Imputation Model for Glucose Values Above the Upper Detection Limit for Continuous Glucose Monitors
Authors' conclusion
An imputation model for glucose values above the upper detection limit of CGMs was developed and validated in various populations.
Study context
- Population
- Inpatients with T2D and outpatients with T1D or T2D using CGM
- Participants
- 85 / 705 / 27
- Comparison
- Imputed vs uncensored CGM data
- Hospitalized T2Dn = 85
- Outpatients T1Dn = 705
- Outpatients T2Dn = 27
- Outcome
- Bias, MSE and R2 of mean glucose, SD and CV after imputation vs uncensored CGM data
Methodological detailsBayesian nonparametric latent Gaussian process regression model, intentional right censoring of top 5%, 10%, 20% and 30%
- Bayesian nonparametric latent Gaussian process regression model
- Intentional right censoring of top 5%, 10%, 20% and 30%
Performance
As reported in the abstract
Bias after imputation, 5% to 30% right censoring
| Test or model | Bias on mean glucose | Bias on SD | Bias on coefficient of variation (CV) |
|---|---|---|---|
| Hospitalized T2D | −0.012 to 0.362 | −0.024 to 0.226 | −0.207 to 1.543 |
| Outpatients T1D | −0.018 to 0.485 | −0.033 to 0.381 | −0.316 to 2.609 |
| Outpatients T2D | −0.008 to 0.130 | −0.016 to 0.138 | −0.222 to 1.721 |
Bias on mean glucose
Hospitalized T2D−0.012 to 0.362
Outpatients T1D−0.018 to 0.485
Outpatients T2D−0.008 to 0.130
Bias on SD
Hospitalized T2D−0.024 to 0.226
Outpatients T1D−0.033 to 0.381
Outpatients T2D−0.016 to 0.138
Bias on coefficient of variation (CV)
Hospitalized T2D−0.207 to 1.543
Outpatients T1D−0.316 to 2.609
Outpatients T2D−0.222 to 1.721
Original publication
Imputation Model for Glucose Values Above the Upper Detection Limit for Continuous Glucose Monitors. Diabetes Technol Ther. Published 21 May 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →