Personalized insulin dosing using reinforcement learning for high-fat meals and aerobic exercises in type 1 diabetes: a proof-of-concept trial
Bibliographic record
Abstract
In type 1 diabetes, high-fat meals require more insulin to prevent hyperglycemia while meals followed by aerobic exercises require less insulin to prevent hypoglycemia, but the adjustments needed vary between individuals. We propose a decision support system with reinforcement learning to personalize insulin doses for high-fat meals and postprandial aerobic exercises. We test this system in a single-arm 16-week study in 15 adults on multiple daily injections therapy (NCT05041621). The primary objective of this study is to assess the feasibility of the novel learning algorithm. This study looks at glucose outcomes and patient reported outcomes. The postprandial incremental area under the glucose curve is improved from the baseline to the evaluation period for high-fat meals (378 ± 222 vs 38 ± 223 mmol/L/min, p = 0.03) and meals followed by exercises (−395 ± 192 vs 132 ± 181 mmol/L/min, p = 0.007). The postprandial time spent below 3.9 mmol/L is reduced after high-fat meals (5.3 ± 1.6 vs 1.8 ± 1.5%, p = 0.003) and meals followed by exercises (5.3 ± 1.2 vs 1.4 ± 1.1%, p = 0.003). Our study shows the feasibility of automatically personalizing insulin doses for high-fat meals and postprandial exercises. Randomized controlled trials are warranted. Individuals with type 1 diabetes face challenges in determining the insulin doses they need before high-fat meals and meals followed by aerobic exercises. Here, the authors show that it is clinically feasible to automatically determine these insulin doses for each individual using a reinforcement learning algorithm.
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.
How this classification was reachedexpand
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".