“…Prior studies have utilized clinic 10,25 or otherwise restricted samples, 9,12 enrolled unusually hypertensive subgroups, 25,26 lacked normotensive controls, 27 and/or did not covary potential confounding factors such as antihypertensive treatments, 8 socioeconomic status, 10,12,27 body mass index (BMI), 26,28 and concurrent chronic disease. 10,25 Population-based studies with adequate controls were either small 11,13 or lacked independent blood pressure measurements, 8,15 and none have enrolled sufficient numbers of racial/ethnic subgroups to evaluate the consistency of labelling effects across non-Hispanic white, non-Hispanic black, and Hispanic subpopulations. Finally, in order to distinguish labelling effects from the effects of hypertension status it is necessary to examine large samples including hypertensive and normotensive individuals, both labelled and unlabelled (or mislabelled, as the case may be).…”