Hypertension/2004/033118: Fields, Burt, Cutler, Hughes, Roccella, Sorlie

Hypertension/2004/033118: Fields, Burt, Cutler, Hughes, Roccella, Sorlie

ONLINE SUPPLEMENT: Methods

Full Title:The Burden of Adult Hypertension in the United States, 1999-2000:

A Rising Tide

Authors:Larry E. Fields, Vicki L. Burt, Jeffery A. Cutler, Jeffery Hughes,

Edward J. Roccella, and Paul Sorlie

Affiliations:(LEF) Office of the Secretary’s Office of Public Health and Science

U.S. Department of Health and Human Services, Washington, DC 20201

and Cardiovascular Division, Department of Medicine,

Washington University School of Medicine, St. Louis, MO 63110

(VLB) National Center for Health Statistics, Centers for Disease Control and

Prevention, U.S. Department of Health and Human Services,

Hyattsville, MD 20782

(JAC, EJR, PS) National Heart, Lung, and Blood Institute, National

Institutes of Health, U.S. Department of Health and Human Services,

Bethesda, MD 20824

(JH) The Orkand Corporation, Falls Church, VA 22043

Short Title:Adult Hypertension Burden in the United States

Corresponding Author: Larry E. Fields, MD, MBA, FACC

Senior Executive Advisor to the Assistant Secretary for Health

U.S. Department of Health and Human Services

200 Independence Avenue

Washington, DC 20201

202-690-7694; 202-690-6960 (fax);

Revised: references 05/11/04 per request of the editor

Word Count: Total (1604); title page (166), methods (1026), references (412)

Methods

A stratified multi-stage probability methodology is used in NHANES to survey a representative sample of the non-institutionalized, civilian U.S. population. A detailed description of the blood pressure protocol used for NHANES 1999-2000 was recently published.[1] Based on formal evaluation of cuff width/arm circumference ratio, end digit preferences, and observer agreement, technical and observer error contributed minimally to BP variability. For this analysis of NHANES 1999-2000, BP was measured on 4,531 persons 18 years of age, including 2,279 males and 2,252 females. Hypertension is defined as a systolic BP of at least 140 mm Hg, a diastolic BP of at least 90 mm Hg, or taking antihypertensive medication (conventional criteria). To more completely estimate total prevalence values, individuals were also classified using nonconventional criteria. These persons were not classified as hypertensive using conventional criteria but were told at least twice by a physician or other health professional that they had high blood pressure. Prevalence estimates for 1988-94 (n= 16,351) and 1999-2000 are based on means of three BP measurements taken on one occasion in a mobile examination center by a physician. Variability of duplicate BP readings has been previously evaluated.[2] Pregnant respondents were not included.

Hypertension prevalence from NHANES for persons 18 years of age is reported as age-adjusted or crude values with variances estimated using replicate weight Jackknife design in SUDAAN for 1999-2000 and Taylor Series (WR) for 1988-1994 (Research Triangle Institute for SUDAAN). Age-adjustment is to the Year 2000 Standard.[3] This adjustment permits comparison of populations that may differ in age distribution. Crude age-specific prevalence data are used in this study to estimate the number of persons with hypertension and the overall adult prevalence percent in 1999-2000.

The U.S. census is used for Congressional apportionment and federal allocation purposes but typically leads to undercounting of the population.[4],[5],[6],[7]Awareness of this tendency is necessary for complete interpretation of reported population-based estimates.[8],[9] For this study, age-specific U.S. resident population estimates were obtained from the bridged-race intercensal file developed by the U.S. Census Bureau in collaboration with the National Center for Health Statistics (Table icen1999) and U.S. Census Bureau July 1, 2000 adult population estimates (NA-EST2002-ASRO-01).[10],[11]For the 1999 estimates, interpolated weights were used to adjust the number of persons 20 years of age to the number estimated to be 18 years of age.

The 1999 and 2000 adult population estimates were averaged and the percentage of persons in each age-specific group relative to the total population 18 years of age and over calculated. The total percent hypertension prevalence was computed using the age-specific hypertension prevalence proportions and U.S. averaged adult population data. The formula used for estimating the number of persons with hypertension living in the U.S. was: [the average of 1999 and 2000 age-specific populations] x [1999-2000 age-specific hypertension prevalence proportions] summed across all age groups. In the absence of information on hypertension for the total U.S. population, rates based on non-pregnant and non-institutionalized U.S. civilians were applied to resident population values to generate an estimate of the total burden. While the assumption that prevalence rates calculated for the civilian, non-institutionalized, non-pregnant population apply to the resident population may not be completely accurate, the estimate of burden obtained will be closer to the true burden than would be the case if estimates of the civilian, non-institutionalized, non-pregnant population were used. The 1990 civilian, non-institutionalized U.S. population is approximately 2% smaller than the resident U.S. population for the same time period. Using this difference as a basis for adjustment, the estimated number of resident U.S. adults with hypertension in 1988-1994 would be about 51 million. Accordingly, the approximately 65 million adults estimated to have hypertension in 1999-2000 would be about 28% higher than the adjusted value of about 51 million adults for1988-1994. The percentage difference between adjusted and unadjusted hypertension numbers is small and without a material impact on hypertension burdens reported herein. Sex-specific burdens of hypertension were estimated for both sexes by age.

Percent, number and standard error (SE) are reported. Prevalence data were generated using SAS version 8.2 and SUDAAN release 8.0, with the exception of the total adult prevalence estimate described above. Independent group t-tests between means and one-way factorial analysis of variance (ANOVA) methodologies were used to determine the significance of differences between populations or of a trend, respectively. P values below 0.05 were considered significant.

The average of 1999 and 2000 population estimates for the non-Hispanic white and black populations, Mexican American and all others are based on a) the bridged-race intercensal file developed by the U.S. Census Bureau in collaboration with the National Center for Health Statistics (Table icen1999), b) the 2000 Census estimate of the proportion of the Hispanic population that is Mexican American (Table DP-1) and c) U.S. Census Bureau July 1, 2000 adult population estimates (Tables NA-EST2002-ASRO-02 and NA-EST2002-ASRO-03).10,11 The number of resident non-Hispanic white or black, Hispanic or Latino, and all other persons identifying with another racial-ethnic population are available in the 1999 bridged-race intercensal file as well as in the 2000 files used in this study. Using icen1999, NA-EST2002-ASRO-02 and NA-EST2002-ASRO-03 files, the Hispanic population in that file is categorized as Mexican-American based on the proportion of the Hispanic population that are identified in the U.S. Census Bureau Table DP-1 (0.5846). The remaining Hispanic population is apportioned to the all other race-ethnicity population count. This process was necessary because the 1999 bridged file does not include estimates for the Mexican American population. In addition, the 1999 file codes race according to the 1977 OMB standard while 2000 files use the 1997 standard which allows for reporting the of multiple races. As only 1.4 percent of non-Hispanic populations in this age group report more than one race, this inconsistency will not affect the population estimates.

To estimate the number of adults with hypertension by sex, race and ethnicity, a two-step process was used. First, crude prevalence rates generated from the 1999-2000 NHANES sample population using SAS and SUDAAN were applied to 1999-2000 U.S. population estimates to compute weights for each sex, race and ethnicity category. Second, computed weights were applied to Table 1 hypertension totals for both sexes.

References

1

[1] Ostchega Y, Prineas RJ, Paulose-Ram R, Grim CM, Willard G, Collins D. National Health and Nutrition Examination Survey 1999-2000: effect of observer training and protocol standardization on reducing blood pressure measurement error. J Clin Epidemiol. 2003;56(8):768-774.

[2] Armitage P, Fox W, Rose GA, Tinker CM. The variability of measurements of casual blood pressure. II. Survey experience. Clin Sci. 1966;30(2):337-344.

[3] Anderson RN, Rosenberg HM. Age standardization of death rates: implementation of the year 2000 standard. Natl Vital Stat Rep. 1998; 47(3):1-16, 20.

[4]United States Government. Democracy in Action. R.C. Remy, author. Glencoe/ McGraw-Hill. New York, New York. 1998. p. 99.

[5] US Census Bureau of the US Department of Commerce. Methodology: Resident Population Estimates of the United States by Age and Sex. Population Estimates Program, Population Division. Intfile2-2.txt. Internet Release Date: April 11, 2000.

[6] US Census Bureau of the US Department of Commerce. Methodology: 1990 Census Base Population Corrections, Adjustments, and Modifications. Population Estimates Program, Population Division. cenadj.txt. Internet Update Release Date: May 2000.

[7]Measuring a Changing Nation: Modern Methods for the 2000 Census. The National Academy of Sciences. . Committee on National Statistics. Division of Behavioral and Social Sciences and Education. National Research Council. The National Academy Press. 1999. p.1.

[8] Freedman DA, Wachter KW. Methods for Census 2000 and Statistical Adjustments. January 2004. Technical Report No. 652.

[9]Proceedings, Third Workshop Panel to Review the 2000 Census. The National Academy of Sciences. Committee on National Statistics. Division of Behavioral and Social Sciences and Education. National Research Council. National Academy Press. 2001.

[10] US Department of Commerce, US Census Bureau. Population Division. Population estimates by national characteristics. Table 1. Internet Release Date: June 18, 2003; Accessed January 28, 2004. Table 2. Table 3. Accessed January 28, 2004. Table DP-1. Accessed January 28, 2004.

[11] U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Division of Data Services with the Population Estimates Program of the U.S. Census Bureau with support from the National Cancer Institute. National Vital Statistics System. U.S. Census populations with bridged race categories. Bridged-race intercensal population estimates for July 1, 1999. 2003. Accessed February 29, 2004.