MACHINE LEARNING IN HYPERTENSION DETECTION: A STUDY ON WORLD HYPERTENSION DAY DATA

Objective: Many modifiable and non-modifiable risk factors have been associated with hypertension. However, current screening programs are still failing in identifying individuals at higher of Given the major impact high blood pressure on cardiovascular events mortality, there is an urgent need to find new strategies improve hypertension detection. We aimed explore whether a machine learning (ML) algorithm can help individual predictors hypertension.Design method: analysed data set generated by questionnaires administered during World Hypertension Day from 2015 2019. Demographic information (age, sex, BMI), self-reported cardio- vascular (hypertension, diabetes, smoking, cholesterol, kidney disease, family history diseases), sleep complaints (snoring, witnessed apneas, daytime sleepiness) prior diseases (previous cardio cerebrovascular events, previous myocardial infarction) as well about awareness its health consequences were collected through questionnaire. computed performance (sensitivity specificity) standard programmes recommended Health Organisation five different ML algorithms, exploiting balancing techniques (oversampling undersampling). Results: A total 20206 (mean age 50.89 years, 51.65% females) included for analysis. Results show that gain sensitivity reflects loss specificity, bringing scenario where not configuration which properly outperforms against others (Table 1). Random Forest provides most interesting balance between specificity greater Area Under Curve compared Conclusions: Detection population level remains challenging approach could making more precise cost effective, when based accurate collection. More studies needed identify features be acquired further performances models. Table 1: Performance test three experimented schemes, original, oversampling undersampling training set, medical protocol evaluation.