1
|
Prendin F, Pavan J, Cappon G, Del Favero S, Sparacino G, Facchinetti A. The importance of interpreting machine learning models for blood glucose prediction in diabetes: an analysis using SHAP. Sci Rep 2023; 13:16865. [PMID: 37803177 PMCID: PMC10558434 DOI: 10.1038/s41598-023-44155-x] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/23/2023] [Accepted: 10/04/2023] [Indexed: 10/08/2023] Open
Abstract
Machine learning has become a popular tool for learning models of complex dynamics from biomedical data. In Type 1 Diabetes (T1D) management, these models are increasingly been integrated in decision support systems (DSS) to forecast glucose levels and provide preventive therapeutic suggestions, like corrective insulin boluses (CIB), accordingly. Typically, models are chosen based on their prediction accuracy. However, since patient safety is a concern in this application, the algorithm should also be physiologically sound and its outcome should be explainable. This paper aims to discuss the importance of using tools to interpret the output of black-box models in T1D management by presenting a case-of-study on the selection of the best prediction algorithm to integrate in a DSS for CIB suggestion. By retrospectively "replaying" real patient data, we show that two long-short term memory neural networks (LSTM) (named p-LSTM and np-LSTM) with similar prediction accuracy could lead to different therapeutic decisions. An analysis with SHAP-a tool for explaining black-box models' output-unambiguously shows that only p-LSTM learnt the physiological relationship between inputs and glucose prediction, and should therefore be preferred. This is verified by showing that, when embedded in the DSS, only p-LSTM can improve patients' glycemic control.
Collapse
Affiliation(s)
- Francesco Prendin
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Jacopo Pavan
- Department of Information Engineering, University of Padova, Padova, Italy
- Department of Psychiatry and Neurobehavioral Sciences, Center for Diabetes Technology, University of Virginia, Charlottesville, VA, USA
| | - Giacomo Cappon
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Simone Del Favero
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Giovanni Sparacino
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Andrea Facchinetti
- Department of Information Engineering, University of Padova, Padova, Italy.
| |
Collapse
|
3
|
Cappon G, Pighin E, Prendin F, Sparacino G, Facchinetti A. A Correction Insulin Bolus Delivery Strategy for Decision Support Systems in Type 1 Diabetes. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2021; 2021:1832-1835. [PMID: 34891643 DOI: 10.1109/embc46164.2021.9630052] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/14/2023]
Abstract
Management of type 1 diabetes (T1D) requires affected individuals to perform multiple daily actions to keep their blood glucose levels within the safe rage and avoid adverse hypo-/hyperglycemic episodes. Decision support systems (DSS) for T1D are composite tools that implement multiple software modules aiming to ease such a burden and to improve glucose control. At the University of Padova, we are developing a new DSS that currently integrate a smart insulin bolus calculator for optimal insulin dosing and a rescue carbohydrate intake advisor to tackle hypoglycemia. However, a module specifically targeting hyperglycemia, that suggests the administration of corrective insulin boluses (CIB), is still missing. For such a scope, this work aims to assess a recent literature methodology, proposed by Aleppo et al., which provides a simple strategy for dealing with hyperglycemia. The methodology is tested retrospectively on clinical data of individuals with T1D. In particular, here we leveraged a novel in silico tool that first identifies a non-linear model of glucose-insulin dynamics on data, then uses such model to simulate and compare the glucose trace obtained by "replaying" the recorded scenario and the glucose trace obtained using the CIB delivery strategy under evaluation. Results show that the CIB delivery strategy significantly reduce the percentage of time spent in hyperglycemia (-15.63%) without inducing any hypoglycemic episode, demonstrating both safety and efficacy of its use. These preliminary results suggest that the CIB delivery strategy proposed by Aleppo et al. is a promising candidate to be included in our system to counteract hyperglycemia. Future work will extensively evaluate the methodology and will compare it against other competing approaches.
Collapse
|
4
|
Camerlingo N, Vettoretti M, Del Favero S, Facchinetti A, Sparacino G. Mathematical Models of Meal Amount and Timing Variability With Implementation in the Type-1 Diabetes Patient Decision Simulator. J Diabetes Sci Technol 2021; 15:346-359. [PMID: 32940087 PMCID: PMC7925444 DOI: 10.1177/1932296820952123] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/17/2022]
Abstract
BACKGROUND In type 1 diabetes (T1D) research, in-silico clinical trials (ISCTs) have proven effective in accelerating the development of new therapies. However, published simulators lack a realistic description of some aspects of patient lifestyle which can remarkably affect glucose control. In this paper, we develop a mathematical description of meal carbohydrates (CHO) amount and timing, with the aim to improve the meal generation module in the T1D Patient Decision Simulator (T1D-PDS) published in Vettoretti et al. METHODS Data of 32 T1D subjects under free-living conditions for 4874 days were used. Univariate probability density function (PDF) parametric models with different candidate shapes were fitted, individually, against sample distributions of: CHO amounts of breakfast (CHOB), lunch (CHOL), dinner (CHOD), and snack (CHOS); breakfast timing (TB); and time between breakfast-lunch (TBL) and between lunch-dinner (TLD). Furthermore, a support vector machine (SVM) classifier was developed to predict the occurrence of a snack in future fixed-length time windows. Once embedded inside the T1D-PDS, an ISCT was performed. RESULTS Resulting PDF models were: gamma (CHOB, CHOS), lognormal (CHOL, TB), loglogistic (CHOD), and generalized-extreme-values (TBL, TLD). The SVM showed a classification accuracy of 0.8 over the test set. The distributions of simulated meal data were not statistically different from the distributions of the real data used to develop the models (α = 0.05). CONCLUSIONS The models of meal amount and timing variability developed are suitable for describing real data. Their inclusion in modules that describe patient behavior in the T1D-PDS can permit investigators to perform more realistic, reliable, and insightful ISCTs.
Collapse
Affiliation(s)
- Nunzio Camerlingo
- Department of Information Engineering,
University of Padova, Padova, Italy
| | - Martina Vettoretti
- Department of Information Engineering,
University of Padova, Padova, Italy
| | - Simone Del Favero
- Department of Information Engineering,
University of Padova, Padova, Italy
| | - Andrea Facchinetti
- Department of Information Engineering,
University of Padova, Padova, Italy
| | - Giovanni Sparacino
- Department of Information Engineering,
University of Padova, Padova, Italy
| |
Collapse
|
5
|
Kalra S, Shaikh S, Priya G, Baruah MP, Verma A, Das AK, Shah M, Das S, Khandelwal D, Sanyal D, Ghosh S, Saboo B, Bantwal G, Ayyagari U, Gardner D, Jimeno C, Barbary NE, Hafidh KA, Bhattarai J, Minulj TT, Zufry H, Bulugahapitiya U, Murad M, Tan A, Shahjada S, Bello MB, Katulanda P, Podgorski G, AbuHelaiqa WI, Tan R, Latheef A, Govender S, Assaad-Khalil SH, Kootin-Sanwu C, Joshi A, Pathan F, Nkansah DA. Individualizing Time-in-Range Goals in Management of Diabetes Mellitus and Role of Insulin: Clinical Insights From a Multinational Panel. Diabetes Ther 2021; 12:465-485. [PMID: 33367983 PMCID: PMC7846622 DOI: 10.1007/s13300-020-00973-0] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 09/17/2020] [Accepted: 11/19/2020] [Indexed: 02/07/2023] Open
Abstract
Diabetes mellitus is a global health concern associated with significant morbidity and mortality. Inadequate control of diabetes leads to chronic complications and higher mortality rates, which emphasizes the importance of achieving glycemic targets. Although glycated hemoglobin (HbA1c) is the gold standard for measuring glycemic control, it has several limitations. Therefore, in recent years, along with the emergence of continuous glucose monitoring (CGM) technology, glycemic control modalities have moved beyond HbA1c. They encompass modern glucometrics, such as glycemic variability (GV) and time-in-range (TIR). The key advantage of these newer metrics over HbA1c is that they allow personalized diabetes management with person-centric glycemic control. Basal insulin analogues, especially second-generation basal insulins with properties such as longer duration of action and low risk of hypoglycemia, have demonstrated clinical benefits by reducing GV and improving TIR. Therefore, for more effective and accurate diabetes management, the development of an integrated approach with second-generation basal insulin and glucometrics involving GV and TIR is the need of the hour. With this objective, a multinational group of endocrinologists and diabetologists reviewed the existing recommendations on TIR, provided their clinical insights into the individualization of TIR targets, and elucidated on the role of the second-generation basal insulin analogues in addressing TIR.
Collapse
Affiliation(s)
- Sanjay Kalra
- Department of Endocrinology, Bharti Hospital and BRIDE, Karnal, Haryana, India.
| | - Shehla Shaikh
- Department of Endocrinology, KGN Institute of Diabetes and Endocrinology, Mumbai, Maharashtra, India
| | - Gagan Priya
- Department of Endocrinology, Fortis Hospital, Chandigarh, Punjab, India
| | - Manas P Baruah
- Department of Endocrinology, Excel Hospital, Guwahati, Assam, India
| | - Abhyudaya Verma
- Endocrine Division, Index Medical College, Indore, Madhya Pradesh, India
| | - Ashok K Das
- Department of Endocrinology and Medicine, Pondicherry Institute of Medical Sciences, Puducherry, India
| | - Mona Shah
- HARMONY Endocrine Diabetes and Metabolic Clinic, Vadodara, Gujarat, India
| | - Sambit Das
- Department of Endocrinology, Apollo Hospitals, Bhubaneswar, Odisha, India
| | - Deepak Khandelwal
- Department of Endocrinology and Diabetes, Maharaja Agrasen Hospital, New Delhi, India
| | - Debmalya Sanyal
- Department of Endocrinology, KPC Medical College, Kolkata, West Bengal, India
| | - Sujoy Ghosh
- Department of Endocrinology and Metabolism, IPGMER, Kolkata, West Bengal, India
| | - Banshi Saboo
- Dia Care, Diabetes Care and Hormone Clinic, Ahmedabad, Gujarat, India
| | - Ganapathi Bantwal
- Department of Endocrinology, St. John's Medical College and Hospital, Bangalore, Karnataka, India
| | - Usha Ayyagari
- Department of Endocrinology, Apollo Sugar Clinics, Chennai, Tamil Nadu, India
| | - Daphne Gardner
- Department of Endocrinology, Singapore General Hospital, Singapore, Singapore
| | - Cecilia Jimeno
- Department of Endocrinology, Philippine Society of Endocrinology, Diabetes and Metabolism, Manila, Philippines
| | - Nancy E Barbary
- Department of Endocrinology, Ain Shams University, Cairo, Egypt
| | - Khadijah A Hafidh
- Department of Endocrinology, Rashid Hospital-Dubai Health Authority, Dubai, UAE
| | - Jyoti Bhattarai
- Department of Endocrinology, Metro Kathmandu Hospital, Kathmandu, Nepal
| | - Tania T Minulj
- Department of Endocrinology, General Hospital Karyadi, Semarang, Indonesia
| | - Hendra Zufry
- Department of Endocrinology, General Hospital Zainoel Abidin, Aceh, Indonesia
| | | | - Moosa Murad
- Department of Internal Medicine, Indira Gandhi Memorial Hospital, Malé, Maldives
| | - Alexander Tan
- Department of Endocrinology, Sunway Medical Centre, Kuala Lumpur, Malaysia
| | - Selim Shahjada
- Department of Endocrinology, BSMMU (Bangabandhu Sheikh Mujib Medical University Hospital), Dhaka, Bangladesh
| | - Mijinyawa B Bello
- Department of Endocrinology, Gwarinpa District Hospital, Abuja, Nigeria
| | - Prasad Katulanda
- Department of Endocrinology, University of Colombo, Colombo, Sri Lanka
| | - Gracjan Podgorski
- Department of Endocrinology, Greenacres Hospital, Port Elizabeth, South Africa
| | | | - Rima Tan
- Department of Endocrinology, FEU-NRMF Medical Center, Quezon City, Philippines
| | - Ali Latheef
- Department of Endocrinology, Indira Gandhi Memorial Hospital, Malé, Maldives
| | | | | | | | - Ansumali Joshi
- Department of Endocrinology, Kathmandu Diabetes and Thyroid Center, Kathmandu, Nepal
| | - Faruque Pathan
- Department of Endocrinology, Bangladesh Institute of Research and Rehabilitation for Diabetes, Endocrine and Metabolic Disorders (BIRDEM), Dhaka, Bangladesh
| | - Diana A Nkansah
- Department of Endocrinology, 37 Military Hospital Accra, Accra, Ghana
| |
Collapse
|
6
|
Francescato MP, Ajčević M, Accardo A. Carbohydrate Requirement for Exercise in Type 1 Diabetes: Effects of Insulin Concentration. J Diabetes Sci Technol 2020; 14:1116-1121. [PMID: 30767503 PMCID: PMC7645145 DOI: 10.1177/1932296819826962] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 12/16/2022]
Abstract
Physical activity is a keystone of a healthy lifestyle as well as of management of patients with type 1 diabetes. The risk of exercise-induced hypoglycemia, however, is a great challenge for these patients. The glycemic response to exercise depends upon several factors concerning the patient him/herself (eg, therapy, glycemic control, training level) and the characteristics of the exercise performed. Only in-depth knowledge of these factors will allow to develop individualized strategies minimizing the risk of hypoglycemia. The main factors affecting the exercise-induced hypoglycemia in patients with T1D have been analyzed, including the effects of insulin concentration. A model is discussed, which has the potential to become the basis for providing patients with individualized suggestions to keep constant glucose levels on each exercise occasion.
Collapse
Affiliation(s)
- Maria Pia Francescato
- Department of Medicine, University of Udine, Udine, Italy
- Maria Pia Francescato, MD, Department of Medicine, University of Udine, p. le M. Kolbe 4, 33100 Udine, Italy.
| | - Miloš Ajčević
- Department of Engineering and Architecture, University of Trieste, Trieste, Italy
| | - Agostino Accardo
- Department of Engineering and Architecture, University of Trieste, Trieste, Italy
| |
Collapse
|
7
|
Cappon G, Facchinetti A, Sparacino G, Georgiou P, Herrero P. Classification of Postprandial Glycemic Status with Application to Insulin Dosing in Type 1 Diabetes-An In Silico Proof-of-Concept. SENSORS 2019; 19:s19143168. [PMID: 31323886 PMCID: PMC6679291 DOI: 10.3390/s19143168] [Citation(s) in RCA: 11] [Impact Index Per Article: 2.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 06/21/2019] [Revised: 07/15/2019] [Accepted: 07/16/2019] [Indexed: 02/06/2023]
Abstract
In the daily management of type 1 diabetes (T1D), determining the correct insulin dose to be injected at meal-time is fundamental to achieve optimal glycemic control. Wearable sensors, such as continuous glucose monitoring (CGM) devices, are instrumental to achieve this purpose. In this paper, we show how CGM data, together with commonly recorded inputs (carbohydrate intake and bolus insulin), can be used to develop an algorithm that allows classifying, at meal-time, the post-prandial glycemic status (i.e., blood glucose concentration being too low, too high, or within target range). Such an outcome can then be used to improve the efficacy of insulin therapy by reducing or increasing the corresponding meal bolus dose. A state-of-the-art T1D simulation environment, including intraday variability and a behavioral model, was used to generate a rich in silico dataset corresponding to 100 subjects over a two-month scenario. Then, an extreme gradient-boosted tree (XGB) algorithm was employed to classify the post-prandial glycemic status. Finally, we demonstrate how the XGB algorithm outcome can be exploited to improve glycemic control in T1D through real-time adjustment of the meal insulin bolus. The proposed XGB algorithm obtained good accuracy at classifying post-prandial glycemic status (AUROC = 0.84 [0.78, 0.87]). Consequently, when used to adjust, in real-time, meal insulin boluses obtained with a bolus calculator, the proposed approach improves glycemic control when compared to the baseline bolus calculator. In particular, percentage time in target [70, 180] mg/dL was improved from 61.98 (±13.89) to 67.00 (±11.54; p < 0.01) without increasing hypoglycemia.
Collapse
Affiliation(s)
- Giacomo Cappon
- Department of Information Engineering, University of Padova, 35131 Padova (PD), Italy
| | - Andrea Facchinetti
- Department of Information Engineering, University of Padova, 35131 Padova (PD), Italy
| | - Giovanni Sparacino
- Department of Information Engineering, University of Padova, 35131 Padova (PD), Italy
| | - Pantelis Georgiou
- Department of Electrical and Electronical Engineering, Imperial College London, London W5 5SA, UK
| | - Pau Herrero
- Department of Electrical and Electronical Engineering, Imperial College London, London W5 5SA, UK.
| |
Collapse
|
8
|
Vettoretti M, Cappon G, Acciaroli G, Facchinetti A, Sparacino G. Continuous Glucose Monitoring: Current Use in Diabetes Management and Possible Future Applications. J Diabetes Sci Technol 2018; 12:1064-1071. [PMID: 29783897 PMCID: PMC6134613 DOI: 10.1177/1932296818774078] [Citation(s) in RCA: 52] [Impact Index Per Article: 8.7] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 12/20/2022]
Abstract
The recent announcement of the production of new low-cost continuous glucose monitoring (CGM) sensors, the approval of marketed CGM sensors for making treatment decisions, and new reimbursement criteria have the potential to revolutionize CGM use. After briefly summarizing current CGM applications, we discuss how, in our opinion, these changes are expected to extend CGM utilization beyond diabetes patients, for example, to subjects with prediabetes or even healthy individuals. We also elaborate on how the integration of CGM data with other relevant information, for example, health records and other medical device/wearable sensor data, will contribute to creating a digital data ecosystem that will improve our understanding of the etiology and complications of diabetes and will facilitate the development of data analytics for personalized diabetes management and prevention.
Collapse
Affiliation(s)
- Martina Vettoretti
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Giacomo Cappon
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Giada Acciaroli
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Andrea Facchinetti
- Department of Information Engineering, University of Padova, Padova, Italy
| | - Giovanni Sparacino
- Department of Information Engineering, University of Padova, Padova, Italy
- Giovanni Sparacino, PhD, Department of Information Engineering University of Padova, Via G. Gradenigo 6B, Padova, 35131, Italy.
| |
Collapse
|