Global healthcare technology giant Roche (Roche) and technology companiesIBMRecently announced,Jointly developed by both partiesAIDriven Blood Glucose Management Solutions——Accu-Chek SmartGuide PredictThe application has officially landed.The system integrates continuous glucose monitors (CGMs) Real-time data and machine learning algorithms provide precise blood glucose fluctuation predictions for diabetic patients, marking a new phase of "proactive prevention" in digital diabetes management.

Roche in2024Annual Advanced Technologies and Treatments for Diabetes (ATTD) The event was first launched during a session of the conference.Accu-Chek SmartGuide。2024Year7Month, thisAIDriven continuous glucose monitor receivedCELogo certification. With the help ofPredictWith the application, users can better manage blood glucose control. The app uses predictive algorithms driven by real-time blood glucose values and provides relevant insights to reduce the risk of hypoglycemia or hyperglycemia.

Accu-Chek SmartGuide PredictThe core lies in its three major predictive function modules,All are based on Roche andIBMJointly Developed byAIAlgorithm Framework:

• Blood Glucose Prediction (Glucose Predict）

Through analysisCGMsEvery5Real-time blood glucose values collected per minute, combined with user input data such as diet and exercise, dynamically generate future2Hourly Blood Glucose Fluctuation Curve, Prediction Accuracy Rate Reaches92%(Clinical validation data), improved compared to traditional monitoring systems40%。

• Hypoglycemia Alert (Low Glucose Predict）

The system can be advanced.30Minutes to identify hypoglycemia risk (blood glucose ≤3.9 mmol/L), with a false positive rate lower than5%. In the pilot at the University Hospital of Zurich, Switzerland, this feature reduced the incidence of nocturnal hypoglycemic events.28%。

• Risk Assessment for Nocturnal Hypoglycemia (Night Low Predict）

For diabetic patients"Vulnerable period," the system analyzes before bedtime7Hourly blood glucose trends, insulin dosage, and basal metabolic rate provide personalized prevention recommendations (such as adjusting bedtime snack amounts or insulin pump parameters), improving user compliance.65%。

The system has tremendousClinical Value，Clinical Blood Glucose Monitoring CompletedFrom"Data Silos" to "Decision-Making Closed Loop"The process.

The system'sThe innovation lies in breaking the tradition.CGMs"The Limitations of 'Monitoring Only, No Intervention':"

• Personalized Intervention:AIThe algorithm can learn from users' historical data and dynamically adjust the parameters of the prediction model. For example, for users who are sensitive to a sharp drop in blood sugar after exercise, the system will provide early warnings.1Hourly reminder to replenish carbohydrates.

• Doctor-Patient Collaboration: Predictive data can be synchronized to the doctor's platform, supporting remote adjustment of treatment plans. Clinical studies from the German Diabetes Center show that patients using this system have improved glycated hemoglobin (HbA1c).HbA1c) Average decrease0.8%, Achievement Rate Increased35%。

• Safety Redundancy: System AdoptionIBMBlockchain technology ensures data privacy and avoids algorithmic misjudgment through multiple verification mechanisms, guaranteeing the reliability of clinical decision-making.

Morriz, Head of Roche Information Solutions·Hartman (Moritz Hartmann) pointed out: "This cooperation has validated 'medical+AI’Synergistic Effect——We Only Use18months to complete the process from algorithm development toCEThe entire certification process is shortened compared to the industry average cycle.40%。”

IBMChristian Keller, General Manager of Switzerland (Christian Keller) emphasized the key to technology implementation: "healthcareAI"The core is not about showing off skills, but about building a trustworthy closed loop. We have designed 'three layers of protection' for the system: data encryption, algorithm explainability, and clinical validation closed loop, ensuring that every predictive recommendation is evidence-based."

According to Roche,The next-generation system will integrate wearable device data (such as heart rate variability, skin conductance) to further expand predictive dimensions. Meanwhile, both parties plan to collaborate globally.50Home Diabetes Center Builds Real-World Evidence Database, PromotingAIThe model continues to be optimized.

"The ultimate goal of diabetes management is to make patients forget the existence of the disease," Hartman said. "ThroughAI"Predictions suggest that we are turning this vision into reality." With the promotion of this system in the European market, a technology-driven revolution in diabetes management has quietly begun.