AI in Healthcare: Predicting Disease Progression

In the world of AI, most recent discussions around healthcare focus on early diagnosis. Machine Learning and Artificial Intelligence are amazing tools and can significantly improve prognosis via early detection. The best two examples I can give you are chronic hypertension and chronic kidney disease. Both conditions are what we call silent killers. In other words, their symptoms often go unnoticed until the disease has progressed to the late stages, and the prognosis is poor.

Now that being said, although early diagnosis is important there is an even more amazing application for AI in healthcare that gets less attention than it merits.

What if AI could predict how a disease will progress before it becomes acute, allowing healthcare providers to intervene at the right time with the right treatments?

Now personally as a healthcare professional… That makes me feel giddy!

The Shift from Diagnosis to Prediction

Traditionally, AI has been used in healthcare to:

✅ Detect anomalies in medical imaging
✅ Flag potential disease risks
✅ Assist in diagnosis through symptom analysis

These are all reactive implementations.

Proactive implementations are another story.

Chronic diseases such as diabetes and kidney disease require continuous monitoring of disease progression for the best outcomes. AI-powered predictive analytics through dynamic AI architecture can be made to anticipate disease progression, helping physicians and their patients make proactive decisions before an illness reaches a critical stage.

💡 Example: AI can analyze patterns in continuous blood sugar monitoring, insulin dosage and vital signs to predict when a diabetic patient is at risk of developing complications—before symptoms appear.

AI can analyze medical notes, blood tests, vital signs, imagery, and progression patterns to alert a physician when a chronic kidney disease patient is at risk of progressing from stage 2 to 3 kidney disease.

The Data Advantage: AI’s Key to Predictive Medicine

What makes this possible? AI thrives on vast, interconnected datasets. By analyzing:
📌 Real-time patient vitals from wearable devices
📌 Lab results over time
📌 Medication adherence data
📌 Social determinants of health (e.g., access to care, lifestyle factors)

📌And so much more

AI models can detect subtle correlations that humans might miss.

With chronic hypertension, patients are encouraged to take their blood pressures home daily. If and when, their blood pressure spikes beyond a certain level, they are asked to let their physicians know.

Ms. Gagnon’s physician set that number to 160 systolic.

On Thursday, her blood pressure is 138 systolic, way below the threshold her physician has set. The AI model analyzes that although 138 is lower, Ms. Gagnon’s BP has increased steadily by small increments over the past 5 days, indicating that something isn’t quite right. As a result, her physician receives a notification, and he recommends a treatment change immediately to get Ms. Gagnon to optimal health.

🔹 Case Study: Some hospitals are already using AI to predict sepsis onset hours before symptoms appear, leading to faster interventions and better outcomes, in many cases saving lives.

In the world of chronic disease management, a switch of focus to predictive disease management could not only save thousands of lives but also dramatically increase the quality of life of many patients.

Physicians carry an enormous mental and emotional load when treating chronic illnesses, constantly wondering if they could have anticipated certain progressions after the fact or simply wondering if they truly ‘turned every rock’ during their last appointment with their patients. Intelligent predictive disease management via AI could completely shift that narrative and allow physicians to make optimal treatment decisions at exactly the right time every time.

From an outside view this may not seem super exciting, but for healthcare professionals… Disney land doesn’t even come close.