The Use of Machine Learning Detection of Lipohypertrophy to Improve Glycemic Variability
The Use of Machine Learning Detection of Lipohypertrophy to Improve Glycemic Variability
Lipohypertrophy is swelling of the fatty tissue located below the skin ("subcutaneous tissue") where many patients with diabetes inject their insulin. Lipohypertrophy can sometimes be felt as firm swelling, lumps or small bumps near insulin injection sites. Previous studies have shown that injecting insulin into areas of lipohypertrophy can affect how insulin is absorbed, and can increase insulin requirements in patients. New data suggest that lipohypertrophy can be detected using ultrasound technology. The ultrasonographic presence of changes to the subcutaneous tissue without swelling that can be felt ("subclinical lipohypertrophy") and the effect of injecting insulin into these sites is unknown.
100 people will participate in the Phase 1 of this study. In the second phase of the study, 40 patients identified with subclinical lipohypertrophy in Phase 1 will be asked to participate in the randomized study using crossover design by checking your glucose levels.
Purpose:
In Phase 1: to use computer based technology to detect lipohypertrophy on portable ultrasound images.
In Phase 2: to demonstrate that injecting insulin into areas to be free of lipohypertrophy will have better glucose control as compared to injecting into areas demonstrated to have lipohypertrophy.
Eligibility:
You can participate in this study if:
You should not participate in this study if:
Inclusion Criteria:
Exclusion Criteria: