Scripps Research and Stanford Medicine are collaborating with wearables vendor Fitbit to see if the data its devices collect can help detect, track and manage infectious diseases such as COVID-19.
The Stanford Healthcare Innovation Lab has launched a study to see whether data collected from wearables, including Fitbit devices, can be leveraged to predict the onset of infectious diseases such as COVID-19 before actual symptoms being to appear. The study relies on data such as heart rate, skin temperature and blood oxygen saturation.
The announcement follows on work by Scripps Research Translational Institute looking at how wearables might help manage flu epidemics. Its research suggests that wearables data can be used to let health officials know about emerging flu outbreaks in real time.
The Scripps study, which appears in The Lancet Digital Health, used de-identified data from Fitbit devices equipped with sleep- and heart-rate tracking capabilities. Scientists were able to demonstrate that using this data, they could improve predictions regarding influenza-like illness at the state level relative to CDC data.
To conduct the study, the Scripps team analyzed de-identified data from Fitbit devices covering two years from 200,000 people in the U.S. Between March 2016 and March 2018, these users generated 13.3 million valid data points.
When researchers added aggregated Fitbit data to the mix into a model including CDC’s surveillance data for flu-like illness from three weeks prior, they were able to make big improvements to real-time predictions at the state level.
If wearables data can be used to track disease outbreaks, it will certainly be useful. Any dataset we can use to better understand the progress of widespread infectious diseases is worth keeping on tap. In fact, it’s worth noting that wearables offer a population-level view that is quite difficult to come by otherwise.
That being said, we should bear in mind that wearable measures still aren’t precise enough to meet the demands of other research efforts. It continues to be the case that such devices typically don’t produce valid, reliable data that can be relied upon for clinical-grade data collection.
The reality is that wearables continue to be more of a consumer hobbyist tool than a true medical device. While consumer users might benefit from understanding broad trends in the personal data they collect, clinicians are rightfully skeptical that individual readings are accurate enough to be used in making care decisions.
Still, we need to do anything we can to better address the spread of COVID-19 and future viruses, and it seems wearables data might play a part. I’ll be interested to see researchers like those at Scripps and Stanford can accomplish here.