Background

Characterising and caring for patients with life-threatening new and emerging infectious diseases is a global priority. New technologies – wearable devices, cloud computing, and AI/machine learning – have the potential to transform how we characterise and care for these diseases. Ho Chi Minh City is suffering rapidly rising numbers of COVID-19 cases (>2000/day), which is placing enormous strain on its health care system.

The primary challenge is the rapid identification and optimal management of those with severe disease and low blood oxygen (hypoxia). There is a severe shortage of basic monitoring equipment outside of intensive care units (ICUs), which is causing fatal delays in the identification and treatment of patients needing oxygen and ICU care.

A Vietnamese tech company, iPARAMED, has developed a remote monitoring platform (see figure 1). It enables real-time monitoring of oxygen saturation (SpO2), respiratory rate, temperature, and ECG, using affordable, battery-operated, FDA-approved, wearable devices. Data are monitored centrally, enabling clinical staff to identify those with clinical warning signs and hypoxia rapidly. There is an urgent need to develop, evaluate and, if successful, implement such a system in hospital wards in Viet Nam, which can be scalable to monitor many thousands of patients safely and continuously. Such a system could rapidly be extended to multiple units caring for severely ill and critical care patients in LMICs in SE Asia, Africa, and elsewhere.

Objectives

We propose a 6-month pilot project, starting immediately, with the following:

• iPARAMED will immediately import the wearable devices (n=300) to Ho Chi Minh city;
• iPARAMED technical team will rapidly deploy the platform and customise the process within the Hospital for Tropical Disease, the major Government COVID-19 treatment centre in the city. NB. While the programme will use the vivalink devices, the iPARAMED platform will be configured to be agnostic to the device manufacturer to ensure low-risk scalability in future;
• Data capture from around 150 patients with COVID-19 at any one time across 3 (non-ICU) wards at the hospital over the next six months;
• The platform will capture and store continuous, high-rate physiological data from all patients;
• Clinical and biological parameters (routine blood tests, plus storage of serial serum and nasopharyngeal throat swab specimens) will also be collected through standard ISARIC protocols;
• These data and specimens will be used to investigate whether machine learning can characterise heterogeneities in disease severity and to develop prototype clinical decision support systems and biomarkers that enable the early optimal identification of those patients who require further respiratory support and mechanical ventilation;
• Link with and extend existing tools for data capture (ISARIC/pandemic sciences centre, Oxford) and develop technology-based tools for disease characterisation and management for use in pandemics.

The project objective will be to provide proof of principle that a team within a low- and middle-income setting can set up a unique remote monitoring platform during a pandemic that will immediately aid clinical care and capture data for innovative research methodologies that improve clinical outcomes.

The latter will demonstrate the development of clinical decision support systems by AI/machine learning within such settings at scale. If shown to be effective, the platform will be scaled to the rest of the city and potentially the whole country.