistinguishable in, the Negro, These differences in pressure in regard to sex appear somewhat smaller generally in the Negro than in the white, principally because of the greater difference in white persons after middle life when the female systolic curve becomes much higher than the male,

-704. The pulse pressure curves reflect the same general relationships in regard to configuration, The sex differences are now apparent in the white and less distinguishable in the Negro. The Nassau data were then compared to other studies of similar nature in other parts of the world. The point of greatest interest is that the blood pressure curves for the races are quite comparable to those of similar racial groups studied elsewhere. This suggets that there is not an environmental factor present here (for example, salt intake) which is not present elsewhere which may possibly elevate blood pressures. Although these people, both white and Negro, have been somewhat of a geographic and genetic isolate for more than 100 years, they still retain the blood pressure features characteristic of their race. It is still possible, of course, that an environmental factor is common to all populations studied, which in some way affects blood pressure, On the basis of the comparisons mentioned, however, it seems more logical to think in terms of a true racial difference rather than an environmental difference,

CHAPTER VIII ANALYSIS OF SPECIFIC BREAK POINTS OF BLOOD PRESSURE Data shown in the previous chapter illustrate again the belief of Pickering(l5) regarding the continuous distribution of blood pressure and the lack of any indication that hypertensive persons differ in any way from "normal" persons except in degree. There was no break in the distribution to suggest two separate populations, ego, hypertensive and normal, Nevertheless, to describe proportions of the population above certain arbitrary break points of blood pressure may be helpful as a further desQriptive measure of these distributions. Break points, eog., 150mm. systolic or 100mm. diastolic, are used clinically in orientation of thinking and in therapy, There are, of course, differences in thinking in this regard and differing reference points are used, Perhaps the most commonly used cutting points are the 150/100 and 1.40/90 readings, singly or in combinationo These will be used in the present analysiso However, in the Appendix Table A the full distribution of systolic by diastolic readings are shown in order that later investigators may choose almost any point of reference desired. In these tables, the total data were classified by 5mm. diastolic groups and lOmmn systolic groupso Each five year age group in each race is tabled separately. A, Break Points Considered Singly The tables were then studied in regard to specific break points, e.go, all. over 140 systolicO The results are shown in Table 8,1 and shown graphically in Figure 8l.1 It is seen that in these figures, the general configuration of the lines mimics those in the -71

-72Table 8 1 PROPORTIONS IN AGE-SEX GROUPS WITH DIASTOLIC BLOOD PRESSURES GREATER THAN 90 AND 100 Negro Race, Nassau, 1958 Diastolic Break-Points (Colored Race) (Unadjusted) Negro Male 90 or over 100 or over N Age Number Prop. Number Prop. 6-9 0 0 0 0 154 10-14 2 1,40% 0 0 143 15-19 13 9.35% 3 2,16% 139 20-24 25 20,32% 11 8.94% 123 25-29 31 31.63% 14 14.29% 98 30-34 19 32.76% 12 20.70% 58 35-39 33 60.00% 15 27.27% 55 40-44 32 47.76% 20 29.85% 67 45-49 45 67.16% 22 32.84% 67 50-54 21 56,76% 11 29735% 37 55-59 8 50.00% 3 18.75% 16 6o+ 17 51.52% 8 42424 33 246 119 990 Negro Female 6-9 1 0.60% 0 0 166 10-14 3 2.16% 0 0 139 15-19 12 7.84% 4 2.61% 153 20-24 26 15.29% 10 5.88% 170 25-29 35 27.34% 16 12,50% 128 30-34 25 27.07% 13 15.12% 86 35-39 50 45.87% 18 16.51% 109 40-44 41 47.13% 29 33533% 87 45-49 37 49533% 19 25,33% 75 50-54 28 60.87% 17 36.96% 46 55-59 21 70.00% 10 33.33% 30 60+ 29 47.54% 19 31o15% 61 308 155 1250

-73Table 8ol (Cont d) PROPORTIONS IN AGE-SEX GROUPS WITH SYSTOLIC BLOOD PRESSURES GREATER THAN 140 AND 150 White Race, Nassau, 1958 Systolic -Break-Points (-White Race) White Male 140 or over 150 or over N Age Number Prop. Number Propo 6-9 0 0 0 0 45 10-14 1 35o57% 1 3o57% 28 15-19 8 24.24% 3 9.09% 33 20-24 10 29.41% 3 8.82% 34 25-29 8 26.67% 0 10 00% 30 50-34 8 21o05% 0 0 38 35-39 5 22a73% 2 9o09% 22 40-44 7 31,82% 5 22o73% 22 45-49 7 3 Soo0 5 25 00% 20 50-54 7 5000o% 4 28.57% 14 55-59 4 36o36% 3 27.27% 11 60+ 15 53 57% 10 35.71% 28 80 39 325 White Female 6-9 0 0 0 Q 47 10-14 2 5 56% 1 2o78% 36 15-19 2 6.45% 0 0 31 20-24 5 11o635% 6o 98% 43 25-29 2 5.13% 1 2,56% 39 30-34 4 12o12% 3 9.09% 33 35-39 5 18o52% 3 11,11o 27 40-44 15 60,00o 11 4400oo 25 45-49 9 45o00% 7 35o00% 20 50-54 8 53o33% 7 46.67% 15 55-59 10 62 50%o 8 50,o0o 16 60+ 20 66.67% 13 45o33% 30 82 57 362

-74Table 8.1 (Cont'd) PROPORTIONS IN AGE-SEX GROUPS WITH SYSTOLIC BLOOD PRESSURES GREATER THAN 140 AND 150 Negro Race, Nassau, 1958 Systolic Break-Points (Colored Race) (Unadjusted) Negro Male 140 or over 150 or over N Age Number Prop. Number Prop. 6-9 1 0.65% 0 0 154 10-14 5 3550% 0 0 143 15-19 20 14.39% 11 7.91% 139 20-24 27 21,95% 11 8.94% 123 25-29 36 31.63% 19 19.39% 98 30-34 17 29.31% 13 22.41% 58 35-39 24 43.64% 16 29,09% 55 40-44 29 43.28% 18 26.87% 67 45-49 41 61.19%g 27 40o30% 67 50-54 23 62.16% 18 48.65% 37 55-59 9 56.25% 8 5o.oo.16 60o+ 27 8182% 22 66.67% 33 Negro Female 6-9 1 0.60% 0 0 166 10-14 1 0.70% 0 0 139 15-19 9 5.88% 5 3527% 153 20-24 21 12.35% 9 5.29% 170 25-29 27 21.09% 14 10.94% 128 30-34 27 31.40% 16 18.60% 86 35-39 47 43,12% 25 22.94% 109 40o-44 48 55.17% 57 42,53% 87 45-49 43 573553% 33 4400oo% 75 50-54 29 63.04% 24 52.17% 146 55-59 26 86.66% 24 8o.00o% 30 6o+ 53 86.89% 51 83.61% 61 55332 238 1250

-75Table 8.1 (Contad) PROPORTIONS IN AGE-SEX GROUPS WITH DIASTOLIC BLOOD PRESSURES GREATER THAN 90 AND 100 White Race, Nassau, 1.958 Diastolic Break-Points (White Race) White Male 90 or over 100 or over N Age Number Prop. Number Prop. 6-9 0 0 0 0 45 o10-14 o0 0 0 28 15-19 1 35.03 0 0 5533 20-24 3 8.82% 1 2.94% 34 25-29 3 o10.00oo% 0 0 30 30-34 6 15.79% 2 5.26% 38 35-39 5 22.72% 2 9.09% 22 40-44 7 31.82% 3 13653% 22 45-49 6 30,00% 1 5.00% 20 50-54 5 35571% 1 7.14% 14 55-59 3 27.00% 0 0 11 6o+ 6 21,43% 4 14.29% 28 45 14 325 White Female 6-9 2 4.26% 0 0 47 10-14 2 5~ 56% 0 0 36 15-19 0 0 0 0 31 20-24 8 18.60% 2 4.65% 43 25-29 4 10.26% 1 2.57% 39 30-34 5 15.15% 3 9.09% 33 35-39 6 22.22% 1 3.70% 27 40-44 153 52.oo00% 6 24.00% 25 45-49 8 4o,oo00% 1.5.00% 20 50-54 7 46.67% 1 6.67% 15 55-59 7 435.75% 4 25.00% 16 6o+ 10 55333% 2 6.67% 30 72 25 5^2

-7690 80 - PROPORTION OVER 150 mm SYSTOLIC PRESSURE — A NEGRO MALE * NEGRO FEMALE 70 - -- WHITE MALE 0 —- WHITE FEMALE 60 50 400 - 20 10- 10 20 30 40 50 60 AGE 50 - -- NEGRO MALE PROPORTION OVER 100 mm - * NEGRO FEMALE DIASTOLIC PRESSURE 40 - WHITE MALE o — WHITE FEMALE 30 20 20 10 10 20 30 40 50 60 AGE Figure 8.1 Proportion of Persons Within Age-Sex-Race Group with Blood Pressure Over Arbitrary Break Points Negro and White Race, Nassau, 1958.

-77graphs on'the means. However crossover points are less evident, particularly in the white race. It was seen that a composite rate over all ages for these points would be preju.dic.ed by the differences in. age and sex composition between the two raceso Accordingly, the data were standardized for age and sex, The white sample population was taken as a standard and the Negro population was adjusted according to the white age and sex proportions, Then, the "expected" numbers by break point in. each age group. The:new total.s thus generated were then used for racial comparisons In Tables 802 through 804 the race and sex specific rates for break points are compared. The differences between compared rates are subjected to the test:for the standard error of a difference. R. - R2 4P1.q1 P2q2 n! n2 When one or botlh of the rates are adjusted rates. the formula for standard error of an adjusted ratie is u.sed~ SE adjusted rate w2 P? qi where mi w- _ or tthe adjusted sample population proportions for each age Eni goup. thF-,e observed proportion over a specific break point9 and mi:= the original group r. on which P. is based, It is seen that in the racial comparisons, Table 8.2, the overall rates for Negroes are consistently higher than whites, The only difference of borderline significance is for males at the 140mm,

-78Table 8.2 RACE DIFFERENCES FOR PROPORTIONS ABOVE SPECIFIC BREAK-POINTS Negro and White Races, Nassau, 1958 Race Differences (All Ages IncLuded) Break-Point Negro White Differ- S.Eo Diff./S.E. (mm. Hg.) Male Male ence Diff. Diff, Over 140S 30,66 26.23 4,43 3505 1o45 150S 21.23 12.79 8.44 2.49 3.39 * 90D 29o15 14.75 14,4o 2.71 5 31 * t 10OD 14,53 4,59 9,94 1.90 5.23 * Break-Point Negro White Differ- SoE Diffo/S.E. (mmo Hgo) Female Female ence Diff, Diff, Over 140S 30.81 22,65 8 16 2.46 -332 * 150S 23543 15o75 7o68 2,41 3519 * 90D 26.99 19o89 7,10 2o42 2o93 * " 100D 13584 6.35 7~49 1,67 4.49 * Race Differences (Adults over 25 years1 Break-Point Negro White Differ- SEo Diffo/S.E. (mm. Hgo) Male Male ence Diff, Diff. Over 140S 48.19 32.97 15.22 3587 3593 " 150S 35o41 173.0 18o11 3 22 5 62 90D 47.15 22,16 24.99 3555 7.04 " 100D 24 35 7o03 17.32 2,39 7o25 Break-Point Negro White Differ- SoE. Diff/ S,E, (mm. Hg.) Female Fem.le. ence Diff. Diff. Over 140S 51.76 355.61 1615 3o46 4 67 i 150S 40o61 25 85 14o76 3521 4,6o 90D 43 66 29.27 14 39 3 37 4,27 t lOOD 23521 10, 24 12,97 2o35 5.52 Note: 1) Age-sex standardized rates calculated for Negros, 2) * Indicates statistical significance at 5% levelo

-79Table 8,3 SEX DIFFERENCES FOR PROPORTIONS ABOVE SPECIFIC BREAK POINTS, ALL AGES COMBINED, NEGRO AND WHITE RACES Nassau, 1958 Sex Differences (All Ages Included) Break Point Negro Negro Difference S.EDiff. Diff. Male Female SE.Diff, Over 140S 30.66 30.81 0.15 2.14 0.07 " 150S 21,23 23543 2.20 1.96 1.12 90D 29 15 26.99 2.16 2.24 0 96 100D 14.53 13584 0.69 1.85 0.37 Break Point White White Difference S.EDiff. Diff. Male Female S.E.Diff. Over 140S 26.23 22o65 3.58 3.28 1,09 150S 12.79 15 75 2,96 2.86 1.03 90D 14,75 19,89 5o14 2.87 1.79 100D 4.59 6535 1.76 1o73 1.02 NOTE: No significant differences at 5% level.

-80Table 8.4 SEX DIFFERENCES FOR PROPORTIONS ABOVE SPECIFIC BREAK POINTS ACCORDING TO BROAD AGE GROUP (PRE-MIDDLE-LIFE AND POST-MIDDLE LIFE) Negro and White Races, Nassau, 1958 Age Category White White Differ- SE. C.R. Group Male Female ence Diff. 6-39 over 21.74 7.81 13.93 3.19 4.37 * 40-6o+ 140S 31.58 58.49 26.91 6.75 3599 * 6-39 over 5.22 430 0.92 - 40-60+ 150S 28.42 43.40 14.98 6.68 2.24 * 6-59 over 7.83 10.55 2.72 2.61 1,04 40-60+ 90D 28.42 42.45 14.03 6.67 2,10 * 6-39 over 2.17 2.73 0.56 40-60+ 100D 9.47 15.10 5.63 4.59 1.23 Age Category Negro Negro Differ- S.E. C.R. Group Male Female ence Diff. 6-39 over 16.88 13.99 2.89 1.73 1.67 40o-6+ 140S 56.36 66.56 10.20 4,31 2.37 * 6-39 over 9.10 7.26 1.84 1.04 1.76 40-60+ 150S 34.09 56.52 22.43 3.19 7.03 * 6-39 over 15.97 15.98 0.01 40-6o+ 90D 55.91 52.17 3.74 4.43 0.84 6-39 over 7.14 6.41 073 - - 40-60+ 1OOD 29.09 31,44 2.35 4.07 0.58 * = Significant difference at 5% level.

-81systolic break point. It is also generally true that differences and significance values are greater at the diastolic break points, particularly at 1.00mm. diastolico These rates are computed on the adjusted Negro population proportions. In, Table 8.2 the race differences are calculated only for adults over 25 years of age wwhich is the area of greatest concern in hypertensive disease, and overall differences are seen to be highly significant at all break points l.istedo Differences between sexes of the same race were analyzed in regard to overall rates and also in terms of pre-middle life and postmiddle life (at point of crossover) In Table 8~4 the latter is shown and the finding of some differences between the sexes after the crossover at age 40 (approximately) is verified. In, the white race, statistically significant differences were found in the "over 140 systolic" break. males having greater proportions prior to age 40, the females after that point in. life. In the "over 150 systolic" and "over 90 diastolic "t female proportions are shown, to be higher than. the male in upper part of the life span., bu-tt in t1he 9.ower ages significance of difference is losto In the Negro race a similar relationship holds. However, the only differences of statistical significance are in the systolic break points where femnale proportions are higher in the upper age group than the male O There is a tendency toward the reverse in the younger ages, but of bo.rderline significance The reduced sex difference by these age groupings in the diastolic break points is what may be expected from the distribution previously showno However, in Table 8o03 the same proportions by break point are cal.culated for the total sample, ignoring ageo T:he inverse

-82differences by the above age grouping was sufficient to render the overall sex difference by race negligibleo A. Study of Break Points in Combination It would be of interest to ascertain from these data if a relationship exists between systolic and diastolic readings in the sense that if systolic pressure is greater than an arbitrary point, the diastolic pressure is also higho The opposite situation may also be studied. It is difficult to compare proportions of -:persions with both a "high" systolic pressure and a "high" diastolic pressure in regard to racial differences because of the varying distributions Of blood pressure between the races It is apparent that if one race demonstrates a great number of persons with very high systolic pressures that the number of high diastolic pressures would also be proportionately higho Considering the races separately, then, in Table 8.5 is shown the number of persons by race and sex who show elevation by both systolic and diastolic criteria (e.g. 150S and also 10OD, 140S and also 90D). The table shows in the first portion these numbers for all persons in the sarlmpe race-sex groups, and in the second portion, only those persons under 55 years. The second portion was calculated so that the data might more truly represent essential hypertension rather than arteriosclerotic blood pressure elevationsQ These rates are unadjusted, since they are not compared across race0

_83Rates then are calculated in the following manner- Of the persons recorded as both 150/1 00, the proportion is determined of all persons in the total group who showed at least 150 systolic pressure when this break point was considered separately. This combination break point is then considered in relation to all persons showing at least 100 diastolic pressure when this break point was considered separately. The same procedure is followed for the 140/90 combinationo Considering first the white female, it is seen in Table 805 that of those recorded as both systolic and diastolic readings in excess of 150/100, this number represents only 53353 of all white females who showed at least 150 in systolic p uessure when this reading was considered alone. Therefore, in the white female, twothirds of persons with systolic readings over 150 do not also have diastolic readings in excess of 100o The reverse of this relationship would be that of white females recorded as both pressures in excess of 150/100, compared with the total white females with at least 100 diastolic pressureo The combination break point represents 82o6% of all white females who showed at least 100 in diastolic pressure when this reading was considered alone. Therefore, less than 18% of persons with diastolic pressures over 100 do not have systolic readings in excess of 150. This relationship is maintained for the 140/90 combinationo It is also maintained, although slightly diminished, when the older persons are excludedo In the white male the pattern is similar in all respectso

-84Table 8.5 PROPORTION OF PERSONS IN RACE-SEX GROUPS WITH BOTH SYSTOLIC AND DIASTOLIC PRESSURE ELEVATIONS, IN RELATION TO EITHER PRESSURE ELEVATED SINGLY, (ALL AGES), NASSAU, 1958 Of those 150/100 Of those 140/90 Race-Sex Persons Persons Prop. of Prop. of Prop. of Prop. of Group at Least at Least Total Total Total Total 150 S 140 S in Sample in Sample in Sample in Sample and 100 D and 90 D Over 150 S Over 100 D Over 140 S Over 90 D Negro Male 94 178 64.835 78.99% 70.07% 72.36% Negro Female 131 241 55.04% 84.5-2% 72.59% 78.25% White Male 12 32 30.77% 85.71% 40.00 71.11% White Female 19 49 33.33% 82.61% 59.76% 68.06% All Persons Under 55 Years of Age Negro Male 84 154 73.04% 77.78% 70.64% 69.68% Negro Female 103 191 63.19% 81.75% 75.49% 74.03% White Male 8 24 30.77% 80.00% 39.34% 66.67% White Female 13 32 36.11% 76.47% 61.54% 58.18%

In the Negro, distributions of systolic and diastolic pressures are considerably higher and the above relationship cannot be demonstrated clearly at these break point combinations. From this discussion of break points in combination it is seen that the commonly used combination points, eg., 140S and/or 90D, and 150S and/or 100D are somewhat fallacious, Proportionately more persons reach these systo.ic points without also reaching the corresponding diastolic points than is true of the reverse situationo This is true even when older persons are excluded. The above relationship suggests that diastolic pressure elevated to these points (90 or 100) should carry with them higher systolic values than u.sually set if the same clinical interpretation is to applyo Perhaps 150/90 and 160/100 would be more realistic if these ponts were to onvey smiar significance in, terms th o th verall distribution of blood pressure. In summary, it has been shown, by analysis of break. points that~ lo Over the entire sample, Negroes show a greater proportion of persons with systolic and diastolic pressures over all break points chosen t'han do whites, This is even more true when the diluting effect of persons under 25 years is excluded, 2o Sex differences within race in regard to break points show no significant differences when viewed in total. over the entire age span. When divided at age 40 (cross over point) white females are seen to have the higher pressures than males after 40 years while males have higher below 40 years.

-86Only a slight white female increase over white males in diastolic pressure is seen, this being mainly at the 90 diastolic break point. No significant sex differences for diastolic pressure were seen in the Negro race in any age grouping or over the entire age span. 30 Considering break points of blood pressure in combination, eogo, 150S and/or 100D, as commonly used, it appears that 150 and/or 90 and 160 and/or 100 would more closely equate these values in terms of the overall distribution of blood pressure

CHAPTER IX RELATIONSHIP OF BLOOD PRESSURE TO BODY BUILD The assessment of degree of obesity is an extremely difficult problem, even in the individual. It is even more difficult to measure in a population. The measurement of the proportion of body weight which is adipose rather than muscular or skeletal, is done best by a difficult laboratory method involving water immersion. Other methods, such as measurement of subcutaneous fat by calipers is relatively simple but requires special instruments, Within the scope of this study only simple measurements of height and weight were possible. Height and weight, considered separately in regard to blood pressure, would appear to yield unrewarding results since a relationship of blood pressure would more likely be to the entire body buildo Boe, et alo( ) conducted a thorough anthropometric study in this regard yielding little of note in regard to height and weight separately. Many methods have been proposed to combine height and weight measurements into an index of body build, An index of this type should yield a constant figure for comparable body builds disregarding heighto The index used in this study is HT//yWT, measured in inches and pounds, This has been referred to as the "ponderal indexo" The difficulty in comparing indices lies in the problem of stating comparability between similar builds but of differing height. In effect, one places a human being in a cylinder and judges weight per unit of height. However, it may well be that a taller person tends more toward ectomorphy on the -87

-88average, and therefore, one might expect a slight rise in the index in the upper extremes of height. Boe compared different indices in an attempt to find the best descriptive index, finally using WT/HT2, measured in centimeters and kilograms. He refers to a variant of the index used in the present study, stating that this index tends to fall slightly with increasing height. Disregarding the controversy about the best possible index, however, for the present purposes it was felt that the index used was sufficiently valid. It must also be said that this or any index so contrived is not a direct indicator of obesity or overweighto For example, a person with a broad skeletal structure with heavy musculature may have the same index as a person of the same height but with a light frame and considerable obesity. However, this does not render the use of the index entirely invalid. Although occasionally a heavy muscular person will confuse the picture, in the majority of persons, a lowering of the index indicates a greater obesity for a given height. It is difficult to compare across race in regard to this index because of the factor of different types of buildo The Negro appears to be somewhat heavier in frame, on the average, and comparatively more muscular. In spite of this, the white man shows (Figure 9.1) a lower (heavier) mean index than the Negro male throughout most of the age span (except middle life) This tends to indicate that this heaviness in the white is probably a greater relative obesityo Figure 9.1 also shows that the white female roughly parallels the curve for the Negro female, but the Negro women are slightly heavier than the white women, except in childhoodo It seems probable that the women

-89Table 9.1 RELATIONSHIP OF HEIGHT-WEIGHT INDEX (HT/ fWT) TO BLOOD PRESSURE WITHIN AGE-SEX GROUPS White Race, Nassau, 1958 Males Females (Non-pregnant) Age 6-9 Pulse No. No. Ht.Wt. Systolic Diastolic Pressure in Ht.Wt. Systolic Diastolic Pressure in Group Index Mean Mean Mean Group 10 - - - - 10 - - - - 11 104.3 62.8 41.4 18 11 107.1 70.6 36.5 22 12 103.2 59.4 46.5 22 12 102.5 61.6 40.9 16 13 - - - 13 - - - 14 - - - - 14- - - - Age 10-19 10 132.8 8o.o 52.8 5 10 - - - - 11 118.8 64.4 54.4 50 11 120.1 66.2 54.0 22 12 129.1 69.3 59.9 7 12 115.1 67.0 48.3 48 13 - - - - 13 104.0 68.0 36.0 1 14 - - - - 14 - - - - Age 20-29 10 140.0 79.3 60.7 6 10 134.1 83.2 52.6 12 11 123.4 73.7 57.4 47 11 118.5 72.8 45.7 69 12 114.7 65.3 49.3 3 12 108.0 76.0 32.0 1 13 - - - 13 - - 14 - - - - 14 - - - - Age 30-39 10 - - - 10 123.2 81.4 41.7 13 11 128.3 80.2 48.1 41 11 119.1 74.7 43.1 45 12 117.4 70.5 46.9 24 12 114.7 80.7 34.0 3 13 - - - - 13 - - - 14 - - - - 14 - - - - Age 40-49 10 - -- - 10 51.1 92.6 58.5' 19 11 132.6 81.9 50.8 27 11 140.6 84.5 56.0 24 12 133.5 82.8 51.5 17 12 157.0 78.0 79.0 2 13 - - - - 13- - - - 14- - - -14- - - Age 50-59 10 126.1 82.2 43.9 7 10 157.6 91.2 66.4 15 11 141.3 84.3 57.1 19 11 146.8 86.3 62.7 28 12 - - - - 12- - - - 13 -- - 13 - - - - 14 - - -- 14 - - - - Age 60 + 10 - - - - 10 169.2 84.8 85.2 13 11 137.9 77.9 60.0 16 11 160.2 81.6 78.6 29 12 155.3 83.4 71.8 12 12 148.0 68.7 79.3 3 13 - - - - 13 - - - 14 - - 14- - - -

-90Table 9.2 RELATIONSHIP OF HEIGHT-WEIGHT INDEX (HT/;WT) TO BLOOD PRESSURE WITHIN AGE-SEX GROUPS Negro Race, Nassau, 1958 Males Females (Non-pregnant) Age 6-9 Pulse No. Pulse No. Ht.Wt. Systolic Diastolic Pressure in HtVWt. Systolic Diastolic Pressure in Index Mean Mean Mean Group Index Mean Mean Mean Group 10 100.0 60.0 40.0 1 10 - - - - 11 - - - - 11 99.5 42.0 52.5 2 12 101.2 64.5 36.7 40 12 102.4 65.7 36.6 30 13 100.8 62.9 37.8 78 13 102.4 64.3 38.1 92 14 95.0 63.0 32.0 2 14 101.3 64.2 37.1 9 Age 10-19 10 - - - - 10 188.0 120.0 68.0 1 11 135.5 73.0 62.5 2 11 126.2 68.4 57.8 9 12 119.6 72.8 46.8 57 12 117.0 69.0 48.0,92 13 114.4 67.7 46.7 202 13 110.0 67.9 42.1 153 14 108.5 67.8 40.8 29 14 110.2 65.9 44.3 27 Age 20-29 10 - - -- 10 188.o 120.0 68.0 1 11 136.9 88.5 47.9 12 11 128.8 81.8 46.9 39 12 131.5 80.2 51.3 97 12 123.2 77.7 45.3 131 13 128.6 79.7 48.7 115 13 122.5 78.1 44.3 92 14 111.8 70.7 41.2'6 14 116.5 72.0 44.5 4 Age 30-39 10 - - - - 10 138.6 89.4 49.2 13 11 146.1 98.7 51.0 11 11 139.3 90.3 49.0 43 12 140.6 88.8 51.9 54 12 134.8 86.5 48.4 75 13 133.5 85.9 47.7 47 13 132.5 80.9 51.8 36 14 184.0 110.0 74.0 2 14 111.3 68.7 42.7 3 Age 40-49 10 139.0 81.o 58.0 2 10 159.3 98.3 61.0 11 11 148.8 96.5 52.2 13 11 160.6 95.5 64.9 56 12 142.1 93.7 48.4 63 12 146.2 90.8 55.3 75 13 140.3 90.4 61.0 54 13 144.7 89.9 54.7 26 14 142.0 88.0 54.0 2 14 140.3 78.0 62.3 4 Age 50-59 10 - - - 10 172.5 95.5 76.8 13 11 181.3 100.7 80.7 6 11 171.1 96.5 74.6 20 12 150.7 92.5 58.2 31 12 156.1 91.7 66.1 37 13 144.6 87.3 57.2 26 13 158.4 91.1 61.0 16 14 - - - - 14 - - - Age 60 + 10 174.0 84.0 90.0 1 10 196.8 95.0 101.8 4 11 170.9 98.3 73.0 7 11 186.1 94.6 92.1 17 12 186.1 90.5 95.6 8 12 171.0 88.5 82.5 22 13 151.6 84.0 67.6 17 13 172.5 85.3 87.3 18 14 210.0 110.0 100.0 2 14 153.0 90.0 63.0 2

-9111.8 b 3 12.0 - - WHITE MALE._ _ 0 —- WHITE FEMALE 0 / *-A* NEGRO MALE NEGRO FEMALE 12.2 12.4 12.6 12.8,. 13.0 13.2 13.4 I I I I I I 10 20 30 40 50 60 AGE Figure 9.1 Mean of Height-Weight Index by Age. Negro and White Races. Nassau, 1958.

-92of both races are nearly equally obese and relatively obese compared to the men of their respective raceso In regard to variations in blood pressure corresponding with changes in the ponderal index, comparisons can only properly be made within age, sex and race specific groupso Table 9,2 demonstrates these comparisons for the Negro raceo As the ponderal index increases from the heavy to the lean (10-14), the mean systolic, diastolic and pulse pressure are recorded~ If only those groups with larger.than 5 persons are considered, it is seen that throughout the age span and for both sexes there is, as a rule, an orderly elevation of systolic and diastolic blood pressure as the index reflects increasing heaviness of build. These increases are of similar magnitude to those seen with arm girth, from about 5-25 or more mm, systolic pressure and from about 5-15 mmo diastolic pressureo The magnitude of the change is not detectably related to age. Data for the white race is unfortunately not as distincto Number of persons were smaller; the white persons tended to be less variable in regard to this index (only one white person in the total had an index of 15 or over), When viewed by finer divisions of the index, differences are not obvious, In Table 9ol the mean pressures are calculated in the same index groups as the Negro, It; appears that the same trend is present, but these data are inadequate for verification of this trendo In summary, data have been presented regarding the heightweight index and its relation to blood pressureo The following points were noted:

-931 T.:he mean height-weight index'by age shows that women of both. races are considerably heavier in build than. their mal.e counterpartso 20 The curve on tih.ese.means by age is similar for women of both races and men of bot.h races, at...hough Negro womrn are genera..l.y heavaer t. han white women and'the reverse t;rue for mal.es, white males being heavier than. Negroes 30 Analysis of mean blood pressure for leanness or heaviness with i. age group for Negroes showed a definite trend to higher pressu.res in the heavier body buildso Suc-h a trend co-l.d not be shown for whites in t-..ls study becase oinadeq'u.ate n'ztmbers, a l;ho-.gh a su;jgg estLion.. of a s.imi.ar trend was presernt 4o..he por.n.de:ral. index used in, the above anal.ysis represents hbeaviness per u.n'l.. of height and in..o sense may this index be juiged as a m.eas"re of obes.itS, per se,

CHAPTER X RELATIONSHIP IN BLOOD PRESSURE TO ARM GIRTH It has long been felt that obesity is an important factor in elevating human blood pressure, Certainly there appears to be ample physiological grounds for this belief since added fatty tissue causes increased work for the heart, an increase in stroke volume and an in, crease in at least the systolic aspect of blood pressure In addition to this, the excess adipose tissue in the arm would seem to alter the mechanical process of compressing the arm and the artery under the sphygmomanometer cuff. Both of these effects have been postulated in greater or lesser degree by many investigatorso (15) Pickering finds that the increase in blood pressure is fairly well correlated with arm girth up to the age of 50 Hamiltong (6) Pickering, Roberts and Sowry suggest correction factors for increase in arm girth to obtain a more nearly true reading. However, use of this correction factor seems to indicate that it tends to overcorrect, In Figure 10.o is seen the progression of the mean arm girth in regard to age, sex and racea This demonstrates that in both races there is little progression in mean arm girth after adulthood and an actual mean decrease in size after age 50o Although the curves are quite similar in magnitude, Negro men and women are seen to have quite consistently larger arms than theLr white counterparts. However, as seen in the preceding section, the height-weight index (heaviness of build) study shows that white men at nearly all ages are heavier -94

-9512 o II o 10 z I,9 // - WHITE MALE r /X/ 0 —-O WHITE FEMALE /z | -— A *NEGRO MALE c< df * —-e NEGRO FEMALE 8 _ 7 - 6 1 1 1 I I I 10 20 30 40 50 60 AGE Figure 10.1 Mean Arm Girth by Age. Negro and White Races. Nassau, 1958.

-96in build than Negro men, whereas Negro women are generally heavier than white women. The Negro men then, would seem to have larger arms by virtue of muscularity rather than adipose tissue. This seemed to be the case by clinical impression. PickeringIs curves of mean arm girth are quite similar to the present data. Comparing the configuration of the curve with the systolic pressure curve (sigmoid in shape) for the white race it is reasonable to suppose that a positive correlation exists between them until middle life. However, it seems highly improbable that such a relationship exists for the Negro race since the systolic curve for Negroes rises so steeply. Disregarding means of arm girth by age, a more valid approach seems to be to study each age-sex-race group separately, comparing mean blood pressure by arm girth within these groups. This analysis is shown in Tables 10ol and 10.2o When first computed by single inch intervals, the numbers in groups were small and obscured a possible trendo The data are shown in two inch intervals with mean systolic, diastolic and pulse pressure for eacho For an analysis of this kind it would have been better to have greater numbers within groups for a more accurate idea of trend, particularly in the white race. For age-sex units in the Negro race these data show a rather remarkable trend toward higher systolic and diastolic pressures as arm girth increases. This is true not only in the pre-adult (growth) phase but is also true in the Negro male until age 60. At that point numbers become smaller and pressure more variableo In the Negro female this relationship is identifiable throughout the age span, including those over 60 years.

-97Table 10.1 RELATIONSHIP OF ARM GIRTH (INCHES) TO BLOOD PRESSURE WITHIN AGE-SEX GROUPS Males White Race, Nassau, 1958 Females Age 6 - 9 Pulse No. Pulse No. Arm Systolic Diastolic Pressure in Arm Systolic Diastolic Pressure in Girth Mean Mean Mean Group Girth Mean Mean Mean Group 6-7 102.5 60.4 44.0 31 6-7 102.0 65.0 37.0 22 8-9 106.1 61.6 44.5 10 8-9 109.0 69.7 39.3 14 10-11 - - - - 10-11 114.0 58.0 56.o 1 12-13 - - - - 12-13 112.0 76.0 36.0 1 Age 10-19 6-7 103.7 59.8 44.0 8 6-7 104.7 65.6 39.1 11 8-9 116.6 67.0 49.6 23 8-9 116.5 65.6 51.2 41 10-11 127.5 67.3 60.2 29 10-11 120.8 69.5 51.3 17 12-13 128.7 60.7 68.0 3 12-13 144.0 74.0 70.0 2 14-15 162.0 94.0 68.0 1 14-15 - - - Age 20-29 7-8 - - - - 7-8 111.0 68.0 44.0 6 9-10 131.3 70.8 60.1 27 9-10 116.9 73.2 43.5 50 11-12 128.7 75.0 53.7 28 11-12 127.8 74.8 53.0 22 13-14 158.0 94.0 64.0 1 13-14 128.0 84.7 43.3 3 15-16 170.0 110.0 60.0 1 15-16 190.0 108.0 82.0 1 Age 30-39 7-8 111.0 60.0 51.0 2 7-8 105.5 67.5 38.0 4 9-10 122.3 75.5 46.8 25 9-10 116.7 73.2 42.1 39 11-12 125.0 77.1 47.9 34 11-12 129.5 85.8 43.7 13 13-14 137.0 88.0 39.0 4 13-14 125.0 81.5 43.5 4 Age 40-49 8-9 144.6 88.6 55.9 7 8-9 145.6 83.4 62.3 8 10-11 129.4 77.0 52.3 25 10-11 151.3 92.0 59.3 21 12-13 133.6 87.9 45.7 12 12-13 147.8 89.3 58.5 13 14-15 - - - - 14-15 154.0 94.0 6o.o 1 16-17 - - - - 16-17 140.0 90.0 50.0 1 Age 50-59 9-10 136.7 83.6 53.1 10 9-11 148.0 85.2 61.8 17 11-12 138.9 84,5 54.4 17 11-12 150.0 88.1 64.6 22 13-14 - - - 13-14 170.0 96.6 73.3 3 15-16 - - - 15-16 148.0 88.0 60.o 1 Age 60 + 7-8 133.3 79.7 53.7 3 7-8 166.1 72.1 95.4 7 9-10 151.1 81.9 69.2 18 9-10 150.1 79.6 70.5 27 11-12 134.1 75.0 59.1 8 11-12 170.8 83.8 87.0 10 13-14 - - - - 13-14 178.7 90.7 88.0 3

-98Table 10.2 RELATIONSHIP OF ARM GIRTH (INCHES) TO BLOOD PRESSURE WITHIN AGE-SEX GROUPS Negro Race, Nassau, 1958 Males Females Age 6-9 Pulse No. Pulse No. Arm Systolic Diastolic Pressure in Arm Systolic Diastolic Pressure in Girth Mean Mean Mean Group Girth Mean Mean Mean Group 5-6 99.6 63.5 36.0 62 5-6 100.1 63.3 36.5 63 7-8 102.1 63.3 38.8 59 7-8 104.2 65.0 39.1 69 Age 10-19 6-7 103.0 64.8 38.1 67 6-7 101.0 67.4 38.3 44 8-9 112.1 66.0 46.2 108 8-9 111.8 66.3 45.4 163 10-11 122.7 73.5 49.6 98 10-11 118.3 71.3 47.2 77 12-13 135.7 74.8 60.9 16 12-13 138.0 84.3 53.7 6 Age 20-29 7-8 126.3 84.0 42.3 3 7-8 110.0 71.2 38.8 17 9-10 127.3 77.8 49.0 67 9-10 120.4 74.6 45.7 182 11-12 129.6 80.3 49.4 145 11-12 124.9 77.8 37.0 94 13-14 142.9 87.7 55.1 15 13-14 129.8 88.5 48.3 15 15-16 - - - - 15-16 175.7 88.7 87.0 3 Age 30-39 7-8 -.. - 7-8 140.8 78.0 62.8 5 9-10 139.2 89.1 50.1 28 9-10 134.7 84.2 49.7 72 11-12 136.7 86.4 50.9 65 11-12 131.7 84.8 47.0 77 13-14 145.6 96.3 49.2 21 13-14 141.6 89.3 52.2 25 15-16 - - - - 15-16 144.0 92.4 51.6 5 Age 40-49 8-9 140.0 87.8 52.2 9 8-9 139.6 85.2 54.5 27 10-11 141.4 93.4 57.8 71 10-11 149.5 90.0 55.4 80 12-13 141.6 93.4 48.2 47 12-13 158.2 95.0 63.2 59 14-15 157.3 91.7 58.0 6 14-15 156.6 98.5 58.1 2 16-17 168.0 104.0 64.0 1 16-17 189.0 102.0 87.0 1 Age 50-59 9-10 142.7 85.9 56.8 18 9-10 157.1 90.7 63.1 30 11-12 152.1 90.8 61.3 38 11-12 156.2 91.7 64.4 28 13-14 161.6 103.8 57.9 8 13-14 184.6 102.3 82.4 11 15-16 - - - - 15-16 188.3 97.8 90.5 6 Age 60 + 7-8 186.7 100.0 86.7 3 7-8 171.2 91.2 80.0 5 9-10 149.1 84.6 84.5 13 9-10 168.0 84.3 83.9 23 11-12 173.4 82.7 90.7 12 11-12 179.3 97.0 82.3 14 13-14 147.9 105.3 76.3 6 13-14 189.4 89.7 99.7 7 15-16 - - - - 15-16 178.3 99.3 79.0 3

-99 - The corresponding data for the white race are less reliable because of small numberso However, the same trend is seen in both sexes for both systolic and diastolic pressures up to the age of 40, After this point the relationship is obscured but its presence is not ruled outo The magnitude of the difference in blood pressure between slender arms and heavy arms is variable, but most frequently in a range of 10-20 mm, systolic pressure and 5-10 mmo diastolic pressure in both races where the trend is observedo This difference is measured only between groups with relatively larger numbers of greater validity, From these data it cannot be said that a large arm connotes absolutely a higher than average blood pressure, Variation is too great, as was seen in the original ungrouped data. However, it seems justifiable to say that within an age-sex group, particularly for Negroes, persons with a large arm would on the average be expected to yield a blood pressure somewhat higher than a person with a slender arm, throughout the age span, For the white race these data cannot demonstrate such a relationship beyond the age of 40 years, The data cited here consider the effect of arm girth separately, although it is undoubtedly very closely related with body build,

CHAPTER XI PARITY IN RELATION TO BLOOD PRESSURE Several studies in recent years have attempted to assess the effect of parity on blood pressure. Notable of these is Humerfelt and (9) (13) Wedervang and Miall and Oldham( It has long been felt that nothing adverse should happen to the blood pressure in normal. pregnancy. If pre-eclampsia should occur, this, of course, will cause blood pressure elevation during the third trimester of pregnancy. At the termination of pregnancy the elevated blood pressure usually returns to normal within a relatively short time. Occasionally, however, the blood pressure remains elevated for long periods or even for lifeo It is a matter of conjecture whether cases of this sort represent renal hypertension as a result of the toxemia or if this represents the onset of essential hypertension occurring in the course of pregnancy, Local practitioners state that toxemia was until very recently a rather serious problem on the island but one which has been greatly reduced in the past several. yearso Attempts were made at obtaining history of pre-eclampsia or eclampsia by interview with the woman seen during the present study but the data gained were felt to be too inaccurate for further study. In view of this situation it seems somewhat contrary to logic to suppose that a population study of women in regard to pregnancy histories should reveal a lowering of blood pressure with increasing multiparityo -100

Miall and Oldham's work is most comparable with the present data since their parity and blood pressure comparisons include women 15-45 years. Humerfelt and Wedervang studied blood pressure of women only at the end of the reproductive period. However, both studies report essentially the same trend of lowering of blood pressure with increasing multiparity. Single women have the highest pressures (both systolic and diastolic) Pressure is then gradually lowered progressively for nulliparous married, those with one child and those with two or more childreno Remarkabl.y this same relationship was also seen for fathers in these families. Although the present study was not designed with this specific goal in mind, the available data were analyzed for this effect, The original data collection sheet provided space for only three pregnancy histories for women in a given household, If more than three were present, the pregnancy histories were frequently unrecordedo However, of a total of 826 Negro women included in the sample over 20 years of age, 676 had pregnancy histories taken,, This9 of course, allows for some statistical criticism, yet it is felt that if a genuine trend were present it would probably reveal itself. The data are presented in Table 11o.1 showing mean systolic and diastolic pressutre for parity groups within age groups. One is struck not by a trend in one direction or another, but the rather remarkable homogeneity over the parity groups within age grouping, These data would not support a hypothesis of either increasing or decreasing blood pressure with increasing parity.

-102Table 11.1 RELATIONSHIP OF PARITY TO BLOOD PRESSURE WITHIN AGE GROUP Negro Females, Nassau, 1958 Age 20-29 Parity Systolic Diastolic Number In Mean Mean Group 0 127.7 81.7 38 1-2 122.9 78.0 89 3-4 125.8 81.5 66 5-6 120.1 79.8 19 7-8 123.8 77.6 5 9+ - - - Age 30-39 0 129.5 81.9 59 1-2 139.1 86.9 37 3-4 130.9 85.4 42 5-6 141.0 89.6 27 7-8 135.7 86.4 14 9+ 134.5 88.0 11 Age 40-49 0 163.5 96.0 21 1-2 148.5 91.6 51 3-4 151.4 91.3 36 5-6 149.7 91.6 27 7-8 150.0 93.6 16 9+ 154.8 94.9 9 Age 50-59 0 165.8 90.0 8 1-2 151.4 92.0 14 3-4 165.9 96.9 17 5-6 161.9 95.5 15 7-8 167.8 90.8 5 9+ 169.7 91.2 16 Age 60+ 0 166.3 86.7 9 1-2 165.6 87.5 12 3-4 181.3 90.8 8 5-6 184.4 88.9 9 7-8 177.3 94.3 6 9+ 179.3 94.5 8

l-03A further finding of interes t, however, is the study of the blood pressure of the 73 Negro women pregnant at the time of the study, Table 11ol shows that the pregnant women had cor.sistently lower blood pressures (both systolic and diastol.ic) then the mean for all women in the respective age groups (pregnant women included)o Although one may not; draw conclusions from. this small groua.p, this findings is provocative It has been postulated t+hat blood pressure is sl.igbtly reduced in normal pregnancy, and previolus..y hypertensive women may frequently show greater reductionso The present findings are consistent with. the impression of reduced blood pressure in pregnancy. Significa.nce of t;hese findings in regard to sex differences seen in the distributions by age is u.nceretair_.;O Ta7..be.11,2 MEAN BLOOD PRESSURES BY AGE GROUP IFORE PREGNANT WOMEN COMPARED TO MEAN BLOOD PRESSURE FOR ALL WOMEN (INCLEUDING PREGN'ANT)1? NEGRO RACE NASSAU, 1.958 Systolic Mean Diastol.ic Mean Age Number Groups All Women Pregnant All Women Pregnant in Group in. Sample Women in Sample Women (Pregnant) o10-9 i.5.4 l-4 9 69. 62,6 9 20-29 123 76 1150 76 7 6,1. 9 43 30-39 1367o7 1.74, 9 86 74o7 15 40-49 i50o 51 110.0 91o20 66 7 6

CHAPTER XII RELATIONSHIP OF PULSE RATE TO BLOOD PRESSURE The pulse rate measurement was shown to be less reliable in regard to differences between observers than was the blood pressure determination, This fact may be partially accounted for by the use of 15 second pulse countso However, the pulse counts are probably sufficiently accurate for comparison. It has been felt that the pulse in relation to blood pressure has only one connotation, that of a systolic increase with excitement or emotion. However, Remington(7) has seen that children of hypertensive parents have lower pulse rates than do children of normotensives. Corcoran and Schneckloth(4 ) studying Negroes in Sto Kitts, Leeward Islands, show a tendency by age group to higher systolic and diastolic pressures as the pulse increaseso For the present data, Table 12ol demonstrates that by agesex-race group, no visible trend for increasing pulse is apparent except for a slight increase in pressures with increasing pulse in the young adult, Since the St. Kitts study was done in a clinic situation and the present study was done in the home, the factor of emotion may have played a greater part in the former0 "10 4c

-105Table 12.1 RELATIONSHIP OF PULSE RATE TO BLOOD PRESSURE WITHIN AGE-SEX GROUPS Negroes, Nassau, 1958 Males Females Age 6-9 No. No. Pulse Systolic Diastolic in Pulse Systolic Diastolic in Mean Mean Group Mean Mean Group 55-64 91.0 38.0 2 55-64 - - - 65-74 104.0 54.7 5 65-74 97.3 65.3 3 75-84 100.7 65.8 25 75-84 99.7 59.5 27 85-94 100.3 66.7 24 85-94 102.8 66.0 33 95+ 101.2 62.4 65 95+ 103.2 65.$ 70 Age 10-19 55-64 116.6 70.4 25 55-64 117.3 70.7 3 65-74 117.6 70.1 48 65-74 116.4 69.7 18 75-84 115.4 67.7 113 75.84 112.0 68.0 106 85-94 114.0 68.4 54 85.94 113.4 68.4 71 95+ 110.7 68.9 47 95+ 11.6 57.5 93 Age 20-29 55-64 125.5 77.8 39 55-64 121.3 78.9 16 65-74 130.6 83.0 58 65-74 124.8 80.3 39 75-84 129.6 79.6 79 75-84 120.7 75.6 119 85-94 131.0 80.4 33 85-94 120.5 72.7 74 95+ 154.2 76.7 19 95+ 126.8 77.6 61 Age 30-39 55-64 130.4 82.4 15 55-64 138.1 85.3 9 65-74 137.1 89.3 17 65-74 133.8 84.9 19 75-84 134.3 85.8 46 75-84 133.5 86.1 74 85-94 146.4 95.3 19 85-94 133.8 83.6 42 95+ 152.7 95.1 17 95+ 136.5 85.5 41 Age 40-49 55-64 140.7 91.0 6 55-64 149.4 97.7 7 65-74 143.3 92.1 19 65-74 157.5 95.0 31 75-84 139.1 92.5 46 75-84 144.5 89.8 64 85-94 140.7 92.3 30 85-94 151.4 92.8 40 95+ 146.6 90.0 33 95+ 151.6 89.0 36 Age 50-59 55-64 152.8 87.6 5 55-64 161.0 88.3 6 65-74 144.7 90.1 15 65-74 174.3 98.6 8 75-84 147.0 88.3 22 75-84 150.4 89.6 30 85-94 161.9 96.9 14 85-94 158.9 94.3 9 95+ 148.0 91.1 7' 95+ 170.9 94.8 23 Age 60-69 55-64 -- - 55-64 173.5 94.5 4 65-74 148.0 90.9 7 65-74 169.2 96.8 5 75-84 183.6 97.8 9 75-84 161.2 88.0 10 85-94 136.0 84.5 4 85-94 158.8 86.2 9 95+ 150.4 74.0 5 95+ 173.7 85.7 6 Age 70 + 55-64 180.7 86.0 3 55-64 154.o 66 1 65-74 160.7 84.0 3 65-74 186.4 93.8 10 75-84 202.0 97.0 2 75-84 192.9 91.8 11 85-94 222.0 108.0 2 85-94 165.5 79.0 4 95+ - - - 95+ 209.8 102.5 4

CHAPTER XIII SUMMARY OF THE STUDY The purpose of this study has been primarily to describe the distributions and characteristics of the blood pressure in a population, comparing the Negro and white races, age and sex specifically. Since the entire population could not be studied, the approach to study which is felt to provide the greatest likelihood of valid results lies in the implementation of adequate sampling techniques, The use of these techniques is finding greater use in the epidemiology of chronic diseaseo Reasons for choice of site for the study were notedo These included the factors of geographic and genetic isolation, the reports of higher levels of blood pressure in the people, and the possibility of an environmental factor (salt intake) which might contribute to the elevation of blood pressure. The sampling method was discussed and measures to judge validity were described with the conclusion that the sample is probably a valid one, The technique of the examinations were standardized and variations between examiners were evaluated so that the information gathered would be as comparable as possible. It was found that differences between examiners were acceptably smallo Comparisons were made between two blood pressure determinations performed about one minute apart, Differences were minimal However, first reading casual blood pressures were used in the analysis. Both -106

-107 fourth and fifth phases of diastolic pressure (muffling of sounds and disappearance of sounds, respectively) were measured, Difference between them was found to be greater in the young and on the average smaller than previously thought (only 3-4kmm.) Variability (standard errors) of fourth and fifth phases was essentially equal, The fifth phase of diastolic pressure was used in the analysiso Distributions of systolic and diastolic mean pressures have been presented along with measures of variability made possible by the sampling procedure. These measures included standard errors of means and standard errors of the individual age-sex-race distributions. It was found that a definite difference was apparent between the Negro and white races. For the white race, the means of systolic pressure by age group showed a sigmoid curve of increase as age progressed, with a "plateau" of relative stability in early adulthood, The Negro enjoyed no such respite. The means for Negroes of systolic pressure progressed upward inexorably from the youngest to the oldest age groups. Curves of diastolic means by age showed no great difference in configuration between. the races, both being roughly parabolic in shape with Negros slightly higher than the white raceo Distributions of both systolic and diastolic pressure for age, sex, and race specific groups demonstrated again the continuous nature of the curves, with no evidence of bimodality or any clear-cut separation between normal and abnormal

-io8In regard to sex differences, the systolic blood pressure curves for both races show that females in early childhood have higher pressures than the males~ At about age 13 years the average systolic pressure of males exceeds that of females and continues to do so to age 38 years when female blood pressures again become the greater. Rather remarkably, these cross points on the curves are. almost identical for the two races, Sex differences in diastolic blood pressure are less distinct but still present in the white race, but are minimal in the Negro. These distributions were compared with other studies of similar nature for both white and Negro races. It was seen that the blood pressure distributions for each race were quite comparable to those of other studies on similar racial stocks in other parts of the world. This observation tends to speak against the likelihood of a local environmental factor such as a high salt intake in causation of elevated blood pressures. Blood pressure was then compared in another fashion than that of arithmetic means, that of determination of proportionsi within the age-sex-race specific groups which have pressure readings above fixed break points on the blood pressure scale. Findings by this approach were: 1) The Negro race again showed significantly higher proportions of persons with elevated blood pressures than did white persons at all break points chosen.

-1092) Sex differences within race, when viewed in total over the entire age span, showed no significant differences. However, women have significantly higher systolic pressures than men after age 40 in both races. Since the total effect for sex was not significant, the expected compensating difference in favor of mean under 40 is less significant. Sex differences in diastolic pressure were much less evident, as expected from the study of means, over the entire age span or at any age grouping, with one exception. White females over 40 years have higher proportions of persons with greater than 90 diastolic pressureo This relationship was not true for the 100 diastolic reference point. 3) Data in regard to break points of systolic and diastolic pressure suggest that commonly used criteria, eog. 150S and/or 100D, do not carry equal weight in terms of overall distribution. Perhaps 160S and/or 100D, and 150S and/or 90D would more nearly equate these valueso The data were also studied to determine the effect of varying anthropometric measurements such as girth of the arm and body buildo It was demonstrated that within age-sex groups there was a definite trend toward higher pressures in heavier arms and in heavier builds, particularly in the Negro.

Blood pressure in relation to parity was seen to show no recognizable trend in the Negro race from nulliparity to multiparity, However, although only 73 Negro were pregnant at the time of the study, their mean blood pressures by age group were consistten.tly lower than the mean pressures for all women of their,respective age groupso The import of this finding is unclear Relationship of blood pressure to pulse rate was also investiated. Except for a slight trend toward higher pressures with higher rates in young adults, no real effect was'noted. The association of blood pressure findings with degree of salinity of water was not analyzed, partly because of the indication mentioned that a local environmental factor was probably not operative. Also, the water samples were drawn from units within strata, whereas the analysis was centered on post-stratification variables'. The water samples did show a relatively high salt content and this may be investigated in the subsequent analysiso

CHAPTER XIV DISCUSSION The goal of this study has been to describe, as accurately as possible, the distribution of blood pressure in a population and to draw comparisons between two races, noting differences by age and sex~ The findings prompt speculations on. what possible implications may be inherent therein, beyond the purely descriptive aspecto At the outset, one feels constrained to dispute the view of Pickering( ) that essential hypertension is not a disease since blood pressure, like height, is a continuously distributed variable in the population. Distributions of blood pressure increase progressively with age, becoming more positively skewed as age progresseso Pickering's concept that no "disease" exists is a statistical on.e,9 based on. the observation that at a given age the distribution curve shows no natural division point nor the presence of two peaks, one denoting the mode of "norm.al" persons, and. the second the mode of abnormal (or "hypertensive") persons. This reasoning seems'unduly restrictive and does not provide any clues to the un.derlying pathological process. A new concept of the meaning of these distributions is needed and will be offered, However,9 we must first reiterat.e that the word "disease" is, and should be, a clinical determinationo Moreover, it is well known that the disease "hypertension" when established definitely, carries an increased risk of death or disability, Hence, from the distribution of blood pressures in. a population it is difficult; to argae that no disease entity exists o -1..11

-.1 2 Is there any reason to believe that separation of a pathological process by inspection of the distributon of0 a normal variable such as blood pressure is necessarily dependent apon. the presence of a bimodal separation? Although it is conceivabl.e that some physiological variables may behave in this manner, such a separation wo-uld imply a -rather marked change in value with initiation of the patholl.ogical process such that it is clearly demarcated from the normal. This is not the case ir essential h ypertension0 Rather, essential hypertension would appear t;o be usually of gradual onset, with considerable overlap of "normal" values in the early stages, It is this overlap which may explain the positivekly skewed, unimodal distribution curves seeno An analogy might be drawn to the range of blood sugar values seen in diabetes mellitus, One cannot deny the existence of the disease, yet if no diagnostic aids were available other than a single determination of blood sugar, it would be extremely difficult to set a value above which diabetes is present and below which lies normalityo One cannot say when or how essential hypertension beginso No diagnostic tool is yet available which is analogous to the glucose tolerance test in diabetes, Assuredly, many tests of blood pressure reactivity are known, but no standardized procedure exists to classify a person at a given age, sex, race, and blood pressure as either hypertensive, pre- ypertensive, or normalo If it is true tbat hypertension. is a disease which has a beginning and an end, it may be postu.l.ated t.hat the continuous distribution curves of blood pressure seen must be a composite of two basic distributions, the "physiological" and the "pathological., as seen in the accompanying diagramo Both underlying curves may approximate bell-shaped distributions, with both high and low valueso These curves overlap such that a person with a given blood pressure may be in the "physiological"' distribition, while another at the

-1!53 same pressure may actuaally be in the initiation of his pathological processo It may well be possible in. the futu.re to formulate mathematical models to describe these two underlying distribsutions. It would then be feasible to attempt to separate these two distributions by a longitudinal study over several years to verify the postulates suggested regarding these distributionso If this could be done, the basic framework would be already available when a more precise method of identification. of the susceptible person is foundo At first glance one might question the justification for laborious efforts to describe the susceptible portion of a population before specific etiology and treatment; are knrown,'Yet, the great majority, if not all, of diseases of unknown. etiology were first studied in a somewhat similar manner. This is the epidemiol.ogical me..thod. If the susceptible groups are identified, important clues may rise in the search for etiology. Current thinking in. regard to essential hypertension revolves about the varied aspects of pathogenesis, the neural, hmnoral, and renal mechanisms, modified in some way by the presence of a predisposing genetic disturbanceo In spite of demonstrated potential relationships to disease of the above mechanisms the precise pathogenetic pattern is unknowno For example) is vasomotor lability a necessary prodromal stage in. development of disease? Are all persons with vasoaot;or lability at greater risk of oo ~ de 1 |d^ / Physiological Distribution o H )rd (U Cd x / \ X -Cumulation 4 ao, $H / \@ ^Distribution P; LI CO... —I Blood Pressure (m.n.)

-114subsequent hypertension? Do intermittent elevations of blood pressure cause subsequent vascular changes which result in permanent elevations of blood pressure, or is the vascular change the primary pathological alteration? How is pathogenesis of hypertension related to the aging process, and if so, why do some persons with marked signs of aging (as grossly estimated) demonstrate "normal" blood pressure? In regard to aging, it must be re-emphasized that one of the principle positive effects in this study was the relationship of age to blood pressure. One is reminded of the dictum "Age is a matter of arteries —not years)" However, such a phase is not entirely adequateo N merous cases of people with the usual outward physical signs of senility demonstrate blood pressures in the normal range, both for systolic and diastolic pressureO The more usual pattern in the aged is at least a systolic elevation with increasing arteriosclerosis. Such a pattern suggests that blood pressures in arteriosclerosis are a reflection of change primarily in vascular structure resulting in decreased distensibility and increased peripheral resistance rather than a change in the vasomotor apparatus itself. However, the present study also shows that the mean but not necessarily the individual blood pressures increase progressively over all of the age span. Can this be merely a manifestation of progressive arteriosclerosis on the blood vessels? This would not appear to be a satisfactory answer. The formulation of hyperthetical curves as shown provoke one to speculate if the increase in mean blood pressure with age is not an entirely pathological process. We note from the data that at least 30-40k of persons do not ever reach even the minimal breakpoints of 140/90o

-115The increase in means with. age could be explained if one assumed that as age progressed, the proportion of persons leaving the "physiological" distribution to join the "pathological" increased with age. The same generalization would hold regarding overlap of the two distributions, unimodality and positive skewness, The second major point of the present findings is the definite difference seen in the racial comparisons. The Negro tends not only to have higher mean pressures (after adolescence) but the pattern of progression with age is differento One sees a "plateau" effect in young adult life in the white race, while the Negro demonstrates an inexorable progression from the earliest ages to the oldest. Does this difference in pattern simply represent an accelerated process of aging in the Negro? Though it may possibly be true that aging is accelerated in the Negro, (which is extremely difficult to evaluate) one feels that this is not the answero From what has been postulated from the distributions, aging, per se, need not be a primary factor in etiology. Could the difference be due to the inheritance of single or multiple genes which predispose the vascular system to hypertension when acted upon by a specific single stress or a combination of stressful factors? It may be possible that the gene is present in equal proportions in both races with the essential difference being that of time of expression. The time of expression may be determined by an unknown environmental factor. One is reminded of the situation in Huntington's chorea in which time of onset is variableo In this case, a possible inciting factor may be postulated, but the time of physiological surrender to this factor may be

-1l6genetically determined. This "time of expression" may be the point at which the individual person leaves the distribution of "normal" persons to join the "pathological" distribution. This study points out the time of life at which divergence of the mean by age of systolic blood pressure in the two races occurs (the "plateau" effect seen in the white race). This divergence indicates that the probability of initiation of a pathological process differs with time between the races. The probability increases at a regular rate in the Negro, and remains constant or actually decreases in the white person until middle life, at which time probability of initiation of disease increases markedly, Another genetic possibility is that of a greater frequency of the offending gene or genes in the Negro than in the white race. It is personally felt that a differential frequency is not essential to explain the differences in pattern seen, but rather that "time of expression" is the crucial point. Many possibilities of inheritance exist, awaiting investigation. Other factors such as the demonstrated sex differences require elucidation, as well as the effects of parityo It is possible that pregnancy itself may partially explain the lowered pressures of women in the reproductive age, Body build, although undoubtedly of some effect, seems not to be of primary importance. A desirable adjunct to the present data would have been clinical appraisals, particularly of renal disease, Although current knowledge is fragmentary, it is possible that differences in frequency of renal abnormality or susceptibility to infection may in part explain the racial differences seen,

-117What h" been learned regarding that elusive unknown, a possible inciting factor? This was not the basic purpose of the study. Nevertheless, the question of the influence of salt intake would seem to have been partially resolved, at least in its affect on a population, A (4) similar study in St, Kitts showed considerable similarity to the presert study with regard to blood pressure trendso Yet, the St. Kitts Negroes had a low salt intake in food and water while Negroes in Nassau are presumed to be considerably higher in average salt intake~ In conclusion, this study has attempted to delineate differences in patterns of blood pressure in the Negro and white races living in a similar environmento The goal has been to define, in broad terms, susceptible groups of people so that when a more definitive test of individual susceptibility is devised, one may concentrate effort in the more restricted areas in the search for possible inciting factors which may precipitate the hypertensive state Also discussed has been the concept of hypertension as a condition in which the initiation of the disease state may be at, relatively low levels of blood pressure, overlapping blood pressures of persons who are still, in truth, "normal " Identification of these "hidden hypertensives" is the important goal in understanding of the disease, its treatment and prevention.

-118APPENDIX SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING White Males - Age 6-9 (45 Persons) Nassau - 1958 Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 25-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 80-89 1 90-99 3 3 4 2 100-109 1 (020) 1 4 2 4 2 1 2 110-119 5 4 1 1 120-129 1 3 White Males - Age 10-14 (28 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 40 80-89 1 (038) 90-99 1 2 100-109 11 1 2 110-119 1 2 1 2 120-129 1 1 1 2 1 2 1 130-139 1 140-149 150-159 1 White Males - Age 15-19 (33 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 40 100-109 1 110-119 3 1 2 1 120-129 1 1 1 2 1 2 130-139 2 2 1 3 1 t40-149 1 1 2 1 150-159 1 1 160-169 1 White Males - Age 20-24 (34 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 1 1 110-119 3 3 1 1 120-129 2 1 2 2 1 130-139 1 2 3 140-149 1 (032) 1 1 2 1 1 150-159 2 160-169 170-179 1 White Males - Age 25-29 (30 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 110-119 1 2 1 2 1 120-129 3 1 1 2 130-139 1 1 1 1 2 1 1 140-149 1 1 2 1 150-159 1 1 1 White Males - Age 30-34 (38 Persons) Diastolic 0- 4 -9 Systolic 0er 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 1 2 2 1 110-119 1 2 3 1 1 120-129 1 1 1 1 2 1 130-139 4 2 2 1 140-149 1 2 3 2

-119SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING White Males - Age 35-39 (22 Persons) Diastolic Systolic nder 40-44 45-49. 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 3 100-109 1 1 110-119 1 120-129 1 1 2 2 130-139 1 1 2 140-149 1 1 1 150-159 1 160-169 170-179 1 White Males - Age 40-44 (22 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 110-119 2 1 120-129 2 1 1 1 130-139 1 1 1 2 140-149 1 1 150-159 1 1 1 160-169 1 170-179 1 White Males - Age 45-49 (20 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 6o-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 1 2 110-119 1 1 1 1 120-129 2 1 1 130-139 1 1 140-149 1 1 150-159 1 2 1 200-209 1 White Males - Age 50-54 (14 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 6o-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 150-134 135-139 110-119 2 120-129 1 1 130-139 1 2 140-149 1 1 1 150-159 1 1 160-169 1 170-179 1 White Males - Age 55-59 (11 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 1 110-119 1 120-129 1 2 130-139 1 1 140-149 1 150-159 1 160-169 1 200-209 1 White Males - Age 60 or Over (28 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 40 100-109 1 110-119 1 1 120-129 1 2 1 1 130-139 1 1 2 1 140-149 1 1 1 1 1 150-159 1 160-169 1 1 1 1 170-179 1 180-189 1 190-199 1 1 1

-120SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING White Females - Age 6-9 (47 Persons) Diastolic Systolic 40 40-44. 45-49 50-54 55-59 6o-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 80-89 1 1 90-99 1 4 1 4 100-109 1 9 1 4 4 110-119 1 1 2 1 2 2 1 120-129 1 1 12 130-139 1 White Females - Age 10-14 (36 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-1 109 l-114 115-119 120-124 125-129 130-134 135-139 80-89 1 90-99 1 1 100-109 1 1 1 1 2 110-119 3 2 2 3 2 120-129 1 1 3 1 1 130-139 2 3 1 140-149 1 150-159 1 White Females - Age 15-19 (31 Persons) Diastolic Systolic 4-44 4-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-1354 135-139 40 100-109 1 4 1 110-119 1 2 2 1 1 1 120-129 2 1 4 1 130-139 1 4 2 140-149 1 1 White Females - Age 20-24 (43 Persons) Diastolic Systolic: 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 1 100-109 1 1 1 1 2 110-119 1 3 1 3 1 1 1 120-129 1 1 1 4 2 2 2 1 130-139 2 1 1 1 140-149 1 1 150-159 1 160-169 1 1 White Females - Age 25-29 (39 Persons) Diastolic Systolic n0de 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 40 90-99 1 1 100-109 1 2 1 1 1 110-119 1 4 3 3 120-129 1 1 1 3 4 2 130-139 1 1 2 2 140-149 1 190-199 1 White Females - Age 30-34 (33 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 1 (00 1 11 2 1 1 110-119 1 1 4 1 120-129 2 1 2 1 130-139 1 1 4 1 1 140-149 1 150-159 2

-121SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING White Females - Age 35-39 (27 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 2 1 1 1 1 110-119 1 1 2 1 120-129 4 3 1 1 130-139 1 140-149 1 1 150-159 1 160-169 1 1 White Females - Age 40-44 (25 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 1 1 110-119 2 120-129 1 1 130-139 1 1 1 140-149 2 150-159 2 160-169 1 3 170-179 1 180-189 1 200-209 250-259 White Females - Age 45-49 (20 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 1 110-119 1 1 120-129 1 1 1 130-139 1 2 1 1 140-149 1 1 150-159 1 160-169 3 1 170-179 1 210-219 1 White Females - Age 50-54 (15 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 120-129 1 1 1 130-139 2 2 140-149 1 150-159 1 160-169 1 3 170-179 1 1 White Females - Age 55-59 (16 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 1 110-119 1 1 120-129 2 140-149 11 160-169 1 1 170-179 1 180-189 1 190-199 1 200-209 1 260-269 1 White Females - Age 60 or Over (30 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 8o-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 110-119 1 120-129 1 130-139 1 1 2 2 1 140-149 1 3 12 150-159 1 1 2 160-169 1 1 1 170-179 1 (032) 21 180-189 1 200-209 1 240-249 1

-122SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING Negro Males - Age 6-9 (154 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 Systolic 40 80-89 1 (038) 2 2 4 6 90-99 2 (036) 1 1 6 16 4 5 1 100-109 1 (020) 4 6 17 8 17 3 5 110-119 3 4 8 2 9 1 5 120-129 2 1 150-139 1 1 140-149 1 Negro Males - Age 10-14 (143 Persons) Diastolic Systolic 40 4o-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 8o-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 80-89 1 (034) 1 90-99 1 2 3 3 1 4 4 100-109 1 (038) 1 1 7 9 9 9 14 5 2 110-119 6 5 4 6 8 3 5 120-129 1 3 2 6 2 4 1 130-139 1 1 1 1 140-149 2 1 1 1 Negro Males - Age 15-19 (139 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 1 3 2 2 2 110-119 1 3 4 4 12 5 6 2 1 120-129 1 2 2 4 5 9 5 12 6 130-139 1 (032) 1 3 2 6 3 3 3 140-149 1 3 2 2 1 150-159 1 1 1 3 1 1 1 160-169 1 170-179 1 Negro Males - Age 20-24 (123 Persons) Diastolic Systolic U0 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 (038) 1 2 100-109 1 1 1 1 110-119 1 1 4 6 6 3 4 3 1 120-129 1 4 3 5 2 10 5 5 130-139 1 1 2 3 3 6 4 1 1 1 140-149 1 4 4 1 2 1 3 150-159 1 2 1 1 160-169 1 2 190-199 1 200-209 1 Negro Males - Age 25-29 (98 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 1 100-109 1 5 1 110-119 1 3 2 3 4 1 120-129 1 1 3 1 5 2 2 130-139 1 1 4 2 6 5 4 1 140-149 1 1 2 4 4 2 1 1 1 150-159 1 3 2 1 2 1 160-169 1 2 1 2 170-179 2 1 Negro Males - Age 30-34 (58 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 40 100-109 1 1 1 110-119 1 2 3 6 120-129 2 2 5 2 130-139 1 2 1 2 4 1 1 140-149 2 1 1 150-159 1 1 160-169 2 2 1 1 170-179 1 1 (210,140) 180-189 1 1 (230,155) 190-199 1

-123SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING Negro Males - Age 35-39 (55 Persons) Diastolic Systolic 40 4o-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 110-119 1 3 120-129 1 3 3 1 3 1 130-139 1 2 2 6 1 2 140-149 1 1 4 1 1 150-159 1 1 1 160-169 1 2 2 170-179 2 1 180-189 11 1 220-229 1 230-239 1 Negro Males - Age 40-44 (67 Persons) Diastolic__ Systolic U4 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 2 1 1 10-119 1 1 2 1 120-129 1 3 5 3 1 1 130-139 1 1 2 3 2 5 1 140-149 2 1 1 2 1 2 150-159 1 1 1 31 160-169 1 1 1 1 170-179 1 180-189 1 1 1 1(196,142) 200-209 1 210-219 1 1(224,140) 240-249 1 Negro Males - Age 45-49 (67 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 2 100-109 1 110-119 4 1 120-129 1 1 4 2 4 130-139 2 3 1 140-149 1 5 5 2 1 150-159 1 1 2 1 2 160-169 1 2 3 1 1 170-179 1 1 1 1 2 1 180-189 2 1 200-209 1 1(200,140) Negro Males - Age 50-54 (37 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 110-119 1 1 120-129 1 1 1 1 1 130-139 4 1 2 140-149 2 1 1 1 150-159 2 1 160-169 1 1 1 170-179 1 1 1 180-189 1 11 190-199 11 200-209 1 1 230-239 1 Negro Males - Age 55-59 (16 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 1 110-119 1 1 1 1 120-129 1 140-149 1 150-159 1 1 160-169 1 170-179 1 190-199 240-249 1

-124SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING Negro Males - Age 60 or Over (33 Persons) Diastolic Systolic n4 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 110-119 1 120-129 1 1 1 130-139 2 140-149 1 1 3 150-159 1 1 1 1 1 160-169 1 1 3 1 170-179 1 1 180-189 1 1 1 190-199 1 1 210-219 1 1(280,140) 250-259 1 300-309 1 Negro Females - Age 6-9 (166 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 40-49 (032) 70-79 1 80-89 1 (038) 2 3 2 90-99 2 2 6 5 15 9 9 1 2 100-109 1 (036) 2 1 4 4 13 13 12 7 4 1 110-119 1 3 3 3 6 4 6 6 120-129 1 1 1 3 1 1 1 130-139 1 140-149 1 Negro Females - Age 10-14 (139 Persons) Diastolic Systolic U0der 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 80-89 1 1 1 90-99 1 5 3 2 2 100-109 2 2 2 3 11 5 6 3 5 110-119 1 (036) 1 1 3 3 12 7 9 3 6 1 1 120-129 7 3 5 8 2 2 130-139 1 1 1 2 1 1 1 140-149 1 Negro Females - Age 15-19 (153 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 2 1 1 100-109 1 2 2 8 4 6 110-119 1 2 3 7 11 12 6 8 3 1 120-129 1 5 7 4 5 4 7 3 3 130-139 2 3 4 1 3 2 2 1 1 140-149 1 2 1 150-159 1 160-169 1 2 180-189 1 Negro Females - Age 20-24 (170 Persons) Diastolic Systolic er 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 1 1 1 2 2 100-109 2 2 5 2 7 1 4 6 1 110-119 1 2 6 11 8 14 3 7 2 3 120-129 1 2 2 8 7 6 2 3 1 130-139 2 3 2 2 7 1 3 2 140-149 1 2 3 2 3 1 150-159 1 1 1 1 1 160-169 1 170-179 1 1(210,140)

-125SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING Negro Females - Age 25-29 (128 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 80-89 1 90-99 3 100-109 1 2 5 2 2 5 1 2 110-119 1 1 4 5 3 4 8 2 120-129 1 5 9 10 6 4 130-139 2 4 1 7 140-149 2 6 1 4 150-159 1 3 160-169 1 1 1 170-179 1 1 180-189 1 1 230-239 1 1(236,146) 250-2-59 1 Negro Females - Age 30-34 (86 Persons) Diastolic Systolic U4 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 1 1 3 4 110-119 1 2 4 2 1 120-129 2 2 4 9 5 1 130-139 1 5 5 4 1 140-149 1 3 2 1 1 3 150-159 1 1 1 1 1 160-169 1 1 1 1 170-179 1 1 180-189 1 2 1 190-199 1 Negro Females - Age 35-39 (109 Persons) Diastolic Under Systolic Ue 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119'120-124 125-129 130-134 135-139 90-99 1 2 100-109 1 1 110-119 1 2 1 4 1 120-129 2 3 4 8 3 5 130-139 1 2 1 4 2 7 1 1 140-149 1 4 7 5 3 1 1 150-159 5 1 1 1 160-169 1 3 1 1 170-179 1 1 1 180-189 3 1 2 190-199 1 200-209 1 210-219 260-269 1 Negro Females - Age 40-44 (87 Persons) Diastolic Under Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 90-99 1 100-109 3 1 1 110-119 1 1 2 3 4 1 120-129 1 6 1 1 130-139 2 2 3 3 2 140-149 4 2 2 2 1 150-159 1 2 1 2 3 1 160-169 2 1 3 1 170-179 1 1 180-189 1 1 2 2 190-199 1 1 200-209 2 1(215,144) 210-219 1 1 1 230-239 1 250-259 1 1(280,140)

-126SYSTOLIC BY DIASTOLIC (5TH PHASE) BLOOD PRESSURES, CASUAL FIRST READING Negro Females - Age 45-49 (75 Persons) Diastolic Systolic nder 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 100-109 2 110-119 1 1 3 2 3 1 1 120-129 1 2 1 130-139 1 3 3 1 1 140-149 1 3 5 3 2 1 150-159 2 5 1 1 160-169 1 1 1 1 2 1 170-179 1 180-189 1 1 2 1 190-199 1 1 200-209 1 210-219 1 1 1 230-239 2 240-249 1 250-259 1 Negro Females - Age 50-54 ( 46 Persons) Diastolic Systolic 40 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79'0-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 80-89 1 100-109 1 1 110-119 1 120-129 2 1 1 2 1 130-139 2 1 1 140-149 1 2 2 150-159 1 (008) 3 1 160-169 1 2 170-179 1 1 1 1 180-189 1 1 1 190-199 1 200-209 1 1 210-219 1 250-259 1 310-319 1 Negro Females - Age 55-59 (30 Persons) Diastolic Systolic Under 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 Systolic 40 120-129 1 1 1 130-139 1 140-149 1 1 150-159 1 1 160-169 1 2 170-179 1 1 1 1 180-189 1 3 190-199 2 200-209 1 1 210-219 1 220-229 1 1 240-249 1 1 1 1(300,140) Negro Females - Age 60 or Over (61 Persons) Diastolic Under Systolic 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84.E5-89 90-94 95-99 100-104 105-109 110-114 115-119 120-124 125-129 130-134 135-139 120-129 1 1 1 130-139 1 1 1 1 1 140-149 1 1 150-159 1 1 2 3 160-169 1 2 4 1 1 170-179 1 1 1 1 1 1 2 180-189 1 2 2 3 1 190-199 4 1 200-209 1 1 210-219 1 1 1 230-239 1 1(220,150) 240-249 1 1 250-259 1 270-279 1

BIBLIOGRAPHY 1. Boe, J., Humerfelt, S. and Wedervang, F. "The Blood Pressure in a Population" Acta Medica Scandinavica. Supplementum 321: 1-336. 1957. 2. Bordley, J., Connor, C. A. R., Hamilton, W. F., Kerr, W. J., and Wiggers, C. J. "Recommendations for Human Blood Pressure Determinations by Sphygmomanometers". Circulation. 4:503, 1951. 3. Comstock, G. W. "An Epidemiological Study of Blood Pressure Levels in a Biracial Community in the Southern United States." American Journal of Hygiene, 65:271. 1957. 4. Corcoran, A. C. and Schneckloth, R. E. Personal Communication. 5. Hamilton, M., Pickering,. G. W., Roberts, J. A. F. and Sowry, G. S. C. "The Aetiology of Essential Hypertension. 1. The Arterial Pressure in the General Population." Clinical Science. 13:11, 1954 6. Hamilton, M., Pickering, G. W., Roberts, J. A. F. and Sowry, G. S. C. "The Aetiology of Essential Hypertension. 2. Scores for Arterial Blood Pressures Adjusted for Differences in Age and Sex." Clinical Sciences 13:37. 1954. 7. Hamilton, M. Pickering, G. W. Roberts, J. A. F. and Sowry, G. S. C. "The Aetiology of Essential Hypertension. 4. The Role of Inheritance." Clinical Science. 13:273. 1954. 8. Humphries, S. V. "A Study of Hypertensions in the Bahamas." South African Medical Journal. 31: 694. 1957. 9. Humerfelt, S., and Wedervang, F. "Further Analysis of the Bergen Blood Pressure Survey," Acta Medica Scandinavica, 159:489 1957. 10. Keyfitz, N. "Estimates of Sampling Variance When Two Units are Selected From Each Stratum." Journal of the American Statistical Association. 52:503, 1957. 11. Kish, L., (ed) Festinger, L., and Katz, D. "Selection of the Sample," in Research Methods in the Behavioral Sciences, New York: Dryden Press, 1953. 12. Miall, W. E., and Oldham, P. D. "A Study of Arterial Blood Pressure and its Inheritance in a Sample of the General Population." Clinical Science. 14:459. 1955. -127

13. Miall, W. E., and Oldham, P. D. "Factors Influencing Arterial Blood Pressures in the General Population." Clinical Science. 17:409, 1958. 14. Morsell, J. A. "The Problem of Hypertension: A Critical Review of the Literature Dealing with Its Extent." In A Symposium on Essential Hypertension: An Epidemiological Approach. Boston, Wright and Potter (1951) 26. 15. Pickering, G. W. High Blood Pressure. New York: Grune and Stratton, 1955. 16. Pickering, G. W. "The Role of the Genetic Factor in Hypertension." Acta Medica Scandinavica 154 (Supplement 312)(1956) 18-27. 17. Remington, R. D. Comparison of General Tests of Blood Pressure Reactivity. School of Public Health, University of Michigan, Ann Arbor, Michigan, 1958. 18. Report on the Census of the Bahama Islands, taken December 6, 1955. 19. Sharer, C. J. Population Growth of Bahama Islands. University of Michigan, Ann Arbor, Michigan, Film 2181, 1956. 20. Weiss, M. M. and Prusmack, J. "Essential Hypertension in the Negro." American Journal of Medical Sciences. 195:510, 1958.