A low-cost, photoplethysmography-based sensor module capable of predicting complications from dengue

Lead Research Organisation: Imperial College London
Department Name: Electrical and Electronic Engineering

Abstract

This project is aimed at tackling one of the most widespread infectious disease that plagues tropical countries called dengue. In most cases, dengue progress is like flu and only manifests itself in the form of high fever. However, a small amount of population can develop the dangerous dengue shock syndrome (DSS) after the fever has stopped, that can lead to complication and even be fatal.

Due to the seasonal nature of dengue, hospitals in the affected regions get overwhelmed because it is difficult to predict which cases can progress into the DSS. The proposed solution is a low-cost wearable device that could monitor the patients after their discharge from hospital and alarm the doctors if there is a danger of DSS. DSS manifests itself mostly as internal bleeding and can be detected from various indicators. The project objective is to apply a standard non-invasive method of Photoplethysmography (PPG) to monitor patients' levels of concentration of red blood cells in blood, also known as haematocrit, and blood pressure.

The basic PPG system consists of a light emitting diode (LED) and a phototransistor that produces an output current proportional to the intensity of incident light. When light is shone into human tissue, it penetrates the various layers. Provided the subject is at rest, the only changing aspect is the blood volume in arteries due to the heart forcing the blood into them with each contraction. A higher volume of blood during heart contraction means more light is absorbed and less gets to the phototransistor. This creates the standard shaped PPG waveform with a clear indication of heartbeat frequency.

Before creating the final wearable device, a novel method using LEDs with different wavelengths and appropriate software post-processing needs to be developed and applied as haematocrit and blood pressure are not commercially sensed by PPG.

There are, however, studies developing prototypes in a laboratory setting showing the potential of PPG application in both sensor areas. For continuous haematocrit monitoring, a difference in the light absorption at a specific wavelength of the red blood cells is exploited. For this, two light sources with carefully selected wavelengths are shone into the tissue alternating at very high speed and their ratio is calculated. As the level of haematocrit change, the ratio between the wavelengths will be different, pointing to change in overall haematocrit. For blood pressure sensing a more complex approach based on extracting features from the PPG waveform and subsequent machine learning techniques will be used.

The two monitoring systems will be developed in parallel, and once sufficient precision is established in the laboratory setting, further testing with actual patients will be conducted. The project will start with thorough research of existing prototypes for each sensing technology and continue with the development of the first demo device capable of appropriate signal acquisition. After initial lab testing and performance evaluation, the device will be taken to a hospital environment where it will undergo extensive verification with patients of various demographics.

With data acquired, more complex analysis and algorithm tuning will be performed. At this stage, a miniaturization of the demo platform will also be explored, addressing additional challenges about power consumption and patient comfort while the device is operating. The overall aim is to develop a novel, miniature and low-cost PPG sensor combining the haematocrit and blood pressure sensing. In addition, appropriate network communication needs to be designed to alert experts if a high likelihood of DSS is detected.

This project can be categorized as the development of novel healthcare sensor for patients and it aligns with the Clinical technologies (excluding imaging) EPSRC research area.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509486/1 01/10/2016 30/09/2021
2127831 Studentship EP/N509486/1 01/10/2018 31/03/2022 Stefan Karolcik
EP/R513052/1 01/10/2018 30/09/2023
2127831 Studentship EP/R513052/1 01/10/2018 31/03/2022 Stefan Karolcik