A number of iGeneTRAiN are currently implementing wearable health devices and analyzing physiological data to determine the feasibility of early infection detection in transplant patient populations and their respective household members. The aim of the study is to detect infections, such as SARS-CoV-2 and influenza, before symptom onset using existing algorithms as well as a number of newer algorithms designed to take confounders such as medications and pre- and post-transplant complications into account. The studies aim to provide early intervention, which can decrease the spread of infections and often correlates with improved outcomes.
Smart watch devices connect to a centralized server through a smartphone app called MyPHD developed by the Snyder Group and implemented through the Innovation Lab in Stanford. The app was developed to study long-term smartwatch data and is in the process of expanding to accommodate a million individuals. Data from the transplant patients and their families is being funneled into dedicated silos to be analyzed for these studies. All wearable device data is de-identified through a label that links household units together and links to each patient’s background health information. A very short daily questionnaire is sent daily through MyPHD so that the symptoms data is captured and can be integrated with the wearable device datasets.
The infection detection algorithms employed for this study come from previous infection detection studies built on the MyPHD platform. These algorithms use smoothing functions to set baseline physiological parameters, An online detection uses the CUSUM method to identify periods of outlier measurements. More information about the MyPHD framework and the algorithms can be found in these papers:
- The first paper is from 2017 describing the use of wearables for disease detection and other areas of health monitoring: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.2001402
- The second paper is about the early detection of COVID-19 using smartwatch data: https://www.nature.com/articles/s41551-020-00640-6
Both of these papers use wearable data and infections from the general population, while this new study aims to bring these this early infection detection method to transplant patients and their household members specifically. The algorithms in these papers were based primarily on heart rate, heart rate variability, and activity data. The newer smartwatches are becoming more advanced and include many sensor capabilities, such as saturating oxygen and blood pressure. One of the goals of the study is to determine the most effective physiological parameters to factor into the transplant specific algorithms to take full advantage of the wearable devices.