Finding Waldo: The Patient with Wearables
In health informatics, Electronic health records (EHR) is one of the most prominent resource of research data. We are constantly craving for finding something new (wishfully ground braking), some variables that have casual relationship and correlations (hopefully publishable) out of EHR. Working with EHR has already been huge burden considering large cohorts. But now we have an external source for personal health records which will change the way of health informatics research. This data is coming from consumer wearables and smart devices. It is not as structured and reliable as good old EHR, but promising. So does it make it easier to find and retrieve the data we are looking for?
There are many devices for tracking wellness (Fitbit), biometrics (Philips), as well as heart rate, blood pressure, weight and extending to hair analysis tools (Withings). These companies mostly allow personal data extraction as CVS or TXT format. However, it is limited to consumer-level information. Data collection development platforms (Apple Research Kit, Dev.Fitbit) were also released to open up the veils to this new source for research. New but, at the current state, less reliable and open to biases. Current studies revealed accuracy and reliability issues about wearables very well, outlining the preliminary findings with current wearable technologies.
One interesting thing about data collected via wearables is that it transcends the cohort dependency, giving researchers the liberty to go for individual-based researches. A recent research published at PLOS Biology investigated one patient for 2 years with 7 different mobile devices. Authors argued that "many types of continuous physiological and activity information can be collected on a single individual on a long-term basis and can be used to measure, analyze, and guide health-related decisions". Thus, instead of sample size, researchers must look for a patient who is willing to carry/wear the tracker technologies for long time.
On the other hand, new research protocols and methods are required in order to utilize this new data source. Instead of using cohorts (using institutional resources), we need to go for individual data collection, or collaborating with third parties who has a data pool collected from individuals (e.g. fitabase.com). Even though common PHR code of ethics applies in collecting the data, researchers should be aware of possible ethical concerns, using updated consent agreements, getting research permissions and paying attention de-identifying data to avoid revealing user identity acquired from wearables (e.g. location data).
