Predicting deterioration in dengue using a low cost wearable for continuous clinical monitoring

Abstract

Close vital signs monitoring is crucial for the clinical management of patients with dengue. We investigated performance of a non-invasive wearable utilising photoplethysmography (PPG), to provide real-time risk prediction in hospitalised individuals. We performed a prospective observational clinical study in Vietnam between January 2020 and October 2022: 153 patients were included in analyses, providing 1353 h of PPG data. Using a multi-modal transformer approach, 10-min PPG waveform segments and basic clinical data (age, sex, clinical features on admission) were used as features to continuously forecast clinical state 2 h ahead. Prediction of low-risk states (17,939/80,843; 22.1%), defined by NEWS2 and mSOFA < 6, was associated with an area under the precision-recall curve of 0.67 and an area under the receiver operator curve of 0.83. Implementation of such interventions could provide cost-effective triage and clinical care in dengue, offering opportunities for safe ambulatory patient management.

Fig. 1. CONSORT diagram of patient recruitment and flow.

The figure shows the outline for the recruitment process in which n = 250 participants were enroled, and n = 153 included in the analyses. The cohort was randomly split into a training (n = 132) and hold out test (n = 21) cohort.

Fig. 2. An overview of the model architecture and data in the study.

Patients enroled in the study underwent continuous PPG monitoring alongside vital signs measurements to derive a score through NEWS2/mSOFA systems. Each fixed window of PPG was fed into the model which consisted of a convolutional, recurrent and/or attention layers in order to generate predictions 2 h into the future. Auxiliary clinical data were inserted through embedding. A multi-task learning approach was used to improve performances.