I’m often struck by how data collection through patient sensors varies from one trial to another. At one end of the spectrum, I’ve seen a sponsor define a very specific data collection and analysis process, whereby patient blood pressure data was collected at sites, kiosks, and at home through a mobile device; each single data point was integrated with Calyx EDC for analysis. At the other end of the spectrum, a different trial recorded activity and heart rate continuously through a wearable activity monitor; despite the continuous recordings, the only data points that were reported daily were the lowest, highest, and average heart rates for each patient. This latter example raises the question: what happened to the rest of the data and how are we using it?
The Clinical Trials Transformation Initiative (CTTI) recommendations on the Use of Mobile Technologies in Clinical Research (July 2018) include one particular recommendation that “only data that are necessary to meet the objectives of the trial should be collected.”There are some nuances within the recommendations, especially around exploratory endpoints, that may require the capture of data without knowing which data points will be the most valuable. However, CTTI is very clear that “when the study endpoints are well understood, CTTI recommends against speculative “data fishing.”
“It seems inevitable that, as a norm, patients will have to wear wearable devices to take part in clinical trials in the future.”
– Sylvain Berthelot, Solution Consultant Director, Calyx
Going back to my example above about activity and heart rate monitoring, how do we set the boundary between strictly required data and “data fishing”? Considering that patients wear the device, no matter what data points are analyzed, is it frowned upon to collect more data points than strictly needed?
Although collecting more data than required doesn’t “hurt,” we must aim to not increase the patient burden. However, according to Kaiser Associates and Intel, 70% of clinical trials will incorporate sensors by 2025. So, it seems inevitable that, as a norm, patients will have to wear wearable devices to take part in clinical trials in the future.
What will we do with all this data? Will we store it in the attic, just in case, and forget about it? Or does it open new opportunities we are not aware of yet? For example, will we be able to access data in the future for retrospective analysis, or could it be used to refine patient selection for synthetic arms?
With the increasing use of real-world evidence and past trial data to build synthetic arms, we should ask ourselves what type of data we may need in the future. As data collected through wearable devices become primary data points, will the absence of at-home data points in current trials limit our future ability to build synthetic arms based on past trial data? Now going back to the recommendation from CTTI, some may feel that it is a limitation not only to what we can analyze now but potentially to how we could use the data in the future.
We are much more aware of the potential clinical trial data represents than we have been in the past. With the emergence of AI/ML in our industry, we may be about to witness the biggest revolution in clinical trials to date. As long as what we are aiming for benefits patients, and doesn’t increase the burden for those patients taking part in clinical trials, should we limit our ability to collect patient data? Or should we aim to set rules as to which data points should be collected with wearables?
In the world we’re living in today and the direction our industry is heading, there’s much to consider when it comes to the collection of clinical trial data by wearable devices.