Abstract
Wearable technologies promise to redefine assessment of health behaviors, yet their clinical implementation remains a challenge. To address this gap, two of the NIH’s Big Data to Knowledge Centers of Excellence organized a workshop on potential clinical applications of wearables. A workgroup comprised of 14 stakeholders from diverse backgrounds (hospital administration, clinical medicine, academia, insurance, and the commercial device industry) discussed two successful digital health interventions that involve wearables to identify common features responsible for their success. Seven features were identified including: a clearly defined problem, integration into a system of healthcare delivery, technology support, personalized experience, focus on end-user experience, alignment with reimbursement models, and inclusion of clinician champions. Health providers and systems keen to establish new models of care inclusive of wearables may consider these features during program design. A better understanding of these features is necessary to guide future clinical applications of wearable technology.
Introduction
Wearable technology, also known as “wearable devices” or simply “wearables”, generally refers to any miniaturized electronic device that can be easily donned on and off the body, or incorporated into clothing or other body-worn accessories1. While wearables have established utility in the fitness, gaming and entertainment industries, their role in the healthcare environment remains less clear2. To date most commercially available wearables are limited in scope, tracking one or two health-related variables, and have yet to produce accurate measurement of many markers of health status that they attempt to assess such as heart rate variability, nutrition, and mood3,4,5. To the extent that wearables overcome these limitations, they hold much promise towards expanding the clinical repertoire of patient-specific measures, and they are considered an important tool for the future of precision health.
Physical activity information is perhaps the most common measurement provided by current wearable devices, thus it serves as a useful example to review the opportunities and difficulties facing digital health development using wearables. Physical activity is a well-established marker of current health status and future health risks, it is a useful estimate of real-life functional performance6, and it has been tracked in health research using body-worn sensors for many decades7,8,9,10. Wearables for remote digital health and physical activity monitoring have been validated in various settings11,12,13. Given the ubiquity of physical activity monitors, it is surprising their effective incorporation into clinical care remains a challenge, especially in face of the multiple known health benefits of physical activity and the many healthcare scenarios where physical activity information has a clinical use14,15,16. Key challenges preventing inclusion of physical activity monitoring in routine clinical care include the need for data standardization between the many different commercially available devices and sensor locations, and integration of this data into the electronic health record and clinical workflow.
Previous research suggests that digital health programs incorporating health behavior models and personalized coaching are the most successful17. Yet, knowledge is limited regarding all the factors that may drive successful implementation of wearables into the healthcare environment18,19,20. Notwithstanding this gap in knowledge, a rapid proliferation of untested digital health applications has led some to dismiss activity monitoring as a health fad that appeals only to fitness fanatics2,21,22. Deficits in knowledge have also given rise to concerns that wearables could pose health risks by paradoxically reducing healthy behavior rather than promoting it, either through a false assurance of healthy behavior or through discouragement from failure to achieve goals2,23,24,25. At the same time, many reports indicate a beneficial impact on the health of individuals using wearable monitors26,27,28,29,30. Some even indicate that utilization of wearables for clinical care delivery provides health benefits that outpace conventional methods of care6,31,32,33.
These disparate opinions prompted us to seek answers to the current utility of wearables in healthcare. To this end, two of the National Institutes of Health’s Big Data to Knowledge (BD2K) Centers of Excellence—the Mobilize Center at Stanford University34 and the Mobile-Sensor-to-Knowledge Center (MD2K)35—organized a workshop on the clinical application of wearables. This invitation-only workshop was led by a practicing physician with expertise in the analysis and clinical interpretation of wearable sensor data, and included 14 stakeholders from diverse backgrounds including: hospital administration, academia, insurance, clinical medicine, and the commercial wearable device industry. The workshop group reviewed and discussed two different digital health interventions that have successfully integrated wearables into their clinical workflow, and that empirically demonstrated superior health outcomes using the digital platform compared to traditional care. The workgroup included a representative from each system with direct knowledge of the development and implementation of their respective program. The workgroup’s stated goal was to define the features common to the two sample cases from health systems that allowed the integration of wearables into a successful digital health intervention.
The Ochsner Health System and Kaiser Permanente each designed a digital health program for the management of hypertension and diabetes, respectively36,37,38. In both programs, commercially available at-home digital device (glucometer at Kaiser, blood pressure monitor at Ochsner) was customized to integrate with the electronic health record (EHR) to wirelessly transmit measurements directly to the EHR. Within this framework, wearables were only one aspect of an integrated care delivery process. The digital monitors provide useful instructions directly to patients, while updating the EHR with new data and promoting more rapid and efficient communication between clinicians and patients when needed.
Within the Ochsner system, home blood pressure monitoring and health behavior data from additional optional wearable connected devices (including physical activity monitor and scale), was utilized in conjunction with health coaching and medication adjustments to track patient health status. A mobile application interface served as a feedback loop sending important information to the treating physician, who is then able to seek additional contact if a patient is not meeting expected goals, or simply continue to monitor if all is going well. Within Kaiser, all patients enrolled in the program are actively managed by their care team, who uses the integrated data in EHR to drive clinical treatment decisions, performed in tandem with diet and physical activity education and coaching, and preventive care. This newly tested approach with wearable connected home monitors contrasted from usual care where patients follow-up with care providers in clinic at designated appointments. The lag time in usual care was overcome by the connected digital monitors allowing real-time two-way communication between patient and care team. The new platform also differed from usual care by providing participants with direct technical support to personally address technical issues. It also incorporated a health support team to tackle health-focused problems such as lifestyle management, medication compliance, nutrition, and health education. These teams added a personalized component to each program that specifically addressed important patient-level health needs.
Both remote glucose monitoring and blood pressure monitoring digital sensors are commercially available and have been described extensively in the literature39,40,41,42. The Ochsner and Kaiser Permanent health systems used devices vetted by their own systems and found to be acceptable for use in their patient cohorts. Given the variability in commercially available products and ongoing validation of their output, each health system should carefully review the available options prior to deployment. A 90-day prospective assessment of the Ochsner digital hypertension management program revealed that 71% of patients treated with the digital health program achieved target blood pressure control, compared to 31% under usual care25. Kaiser Permanente’s digital glucose monitoring program reduced time for contacting patients in the program via telephone visits by 50%, and therefore effectively doubled clinicians’ capacity to manage patients with diabetes38. Both digital programs had similar features that allowed them to empirically outperform usual care. In this paper, we highlight the workgroup’s findings regarding the common key features between these two programs, which underscore the successful implementation of wearables and digital health programs in two different health systems (Fig. 1).
Evaluating the shared characteristics of these programs provides insight into the emerging role of wearable technologies in the clinical arena that expand beyond the role of physical activity monitoring. These shared characteristics provide a roadmap for future implementation of wearables into digital health interventions. Successfully designed and implemented, these new technologies can produce measurable benefits, which accrue to patients and the health system. We suggest that all medical institutions interested in utilizing and scaling wearable technology in their clinical workflow should consider the common features that we outline below. In order to highlight areas with near term potential, we finish by offering our insights into the expanding future direction of wearable physical activity monitoring in the clinical workspace.
Features for successful implementation of wearables in clinical environments
The workgroup identified many common features between the Kaiser Permanente and Ochsner Health System digital health interventions. Among them, the workgroup agreed that the following common features were directly related to the successful implementation of wearable technology into the digital patient care platform:
Clearly defined role of wearables to address a specific problem behavior for the disease state
In-line with the scientific method, which begins with a hypothesis, implementation of a successful wearable program requires one to establish a priori goals for the target disease state. In other words, there must be a specific problem that one hopes to address with the wearable technology. As an illustration, Kaiser Permanente and Ochsner identified clinician, patient and system specific problems that were limiting their ability to effectively manage patients with diabetes and hypertension, respectively. For example, clinicians indicated that the frequency and reliability of core disease measures (blood glucose and blood pressure) obtained outside of the clinic were highly variable. Both programs design solutions around this using digital technology applications, including wearables. See Table 1 for an expanded view of the many issues addressed by each program. Simply put, the role of the wearable was clearly defined based on a health-specific need; yet, importantly, the wearable was only one part of a larger digital health program.