Value Added

Abstract

We welcome Sorlie's1 appraisal of the Add Health-NHANES discordance. His tenure within the NHLBI,2 its longstanding sponsorship of NHANES, and his intimate knowledge of blood pressure methodology within NHANES indicate his command of NHANES' strengths. The combination of a national probability sampling strategy, high response rate, large sample, clinical equipment, physician staff, and standardized data collection protocol is unusual in epidemiologic studies of cardiovascular disease. However, such combinations are not limited to NHANES and do not conclusively indicate methodological superiority. As described in our original paper,3 Add Health is characterized by some distinctly advantageous variations on these design features. The advantages of Add Health's national probability sampling of adolescents from school rosters (vs. residents from households within population sampling units) may be debatable, but school roster-based sampling is an unlikely contributor to potential overestimation of blood pressure and hypertension. Indeed, the school-roster sampling frame may have led to the selection of slightly better educated, socioeconomically advantaged, and healthier Add Health participants (presumably, with lower blood pressures), compared with the small number of adolescents who never attended school. Add Health's sample size and oversampling of typically underrepresented groups also allow for study of potentially important disparities in blood pressure and hypertension among its representative population of young adults (20 times larger than in NHANES). In contrast, NHANES' national household probability sample lacks sufficient data for estimating mean blood pressure or hypertension prevalence in young adult Asian/Pacific Islanders or Hispanics other than Mexican Americans. Moreover, in a convenience sample of 22 adults aged 24–32 years (55% female; 27% black, 27% white, 23% Hispanic, 23% other race/ethnicity), Add Health's oscillometric method of in-home systolic/diastolic blood pressure (SBP/DBP) measurement produced values strongly correlated with and similar to those estimated by a physician using a mercury sphygmomanometer,4 stethoscope, and manual auscultation; Pearson rSBP = 0.80, rDBP = 0.89, and mean (standard deviation) NHANES-Add Health difference in SBP/DBP = −3 (7)/4 (4) mm Hg.5 These findings are consistent with 2008 guidelines—affirming that in-home, oscillometric measures are accurate, reliable, and easy to collect; often lower than in-office, sphygmomanometric measures (like those in NHANES); closer to 24-hour ambulatory measures (that best predict cardiovascular disease); and useful in detecting masked hypertension.6 With these characteristics of blood pressure measurement in mind, we highlight the 2:1 ratio of measured to self-reported hypertension that we3 found in Add Health, because it is in line with expectations that measurement will capture subclinical hypertension unknown to otherwise healthy young adults. Surprisingly, this pattern is reversed in NHANES, where prevalence of measured versus self-reported hypertension is 4% versus 9%.3 We therefore disagree with Sorlie's1 rationale for “stick[ing] with NHANES as the best” alternative for estimating the true mean blood pressure and prevalence of hypertension in US adults aged 24–32 years. Investigators can overcome sparse-to-no information within this NHANES age group by using the quality-controlled blood pressure, medical history, and prescription medication data collected at Add Health Wave IV.7 Within-study analyses of these publicly available data are already yielding important insights into evolution of health disparities among young adults that NHANES is unable to provide. ACKNOWLEDGMENTS This research uses data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth).