How a Sensor Checks Itself for Accuracy Using Other Body Data

This patent describes a smart system that uses additional body measurements or relationships between different body chemicals to figure out if an analyte sensor, like a glucose monitor, is giving wrong readings and then fixes it or tells you to replace it.

What it covers

The apparatus includes an analyte sensor, memory, and processor (Claim 1). The processor monitors a patient's analyte, such as glucose (Claim 2), and also gathers other measured sensor data indicative of the patient's physiological state, like oxygen levels or temperature (Claim 3). Based on this physiological state, the system determines what the analyte reading *should* be (expected analyte data), potentially using historical data mapping (Claim 4). If the measured analyte data deviates significantly from this expected data (Claim 5), the system calculates a 'correction factor' (Claim 1, 8) that indicates a calibration error. If recalibration is possible, the system adjusts the sensor, perhaps by changing its sensitivity (Claim 9). If recalibration is not possible, it recommends that the patient replace the sensor (Claim 1). Another approach involves using two different analyte sensors, for example, for glucose and lactate (Claim 10, 12). The system learns the usual relationship between these two analytes and then checks if their current relationship matches the historical one. A substantial mismatch indicates a calibration error in one of the sensors, prompting recalibration (Claim 10). For example, if a continuous glucose monitor (CGM) is showing unusually low glucose readings while the patient's body temperature is high, the system might detect a calibration error and adjust the CGM's readings.

What it doesn't cover

• —Does not cover simple recalibration based solely on a single reference measurement (e.g., a finger-prick blood test) without considering additional physiological data or relationships between analytes.
• —Does not cover systems that only monitor a single analyte without cross-referencing with other measured sensor data indicative of a physiological state or a second analyte.
• —Does not cover systems that detect sensor errors but do not attempt recalibration or recommend replacement to the patient.
• —Does not cover determining expected analyte data based on user-inputted lifestyle factors (like diet or exercise) unless these are derived from 'other measured sensor data indicative of a physiological state'.

The clever bit

The actual novelty is using concurrently measured physiological data (like temperature or oxygen levels, or the relationship between two different analytes) to predict what the primary analyte reading *should* be, and then automatically correcting the sensor's calibration or recommending replacement if there's a significant difference.

Why it matters

This patent is significant for improving the reliability and accuracy of continuous monitoring devices, especially in healthcare. For patients relying on these devices, such as those with diabetes using continuous glucose monitors (CGMs), accurate readings are crucial for managing their health. By automatically detecting and correcting calibration errors, or prompting replacement, it reduces the risk of incorrect medical decisions based on faulty data. Dexcom, the assignee, is a major player in the CGM market.

Real-world examples

• 1.Dexcom G-series continuous glucose monitors
• 2.Abbott FreeStyle Libre systems
• 3.Future multi-analyte wearable sensors
• 4.Integrated health monitoring systems