Establishment of predictive model for blood pressure elevation in outpatients with hypertension
10.3969/j.issn.1008-9691.2019.04.012
- VernacularTitle:门诊高血压患者血压升高预测模型的建立
- Author:
Li WANG
1
;
Junyan GUO
Author Information
1. 西藏民族大学附属医院内二科
- Keywords:
Pulse wave transit time;
Wrist smart watch hematomanometer;
Blood pressure prediction
- From:
Chinese Journal of Integrated Traditional and Western Medicine in Intensive and Critical Care
2019;26(4):426-428
- CountryChina
- Language:Chinese
-
Abstract:
Objective To establish the regression equation of blood pressure in population based on the pulse wave transit time (PWTT) and verify its accuracy. Methods A convenient sampling method was used to collect gender, age, arm circumference, arm length, PWTT, history of hypertension, and body height, etc. of 4 121 outpatients' information from the Hainan Branch of Chinese PLA General Hospital from June 2016 to May 2018 to establish a binary variable logistic regression equation for blood pressure elevation or not and screen out the influencing factors of time point blood pressure. The accuracy of the equation was then verified in 252 outpatients. Results Logistic regression analysis showed that PWTT, gender, age, history of hypertension present or not, body height might be the influencing factors of blood pressure elevation [odds ratio (OR) values were 0.995, 0.530, 0.980, 107.128, 0.979, 95% confidence interval (95% CI) were 0.991-0.999, 0.405-0.695, 0.971-0.989, 73.935-155.223, and 0.962-0.996, respectively, all P < 0.05]. The classification prediction equation: Ln [P / (1-P)] = 2.087-0.005×PWTT-0.635×gender (man = 1, woman = 2)-0.02×age + 4.674×hypertension history present or not-0.021×body height [P indicates the probability of a positive result: systolic blood pressure≥140 mmHg (1 mmHg = 0.133 kPa) and/or diastolic blood pressure≥90 mmHg]. The overall test results showed that χ2 = 1 835.305, P < 0.05, statistically significant; The goodness of fit test results:χ2 = 5.881, P > 0.05, the data was conform to the equal proportion distribution. Model-confirmed results showed that in patients with a history of hypertension, the probability of predicting accuracy was 95%-99%. In patients without a history of hypertension, the probability of predicting accuracy was 45%-89%. Conclusion The predictive value of blood pressure elevation with this model in patients with a history of hypertension is superior to that of patients without the history of hypertension.
