Decision-Tree Analysis to Predict Blood Pressure Control Status Among Hypertension Patients Taking Antihypertensive Medications

약물복용 중인 고혈압 환자의 혈압관리양상 예측을 위한 의사결정나무분석

Purpose: This study was performed to analyze the levels of blood pressure and to identify good or poor blood pressure control (BPC) groups among hypertension patients. The study was based on the Korea National Health and Nutrition Examination Survey (KNHANES VI and VII) conducted from 2013 to 2016. Methods: The sociodemographic and clinical data of 4,151 Korean hypertension patients aged 20-79 years and who were taking antihypertensive medications was extracted from the KNHANES VI and VII database. Descriptive statistics for complex samples and a decision-tree analysis were performed using the SPSS WIN 24.0 program. Results: The mean age was $62.46{\pm}0.21years$. The mean systolic blood pressure (SBP) was $128.07{\pm}0.28mmHg$, and the diastolic blood pressure (DBP) was $76.99{\pm}0.21mmHg$. 71.9% of participants showed normal blood pressure (SBP < 140mmHg and DBP < 90mmHg). From the decisiontrees analysis, the characteristics of participants related to good BPC group were presented with 9 different pathways same as those from the poor BPC group. Good or poor BPC groups were classified according to the patients' characteristics such as age, living status, occupation, education, hypertension diagnosis period, numbers of comorbidity, perceived health status, total cholesterol, high density lipoprotein-cholesterol, alcohol drinking per month, and depressive mood. Total cholesterol level (< 201mg/dL or ${\geq}201mg/dL$ cutoff point) was the most significant predictor of the participants' BPC group. Conclusion: This decision-tree model with the 18 different pathways can form a basis for the screening of hypertension patients with good or poor BPC in either clinical or community settings.