Using AI To Scale Care Coordination Beyond Episodic Care

U.S. health care is primarily structured around episodes of care. A patient visits a clinician, is treated, and then isn’t seen or heard from until their next encounter. That structure was more appropriate when acute illness and infectious disease dominated medicine, but now it feels antiquated when it is now marked by chronic illness, limited clinical resources for managing patients, and increasingly complex patient needs.

Over half of U.S. adults have at least one chronic condition, according to the Centers for Disease Control and Prevention, and many have multiple. Today, chronic and mental health conditions contribute to around 90 percent of the nation’s almost $5 trillion in annual health care spending. These conditions should not be treated in isolation. They necessitate continuity, monitoring, and care coordination over time.

The clinical workforce required to provide this level of care is increasingly strained. New federal estimates predict that there will be a national shortage of as many as 141,000 physicians by the late 2030s, assuming trends hold, with primary care and rural communities reporting the most significant gaps. This limited provider availability, paired with high patient volume, leaves many patients with long stretches between meaningful touchpoints, even when their conditions require close follow-up.

According to national figures, there are approximately 320 physician office visits per 100 people each year, or a little over three visits per person per year. That cadence means that there are usually months between encounters, an extended period of time for symptoms to escalate, adherence to care to break down, or social conditions to impact care. Additionally, with complex conditions, patients’ care can be spread out among many different providers. A 2022 study of Medicare beneficiaries found that 4 in 10 experienced highly dispersed ambulatory care, with an average of 13 visits to 7 different practitioners per year.

Care coordination is meant to fill these gaps in care but is limited by human capacity. Physicians must compile and analyze vast medical histories, lab results, imaging, specialist notes, utilization data and other signals, often spread across disparate records and systems. Even if care teams accept that outreach must be proactive, they encounter the ongoing challenge of determining exactly which patients can be treated immediately and which can wait, while also load-balancing large panels of patients and competing clinical priorities.

Artificial Intelligence (AI) can help to solve these problems, but only if the role of technology is framed correctly. We cannot completely replace clinicians with AI and care itself has always been fundamentally human. Trust-building, understanding context, and helping patients with difficult decisions requires empathy and judgment that technology can’t easily replicate. The real opportunity is to use AI to scale those human strengths. AI can process large amounts of data across electronic health records, medical claims, social data sets, remote monitoring devices, and more, to draw up patterns and identify risks that are hard for humans to effectively do at scale.

AI can help spot patients whose illnesses are deteriorating, whose care plans are falling behind, or whose social conditions increase the risk of avoidable utilization. AI can also assist clinicians in effectively overseeing large-scale panels while implementing clinical interventions more efficiently and with higher efficacy; this provides space for deeper patient relationships instead of the more transactional approach of single-visit treatment.

For conditions like heart failure, studies show that higher levels of coordinated care are associated with roughly 10% lower odds of 30-day readmission and 17% lower odds of 30-day mortality. This proves that proactive and continuous management goes hand-in-hand with improving patient care. Stronger care coordination supports better chronic disease management, improves patient outcomes, reduces avoidable hospital admissions, and enables health systems to use limited clinical resources more effectively.

Beyond just clinical implications, AI can extend care into everyday life and the social conditions that shape health outcomes. Today, more than one in five adults without access to a vehicle or public transportation reported skipping necessary healthcare services. This underscores how important care coordination can be for patients who face transportation, mobility, or other socioeconomic barriers. Continual support between visits can vastly improve care and outcomes without adding unnecessary burdens to clinicians.

Looking ahead, healthcare leaders face a clear choice. Continuing to rely on episodic models will deepen workforce strain and hinder patients’ access to medical care. Shifting toward scaled, continuous, coordinated care necessitates new, and often novel, AI infrastructure, but does not require sacrificing the human elements of medicine – the goal is not to automate away care, but to make delivery more personalized, more attentive, more timely, and, in a way, more human.