12/06/21
Conventional intensive monitoring of ICU patients’ vital signs requires high-quality monitors and well-trained staff, both of which are unaffordable and unavailable in limited-resource settings, particularly during health emergencies. Low-cost wearable physiological monitors, such as fitness trackers, offer an affordable solution to monitor physiological data in limited-resource settings.
This project considered both wearable IoT and the connected infrastructure to produce data-driven vital sign monitoring.
Our project’s primary objective was to provide proof-of-principle that a low-cost health-monitoring using low-cost wearable sensors, recognising the infrastructure they are connected to, can be used to develop predictive monitoring systems for the early detection of patient deterioration in limited-resource settings.
We worked in close collaboration with our clinical collaborators at the Oxford University Clinical Research Unit (OUCRU) in Ho Chi Min City, Vietnam, to assess and review wearable devices practical aspects.
This project attempts to demystify crucial barriers to the application of IoT in healthcare settings in limited-resource environments. This became even more important during the recent COVID-19 pandemic when most healthcare systems worldwide were under immense pressure as demand rapidly outstripped supply. The need for remote monitoring is more important than ever as hospitals are learning to cope with a new virus whilst providing ongoing care.
Low-cost wearable devices and the infrastructure supporting these play an essential role in establishing the vital sign data streams for predictive health-monitoring systems. Such systems can allow in-home monitoring that will alert a deteriorating patient to visit the hospital and keep hospital resources available for emergency and ongoing critical care.
This project developed a detailed overview of the rapidly expanding wearable device market landscape. In addition to previously completed desk research about different devices, we interviewed relevant suppliers to assess the collaborative nature of possible future relationships.
We collaborated with policy researchers to assess the upstream policy changes that suppliers implement, which often impact the downstream research. We conducted a literature review of previous related work using wearables in limited-resource settings in low- and middle-income countries. A carefully designed evaluation criteria followed this to assess the usability of devices and interactions with legacy IoT infrastructure.
This was done in collaboration with our clinical partners at Oxford University Clinical Research Unit in Ho Chi Min City, Vietnam.
Through an innovative approach that involved early collaborations with suppliers, this project was able to identify key challenges in the use of wearable to develop IoT solutions for frugal health-monitoring in limited-resource settings. The outcome of the project includes preparing an updated literature review (in draft) and a summary recommendation.
The evaluation criteria developed is intended as a valuable tool to enable researchers to fast-track the selection process and avoiding costly mistakes. Most notably is the importance of supplier agreements that avoid future hardware or firmware changes that may impact ongoing research projects.
As part of the investigation, the Oxford University Clinical Research Unit’s clinical team has identified many advances in the local manufacturing processes of wearable devices and generated recommendations for potential future collaborations.
The outcome of this project highlights an important imbalance in the current supplier and researcher relationship. We hope that the evaluation criteria developed in this work will function as an essential tool for future researchers who wish to consider the use of wearables in their IoT-related work.
Moreover, establishing these agreements earlier in the relationships will save valuable time and resources for ongoing research related to IoT in healthcare by safeguarding them against unintended changes that could impede ongoing data streams that function as a key input to remote health-monitoring systems.
Our immediate dissemination activities include publications of both the literature review (currently in draft) and the proposed evaluation criteria tool.
As part of our ongoing research to use wearable vital sign data from low- and middle-income countries, the CHI Lab will host two global health researchers to focus on using machine learning techniques for this unique data stream.
The immense pressure on healthcare systems worldwide due to the ongoing pandemic greatly enhanced the overall awareness of the importance of digital health innovations. One of the successful outcomes of this project was working closely with clinicians in Vietnam and getting continuous feedback about the practical limitations of the wearables. This is facilitating an important conversation about the future applications of IoT in such healthcare settings.
The ongoing pandemic caused some delays with suppliers, and moving devices around from suppliers to Oxford and then on to Vietnam proved more challenging during this time. In hindsight, it would have been faster for the suppliers to dispatch the devices directly to Vietnam.
Overall limited support (both clinical and administrative) due to the pandemic made progressing work within the healthcare sector a challenge. Compounded with restricted support from the university (due to lockdown working conditions), some of the processes were delayed in rolling out.
This Pitch-In project provided our partners and us with a wonderful opportunity to take a holistic view of the wearables landscape and evaluate the devices without the immediate pressure of data collection. The finding from this would have a long-term impact on supporting ongoing wearables research to avoid costly decisions in future.
Professor David Clifton – University of Oxford
Heloise Greeff – University of Oxford