Always On, Never Distracted: AI’s Role in Protecting Patients

Clinicians are being asked to manage increasingly complex patient profiles with limited time and resources. This challenge is especially pronounced in the care of patients with chronic conditions like heart failure, where the difference between proactive management and a reactive crisis can determine whether a patient recovers or is readmitted to the hospital. Dr. Ries Robinson, a leading voice in digital health innovation and a member of DocGo’s Medical Advisory Board, believes artificial intelligence (AI) is a key part of the answer.

Healthcare systems have developed strong capabilities for managing acute events. However, they often struggle with the sustained vigilance needed for chronic disease management. As Dr. Robinson explained:

“Monitoring five parameters for one patient is something a clinician can do. But if you’re monitoring 20 people, each with five parameters, that becomes extremely difficult. And at scale, say 150 patients and beyond, it becomes impossible.”

This information burden is compounded by the unique nature of each and every patient. What appears normal for one person may indicate a problem for someone else. In conditions such as heart failure, which can result from several distinct causes, each individual may exhibit a different physiological baseline. “Heart failure has many aetiologies,” Robinson noted. “Each can result in a different pattern of what is ‘normal’ for the patient. AI can help by identifying and tracking those individual baselines across multiple parameters. That’s not something the human brain is built to manage.”

Hydration management offers another example: it requires deciphering the choreographed movement of water among intravascular, interstitial, and intracellular spaces, each reshaped minute?to?minute by salt intake, fluid loss, medications, and age. “Effective hydration management demands tools that fuse time?based kinetics, external disturbances, and three?compartment fluid flows into a single, actionable view,” Dr.?Robinson says.

Dr. Robinson sees three key opportunities for AI in patient monitoring:

1. Managing Information Overload
   AI can process large volumes of data from hundreds of patients without fatigue. While human clinicians may miss early warning signs, AI can detect subtle changes and trends that often precede deterioration by continuously analyzing patterns in patient data and flagging gradual shifts that might be imperceptible to human observation.
2. Establishing Individualized Baselines
   Rather than relying on generic population averages, AI can learn what is typical for each patient. This allows alerts to be based on personalized thresholds, supporting earlier and more accurate interventions.
3. Supporting Medication Management
   Heart failure management often involves five separate classes of medication. AI can assist by tracking patient-specific medication response curves, identifying potential conflicts, and helping to ensure optimal combinations over time.

One often overlooked advantage of AI is its consistency. In hospitals, clinical focus shifts over time as institutional priorities change based on patient needs – potentially emphasizing fall prevention, hydration management or infection control. Staff naturally adapt their attention to the current initiative. Dr. Robinson shared a case from Presbyterian Hospital, where a targeted campaign to reduce patient falls initially succeeded. However, as leadership moved to new priorities, fall rates crept back up. “It was like whack-a-mole. People can only focus on one thing at a time. AI doesn’t have that limitation,” he said. Similarly, when experienced staff leave (whether due to retirement, rotation, or reassignment), it can lead to declines in patient outcomes. AI can help by preserving procedural knowledge, patient-specific knowledge, and institutional best practices. It provides continuity that human teams may struggle to maintain.

Although AI can deliver real-time insight and pattern recognition, it does not replace the physician’s role in decision-making. “The physician still plays a critical role, especially at the point of discharge where experience and intuition become critical,” Robinson added. “They’re the only ones with full visibility into comorbidities, the patient’s home support, and socio-economic context. AI can analyze data with dazzling speed, but does not have intuition—it cannot sense the undocumented interactions, emotions, and contextual nuances that make comfort and truly effective care possible.” These non-quantifiable factors still require clinical judgment.