Beyond Expected Patterns in Insulin Needs of People With Type 1 Diabetes: Temporal Analysis of Automated Insulin Delivery Data
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Background: Type 1 diabetes (T1D) is a chronic condition in which the body produces too little insulin, a hormone needed to regulate blood glucose. Various factors such as carbohydrates, exercise, and hormones impact insulin needs. Beyond carbohydrates, most factors remain underexplored. Regulating insulin is a complex control task that can go wrong and cause blood glucose levels to fall outside a range that protects people from adverse health effects. Automated insulin delivery (AID) has been shown to maintain blood glucose levels within a narrow range. Beyond clinical outcomes, data from AID systems are little researched; such systems can provide data-driven insights to improve the understanding and treatment of T1D. Objective: The aim is to discover unexpected temporal patterns in insulin needs and to analyze how frequently these occur. Unexpected patterns are situations where increased insulin does not result in lower glucose or where increased carbohydrate intake does not raise glucose levels. Such situations suggest that factors beyond carbohydrates influence insulin needs. Methods: We analyzed time series data on insulin on board (IOB), carbohydrates on board (COB), and interstitial glucose (IG) from 29 participants using the OpenAPS AID system. Pattern frequency in hours, days (grouped via k-means clustering), weekdays, and months were determined by comparing the 95% CI of the mean differences between temporal units. Associations between pattern frequency and demographic variables were examined. Significant differences in IOB, COB, and IG across temporal dichotomies were assessed using Mann-Whitney U tests. Effect sizes and Euclidean distances between variables were calculated. Finally, the forecastability of IOB, COB, and IG for the clustered days was analyzed using Granger causality. Results: On average, 13.5 participants had unexpected patterns and 9.9 had expected patterns. The patterns were more pronounced (d>0.94) when comparing hours of the day and similar days than when comparing days of the week or months (0.3<d<0.52). Notably, 11 participants exhibited a higher IG overnight despite concurrently higher IOB (10/11). Additionally, 17 participants experienced an increase in IG after COB decreased after meals. The significant associations between pattern frequency and demographics were moderate (0.31≤τ≤0.48). Between clusters, mean IOB (P=.03, d=0.7) and IG (P=.02, d=0.67) differed significantly, but COB did not (P=.08, d=0.55). IOB and IG were most similar (mean distance 5.08, SD 2.25), while COB and IG were most different (mean distance 11.43, SD 2.6), suggesting that AID attempts to counteract both observed and unobserved factors that impact IG. Conclusions: Our study shows that unexpected patterns in the insulin needs of people with T1D are as common as expected patterns. Unexpected patterns cannot be explained by carbohydrates alone. Our results highlight the complexity of glucose regulation and emphasize the need for personalized treatment approaches. Further research is needed to identify and quantify the factors that cause these patterns.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.002 | 0.006 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it