Physical activity is a predictor of survival in COPD. Patients undertaking some level of regular exercise have a lower rate of COPD related admissions and mortality. Increases in daily physical activity have been noted following exacerbations and optimisation of COPD management. There has been a steady uptake in ownership of consumer-based fitness trackers which can capture continuous physiology data over prolonged periods. With advancements in cloud computing, there is now the potential to integrate data from these wearable devices with electronic health record systems. Data can be reviewed by clinicians to monitor physical activity and physiology in patients with COPD, but value-add and clinician capacity are uncertain. Application of machine-learning analyses could generate predictive actionable insights, allowing clinician data review requirements to be focused.
Aim Explore the potential insights to be gained from capturing continuous physiological measurements in COPD patients using commercially available wearable technology.
Method As part of the RECEIVER digital innovation study (NCT04240353), high risk COPD patients were given Fitbit Charge 3 devices linked to a co-designed web app which also captures daily patient reported outcomes and exacerbation events. Exacerbation events, hospital admissions and treatment changes were plotted with daily step counts and daily average heart rate to evaluate potential patterns which would justify more extensive analyses.
Results Data from 32 patients with sustained FitBit recordings were reviewed as part of planned 6 month interim analyses. Average days of available data = 58 (8–147). We identified notable trends in daily step count and heart rate around exacerbation events (figure 1). Increased daily step counts and reduction in heart rate were observed following commencement of home NIV.
Conclusion Notable bio-plausible insights are present, with correlation between Fitbit data, exacerbations and treatment interventions. Further evaluations of the capability of commercially available wearable sensors to track and predict COPD events are indicated. Results from this evaluation have directed the further analyses of the RECEIVER trial data. This includes expansion of the digital connectivity to incorporate intra-day wearable data, integration with patient-reported outcome and clinical summary data, and application of machine-learning algorithms targeting a risk of exacerbation decision support prediction model.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.