Generating real-world evidence on diabetes technology using the CareLink Personal data management system

Area of Science:

• Diabetes technology
• Real-world evidence generation
• Health informatics

Background:

• Medtronic's CareLink Personal (CP) system collects data from MiniMed insulin pump users.
• Establishing trust in real-world evidence (RWE) is crucial for device performance and safety assessments.
• Existing RWE methodologies require validation for specific data platforms.

Purpose of the Study:

• To validate the CareLink Personal (CP) database and its analytical processes for generating reliable real-world evidence (RWE).
• To demonstrate the fitness-for-research of the CP system using MiniMed 780G data.
• To establish trust in RWE derived from the CP system for MiniMed users.

Main Methods:

• Described the end-to-end process from raw data to predefined outcomes for MiniMed system users.
• Assessed CP's fitness-for-research using metrics: population representativeness, endpoint relevance, data granularity, completeness, and consistency.
• Utilized one year of MiniMed 780G system data as a case study for validation.

Main Results:

• The data-to-outcomes process was validated, adhering to established definitions.
• Over 91% of the user population is available for analysis, with >95% of users having CP accounts and 96% providing consent.
• CP data offers >50 endpoints, including Time in Range metrics, with high granularity (263 data points/person/day) and automated uploads (>99%).
• Data gaps were minimal (<1 day for 5/6 users, <1 week for 49/50 users).
• In-silico studies showed consistency with real-world data from diverse populations.

Conclusions:

• Real-world evidence derived from the CareLink Personal system adheres to RWE principles.
• The CP system provides robust, reliable data suitable for assessing MiniMed system performance and safety.
• This validation supports the use of CP data for regulatory and clinical decision-making.