Bayesian analysis supports use of ambulatory blood pressure monitors for screening.
Bayesian analysis supports use of ambulatory blood pressure monitors for screening.
复制标题
贝叶斯分析支持使用动态血压监测仪进行筛查。
DOI：
10.1161/01.hyp.19.2_suppl.ii268
复制
发表时间：
1992
期刊：
影响因子：
8.3
通讯作者：
C. Laffer
中科院分区：
文献类型：
--
作者：
F. Elijovich;C. Laffer
为了评估动态血压监测（ABPM）在高血压筛查中是否有作用，我们对具有不同高血压先验概率（PP）的人群进行了办公室血压（OBP）的贝叶斯分析，作为诊断“测试”。如果收缩压大于140 mm Hg或舒张压大于90 mm Hg，则OBP被认为是阳性测试。高血压诊断的"金标准"选择的日间ABPM截止值为收缩压139 mmHg和舒张压88 mmHg。对72例确诊为高血压（PP = 1）的患者进行了OBP敏感性和特异性测定。停药3周后，OBP为168 +/-3/101 +/-1，ABPM为151 +/-2/94 +/-1 mm Hg。17例患者的收缩期ABPM在正常血压范围内，15例患者的舒张期ABPM在正常血压范围内。OBP假阳性分别为14例和15例。因此，OBP的敏感性和特异性分别为0.8909和0.1765（收缩期）和0.9825和0（舒张期）。使用贝叶斯定理，这些数据不能外推到PP较低的人群。因此，我们根据已发表的血压正常受试者的ABPM系列计算PP = 0的灵敏度和特异性，并将我们的测量结果和这些计算结果用于PP 0.1 - 0.9人群的算术插值。利用贝叶斯定理揭示了PP与OBP阳性预测高血压的S型关系。他们的最佳拟合三次多项式预测，在一般人群中，OBP升高将误诊高血压的概率为46 - 50%（PP = 0.2），但在专业实践中仅为8 - 9%（PP = 0.9）。（250字处删节）
To assess whether there is a role for ambulatory blood pressure monitoring (ABPM) in screening for hypertension, we conducted Bayesian analysis of office blood pressure (OBP) as a diagnostic "test" in populations with different prior probabilities (PP) of hypertension. OBP was considered a positive test if systolic blood pressure was greater than 140 mm Hg or diastolic pressure was greater than 90 mm Hg. Chosen daytime ABPM cutoffs for a "gold standard" diagnosis of hypertension were systolic pressure of 139 and diastolic pressure of 88 mm Hg. Sensitivity and specificity of OBP were determined in 72 patients with established hypertension (PP = 1). After 3 weeks off therapy, OBP was 168 +/- 3/101 +/- 1 and ABPM was 151 +/- 2/94 +/- 1 mm Hg. Systolic ABPM was in the normotensive range in 17 patients and diastolic in 15 patients. OBP was falsely positive in 14 and 15 of these patients, respectively. Thus, sensitivity and specificity of OBP were 0.8909 and 0.1765 (systolic) and 0.9825 and 0 (diastolic). These data cannot be extrapolated to populations with lower PPs for use of Bayes' theorem. Hence, we calculated sensitivity and specificity for PP = 0 from published series of ABPM in normotensive subjects and used our measurements and these calculations in arithmetic interpolations for populations with PP 0.1-0.9. Sigmoid relations between PP and predictability of hypertension by a positive OBP were disclosed by use of Bayes' theorem. Their best-fit cubic polynomials predict that an elevated OBP will misdiagnose hypertension 46-50% of the time in a general population (PP = 0.2) but only 8-9% in a specialty practice (PP = 0.9).(ABSTRACT TRUNCATED AT 250 WORDS)
DOI：
10.1001/jama.1988.03720020027031
发表时间：
1988-01
期刊：
JAMA
影响因子：
--
作者：
T. Pickering;G. James;Charlene Boddie;G. Harshfield;S. Blank;J. Laragh
通讯作者：
T. Pickering;G. James;Charlene Boddie;G. Harshfield;S. Blank;J. Laragh