AI can warn doctors when patients could become ill

Artificial intelligence (AI) is transforming healthcare, providing tools to help doctors make better decisions and save patients. A recent study shows the profound impact artificial intelligence can have in the clinical setting.

This study, conducted by experts at Mount Sinai, shows that AI-generated alerts can significantly improve patient care and outcomes.

Improving patient care through AI alerts

The study’s results showed that patients were 43% more likely to intensify treatment and significantly less likely to die when their healthcare teams received real-time AI alerts. These alerts signaled adverse changes in patients’ health, enabling timely intervention.

“We wanted to find out whether rapid alerts from AI and machine learning trained on a wide variety of patient data can help reduce both the frequency of patients needing intensive care and their risk of death in the hospital,” said lead study author Dr. Matthew A. Levin, director of clinical data science at Mount Sinai Hospital.

“Traditionally, we rely on older manual methods such as the Modified Early Warning Score (MEWS) to predict clinical deterioration.”

“However, our study shows that automated machine learning algorithms that trigger a provider assessment can outperform these previous methods in accurately predicting this decline. Importantly, they enable earlier intervention that could save more lives.”

Real-time AI alerts in clinical settings

The study was a nonrandomized, prospective analysis of 2,740 adult patients admitted to four medical and surgical departments at Mount Sinai Hospital.

The patients were divided into two groups: one received real-time alerts from AI based on a predicted deterioration in their health, while the other did not receive these alerts even though they were generated.

In wards where alarms were suppressed, urgent medical care was still provided to patients whose condition met standard criteria.

The results for the intervention group were promising:

• Patients were more likely to receive medications to support their heart and circulation, suggesting early intervention by doctors.
• Patients were less likely to die within 30 days.

A learning healthcare system

“Our research shows that real-time alerts using machine learning can significantly improve patient outcomes,” noted lead study author David L. Reich.

“These models are accurate and timely tools for clinical decision-making that help us get the right team to the right patient at the right time.”

“We see these as ‘augmented intelligence’ tools that will accelerate the personalized clinical assessments made by our doctors and nurses and initiate treatments that provide greater safety to our patients. These are important steps toward a learning healthcare system.”

Implementation and future prospects

Although the study was terminated early due to the COVID-19 pandemic, the algorithm was implemented in all stepdown units at Mount Sinai Hospital.

These units care for patients whose condition is stable but who still require close monitoring – a critical phase between the intensive care unit and the general hospital area.

A specialized team of intensive care physicians now visits the 15 patients with the highest predictive values ​​every day and provides treatment recommendations to the treating doctors and nurses.

As the algorithm is continuously retrained using larger patient data sets, it improves its accuracy through reinforcement learning.

In addition to this clinical deterioration algorithm, Mount Sinai researchers have developed and implemented 15 additional AI-based clinical decision support tools across the healthcare system.

These advances represent a significant step toward integrating AI into healthcare, improving patient care and paving the way for more innovative solutions in the future.

Improving patient outcomes

As the Mount Sinai study shows, the use of AI in healthcare has enormous potential to improve treatment outcomes.

Real-time AI alerts not only enable timely interventions but also help healthcare professionals make informed decisions quickly.

As AI technology advances, its role in clinical settings will undoubtedly increase, promising a future where healthcare becomes more efficient, effective and patient-focused.

AI tools such as real-time alerts will improve decision-making and enable timely interventions and better treatment outcomes. As AI becomes more integrated into healthcare, it will help healthcare professionals deliver higher quality care.

Continuous improvement of AI algorithms will lead to more accurate predictions and tailored treatments. Ultimately, AI will transform healthcare by making it more responsive and focused on patient needs, leading to improved overall health and well-being.

The study was published in the journal Intensive care medicine.

—–

Like what you read? Subscribe to our newsletter for exciting articles, exclusive content and the latest updates.

Check out EarthSnap, a free app from Eric Ralls and Earth.com.

—–