ORIGINAL ARTICLE Estimating the proportion of young adults on antihypertensive treatment that have been correctly diagnosed T Marshall, R Tennant and WN Harrison Department of Public Health & Epidemiology, University of Birmingham, Edgbaston, Birmingham, UK This paper aims to identify how many young adults on antihypertensive treatment have been misclassified as hypertensive. We identified subjects aged under 35 on antihypertensive treatment, from the Health Surveys for England, 1998–2004. Pretreatment systolic and diastolic blood pressures were calculated by adjusting on-treat- ment blood pressures for the effects of treatment. Treat- ment effects were derived from meta-analysis. Subjects were classified as hypertensive if pretreatment blood pressure was X160/100 mm Hg, or was X140/90 mm Hg in conjunction with high cardiovascular risk. We then identified the proportion of treated subjects on antihyper- tensive treatment who were truly eligible for treatment. From the survey data we identified 65 adults (25 men and 40 women) under 35 on diuretics, b-blockers, angiotensin converting enzyme inhibitors, calcium blockers or other antihypertensives. Average pretreatment blood pressure was 164/100 mm Hg in those eligible for treatment, and 136/79 mm Hg in those not eligible. The analysis indicated that 29.2% of adults aged 16–34 (95% con- fidence interval (CI): 18.6–41.8%) were truly eligible for antihypertensive treatment: 32.0% (95% CI: 14.9–53.5%) of men and 25.0% (95% CI: 12.7–41.2%) of women. A total of 73.7% (14 of 19) of subjects eligible and 41.3% (19 of 46) of subjects not eligible for treatment either had a body mass index430 kg m 2 or kidney disease (v 2 -test P ¼ 0.018). Because of biological variation in blood pressure, most young adults on treatment for hyperten- sion have been misclassified as hypertensive. Most who have been correctly diagnosed are either clinically obese or have kidney disease. Journal of Human Hypertension (2008) 22, 96–101; doi:10.1038/sj.jhh.1002291; published online 13 September 2007 Keywords: blood pressure; measurement error; diagnostic tests Introduction When blood pressure measurement is carried out in healthy adults, it is a type of diagnostic test. Individuals are classified either as hypertensive (eligible for antihypertensive treatment) or normo- tensive. Because of biological variation in blood pressure, like any diagnostic test, blood pressure measurement is subject to misclassification. Those diagnosed as hypertensive therefore include hyper- tensive individuals correctly identified as hyperten- sive (true positives), and normotensive individuals who have been misclassified (false positives). If the prevalence of hypertension in a population is low, false positives may be more common than true positives. Modelling suggests that this may be the case in adults under 35. If a diagnosis of hypertension is based on the average of three blood pressure measurements in adults aged 16–35, we would expect 31% (95% confidence interval (CI) 16–46%) of men on antihypertensives and 26% (95% CI 7–46%) of women on antihypertensives to be true hypertensives. 1 This means that adults under 35 who have been identified as hypertensive, are more likely to be false positives than to be truly hypertensive. This paper investigates the hypothesis that many young adults on antihypertensive treat- ment have been misclassified. It does so by identify- ing a population of young adults on antihypertensive treatment, estimating their likely pretreatment blood pressure and determining the proportion that would have been hypertensive pretreatment. Materials and methods Adults under 35 on antihypertensive treatment were identified from the Health Surveys for England conducted in the seven years from 1998 to 2004. 2 In the surveys, blood pressure was measured in the right arm with the patient seated and at rest for five minutes. Three measurements were taken using an automated blood pressure monitor. In 2003, the Received 8 May 2007; revised 6 August 2007; accepted 8 August 2007; published online 13 September 2007 Correspondence: Dr T Marshall, Department of Public Health & Epidemiology, University of Birmingham, Edgbaston, Birmingham, West Midlands B15 2TT, UK. E-mail: T.P.Marshall@bham.ac.uk Contribution: Tom Marshall had the original idea. Ruth Tennant, Wayne N Harrison and Tom Marshall obtained the data, carried out the analysis and wrote the paper jointly. Journal of Human Hypertension (2008) 22, 96–101 & 2008 Nature Publishing Group All rights reserved 0950-9240/08 $30.00 www.nature.com/jhh