PP.08.16] AUTOMATICALLY CLASSIFYING ESSENTIAL ARTERIAL HYPERTENSION FROM PHYSIOLOGICAL AND DAILY LIFE STRESS RESPONSES

Abstract

Objective: Essential arterial hypertension (EAH) might be correlated with an abnormal response to stressful life events. For understanding the psychological dimensions associated with EAH, we investigate individual reactions to stressful events in everyday life. We aim to assess the combined influence of individual perception of stress and physiological signals by continuously monitoring patients with EAH. Design and method: Two groups of subjects, balanced by gender, were compared: twelve normotensive, and twelve diagnosed with EAH grade I or II controlled by therapy, without organ damage (average age 49,2, average BMI 25,43 Kg/m2, mean office arterial pressure 127,57/83,57, average heart rate 73,5, average number of drugs 1,5). Patients performed ambulatory blood pressure monitoring 24 hours (ABMP) to rule out white coat effect. Controls were checked to rule out hidden hypertension. Blood Volume Pulse (BVP), Heart Rate Value (HRV), Galvanic Skin Response (GSR) and skin temperature were continuously recorded by Empatica E3 device. An iPhone application capable of storing and streaming physiological signals was provided; it elicited annotations from subjects to report their psychological state, perceived and reported stress and workload, and daily activities. We used behavioural analytics algorithms to analyse and classify the recorded data. Results: Two dimensions of emotional response were identified - emotional suppression, and anxious anticipation of stressful events. Subjects with high and low scores in emotional inhibition are equally represented in the two groups, but subjects with high scores on anticipatory anxiety were more represented in the EAH category. The EAH subjects recorded higher percentage of both anticipated and reported stress. Physiological signal streams showed that EAH subjects significantly differ from controls. HRV features showed significant differences between normotensive and hypertensive patient groups. Combining HRV features with other physiological signals such as GSR and BVP provided a high accuracy in distinguishing between the two groups. Conclusions: Hypertension category tends to overestimate stressful events, and are incline towards anticipatory anxiety. They also exhibit different physiological responses to stress when compared to normotensive subjects. This pattern is relevant to the postulated links between hypertension and patterns of emotional response.