Using Wearables to Examine Effects of Long Covid

Feasibility Study on Using Wearables to Examine Effects of Long Covid

‍

Drs. Sandra Webber and Diana Sanchez-Ramirez and their team at the University of Manitoba ran a small pilot study examining the feasibility of using smartwatches to study the effects of Long Covid. They ran an eight-week intervention pilot where participants with long covid were assigned to one of two programs (group Zoom classes or pre-recorded individual exercises sessions). Physiological data was collected between weeks three and five.

Having adopted ActiGraph accelerometers to measure activity in previous studies, Sandra was also interested in using this pilot as somewhat of an unofficial validation study comparing the results from the Garmin Fenix 5 watches to those from the ActiGraph GT3X+. 

‍

Can you give us a bit of background on your recent pilot study?‍

We were interested in collecting physiological data looking at people with symptoms of Long Covid.  In this pilot study, participants were given initial assessments where we collected baseline data, mostly respiratory parameters and vital sign responses, physical function types of tests, and some questionnaires about their symptoms, and then they were given different rehabilitation programs to complete at home. We didn't try to look at differences in the smartwatch/accelerometer data over time, we really just wanted to collect that data in one shot because it was more about the feasibility of using the devices and seeing what we could get with them.

Were you using any other devices to track physiological data?

Yes, we used the Garmin Fenix 5 as well as the ActiGraph GT3X+ accelerometer. Participants were asked to wear the Garmin 24 hours a day, and then the accelerometer they were supposed to wear for waking hours. And they were to wear the devices over the course of 8 days, with both of them on at the same time.

Having worked with ActiGraph devices in the past, I was interested to do a little bit of unofficial validation between the two devices in terms of the step count. Although it wasn't a formal validation study, I am encouraged by the results. We are planning to move on to a more formal validation study this fall, and this study provided good preliminary information.
‍

‍

What are some insights gained so far?‍

One of the metrics we looked closely at was heart rate, which we could collect from the Garmin Fenix 5 at minute intervals with Labfront. Our participants were in their 30’s, 40’s, 50’s and 60’s,  and they all were experiencing Long Covid symptoms three months past the time that they were infected. We looked at their mean heart rates over each 24-hour period, and they were rather high in about four of six people, which was very interesting.

Because we didn’t know exactly what time the participants really went to bed, we estimated sleep time to be from midnight to 6am. And then we called 10am to 10pm ‘active time’ and looked at heart rates over those two time periods. Heart rates were generally high during sleep periods. So again, we're thinking this is interesting and may be something to measure in people with COVID-19. That preliminary data suggests that those with Long Covid are not really getting back to their normal resting state. And that may be driving their mean heart rates over the 24 hours to be kind of high too, because they don't really go down very low in their sleep.

I am encouraged by the results. We are planning to move on to a more formal validation study this fall, and this study provided good preliminary information.

How did the Garmin perform compared to the ActiGraph GT3X+? ‍

Step counts were, as I expected, higher with the Garmin Fenix 5. But I was somewhat encouraged to find that they were only about 20% higher, and they were consistently 20% higher for the six people. That was nice to see- it gives me a lot more confidence in the stability of the Garmin device data for future studies.
‍

Having participants wear smartwatches can be a new challenge for some. How was your adherence during the week of physiological data collection?

‍One of the things we looked at when examining for feasibility was how many minutes participants wore the watch and how many minutes they took it off each day, so we would know what to expect for future studies. Non-wear time ranged from eight minutes to 36 minutes per day. The rest of the day, participants  were very good about wearing the watches and were very adherent. We were pretty impressed and happy to see those numbers, considering the Fenix 5 watches are quite big, and most likely not the most comfortable to sleep in.

Did you gift participants the devices, as an incentive? And if so, could that have factored into the adherence?‍

Those Fenix watches are around $900 a watch, so no we did not. The participants were just interested in wearing the watches. They wanted to see their own data, because they could check on their Garmin Connect app that the data  was being collected. I think they were quite keen to learn more about it.‍

For all six participants, we got data over the time period that we wanted, with participants only removing the watch for 8 to 36 minutes a day. We were impressed with those numbers.

Were there any challenges that you faced working with the devices? 

There weren’t too many surprises, since I’ve been working with watches and accelerometers for quite a while. But we did write very detailed protocols for our research staff. We also used the documentation available in the Labfront Help Center to assist with participant protocols and troubleshooting, which was very helpful.

I was actually pleasantly surprised that we didn't have a lot of difficulties. For all six participants, we got data over the time period that we wanted. I would attribute most of that to the detailed protocols that we wrote,  and also to the research assistant that we had. She made sure she was very familiar with the watches and with the apps before she visited with participants to give them the equipment.  She ran through everything thoroughly with individual participants and was very prepared to assist with any technical issues if they arose.

Any tips to share for future researchers using wearable devices?‍

Well, one thing we did do is collect an additional day of data for feedback reasons. So while on paper we analyzed the data over seven days, we actually measured for eight days, and we dropped the first day. It’s a common thing that researchers do with accelerometers, and so we did it with the Garmins as well. The reason is that people often change their behavior when they know they're being monitored; for example, they become more active. But most of the research has shown that people don't keep that up for more than a day or two. So if you collect data for eight days and  drop the first day, you're probably getting data that is pretty representative of what their normal week would be.
‍

Were you inspired by the results of this pilot to run this study on a larger scale in the future? 

Yes! So we do actually have funding now for running another study, which will be a larger pilot. In our next pilot, we're actually planning to use the Garmin Venu SQ instead of the Fenix 5, and also test the Hexoskin, which is like a vest that people can wear to collect respiratory data with monitors around the chest wall (as well as heart rate and physical activity parameters) 

Unlike the wearability of the Garmins, getting people to wear a vest for more than an hour or two at a time might be a bit of a challenge, but we’re optimistic! And we’ll still have the Garmin Venu SQ data to support our results.