The U.S. healthcare system is excellent at delivering emergency and acute care. It falters at managing chronic conditions such as diabetes, chronic obstructive pulmonary disease and congestive heart failure (CHF). We launched an in-depth ethnographic study to understand why chronic disease management is so difficult and whether digital therapeutics could help.
We shadowed patients recently discharged with chronic condition diagnoses and talked to their physicians. We asked about their issues and wish lists. The process produced clear insights into the capabilities that physicians and patients need to effectively manage the chronic conditions. It also enabled us to identify a persona, Marge, diagnosed with CHF, who reflects the typical patient in our ethnographic study. Marge’s journey (see figure below) spotlights the core challenges of effective chronic condition management.
Patients deal with these issues in managing their chronic illness:
Lack of clear information to help them understand their newly diagnosed condition and how other co-morbidities they have may influence it.
Complicated prescription schedules that are difficult to follow accurately.
Weak support systems such as lack of transport to pharmacies and quality grocery stores and limited connection to other CHF patients.
Transition of care challenges and insufficient post-discharge communication with providers for getting advice for overcoming barriers.
Minimal assistance with making fundamental behavioral changes, such as learning how to cook in healthier ways or becoming more active.
Inability to self-monitor their condition because they don’t have the data or experience to gauge the severity of symptoms.
Providers face these obstacles in supporting chronically ill patients:
Lack of visibility into patients’ journeys after they leave the hospital and between appointments, which can lead to gaps in care.
Inability to provide the 24-hour support patients may require for weeks or months while they adjust to their diagnosis and medications.
Too much of the wrong type of data instead of actionable, just-in-time alerts so they can intervene when patients need them most.
Overwhelming workflows that don’t accommodate standalone health apps or tools.
These findings signaled that a digital therapeutics-based approach to managing CHF would be effective. Digital therapeutics combine disease management algorithms, telemedicine, Internet of Things (IoT) devices and machine learning technology to create evidence-based solutions. They meet clinical safety standards, complement clinical treatment and integrate smoothly into provider workflows, including reimbursement cycles.
We developed a prototype of a conversational AI-based digital therapeutic for CHF. CHF was an optimal test candidate for the prototype because virtually all CHF patients have one or more additional chronic conditions, and those in our study wanted just one tool to manage all their conditions.
The prototype was designed to create a smoother transition of care after patients are discharged. We trained the AI-based system to recognize and respond to different moods. It integrated symptom tracking recommended by American Heart Association guidelines and also vitals tracking through automatic synchronization with IoT devices. It incorporated proven behavior-change models and techniques to handhold the patients to adopt a healthier lifestyle.
In the tests, patients conversed widely with the AI-powered companion. They discussed practical matters, such as monitoring vitals, medical adherence and making healthy meals. They ventured into emotional wellness, discussing managing anxiety and finding a new life purpose. These conversations were contextually appropriate and highly personalized. Many patients told us the education, motivation and support that the AI agent provided would encourage them to commit to new behaviors and achieve better outcomes.
As digital therapeutics help patients achieve these goals, accountable care organizations, health systems, federally qualified health centers and payers will close gaps in care, improve quality and reduce costs. Healthcare organizations could offer digital therapies to their highest risk CHF patients first, then provide them to patients with gateway diseases, such as diabetes, to help prevent progression toward CHF. The key to success will be designing digital therapeutics that put the needs of humans first.