Frontiers in endocrinology · 2024
Prediction of personalised postprandial glycaemic response in type 1 diabetes mellitus.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Xiong Xin, Xue Yuxin, Cai Yunying, He Jianfeng, Su Heng
Аннотация
OBJECTIVES: Patients with type 1 diabetes (T1D) face unique challenges in glycaemic control due to the complexity and uniqueness of the dietary structure in China, especially in terms of postprandial glycaemic response (PPGR). This study aimed to establish a personalized model for predicting PPGR in patients with T1D.
MATERIALS AND METHODS: Data provided by the First People's Hospital of Yunnan Province, 13 patients with T1D, were recruited and provided with an intervention for at least two weeks. All patients were asked to wear a continuous glucose monitoring (CGM) device under free-living conditions during the study period. To tackle the challenge of incomplete data from wearable devices for CGM measurements, the GAIN method was used in this paper to achieve a more rational interpolation process. In this study, patients' PPGRs were calculated, and a LightGBM prediction model was constructed based on a Bayesian hyperparameter optimisation algorithm and a random search algorithm, which integrated glucose measurement, insulin dose, dietary nutrient content, blood measurement and anthropometry as inputs.
RESULTS: The experimental outcomes revealed that the PPGR prediction model presented in this paper demonstrated superior accuracy (R=0.63) compared to both the carbohydrate content only model (R=0.14) and the baseline model emulating the standard of care for insulin administration (R=0.43). In addition, the interpretation of the model using the SHAP method showed that blood glucose levels at meals and blood glucose trends 30 minutes before meals were the most important features of the model.
CONCLUSION: The proposed model offers a heightened precision in predicting PPGR in patients with T1D, so it can better guide the diet plan and insulin intake dose of patients with T1D.
MATERIALS AND METHODS: Data provided by the First People's Hospital of Yunnan Province, 13 patients with T1D, were recruited and provided with an intervention for at least two weeks. All patients were asked to wear a continuous glucose monitoring (CGM) device under free-living conditions during the study period. To tackle the challenge of incomplete data from wearable devices for CGM measurements, the GAIN method was used in this paper to achieve a more rational interpolation process. In this study, patients' PPGRs were calculated, and a LightGBM prediction model was constructed based on a Bayesian hyperparameter optimisation algorithm and a random search algorithm, which integrated glucose measurement, insulin dose, dietary nutrient content, blood measurement and anthropometry as inputs.
RESULTS: The experimental outcomes revealed that the PPGR prediction model presented in this paper demonstrated superior accuracy (R=0.63) compared to both the carbohydrate content only model (R=0.14) and the baseline model emulating the standard of care for insulin administration (R=0.43). In addition, the interpretation of the model using the SHAP method showed that blood glucose levels at meals and blood glucose trends 30 minutes before meals were the most important features of the model.
CONCLUSION: The proposed model offers a heightened precision in predicting PPGR in patients with T1D, so it can better guide the diet plan and insulin intake dose of patients with T1D.