# Cluster explanations and source evidence Collection: `device_events_20k_minilm`. Query: `continuous glucose monitoring problems`. Threshold: 0.4. Saved analysis: `job_9b9d4e2409d646868cb6fb32e12cc32a`. Explanations below are editorial summaries of representative source narratives. They are not replacement automatic labels, validated root causes or counts of distinct incidents. | Cluster | Reports | Editorial explanation | | ------: | ------: | --------------------------------------------------------------------------- | | 0 | 685 | Low readings and an existing field-action reference | | 2 | 317 | Reading discrepancies with products requested for investigation | | 4 | 246 | CGM-versus-meter discrepancies and investigation wording | | 9 | 112 | Erroneous readings, product-return difficulties and investigation templates | | 6 | 83 | Monitor error 42 with reported pump resets | | 3 | 58 | Cartridge alarms with a reported change in delivery method | | 5 | 55 | Malfunction alarms reported as cleared after support troubleshooting | | 10 | 40 | Malfunction alarms with a reported change in delivery method | | 8 | 27 | Intermittent occlusion alarms with continued pump use | | 12 | 20 | Returned-sensor investigations discussing normal termination | | 1 | 16 | Minimum-fill notifications during cartridge loading | | 7 | 16 | Occlusion alarms with a reported change in delivery method | | 11 | 16 | Returned-sensor investigations discussing error-triggered termination | Assigned: 1,691. Unassigned: 890. Total: 2,581. ## Representative reports ### Cluster 0: Low readings and an existing field-action reference Reports: 685. Automatic label: `Fa1002 2025`. Representative original report IDs: `25011860`, `25483494`. Source excerpt: > A CUSTOMER REPORTED RECEIVING LOW GLUCOSE RESULTS FROM AN ABBOTT DIABETES CARE (ADC) DEVICE. THIS IS A KNOWN ISSUE FOR THE SERIAL NUMBER ASSOCIATED WITH THIS COMPLAINT, ADDRESSED BY ADC FIELD ACTION FA1002-2025. IMPACTED PRODUCT WAS DISTRIBUTED TO ANDORRA, AUSTRIA, BELGIUM, CANADA, DENMARK, FINLAND, FRANCE, GERMANY, ITALY, LUXEMBOURG, NETHERLANDS, NEW ZEALAND, NORWAY, SAN MARINO, SPAIN, SWEDEN, SWITZERLAND, UK, US, AND QATAR. ADC FIELD ACTION FA1002-2025 WAS ISSUED ONLY TO IMPACTED COUNTRIES. THIS REPORT IS BEING SUBMITTED AS PART OF RETROSPECTIVE REPORTING RELATED TO FA1002-2025. THERE WAS NO REPORT OF DEATH, SERIOUS INJURY, OR MISTREATMENT ASSOCIATED WITH THIS EVENT. ADC INVESTIGATED THIS ISSUE AND DETERMINED THAT THE SENSOR ASSOCIATED WITH THIS COMPLAINT MAY NOT MEET PRODUCT DESIGN SPECIFICATIONS AND MAY PRODUCE LOW GLUCOSE READINGS WHICH ARE NOT CLINICALLY ACCEPTABLE. THIS PRODUCT HAS BEEN DETERMINED TO BE WITHIN THE SCOPE OF THE INVESTIGATION AND ACTIONS CONDUCTED UNDER ADC FIELD … ### Cluster 2: Reading discrepancies with products requested for investigation Reports: 317. Automatic label: `Mg dl`. Representative original report IDs: `25439888`, `25666163`. Source excerpt: > THE PRODUCT HAS BEEN REQUESTED BACK FOR INVESTIGATION. A FOLLOW-UP REPORT WILL BE SUBMITTED UPON COMPLETION OF INVESTIGATION ACTIVITIES OR IF ADDITIONAL INFORMATION IS OBTAINED. ALL PERTINENT INFORMATION AVAILABLE TO ABBOTT DIABETES CARE HAS BEEN SUBMITTED. A CUSTOMER REPORTED RECEIVING ERRONEOUS GLUCOSE RESULTS FROM AN ABBOTT DIABETES CARE DEVICE. THE CUSTOMER RECEIVED SENSOR SCAN RESULTS OF 58.00 MG/DL COMPARED TO READINGS OF 160.00 MG/DL RESPECTIVELY OBTAINED ON A METER AND THE RESULTS, WHEN PLOTTED ON A PARKES ERROR GRID, FALL INTO THE C ZONE, SHOWING THE DIFFERENCE IN VALUES TO BE CLINICALLY SIGNIFICANT. THE LENGTH OF SENSOR WEAR WAS 3 DAYS. THERE WAS NO REPORT OF DEATH, SERIOUS INJURY, OR MISTREATMENT ASSOCIATED WITH THIS EVENT. ### Cluster 4: CGM-versus-meter discrepancies and investigation wording Reports: 246. Automatic label: `Zone parkes`. Representative original report IDs: `25103203`, `25075186`. Source excerpt: > IT WAS REPORTED THAT AN INACCURACY BETWEEN THE CONTINUOUS GLUCOSE MONITOR (CGM) AND THE BLOOD GLUCOSE (BG) METER OCCURRED. THE SENSOR WAS INSERTED INTO THE ARM, WHICH IS OFF-LABEL USAGE OF THE DEVICE, ON (B)(6) 2026. DATA WAS EVALUATED AND THE ALLEGATION WAS CONFIRMED. THE PROBABLE CAUSE COULD NOT BE DETERMINED VIA DATA. THE REPORTED GLUCOSE VALUES FALL WITHIN THE D ZONE OF THE PARKES ERROR GRID. THE DATA INVESTIGATION FOUND A DIFFERENCE IN VALUES THAT FALL WITHIN THE A ZONE OF THE PARKES ERROR GRID. NO INJURY OR MEDICAL INTERVENTION WAS REPORTED. (B)(4). ### Cluster 9: Erroneous readings, product-return difficulties and investigation templates Reports: 112. Automatic label: `Risk profiles`. Representative original report IDs: `24963058`, `24855202`. Source excerpt: > A CUSTOMER REPORTED RECEIVING ERRONEOUS GLUCOSE RESULTS FROM AN ABBOTT DIABETES CARE DEVICE. THE CUSTOMER RECEIVED SENSOR SCAN RESULTS OF 49.00 MG/DL COMPARED TO READINGS OF 172.00 MG/DL RESPECTIVELY OBTAINED ON A METER AND THE RESULTS, WHEN PLOTTED ON A PARKES ERROR GRID, FALL INTO THE C ZONE, SHOWING DIFFERENCE IN VALUES TO BE CLINICALLY SIGNIFICANT. THE LENGTH OF SENSOR WEAR WAS UNKNOWN. THERE WAS NO REPORT OF DEATH, SERIOUS INJURY, OR MISTREATMENT ASSOCIATED WITH THIS EVENT. REPEATED ATTEMPTS BY ADC TO RETRIEVE THE PRODUCT WERE UNSUCCESSFUL AND/OR THE CUSTOMER DISCARDED THE PRODUCT. THE MOST PROBABLE ROOT CAUSES ASSOCIATED WITH THIS FAILURE MODE ARE DISCONNECTED, FAULTY OR DAMAGED COMPONENTS, SOFTWARE/DATA CORRUPTION, OR MISUSE. HOWEVER, MITIGATIONS ARE IN PLACE TO REDUCE AND PREVENT SUCH ISSUES. ALL VITAL MANUFACTURING STEPS ARE VALIDATED, MONITORED, AND VERIFIED DURING MANUFACTURING TO ENSURE THE SYSTEM IS IN CONFORMANCE WITH THE VERIFIED DESIGN. ALL VITAL FUNCTIONS ARE MONITORED… ### Cluster 6: Monitor error 42 with reported pump resets Reports: 83. Automatic label: `42 pump`. Representative original report IDs: `25283977`, `25184404`. Source excerpt: > IT WAS REPORTED THAT THE CUSTOMER RECEIVED A CONTINUOUS GLUCOSE MONITOR ERROR 42. THE PUMP WAS RESET, AND THE CUSTOMER CONTINUED TO USE THE PUMP FOR INSULIN THERAPY. THERE WAS NO REPORTED ADVERSE IMPACT TO THE CUSTOMER. ### Cluster 3: Cartridge alarms with a reported change in delivery method Reports: 58. Automatic label: `Occurred insulin`. Representative original report IDs: `25010427`, `25315914`. Source excerpt: > IT WAS REPORTED THAT A CARTRIDGE ALARM OCCURRED DURING INSULIN DELIVERY. CUSTOMER REVERTED TO AN ALTERNATE METHOD OF INSULIN THERAPY. THERE WAS NO ADVERSE IMPACT TO THE CUSTOMER¿S BLOOD GLUCOSE LEVEL. ### Cluster 5: Malfunction alarms reported as cleared after support troubleshooting Reports: 55. Automatic label: `Malfunction alarm`. Representative original report IDs: `25136001`, `24792046`. Source excerpt: > IT WAS REPORTED THAT A MALFUNCTION ALARM OCCURRED. DURING TROUBLESHOOTING WITH TECHNICAL SUPPORT, THE MALFUNCTION ALARM WAS CLEARED, AND INSULIN DELIVERY WAS RESUMED SUCCESSFULLY. THERE WAS NO ADVERSE IMPACT TO THE CUSTOMER¿S BLOOD GLUCOSE LEVEL. ### Cluster 10: Malfunction alarms with a reported change in delivery method Reports: 40. Automatic label: `Reported malfunction`. Representative original report IDs: `24925877`, `25204115`. Source excerpt: > IT WAS REPORTED THAT A MALFUNCTION ALARM OCCURRED. THE CUSTOMER REVERTED TO AN ALTERNATE METHOD OF INSULIN THERAPY. THERE WAS NO REPORTED ADVERSE IMPACT TO THE CUSTOMER'S BLOOD GLUCOSE LEVEL. ### Cluster 8: Intermittent occlusion alarms with continued pump use Reports: 27. Automatic label: `Insulin therapy`. Representative original report IDs: `25347858`, `25185050`. Source excerpt: > IT WAS REPORTED THAT ONGOING INTERMITTENT OCCLUSION ALARMS OCCURRED. THERE WAS NO ADVERSE IMPACT TO THE CUSTOMER¿S BLOOD GLUCOSE LEVEL. THE CUSTOMER CONTINUED TO USE THE PUMP FOR INSULIN THERAPY AND HAS AN ALTERNATE METHOD OF INSULIN THERAPY AVAILABLE IF NECESSARY. ### Cluster 12: Returned-sensor investigations discussing normal termination Reports: 20. Automatic label: `Sensor`. Representative original report IDs: `25113169`, `25514981`. Source excerpt: > SENSOR (B)(6) HAS BEEN RETURNED AND INVESTIGATED. THE SENSOR PLUG IS FULLY SEATED AND NO ISSUES WERE OBSERVED ON THE SENSOR PATCH. EXTRACTED DATA FROM THE RETURNED SENSOR USING APPROVED SOFTWARE. THE SENSOR WAS FOUND TO BE IN SENSOR STATE 6 (INDICATING EARLY TERMINATION). SENSOR STATE 6 WITH EVENT LOGS 6 AND 7 ARE AN INDICATION THAT THE PATCH WAS TERMINATED WITHOUT ANY ERROR. THE EVENT LOGS OBSERVED ARE CONSISTENT WITH NORMAL TERMINATION AND INDICATE THE SENSOR HAS REACHED ITS END OF LIFE. VISUAL INSPECTION HAS BEEN PERFORMED ON THE SENSOR PLUG ASSEMBLY, NO FAILURE MODES WERE OBSERVED. PERFORMED AN SMU (SOURCE MEASUREMENT UNIT) TEST TO ENSURE THE SENSOR'S ELECTRONICS WERE FUNCTIONING CORRECTLY AND THE RETURNED UNIT DID NOT HAVE ANY GLUCOSE READING ISSUES. POISE VOLTAGE AND SENSOR THERMISTOR TESTING WERE BOTH WITHIN SPECIFICATION, INDICATING THE SENSOR WAS PROVIDING ACCURATE GLUCOSE READINGS. THE PASSING OF FUNCTIONALITY TESTING IS AN INDICATION THAT THERE WERE NO ISSUES WITH SENSOR FUN… ### Cluster 1: Minimum-fill notifications during cartridge loading Reports: 16. Automatic label: `Insulin load`. Representative original report IDs: `25342278`, `24988111`. Source excerpt: > IT WAS REPORTED THAT A MINIMUM FILL NOTIFICATION OCCURRED AFTER THE USER FILLED THE CARTRIDGE WITH 250 UNITS OF INSULIN DURING THE LOAD SEQUENCE. THE CUSTOMER RE-LOADED THE CARTRIDGE TO RESOLVE THE ISSUE. THERE WAS NO ADVERSE IMPACT TO THE CUSTOMER¿S BLOOD GLUCOSE LEVEL. ### Cluster 7: Occlusion alarms with a reported change in delivery method Reports: 16. Automatic label: `Occlusion alarms`. Representative original report IDs: `25494281`, `25524262`. Source excerpt: > IT WAS REPORTED THAT ONGOING OCCLUSION ALARMS OCCURRED. THERE WAS NO REPORTED ADVERSE IMPACT TO CUSTOMER¿S BLOOD GLUCOSE LEVEL. REPORTEDLY, THE CUSTOMER REVERTED TO AN ALTERNATE METHOD OF INSULIN THERAPY ### Cluster 11: Returned-sensor investigations discussing error-triggered termination Reports: 16. Automatic label: `Sensor`. Representative original report IDs: `24883593`, `24962490`. Source excerpt: > A CUSTOMER REPORTED RECEIVING ERRONEOUS GLUCOSE RESULTS FROM AN ABBOTT DIABETES CARE DEVICE. THE CUSTOMER RECEIVED SENSOR SCAN RESULTS OF 47.00 MG/DL COMPARED TO READINGS OF 180.00 MG/DL RESPECTIVELY OBTAINED ON A METER AND THE RESULTS, WHEN PLOTTED ON A PARKES ERROR GRID, FALL INTO THE C ZONE, SHOWING THE DIFFERENCE IN VALUES TO BE CLINICALLY SIGNIFICANT. THE LENGTH OF SENSOR WEAR WAS 8 DAYS. THERE WAS NO REPORT OF DEATH, SERIOUS INJURY, OR MISTREATMENT ASSOCIATED WITH THIS EVENT. SENSOR (B)(6) HAS BEEN RETURNED AND INVESTIGATED. VISUAL INSPECTION PERFORMED ON THE RETURNED SENSOR PATCH AND NO ISSUE WAS OBSERVED. DATA WAS EXTRACTED USING APPROVED SOFTWARE AND EXTRACTION WAS SUCCESSFUL. SENSOR STATE 7 WITH EVENT LOG 11 AND SENSOR STATE 8 WITH EVENT LOG 9 ARE AN INDICATION THAT THE SENSOR HAD TERMINATED DUE TO AN ERROR RECOGNIZED BY THE SENSOR. THIS IS A PART OF SOFTWARE DESIGN AND IS NOT AN INDICATION OF PRODUCT NON-CONFORMANCE. FUNCTIONALITY TEST WILL BE PERFORMED ON THE SENSOR TO VERI… ## Checked across cluster membership - Cluster 0: all 685 narrative texts contain `fa1002` (case-insensitive literal matching). - Cluster 6: all 83 narrative texts contain `error 42` (case-insensitive literal matching). - Cluster 3: all 58 narrative texts contain `cartridge alarm` (case-insensitive literal matching).