Table 1.

Examples of solutions to problems addressed by each of the seven identified features of successful digital health interventions that include wearables.

Features	Problem(s)	Solution(s)
(1) Clearly defined problem and disease state	Patient domain
	(i) Health literacy	(i) Both programs provided videos and other educational resources on diabetes and hypertension, which were immediately available on mobile phone applications, with automated daily educational emails to patients.
	(ii) Patient adherence	(ii) and (iii) Programs assigned each patient a health coach and also sent daily push notification via phone apps reminding patients to take their medications. Patients were also given access to pill reminder phone apps and where possible, their medication regimen was simplified to facilitate patient adherence.
	(iii) Patient engagement
	(iv) Social isolation
	(v) Limited resources/affordability	(iii) Reward points/tokens were provided via mobile apps for completing daily recommended dose of medications and exercises. Apps and wearable monitors provided a feedback loop between clinical teams, health coach, physician and patients for 2-way communication. This allowed immediate feedback on a patient’s clinical status and progress.
	System domain
	(vi) Misclassification bias
	(vii) Delayed interventions
		(iv) Patients were able to join digital platforms where they could share daily challenges and progress with hypertension and diabetes management with other patients with similar challenges. Patients are able to develop sense of community with shared goals towards more healthy lifestyle choices.
		(v) Patients who could not afford brand medications were switched to generics or less expensive combination agents, and when appropriate and feasible, enrolled in medication assistance programs.
		(vi) At home monitors helped avoid the issue of white-coat hypertension and allowed more accurate monitoring over the course of 1 week. Patients were encouraged to take 3–4 readings per week. If the care team had not received readings for 8 days, patients received an automated text alerting them that a blood pressure reading was needed.
		(vii) Patients with high digital readings were immediately contacted by health coach or a member of the clinical team to address the problem in real time
(2) Incorporation of wearables into an integrated system of delivery	Patient domain
	(i) Health literacy	(i) and (ii) Home glucose and blood pressure monitoring was linked directly to patients’ EHR. Patient-specific behavioral data were generated in the EHR that allowed a health coach to provide appropriate education on diet and physical activity concordant with a patient’s clinical progress.
	System domain
	(ii) Patient engagement
		(i) For patients not meeting expected goals, the care team reached out directly to them to address individual barriers and provide focused digital education, as well as automated medication dosing information with human assistance when needed.
		(ii) Health coaches used integrated EHR data to design health and exercise programs, as well as automated text reminders to help patients work toward their goals and to engage them in daily preventive care.
(3) Technology support bar and service	Patient domain
	(i) Limited resources/affordability	(i–iv) Kaiser Permanente and Ochsner each established onsite technology assistance (respectively, the “Thrive Bar” and “O-Bar”) for technology support and troubleshooting. Custom commercial glucose and blood pressure monitoring smart devices were available for purchase at the Thrive Bar and O-Bar. Patients with poor digital literacy received assistance with setting up digital devices and downloading mobile apps on their phones from the tech team. Patient who could not afford smart phones were enrolled in financial assistance programs to help with access to smart devices. Onsite onboarding increased the likelihood of same-day enrollment in the program and shifted the burden of support and technological knowledge from clinicians to the tech support team. This also cut down loss to follow-up.
	(ii) Digital literacy
	System domain
	(iii) Loss to follow-up
	(iv) Information asymmetry
(4) Personalized experience and (5) Enhanced end-user experience	Patient domain
	(i) Digital literacy	(i) The Thrive and O-bars provided a direct a patient-facing service where patients could pick up information, training, and technical support tailored specifically to meet their health needs in addressing challenges with management their diabetes and hypertension.
	System domain
	(ii) Information asymmetry	(ii) Vetted wellness apps by each health system, with gaming aspects were provided as an option to enhance patient adherence. This also provided care teams and patients full information access, further addressing the issue of information asymmetry. It also allowed real-time response from clinicians and care teams. Weekly graphs on glucose and blood pressure values allowed patients and care teams to remotely monitor clinical progress with blood pressure and glucose levels.
	(iii) Loss to follow-up
	(iv) No shows
		(i–iv) Patients could call a direct line for assistance with lifestyle management, medication compliance, nutrition, and health education issues. This helped to increase patient engagement and provide information specific to patients’ needs.
		(ii) By linking blood pressure and glucose monitoring devices to patients’ EHR, additional clinical data were obtained from the electronic medical record, including serum sodium, potassium, creatinine, estimated glomerular filtration rate, thyroid function tests, and body mass index (BMI). These data were used to create patient phenotypes, which assisted in the design of patient-specific management plans.
		(iii–iv) To minimize no shows and loss to follow-up, patients in the digital program were given the option for virtual visits vs. face-to-face visits, based on patient preference. An option for an additional wearable activity monitor was also recommended. This provided the patient and physician with important information regarding health behavior adjustments that influence the primary outcome. Allowing the users to opt-in or out of these add-on features is another important aspect of end-user design that fostered digital health program adoption and adherence.
(6) Aligned payment and reimbursement models	System domain
	(i) Fiscal sustainability	(i) Most patients in the Kaiser Permanent system are part of a risk-based or capitated reimbursement models, which provides coverage for chronic health management programs inclusive of digital monitoring devices. Patients in either program that could not afford smart devices were provided medical and financial assistance programs to enable patients to participate in the program. The Ochsner health system self-funded its digital hypertension monitoring program in the initial pilot phase of their program. Subsequently, using collected data from at-risk populations, Ochsner was able to negotiate and establish payment models for their digital health intervention. The CMS Current Procedural Terminology (CPT) code 99091 allows unbundling of billing for chronic care services from billing for the collection and interpretation of physiologic- and patient-generated digital health data. There are specific guidelines for reimbursement, and providers meeting the requirements are reimbursed up to $60 per month per patient for a cumulative 30 min of collecting and interpreting data. Other health systems may tailor digital health program in a way that is fiscally sustainable. For example, health systems may encourage patients to consider opening health savings account, which they could use to fund components of a home digital health monitoring program.
(7) Clinician champions and stakeholder support	System domain
	(i) Program optimization	(i–ii) Kaiser Permanent and Ochsner identified physician champions at local clinical sites to beta test the digital health program, provide feedback, and optimize initial clinical use. After initial buy-in, program optimization and pilot demonstrations of the program’s effectiveness, the health systems scaled the program to other sites. Ochsner’s program primarily targeted primary care physicians (PCP), while Kaiser’s targeted clinical care managers. In the Ochsner program, indirect financial incentives for participation was provided by connecting 5% of physician base compensation to quality metrics. Beyond financial compensation, an internal assessment of physician attitudes about the digital hypertension management program revealed that physician adoption was motivated more by the improved efficiency of care than by any perceived financial reward or incentive. At Kaiser, clinicians are tracked on their diabetes performance by monitoring improvements in HbA1c, “touches” (clinical contacts or treatment intensifications), and according to the Healthcare Effectiveness Data and Information Set (HEDIS) measures/ There was no direct financial compensation tied to performance.
	(ii) Program implementation and stakeholder buy-in

The left column lists the seven common features identified in the two successful digital health interventions that involve wearables. The middle column outlines problems within the patient domain that are addressed by each feature. The right column lists specific representative examples of solutions provided for these problems by the two sample digital health programs.