MED-PUMP-010 medical data_error ai_generated true

AI programs an insulin pump with a 24-hour basal rate profile that ignores circadian insulin sensitivity, causing nocturnal hypoglycemia and dawn phenomenon hyperglycemia

ID: medical/insulin-pump-circadian-basal-rate-miss

Also available as: JSON · Markdown · 中文
78%Fix Rate
87%Confidence
1Evidence
2023-09-20First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
Medtronic MiniMed 780G 2.0 active — — —
Tandem t:slim X2 1.5 active — — —
Insight 1.4.2 active — — —

Root Cause

Circadian insulin sensitivity varies: early morning (dawn phenomenon) requires higher basal rate, while midnight to 3 AM requires lower basal rate to prevent hypoglycemia. AI often uses a flat or incorrectly shaped basal profile, ignoring these physiological patterns.

generic

中文

昼夜胰岛素敏感性不同：清晨（黎明现象）需较高基础率，午夜至凌晨 3 点需较低基础率以防低血糖。AI 常使用平坦或形状错误的基础率曲线，忽略这些生理模式。

Official Documentation

https://www.diabetes.org.uk/guide-to-diabetes/managing-your-diabetes/insulin-pumps/basal-rates

Workarounds

  1. 88% success Program a multi-step basal profile with 2-3 hour segments. Example for typical adult: 00:00-03:00: 0.8 U/h, 03:00-06:00: 1.0 U/h, 06:00-09:00: 1.3 U/h, 09:00-12:00: 1.1 U/h, 12:00-18:00: 1.0 U/h, 18:00-24:00: 0.9 U/h. Then fine-tune based on CGM data over 3-5 days. 
    Program a multi-step basal profile with 2-3 hour segments. Example for typical adult: 00:00-03:00: 0.8 U/h, 03:00-06:00: 1.0 U/h, 06:00-09:00: 1.3 U/h, 09:00-12:00: 1.1 U/h, 12:00-18:00: 1.0 U/h, 18:00-24:00: 0.9 U/h. Then fine-tune based on CGM data over 3-5 days.
  2. 92% success Use automated basal rate optimization via hybrid closed-loop algorithm (e.g., Control-IQ or SmartGuard) that adjusts basal every 5 minutes based on CGM trend. This bypasses manual profile errors. 
    Use automated basal rate optimization via hybrid closed-loop algorithm (e.g., Control-IQ or SmartGuard) that adjusts basal every 5 minutes based on CGM trend. This bypasses manual profile errors.

中文步骤

  1. Program a multi-step basal profile with 2-3 hour segments. Example for typical adult: 00:00-03:00: 0.8 U/h, 03:00-06:00: 1.0 U/h, 06:00-09:00: 1.3 U/h, 09:00-12:00: 1.1 U/h, 12:00-18:00: 1.0 U/h, 18:00-24:00: 0.9 U/h. Then fine-tune based on CGM data over 3-5 days.
  2. Use automated basal rate optimization via hybrid closed-loop algorithm (e.g., Control-IQ or SmartGuard) that adjusts basal every 5 minutes based on CGM trend. This bypasses manual profile errors.

Dead Ends

Common approaches that don't work:

  1. Use a single flat basal rate for 24 hours to simplify programming 80% fail

    Flat rate causes nocturnal hypoglycemia (too high at 2 AM) and morning hyperglycemia (too low at 6 AM) in most patients

  2. Set lowest basal rate at midnight and highest at noon, following a simple sine wave 65% fail

    The physiological nadir is 2-3 AM, not midnight; and peak is 6-8 AM (dawn phenomenon), not noon. A sine wave is too simplistic and mismatches real patterns

  3. Use the same basal profile for all patients, adjusting only total daily dose 90% fail

    Circadian patterns vary significantly between individuals; a one-size-fits-all profile causes poor glycemic control