5 emerging technologies in telemedicine

Telemedicine, or telehealth, refers to health care services provisioned via internet and other telecommunication technologies. Particularly right now, with movement restrictions and overburdened hospitals due to the coronavirus disease 2019 outbreak, interest in telehealth has surged. Here is some evidence of this surge in interest, a Google Trends chart showing worldwide searches for telemedicine and telehealth over the past five years:

According to Google Trends, search interest worldwide for "telemedicine" and "telehealth" has surged dramatically.
Google Trends for “telemedicine” (blue) and “telehealth” (red) worldwide searches over the past five years.

In principle, of course, telehealth is not really new. And we probably all know that currently emerging technologies such as fast connectivity (5G, for example), wearable devices, or advanced visualization technologies such as augmented reality or virtual reality are important enablers of telehealth.

But instead of looking at these technologies, I wanted to discover other emerging technologies that might be less obvious but important nevertheless as enablers of telehealth. Please note though that my article here is not intended to be comprehensive by any means. Also, I am no telehealth expert but just a curious layperson in this context. In this article, I’d just like to show you some examples of what I found in ca. 30 minutes’ time.

Here is the breakdown of what I found:

  1. Paper-based microfluidics for immunoassays
  2. A 3D printed smartphone adaptor for nasolaryngoscopy
  3. Edge computing for real-time patient monitoring
  4. Machine learning for vital signs monitoring of home-isolated Covid 19 patients
  5. Machine learning for wheezing detection

But first, let me explain to you how I conducted my search.

How I searched

I used Mergeflow for my search, focusing on the Scientific Publications and Clinical Trials data sets. The Scientific Publications data set includes journals, conference proceedings, etc.. It also includes PubMed, which is one of the, if not the, largest life science and health care data sources in the world.

The Clinical Trials data set includes publications from clinicaltrials.gov. These data are provided by the U.S. National Library of Medicine at the NIH. We also collect trials from the EU Clinical Trials Register, which is maintained by the European Medicines Agency. We have used and explored the Clinical Trials data set previously, in another article.

How can you discover emerging technologies in these data sets?

At Mergeflow we have built an algorithm that matches all contents collected by Mergeflow against semantic models of emerging technologies. So far, we have such semantic models for ca. 200 emerging technologies, and we keep building more of them.

See the list of ’emerging technologies’ semantic models available in Mergeflow

In its user interface, Mergeflow then displays emerging technologies as tag clouds, or in graphs. Here is an emerging technologies tag cloud for my “telemedicine OR telehealth” search in the Scientific Publications data set, for example, once I excluded the “big known” ones like “wearables”, “augmented reality”, or “virtual reality”:

"Emerging technologies" tag cloud in Mergeflow for telehealth and telemedicine documents from the Scientific Publications data set.
“Emerging technologies” tag cloud in Mergeflow for the Scientific Publications data set.

I focused on data from the past year, and used Mergeflow’s emerging technology tags to zoom in on five example findings.

1. Paper-based microfluidics for immunoassays

Schematic diagram of Fu et al's microfluidic paper-based analytical device.
Schematic diagram of Fu et al’s microfluidic paper-based analytical device.

What it is

Paper-based microfluidics is a technology that moves tiny amounts of liquids across cellulose fibers that guide the liquid across diagnostic assays. These assays can then determine e.g. the presence of inflammatory agents. Using this method, a group led by Hao Fu from the Microfluidics and BioMEMS Lab at the University of Toronto has developed a new diagnostic tool that performs at very similar levels as much more complex and expensive conventional diagnostics methods.

Why this is important

Work such as Fu et al’s can help “democratize” otherwise much more expensive and complex medical diagnostics. In other words, besides bringing direct costs down, such methods may not even require a trained health care professional.

Reference details

A paper-based microfluidic platform with shape-memory-polymer-actuated fluid valves for automated multi-step immunoassays.

2. A 3D printed smartphone adaptor for nasolaryngoscopy

Thompson et al's 3d printed smartphone adaptor for a nasolaryngoscope.
Thompson et al’s 3d printed smartphone adaptor for a nasolaryngoscope.

What it is

Remember the last time one of your appliances, e.g. your washing machine, broke? And how the thing that broke was this tiny little plastic thing that costs a few cents to make but that the manufacturer of your appliance charged you EUR 80 for? Yeah. Apparently very similar problems exist with medical devices as well.

A group led by Daniel Thompson at St. Vincent’s Hospital in Melbourne has now developed a 3D printed adaptor so you can use a smartphone to record images from a nasolaryngoscope. A nasolaryngoscope is a device that doctors use for exams of your nose and your throat. This new device then costs $29 AUD, rather than $160 AUD (the cost of the next cheapest commercial alternative product).

Why it is important

Since telemedicine is often particularly relevant to less-well-off communities, bringing down the price of medical devices is really important. And in order to bring down prices, you also have to “sweat the small stuff”, like those relatively simple adaptors.

Reference details

A 3D printed smartphone adaptor for nasolaryngoscopy.

3. Edge computing for real-time patient monitoring

Tele-ICU (Intensive Care Unit) monitoring
Tele-ICU (Intensive Care Unit) monitoring

What it is

Edge computing, or fog computing, is a technology where computation is performed as close as possible to the point where it is needed, rather than on centralized servers. This is particularly important for real-time applications because it essentially removes the response times that a centralized client-server model would incur.

Real-time (remote) patient monitoring has special requirements because particularly in life-critical situations the monitoring data can be bursty (= not a lot of data usually but then a sudden burst of data once the patient’s condition changes). The problem here is that one needs real-time capacity to deal with data bursts while at the same time preserving communication bandwidth requirements. In order to address these conflicting requirements, Sangeetha Pushpan and Bhanumathi Velusamy from Anna University in Coimbatore have developed a new time slot allocation algorithm.

Why it is important

One cannot simply count on 5G or other high bandwidth technologies to be available everywhere. If one wants to do telemedicine in low-bandwidth environments, approaches like the one described here could help.

Reference details

Fuzzy-based dynamic time slot allocation for wireless body area networks.

4. Machine learning for vital signs monitoring of home-isolated Covid 19 patients

An undershirt for vital signs monitoring, tested by the US Air Force
An undershirt for vital signs monitoring, tested by the US Air Force

What it is

In a clinical study, a group led by Lars Wik at Oslo University Hospital is investigating how to use biosensor data from home-isolated Covid 19 patients to detect changes in patients’ conditions that indicate they should be hospitalized. The idea is to use machine learning for detecting these patterns from continuously recorded physiological data.

Why it is important

In order to preserve hospital capacity, it is important not to hospitalize patients unnecessarily. At the same time, of course, it is important to detect signs of deterioration in home-isolated patients so that they can be hospitalized soon enough. Methods such as the ones investigated by Lars Wik and colleagues could provide valuable support during pandemics.

Reference details

Sensor based vital signs monitoring of Covid 19 patients during home isolation.

5. Machine learning for wheezing detection

Asthma affects lungs and airways.
Asthma affects lungs and airways.

What it is

Wheezing is a symptom common in asthma and other chronic obstructive pulmonary diseases. Since wheezing can indicate potentially very serious conditions and can occur at any time, continuous monitoring is important. Obviously though this is only possible at larger scale if it does not require the attendance of health care personnel. Therefore, Hai Chen and colleagues from the Macau University of Science and Technology have developed a machine learning classifier that can detect wheezing from respiratory sounds.

Why it is important

Chronic diseases require continuous monitoring. But currently this is not possible in many situations because it is too resource intensive. Hai Chen et al’s study is an example of a very broad medical use case, which is continuous and at the same time low-resource patient condition monitoring. Other candidate applications in telemedicine could be diabetes, or chronic cardiovascular conditions.

Reference details

Automatic multi-level in-exhale segmentation and enhanced generalized S-transform for wheezing detection.


Featured article image: A remote pre-op cardiology consultation, https://commons.wikimedia.org/wiki/File:Telemedicine_Consult.jpg

Leave a Reply