In children using intensive insulin therapy, a 20-g variation in carbohydrate amount significantly impacts on postprandial glycaemia

Automated Artificial Intelligence-Based Thai Food Dietary Assessment System: Development and Validation

Background

Dietary assessment is a fundamental component of nutrition research and plays a pivotal role in managing chronic diseases. Traditional dietary assessment methods, particularly in the context of Thai cuisine, often require extensive training and may lead to estimation errors.

Objectives

To address these challenges, Institute of Nutrition, Mahidol University (INMU) iFood, an innovative artificial intelligence-based Thai food dietary assessment system, allows for estimating the nutritive values of dishes from food images.

Methods

INMU iFood leverages state-of-the-art technology and integrates a validated automated Thai food analysis system. Users can use 3 distinct input methods: food image recognition, manual input, and a convenient barcode scanner. This versatility simplifies the tracking of dietary intake while maximizing data quality at the individual level. The core improvement in INMU iFood can be attributed to 2 key factors, namely, the replacement of Yolov4-tiny with Yolov7 and the expansion of noncarbohydrate source foods in the training image data set.

Results

This combination significantly enhances the system's ability to identify food items, especially in scenarios with closely packed food images, thus improving accuracy. Validation results showcase the superior performance of the INMU iFood integrated V7-based system over its predecessor, V4-based, with notable improvements in protein and fat estimation. Furthermore, INMU iFood addresses limitations by offering users the option to import additional food products via a barcode scanner, thus providing access to a vast database of nutritional information through Open Food Facts. This integration ensures users can track their dietary intake effectively, with expanded access to over 3000 food items added to or updated in the Open Food Facts database covering a wide variety of dietary choices.

Conclusions

INMU iFood is a promising tool for researchers, health care professionals, and individuals seeking to monitor their dietary intake within the context of Thai cuisine and for ultimately promoting better health outcomes and facilitating nutrition-related research.

Accuracy and Feasibility of Using a Smartphone Application for Carbohydrate Counting Versus Traditional Carbohydrate Counting for Adults With Insulin-Treated Diabetes

BACKGROUND

Patients with insulin-treated diabetes struggle with performing accurate carbohydrate counting for proper blood glucose control. Little is known about the comparative accuracy and feasibility of carbohydrate counting methods.

PURPOSE

The purpose of this study was to determine whether carbohydrate counting using a smartphone application is more accurate and feasible than a traditional method.

THEORETICAL/CONCEPTUAL FRAMEWORK

Based on a conceptual model derived from the Technology Acceptance Model, feasibility was defined as usefulness, ease of use, and behavioral intention to use each method.

METHODS

A standardized meal was presented to 20 adults with insulin-treated diabetes who counted carbohydrates using traditional and smartphone methods. Accuracy was measured by comparing carbohydrate counting estimates with the standardized meal values. Perceived feasibility (usefulness, ease of use, behavioral intention) was measured using rating forms derived from the Technology Acceptance Model.

RESULTS

The number of training and estimation minutes were significantly higher for the traditional method than the smartphone method (Z = -3.83, P < .05; Z = -2.30, P < .05). The traditional method took an additional 1.4 minutes for estimation and 12.5 minutes for training. There were no significant differences in accuracy between traditional and smartphone methods for carbohydrate counting (Wilcoxon signed-rank test, Z = -1.10, P = .28). There were no significant differences between traditional and smartphone methods for feasibility (usefulness, Z = -.10, P = .95; ease of use, Z = -.36, P = .72; or behavioral intention, Z = -.94, P = .35).

CONCLUSION

While both traditional and smartphone methods were found to be similar in terms of accuracy and feasibility, the smartphone method took less time for training and for carbohydrate estimation.

Postprandial glucose metabolism in children and adolescents with type 1 diabetes mellitus: potential targets for improvement

The main goal of therapeutic management of type 1 Diabetes Mellitus (T1DM) is to maintain optimal glycemic control to prevent acute and long-term diabetes complications and to enable a good quality of life. Postprandial glycemia makes a substantial contribution to overall glycemic control and variability in diabetes and, despite technological advancements in insulin treatments, optimal postprandial glycemia is difficult to achieve. Several factors influence postprandial blood glucose levels in children and adolescents with T1DM, including nutritional habits and adjustment of insulin doses according to meal composition. Additionally, hormone secretion, enteroendocrine axis dysfunction, altered gastrointestinal digestion and absorption, and physical activity play important roles. Meal-time routines, intake of appropriate ratios of macronutrients, and correct adjustment of the insulin dose for the meal composition have positive impacts on postprandial glycemic variability and long-term cardiometabolic health of the individual with T1DM. Further knowledge in the field is necessary for management of all these factors to be part of routine pediatric diabetes education and clinical practice. Thus, the aim of this report is to review the main factors that influence postprandial blood glucose levels and metabolism, focusing on macronutrients and other nutritional and lifestyle factors, to suggest potential targets for improving postprandial glycemia in the management of children and adolescents with T1DM.

Associations Between Psychosocial Needs, Carbohydrate-Counting Behavior, and App Satisfaction: A Randomized Crossover App Trial on 92 Adults With Diabetes

To examine whether psychosocial needs in diabetes care are associated with carbohydrate counting and if carbohydrate counting is associated with satisfaction with diabetes applications' usability, a randomized crossover trial of 92 adults with type 1 or 2 diabetes requiring insulin therapy tested two top-rated diabetes applications, mySugr and OnTrack Diabetes. Survey responses on demographics, psychosocial needs (perceived competence, autonomy, and connectivity), carbohydrate-counting frequency, and application satisfaction were modeled using mixed-effect linear regressions to test associations. Participants ranged between 19 and 74 years old (mean, 54 years) and predominantly had type 2 diabetes (70%). Among the three tested domains of psychosocial needs, only competence-not autonomy or connectivity-was found to be associated with carbohydrate-counting frequency. No association between carbohydrate-counting behavior and application satisfaction was found. In conclusion, perceived competence in diabetes care is an important factor in carbohydrate counting; clinicians may improve adherence to carbohydrate counting with strategies designed to improve perceived competence. Carbohydrate-counting behavior is complex; its impact on patient satisfaction of diabetes application usability is multifactorial and warrants consideration of patient demographics such as sex as well as application features for automated carbohydrate counting.

In Silico Validation of Personalized Safe Intervals for Carbohydrate Counting Errors

For patients with Type 1 diabetes mellitus (T1DM), accurate carbohydrate counting (CC) is essential for successful blood glucose regulation. Unfortunately, mistakes are common and may lead to an incorrect dosage of prandial insulin. In this work, we aim to demonstrate that each person has their own limits for CC errors, which can be computed using patient-specific data. To validate the proposed method, we tested it using several scenarios to investigate the effect of different CC errors on postprandial blood glucose. Virtual subjects from the T1DM Simulator were used in a clinical trial involving 450 meals over 90 days, all following the same daily meal plan but with different intervals for CC errors near, below, and above the limit computed for each patient. The results show that CC errors within personalized limits led to acceptable postprandial glycemic fluctuations. In contrast, experiments where 50% and 97.5% of the meals present a CC error outside the computed safe interval revealed a pronounced degradation of the time in range. Given these results, we consider the proposed method for obtaining personalized limits for CC errors an excellent starting point for an initial assessment of patients' capabilities in CC and to provide appropriate ongoing education.

Impact of Carbohydrate Counting Error on Glycemic Control in Open-Loop Management of Type 1 Diabetes: Quantitative Assessment Through an In Silico Trial

BACKGROUND

In the management of type 1 diabetes (T1D), systematic and random errors in carb-counting can have an adverse effect on glycemic control. In this study, we performed an in silico trial aiming at quantifying the impact of different levels of carb-counting error on glycemic control.

METHODS

The T1D patient decision simulator was used to simulate 7-day glycemic profiles of 100 adults using open-loop therapy. The simulation was repeated for different values of systematic and random carb-counting errors, generated with Gaussian distribution varying the error mean from -10% to +10% and standard deviation (SD) from 0% to 50%. The effect of the error was evaluated by computing the difference of time inside (∆TIR), above (∆TAR) and below (∆TBR) the target glycemic range (70-180mg/dl) compared to the reference case, that is, absence of error. Finally, 3 linear regression models were developed to mathematically describe how error mean and SD variations result in ∆TIR, ∆TAR, and ∆TBR changes.

RESULTS

Random errors globally deteriorate the glycemic control; systematic underestimations lead to, on average, up to 5.2% more TAR than the reference case, while systematic overestimation results in up to 0.8% more TBR. The different time in range metrics were linearly related with error mean and SD (R²>0.95), with slopes of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>E</mml:mi><mml:mi>A</mml:mi><mml:mi>N</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>21</mml:mn></mml:mrow></mml:math>, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>S</mml:mi><mml:mi>D</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>07</mml:mn></mml:mrow></mml:math> for ∆TIR, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>E</mml:mi><mml:mi>A</mml:mi><mml:mi>N</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>-</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>25</mml:mn></mml:mrow></mml:math>, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>S</mml:mi><mml:mi>D</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>06</mml:mn></mml:mrow></mml:math> for ∆TAR, and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>E</mml:mi><mml:mi>A</mml:mi><mml:mi>N</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>05</mml:mn></mml:mrow></mml:math>, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:msub><mml:mi>β</mml:mi><mml:mrow><mml:mi>S</mml:mi><mml:mi>D</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mo>+</mml:mo><mml:mn>0</mml:mn><mml:mo>.</mml:mo><mml:mn>01</mml:mn></mml:mrow></mml:math> for ∆TBR.

CONCLUSIONS

The quantification of carb-counting error impact performed in this work may be useful understanding causes of glycemic variability and the impact of possible therapy adjustments or behavior changes in different glucose metrics.

Super Bolus: a remedy for a high glycemic index meal in children with type 1 diabetes on insulin pump therapy?-study protocol for a randomized controlled trial

Background

Postprandial hyperglycemia (PPH) is a common clinical problem among patients with type 1 diabetes (T1D), which is related to high glycemic index (h-GI) meals. The main problem is linked to high, sharp glycemic spikes following hypoglycemia after h-GI meal consumption. There is a lack of effective and satisfactory solutions for insulin dose adjustment to cover an h-GI meal. The goal of this research was to determine whether a Super Bolus is an effective strategy to prevent PPH and late hypoglycemia after an h-GI meal compared to a Normal Bolus.

Methods

A total of 72 children aged 10–18 years with T1D for at least 1 year and treated with continuous subcutaneous insulin infusion for more than 3 months will be enrolled in a double-blind, randomized, crossover clinical trial. The participants will eat a h-GI breakfast for the two following days and receive a prandial insulin bolus in the form of a Super Bolus 1 day and a Normal Bolus the next day. The glucose level 90 min after the administration of the prandial bolus will be the primary outcome measure. The secondary endpoints will refer to the glucose levels at 30, 60, 120, 150, and 180 min postprandially, the area under the blood glucose curve within 180 min postprandially, peak glucose level and the time to peak glucose level, glycemic rise, the mean amplitude of glycemic excursions, and the number of hypoglycemia episodes.

Discussion

There are still few known clinical studies on this type of bolus. A Super Bolus is defined as a 50% increase in prandial insulin dose compared to the dose calculated based on the individualized patient’s insulin-carbohydrate ratio and a simultaneous suspension of basal insulin for 2 h. Our patients reported the best experience with such a combination. A comprehensive and effective solution to this frequent clinical difficulty of PPH after an h-GI meal has not yet been found. The problem is known and important, and the presented solution is innovative and easy to apply in everyday life.

Trial registration

ClinicalTrials.gov NCT04019821

Feasibility Study of an Automated Carbohydrate Estimation System Using Thai Food Images in Comparison With Estimation by Dietitians

Carbohydrate counting is essential for well-controlled blood glucose in people with type 1 diabetes, but to perform it precisely is challenging, especially for Thai foods. Consequently, we developed a deep learning-based system for automatic carbohydrate counting using Thai food images taken from smartphones. The newly constructed Thai food image dataset contained 256,178 ingredient objects with measured weight for 175 food categories among 75,232 images. These were used to train object detector and weight estimator algorithms. After training, the system had a Top-1 accuracy of 80.9% and a root mean square error (RMSE) for carbohydrate estimation of <10 g in the test dataset. Another set of 20 images, which contained 48 food items in total, was used to compare the accuracy of carbohydrate estimations between measured weight, system estimation, and eight experienced registered dietitians (RDs). System estimation error was 4%, while estimation errors from nearest, lowest, and highest carbohydrate among RDs were 0.7, 25.5, and 7.6%, respectively. The RMSE for carbohydrate estimations of the system and the lowest RD were 9.4 and 10.2, respectively. The system could perform with an estimation error of <10 g for 13/20 images, which placed it third behind only two of the best performing RDs: RD1 (15/20 images) and RD5 (14/20 images). Hence, the system was satisfactory in terms of accurately estimating carbohydrate content, with results being comparable with those of experienced dietitians.

Prospective Independent Evaluation of the Carbohydrate Counting Accuracy of Two Smartphone Applications

INTRODUCTION

Smartphone applications (apps) have been designed that help patients to accurately count their carbohydrate intake in order to optimize prandial insulin dose matching. Our aim was to evaluate the accuracy of two carbohydrate (carb) counting apps.

METHODS

Medical students, in the role of mock patients, evaluated meals using two smartphone apps: Foodvisor^® (which uses automatic food photo recognition technology) and Glucicheck^® (which requires the manual entry of carbohydrates with the help of a photo gallery). The macronutrient quantifications obtained with these two apps were compared to a reference quantification.

RESULTS

The carbohydrate content of the entire meal was underestimated with Foodvisor^® (Foodvisor^® quantification minus gold standard quantification = - 7.2 ± 17.3 g; p < 0.05) but reasonably accurately estimated with Glucicheck^® (Glucicheck^® quantification minus gold standard quantification = 1.4 ± 13.4 g; ns). The percentage of meals with an absolute error in carbohydrate quantification above 20 g was greater for Foodvisor^® compared to Glucicheck^® (30% vs 14%; p < 0.01).

CONCLUSION

The carb counting accuracy was slightly better when using Glucicheck^® compared to Foodvisor^®. However, both apps provided a lower mean absolute carb counting error than that usually made by T1D patients in everyday life, suggesting that such apps may be a useful adjunct for estimating carbohydrate content.

Evaluation of Meal Carbohydrate Counting Errors in Patients with Type 1 Diabetes

AIM

Correct estimation of meal carbohydrate content is a prerequisite for successful intensified insulin therapy in patients with diabetes. In this survey, the counting error in adult patients with type 1 diabetes was investigated.

METHODS

Seventy-four patients with type 1 diabetes estimated the carbohydrate content of 24 standardized test meals. The test meals were categorized into 1 of 3 groups with different carbohydrate content: low, medium, and high. Estimation results were compared with the meals' actual carbohydrate content as determined by calculation based on weighing. A subgroup of the participants estimated the test meals for a second (n=35) and a third time (n=22) with a mean period of 11 months between the estimations.

RESULTS

During the first estimation, the carbohydrate content was underestimated by -28% (-50, 0) of the actual carbohydrate content. Particularly meals with high mean carbohydrate content were underestimated by -34% (-56, -13). Median counting error improved significantly when estimations were performed for a second time (p<0.001).

CONCLUSIONS

Participants generally underestimated the carbohydrate content of the test meals, especially in meals with higher carbohydrate content. Repetition of estimation resulted in significant improvements in estimation accuracy and is important for the maintenance of correct carbohydrate estimations. The ability to estimate the carbohydrate content of a meal should be checked and trained regularly in patients with diabetes.

"Counting Carbs to Be in Charge": A Comparison of an Internet-Based Education Module With In-Class Education in Adolescents With Type 1 Diabetes

Carbohydrate counting is an essential component of type 1 diabetes education but can be difficult for adolescents to learn. Because adolescents are avid users of technology, an Internet-based education module was compared with an in-class education session in terms of carbohydrate counting accuracy in adolescents with type 1 diabetes. Adolescent participants displayed increased carbohydrate counting accuracy after attending an in-class education session compared with an Internet-based education module. These results suggest that online education is best reserved as an adjunctive therapy to in-class teaching in this population.

Carbohydrate Counting App Using Image Recognition for Youth With Type 1 Diabetes: Pilot Randomized Control Trial

BACKGROUND

Carbohydrate counting is an important component of diabetes management, but it is challenging, often performed inaccurately, and can be a barrier to optimal diabetes management. iSpy is a novel mobile app that leverages machine learning to allow food identification through images and that was designed to assist youth with type 1 diabetes in counting carbohydrates.

OBJECTIVE

Our objective was to test the app's usability and potential impact on carbohydrate counting accuracy.

METHODS

Iterative usability testing (3 cycles) was conducted involving a total of 16 individuals aged 8.5-17.0 years with type 1 diabetes. Participants were provided a mobile device and asked to complete tasks using iSpy app features while thinking aloud. Errors were noted, acceptability was assessed, and refinement and retesting were performed across cycles. Subsequently, iSpy was evaluated in a pilot randomized controlled trial with 22 iSpy users and 22 usual care controls aged 10-17 years. Primary outcome was change in carbohydrate counting ability over 3 months. Secondary outcomes included levels of engagement and acceptability. Change in HbA1c level was also assessed.

RESULTS

Use of iSpy was associated with improved carbohydrate counting accuracy (total grams per meal, P=.008), reduced frequency of individual counting errors greater than 10 g (P=.047), and lower HbA1c levels (P=.03). Qualitative interviews and acceptability scale scores were positive. No major technical challenges were identified. Moreover, 43% (9/21) of iSpy participants were still engaged, with usage at least once every 2 weeks, at the end of the study.

CONCLUSIONS

Our results provide evidence of efficacy and high acceptability of a novel carbohydrate counting app, supporting the advancement of digital health apps for diabetes care among youth with type 1 diabetes. Further testing is needed, but iSpy may be a useful adjunct to traditional diabetes management.

TRIAL REGISTRATION

ClinicalTrials.gov NCT04354142; https://clinicaltrials.gov/ct2/show/NCT04354142.

Modeling Carbohydrate Counting Error in Type 1 Diabetes Management

Background: The error in estimating meal carbohydrates (CHO) amount is a critical mistake committed by type 1 diabetes (T1D) subjects. The aim of this study is both to investigate which factors, related to meals and subjects, affect the CHO counting error most and to develop a mathematical model of CHO counting error embeddable in T1D patient decision simulators to conduct in silico clinical trials. Methods: A published dataset of 50 T1D adults is used, which includes a patient's CHO count of 692 meals, dietitian's estimates of meal composition (used as reference), and several potential explanatory factors. The CHO counting error is modeled by multiple linear regression, with stepwise variable selection starting from 10 candidate predictors, that is, education level, insulin treatment duration, age, body weight, meal type, CHO, lipid, energy, protein, and fiber content. Inclusion of quadratic and interaction terms is also evaluated. Results: Larger errors correspond to larger meals, and most of the large meals are underestimated. The linear model selects CHO (P < 0.00001), meal type (P < 0.00001), and body weight (P = 0.047), whereas its extended version embeds a quadratic term of CHO (P < 0.00001) and interaction terms of meal type with CHO (P = 0.0001) and fiber amount (P = 0.001). The extended model explains 34.9% of the CHO counting error variance. Comparison with the CHO counting error description previously used in the T1D patient decision simulator shows that the proposed models return more credible realizations. Conclusions: The most important predictors of CHO counting errors are CHO and meal type. The mathematical models proposed improve the description of patients' behavior in the T1D patient decision simulator.

Effect of adherence to carbohydrate counting on metabolic control in children and adolescents with type 1 diabetes mellitus

PURPOSE

Carbohydrate counting provides better glycemic control and flexibility than other food planning methods. Consistent adherence to such a complex method is difficult, especially for youth. However, studies that determine adherence to this method and whether it alters metabolic control are limited. The aim of the current study was to determine adherence to this method and investigate its effect on metabolic control, anthropometric measurements, insulin dose, and energy intake.

METHODS

In this prospective cross-sectional study, 53 children and adolescents with type 1 diabetes mellitus aged 2 to18 years and receiving intensive insulin therapy were trained and followed for 6 months. Demographics, anthropometrics, insulin requirements, hemoglobin A1c (HbA1c), fasting lipids, and food records at baseline and study conclusion were evaluated. At the end of the study patients were divided into adherer and nonadherer groups according to carbohydrate estimate deviations from standardized daily sample menus and calculations for accurate insulin doses. More than 10-g variation in daily consumed carbohydrate amount or failure to decide bolus insulin dose was defined as a nonadherer.

RESULTS

The mean HbA1c, low-density lipoprotein cholesterol, and body mass index standard deviation score changed after the carbohydrate counting training while the mean HbA1c between groups was significant (P<0.05). Total daily insulin doses increased, and the mean high-density lipoprotein cholesterol levels decreased in both groups. There were significant correlations between HbA1c and carbohydrate deviation scores as well as HbA1c and caregiver's education level.

CONCLUSION

Since adherence to carbohydrate counting may affect metabolic control, health professionals should evaluate and monitor carbohydrate counting skills of caregivers and patients in order to improve efficiency.

goFOODTM: An Artificial Intelligence System for Dietary Assessment

Accurate estimation of nutritional information may lead to healthier diets and better clinical outcomes. We propose a dietary assessment system based on artificial intelligence (AI), named goFOOD^TM. The system can estimate the calorie and macronutrient content of a meal, on the sole basis of food images captured by a smartphone. goFOOD^TM requires an input of two meal images or a short video. For conventional single-camera smartphones, the images must be captured from two different viewing angles; smartphones equipped with two rear cameras require only a single press of the shutter button. The deep neural networks are used to process the two images and implements food detection, segmentation and recognition, while a 3D reconstruction algorithm estimates the food’s volume. Each meal’s calorie and macronutrient content is calculated from the food category, volume and the nutrient database. goFOOD^TM supports 319 fine-grained food categories, and has been validated on two multimedia databases that contain non-standardized and fast food meals. The experimental results demonstrate that goFOOD^TM performed better than experienced dietitians on the non-standardized meal database, and was comparable to them on the fast food database. goFOOD^TM provides a simple and efficient solution to the end-user for dietary assessment.

Amount and Type of Dietary Fat, Postprandial Glycemia, and Insulin Requirements in Type 1 Diabetes: A Randomized Within-Subject Trial

OBJECTIVE

The American Diabetes Association recommends individuals with type 1 diabetes (T1D) adjust insulin for dietary fat; however, optimal adjustments are not known. This study aimed to determine 1) the relationship between the amount and type of dietary fat and glycemia and 2) the optimal insulin adjustments for dietary fat.

RESEARCH DESIGN AND METHODS

Adults with T1D using insulin pump therapy attended the research clinic on 9-12 occasions. On the first six visits, participants consumed meals containing 45 g carbohydrate with 0 g, 20 g, 40 g, or 60 g fat and either saturated, monounsaturated, or polyunsaturated fat. Insulin was dosed using individual insulin/carbohydrate ratio as a dual-wave 50/50% over 2 h. On subsequent visits, participants repeated the 20-60-g fat meals with the insulin dose estimated using a model predictive bolus, up to twice per meal, until glycemic control was achieved.

RESULTS

With the same insulin dose, increasing the amount of fat resulted in a significant dose-dependent reduction in incremental area under the curve for glucose (iAUC_glucose) in the early postprandial period (0-2 h; P = 0.008) and increase in iAUC_glucose in the late postprandial period (2-5 h; P = 0.004). The type of fat made no significant difference to the 5-h iAUC_glucose. To achieve glycemic control, on average participants required dual-wave insulin bolus: for 20 g fat, +6% insulin, 74/26% over 73 min; 40 g fat, +6% insulin, 63/37% over 75 min; and 60 g fat, +21% insulin, 49/51% over 105 min.

CONCLUSIONS

This study provides clinical guidance for mealtime insulin dosing recommendations for dietary fat in T1D.

Variation of carbohydrate intake in diabetic children on carbohydrate counting

AIMS

Carbohydrate counting (CC) is a technique for managing diabetes particularly based on the counting of carbohydrates. It allows diabetic patients to vary their amount of carbohydrates from one meal to another by adjusting their insulin dose. The primary objective was to determine the variation of carbohydrate intake (CI) in children on CC.

METHOD

This was a prospective study conducted between 2014 and 2016. We collected the amount of carbohydrates eaten at each meal by 77 diabetic over a period of 28 days (i.e. 8068 data). We analyzed the number and percentage of significant CI variation rates from one day to another, both for the whole day and for each meal. The CI variation rate was deemed significant if it was greater than or equal to 30%.

RESULTS

The percentage of significant CI variation rates was 30% at the daily level, 34% for breakfast, 44% for lunch and dinner, and 53% for snack. The percentage of significant variation rates varied according to age, treatment and occurrence of events.

CONCLUSION

Children varied their CI significantly from one meal to another more than one in three times. CC offers flexibility and a better quality of life for children using this method.

Accuracy of Bolus and Basal Rate Delivery of Different Insulin Pump Systems

BACKGROUND

Insulin pumps are used for basal rate and bolus insulin delivery in patients with diabetes. In this in vitro study, accuracy of delivery of different commercial insulin pumps was evaluated.

MATERIALS AND METHODS

Accuracy of 10 different insulin pump systems (5 durable pumps with different insulin infusion sets and 1 patch pump) was tested with a microgravimetric method. Mean bolus accuracy of 25 successive 1 U boluses and of 12 successive 10 U boluses was assessed, and delivery time for 10 U boluses was measured. Basal rate accuracy at 1.0 U/h was evaluated for 72 h and for individual 1-h windows.

RESULTS

Mean bolus delivery was within ±5% of target for both tested bolus sizes, but precision of individual boluses was higher with the larger boluses. Delivery times varied between the different pump models but agreed with the specifications of the respective manufacturers. Regarding basal rate accuracy, the total deviation for 72 h was very small in all pumps; however, larger deviations were observed during the first 12 h. For the patch pump, large variations between individual 1-h windows were observed.

CONCLUSIONS

In general, all compared insulin pump systems showed a similar level of accuracy. Differences, especially between durable pumps and the patch pump, were observed when considering each hour of basal rate delivery separately.

A Two-Level Food Classification System For People With Diabetes Mellitus Using Convolutional Neural Networks

Accurate estimation of food's macronutrient content for people with Diabetes Mellitus (DM) is of great importance, as it determines postprandial insulin dosage. This paper introduces a classification system for food images that is adjusted to the nutritional needs of people with DM. A two-level image classification scheme, exploiting Convolutional Neural Networks (CNNs), is proposed, in order to classify an image in one of eight broad food categories with similar macronutrient content and then assign it to a specific food within that category. To this end, a visual dataset, namely NTUA-Food 2017, has been designed, consisting of 3248 images organized in eight broad food categories of totally 82 different foods. Moreover, a novel evaluation metric is proposed, which penalizes classification errors proportionally to the discrepancy in postprandial blood sugar levels between the actual and predicted class. The proposed system achieves 84.18% and 85.94% classification accuracy at the first and second level of classification, respectively, on the NTUA-Food 2017 dataset. The algorithm developed for the first level of classification on the NTUA-Food 2017 dataset improves classification accuracy on the benchmark Food Image Dataset (FID) to 97.08% outperforming previous approaches. The algorithm's mean error in terms of carbohydrate content estimation on the NTUA-Food 2017 dataset is less than 2 g per food serving.

One potato, two potato,… assessing carbohydrate counting accuracy in adolescents with type 1 diabetes

BACKGROUND/OBJECTIVE

Carbohydrate (CHO) counting is a recommended daily practice to help manage blood glucose levels in type 1 diabetes. Evidence suggests that CHO estimates should be within 10 to 15 g of the actual meal for optimal postprandial blood glucose control. The objective of this study was to assess accuracy of CHO counting in adolescents with type 1 diabetes.

METHODS

Adolescents (aged 12-18 years) with type 1 diabetes who self-identified as regular CHO counters were recruited from the SickKids Diabetes Clinic, Toronto, Canada. Adolescents completed the PedsCarbQuiz (PCQ) and estimated CHO content of test trays (three meals and three snack trays) that were randomly assigned. Analyses were conducted to identify factors associated with accuracy of counting and CHO counting knowledge (PCQ score).

RESULTS

A total of 140 adolescents (78 females, mean age 14.7, SD = 1.8) participated. The average PCQ score was 81 ± 10%. Forty-two percent of adolescents were accurate in estimating meal trays (ie, within 10 g of the actual CHO content), 44% estimated inaccurately (within 10-20 g), while 14% were significantly inaccurate counters (>20 g variation). PCQ scores were higher in teens who CHO counted accurately than in those with significant inaccuracy (>20 g) (P < 0.05), and a longer duration of diabetes corresponded significantly with a lower PCQ score. No demographics correlated significantly with CHO counting accuracy.

CONCLUSIONS

Less than half of the teens in our study were accurate CHO counters. These results indicate the need for regular clinical accuracy check and reeducation.

A Comparative Study on Carbohydrate Estimation: GoCARB vs. Dietitians

GoCARB is a computer vision-based smartphone system designed for individuals with Type 1 Diabetes to estimate plated meals’ carbohydrate (CHO) content. We aimed to compare the accuracy of GoCARB in estimating CHO with the estimations of six experienced dietitians. GoCARB was used to estimate the CHO content of 54 Central European plated meals, with each of them containing three different weighed food items. Ground truth was calculated using the USDA food composition database. Dietitians were asked to visually estimate the CHO content based on meal photographs. GoCARB and dietitians achieved comparable accuracies. The mean absolute error of the dietitians was 14.9 (SD 10.12) g of CHO versus 14.8 (SD 9.73) g of CHO for the GoCARB (p = 0.93). No differences were found between the estimations of dietitians and GoCARB, regardless the meal size. The larger the size of the meal, the greater were the estimation errors made by both. Moreover, the higher the CHO content of a food category was, the more challenging its accurate estimation. GoCARB had difficulty in estimating rice, pasta, potatoes, and mashed potatoes, while dietitians had problems with pasta, chips, rice, and polenta. GoCARB may offer diabetic patients the option of an easy, accurate, and almost real-time estimation of the CHO content of plated meals, and thus enhance diabetes self-management.

The influence of food order on postprandial glucose levels in children with type 1 diabetes

OBJECTIVE

To evaluate the effect of the order of intake of carbohydrates, protein, and fat on postprandial glucose levels in children with type 1 diabetes (T1D). Our hypothesis was that postprandial glucose levels would be lower when fat and protein are consumed prior to carbohydrates, compared to a meal where all macronutrients are combined.

METHODS

A randomized, open-label, within-subject crossover study was conducted. Twenty patients aged 7 to 17 years diagnosed with T1D for >1 year consumed 2 isocaloric meals (with similar composition) in random order. In 1 meal, the protein and fat part was consumed 15 minutes prior to the carbohydrates (test meal). In the other meal, all macronutrients were consumed together (standard meal). Capillary blood glucose measurements and continuous glucose monitoring system were used to assess multiple glucose levels during a 3-hour postprandial period.

RESULTS

Overall, mean glucose levels were 1 mmol/L lower following the test meal compared to the standard meal (9.30 ± 3.20 vs 10.24 ± 3.35 mmol/L; P < .001). No significant difference in peak glucose was found. Glucose excursions were 1.5 and 1 mmol/L lower at 30 and 120 minutes following the test meal. A reduction in the total time period in which glucose levels exceeded 10 and 12 mmol/L of 28.7 (P = .001) and 22.3 minutes (P = .004), respectively, after the test meal was found.

CONCLUSIONS

This study shows that consumption of protein and fat prior to carbohydrates results in lower postprandial glucose levels and reduced glycemic variability in children with T1D.

Young children with type 1 diabetes can achieve glycemic targets without hypoglycemia: Results of a novel intensive diabetes management program

BACKGROUND

Young children with type 1 diabetes (T1D) present unique challenges for intensive diabetes management. We describe an intensive diabetes program adapted for young children and compare glycemic control, anthropometry, dietary practices and insulin regimens before and after implementation.

METHODS

Cross sectional data from children with T1D aged ≥0.5 to <7.0 years attending the John Hunter Children's Hospital (JHCH), Australia in 2004, 2010 and 2016 were compared. Outcome measures were glycemic control assessed by hemoglobin A_1c (HbA_1c ); severe hypoglycemia episodes; body mass index standard deviation scores (BMI-SDS); diabetes ketoacidosis (DKA) episodes; and insulin regimen-twice daily injections, multiple daily injections, or continuous subcutaneous insulin infusion.

RESULTS

Mean HbA_1c declined by 12 mmol/mol over the study period (P < .01). The proportion of children achieving a mean HbA_1c  < 58 mmol/mol increased significantly from 31% in 2004 to 64% in 2010 (P < .01), and from 64% in 2010 to 83% in 2016 (P = .04). The mean BMI-SDS was significantly lower in 2010 when compared with 2004 (P<.01); however, this trend plateaued between 2010 and 2016 (P = .97). Severe hypoglycemia and DKA occurred infrequently. The prevalence of overweight or obesity increased from 2010 to 2016 (P = .03).

CONCLUSIONS

The JHCH intensive diabetes management program has resulted in 83% of young children in 2016 achieving target glycemia without an increase in severe hypoglycemia or DKA. Overweight remains a challenge in this population warranting action to reduce weight and protect these children from future obesity-related health risks.

Accuracy of Automatic Carbohydrate, Protein, Fat and Calorie Counting Based on Voice Descriptions of Meals in People with Type 1 Diabetes

The aim of this work was to assess the accuracy of automatic macronutrient and calorie counting based on voice descriptions of meals provided by people with unstable type 1 diabetes using the developed expert system (VoiceDiab) in comparison with reference counting made by a dietitian, and to evaluate the impact of insulin doses recommended by a physician on glycemic control in the study&rsquo;s participants. We also compared insulin doses calculated using the algorithm implemented in the VoiceDiab system. Meal descriptions were provided by 30 hospitalized patients (mean hemoglobin A1c of 8.4%, i.e., 68 mmol/mol). In 16 subjects, the physician determined insulin boluses based on the data provided by the system, and in 14 subjects, by data provided by the dietitian. On one hand, differences introduced by patients who subjectively described their meals compared to those introduced by the system that used the average characteristics of food products, although statistically significant, were low enough not to have a significant impact on insulin doses automatically calculated by the system. On the other hand, the glycemic control of patients was comparable regardless of whether the physician was using the system-estimated or the reference content of meals to determine insulin doses.

Rapid-acting and Regular Insulin are Equal for High Fat-Protein Meal in Individuals with Type 1 Diabetes Treated with Multiple Daily Injections

INTRODUCTION

The fat and protein content can impact late postprandial glycemia; therefore, prolonged insulin boluses for high-fat/-protein meals are recommended for patients with type 1 diabetes on insulin pump therapy. It is not clear how to translate these findings to multiple daily injection (MDI) therapy. We hypothesized that regular insulin with a slower onset and a longer duration of action might be advantageous for such meals.

METHODS

Twenty-five patients with well-controlled type 1 diabetes (mean HbA1c 6.8%, 51 mmol/mol, no episodes of hypoglycemia) on MDI therapy, aged 27.9 ± 4.3 years and well trained in flexible intensive insulin therapy, were given three test breakfasts with the same carbohydrate (CHO) content. The amount of fat and protein was low (LFP) or high (HFP). For LFP meals, patients received a rapid-acting insulin; for HFP meals, a rapid-acting or regular insulin was given in individual doses according to the CHO content and individual insulin-CHO ratios. Postprandial glycemia was determined by 6-h continuous glucose monitoring.

RESULTS

Acute postprandial glucose levels measured for 2 h were similar after LFP and two HFP meals (7.8 ± 2.0, 8.1 ± 2.1, 8.0 ± 1.9 mmol/l). Late postprandial glycemia measured from 2 to 6 h was significantly lower after the LFP meal (6.7 ± 1.8 mmol/l, p < 0.05) than after the HFP meals, but there was no difference between the rapid-acting or regular insulin on HFP days (8.6 ± 2.6 and 8.9 ± 2.8 mmol/l, NS).

CONCLUSION

The preliminary results of this study indicate no benefit to cover fat-protein meals with regular insulin in individuals with type 1 diabetes treated with MDI.

Endocrine and metabolic consequences due to restrictive carbohydrate diets in children with type 1 diabetes: An illustrative case series

Low carbohydrate diets for the management of type 1 diabetes have been popularised by social media. The promotion of a low carbohydrate diet in lay media is in contrast to published pediatric diabetes guidelines that endorse a balanced diet from a variety of foods for optimal growth and development in children with type 1 diabetes. This can be a source of conflict in clinical practice. We describe a series of 6 cases where adoption of a low carbohydrate diet in children impacted growth and cardiovascular risk factors with potential long-term sequelae. These cases support current clinical guidelines for children with diabetes that promote a diet where total energy intake is derived from balanced macronutrient sources.

Wearable Continuous Glucose Monitoring Sensors: A Revolution in Diabetes Treatment

Worldwide, the number of people affected by diabetes is rapidly increasing due to aging populations and sedentary lifestyles, with the prospect of exceeding 500 million cases in 2030, resulting in one of the most challenging socio-health emergencies of the third millennium. Daily management of diabetes by patients relies on the capability of correctly measuring glucose concentration levels in the blood by using suitable sensors. In recent years, glucose monitoring has been revolutionized by the development of Continuous Glucose Monitoring (CGM) sensors, wearable non/minimally-invasive devices that measure glucose concentration by exploiting different physical principles, e.g., glucose-oxidase, fluorescence, or skin dielectric properties, and provide real-time measurements every 1–5 min. CGM opened new challenges in different disciplines, e.g., medicine, physics, electronics, chemistry, ergonomics, data/signal processing, and software development to mention but a few. This paper first makes an overview of wearable CGM sensor technologies, covering both commercial devices and research prototypes. Then, the role of CGM in the actual evolution of decision support systems for diabetes therapy is discussed. Finally, the paper presents new possible horizons for wearable CGM sensor applications and perspectives in terms of big data analytics for personalized and proactive medicine.

Accurate Carbohydrate Counting Is an Important Determinant of Postprandial Glycemia in Children and Adolescents With Type 1 Diabetes on Insulin Pump Therapy

BACKGROUND

Carbohydrate (CHO) counting is a key nutritional intervention utilized in the management of diabetes to optimize postprandial glycemia. The aim of the study was to examine the impact of accuracy of CHO counting on the postprandial glucose in children and adolescents with type 1 diabetes on insulin pump therapy.

METHODS

Children/adolescents with type 1 diabetes who were on insulin pump therapy for a minimum of 6 months are enrolled in the study. Patients were instructed to record details of meals consumed, estimated CHO count per meal, and 2-hour postprandial glucose readings over 3-5 days. Meals' CHO contents were recounted by an experienced clinical dietician, and those within 20% of the dietician's counting were considered accurate.

RESULTS

A total of 30 patients (21 females) were enrolled. Age range (median) was 8-18 (SD 13) years. Data of 247 meals were analyzed. A total of 165 (67%) meals' CHO contents were accurately counted. Of those, 90 meals (55%) had in-target postprandial glucose ( P < .000). There was an inverse relationship between inaccurate CHO estimates and postprandial glucose. Of the 63 underestimated meals, 55 had above-target glucose, while 12 of the 19 overestimated meals were followed by low glucose. There was no association between accuracy and meal size (Spearman's rho = .019).

CONCLUSION

Accuracy of CHO counting is an important determining factor of postprandial glycemia. However, other factors should be considered when advising on prandial insulin calculation. Underestimation and overestimation of CHO result in postprandial hyperglycemia and hypoglycemia, respectively. Accuracy does not correlate with meal size.

Impact of ELKa, the Electronic Device for Prandial Insulin Dose Calculation, on Metabolic Control in Children and Adolescents with Type 1 Diabetes Mellitus: A Randomized Controlled Trial

Background. The ELKa system is composed of computer software, with a database of nutrients, and a dedicated USB kitchen scale. It was designed to automatize the everyday calculations of food exchanges and prandial insulin doses. Aim. To investigate the influence of the ELKa on metabolic control in children with type 1 diabetes mellitus (T1DM). Methods. A randomized, parallel, open-label clinical trial involved 106 patients aged <18 years with T1DM, HbA_1C ≤ 10%, undergoing intensive insulin therapy, allocated to the intervention group, who used the ELKa (n = 53), or the control group (n = 53), who used conventional calculation methods. Results. After the 26-week follow-up, the intention-to-treat analysis showed no differences to all endpoints. In per protocol analysis, 22/53 (41.5%) patients reporting ELKa usage for >50% of meals achieved lower HbA_1C levels (P = 0.002), lower basal insulin amounts (P = 0.049), and lower intrasubject standard deviation of blood glucose levels (P = 0.023) in comparison with the control. Moreover, in the intervention group, significant reduction of HbA_1C level, by 0.55% point (P = 0.002), was noted. No intergroup differences were found in the hypoglycemic episodes, BMI-SDS, bolus insulin dosage, and total daily insulin dosage. Conclusions. The ELKa system improves metabolic control in children with T1DM under regular usage. The trial is registered at ClinicalTrials.gov, number NCT02194517.

Food exchange estimation by children with type 1 diabetes at summer camp

BACKGROUND

As exchange counting poses difficulty for children with type 1 diabetes (T1D) attending diabetes camps, they often guesstimate food amount without performing an exchange calculation. The aim of the study was to compare the accuracy of estimation with exchange counting using the mobile food exchange calculator (MFEC).

METHODS

During a summer camp, 25 children with T1D on pumps estimated the number of carbohydrate (CE) and fat/protein exchanges (FPE) appropriate for main meals. Afterwards, the number of exchanges was counted with MFEC and electronic scales.

RESULTS

There was a difference between CE (p<0.0001) and FPE (p<0.0001) estimations and counting using MFEC. The youth miscalculated the true values of ≥1 CE and ≥1 FPE by 31% and 23%, respectively. They more often underestimated than overestimated CE and FPE (p<0.0001). The estimation error increased with younger age.

CONCLUSIONS

Carbohydrate counting caused significant error in the exchange number. The use of MFEC facilitates correct exchange calculation. Patients should weigh food and calculate exchanges themselves using mobile applications.

Carbohydrate Estimation by a Mobile Phone-Based System Versus Self-Estimations of Individuals With Type 1 Diabetes Mellitus: A Comparative Study

Background

Diabetes mellitus is spreading throughout the world and diabetic individuals have been shown to often assess their food intake inaccurately; therefore, it is a matter of urgency to develop automated diet assessment tools. The recent availability of mobile phones with enhanced capabilities, together with the advances in computer vision, have permitted the development of image analysis apps for the automated assessment of meals. GoCARB is a mobile phone-based system designed to support individuals with type 1 diabetes during daily carbohydrate estimation. In a typical scenario, the user places a reference card next to the dish and acquires two images using a mobile phone. A series of computer vision modules detect the plate and automatically segment and recognize the different food items, while their 3D shape is reconstructed. Finally, the carbohydrate content is calculated by combining the volume of each food item with the nutritional information provided by the USDA Nutrient Database for Standard Reference.

Objective

The main objective of this study is to assess the accuracy of the GoCARB prototype when used by individuals with type 1 diabetes and to compare it to their own performance in carbohydrate counting. In addition, the user experience and usability of the system is evaluated by questionnaires.

Methods

The study was conducted at the Bern University Hospital, “Inselspital” (Bern, Switzerland) and involved 19 adult volunteers with type 1 diabetes, each participating once. Each study day, a total of six meals of broad diversity were taken from the hospital’s restaurant and presented to the participants. The food items were weighed on a standard balance and the true amount of carbohydrate was calculated from the USDA nutrient database. Participants were asked to count the carbohydrate content of each meal independently and then by using GoCARB. At the end of each session, a questionnaire was completed to assess the user’s experience with GoCARB.

Results

The mean absolute error was 27.89 (SD 38.20) grams of carbohydrate for the estimation of participants, whereas the corresponding value for the GoCARB system was 12.28 (SD 9.56) grams of carbohydrate, which was a significantly better performance ( P=.001). In 75.4% (86/114) of the meals, the GoCARB automatic segmentation was successful and 85.1% (291/342) of individual food items were successfully recognized. Most participants found GoCARB easy to use.

Conclusions

This study indicates that the system is able to estimate, on average, the carbohydrate content of meals with higher accuracy than individuals with type 1 diabetes can. The participants thought the app was useful and easy to use. GoCARB seems to be a well-accepted supportive mHealth tool for the assessment of served-on-a-plate meals.

Influence of dietary protein on postprandial blood glucose levels in individuals with Type 1 diabetes mellitus using intensive insulin therapy

AIM

To determine the effects of protein alone (independent of fat and carbohydrate) on postprandial glycaemia in individuals with Type 1 diabetes mellitus using intensive insulin therapy.

METHODS

Participants with Type 1 diabetes mellitus aged 7-40 years consumed six 150 ml whey isolate protein drinks [0 g (control), 12.5, 25, 50, 75 and 100] and two 150 ml glucose drinks (10 and 20 g) without insulin, in randomized order over 8 days, 4 h after the evening meal. Continuous glucose monitoring was used to assess postprandial glycaemia.

RESULTS

Data were collected from 27 participants. Protein loads of 12.5 and 50 g did not result in significant postprandial glycaemic excursions compared with control (water) throughout the 300 min study period (P > 0.05). Protein loads of 75 and 100 g resulted in lower glycaemic excursions than control in the 60-120 min postprandial interval, but higher excursions in the 180-300 min interval. In comparison with 20 g glucose, the large protein loads resulted in significantly delayed and sustained glucose excursions, commencing at 180 min and continuing to 5 h.

CONCLUSIONS

Seventy-five grams or more of protein alone significantly increases postprandial glycaemia from 3 to 5 h in people with Type 1 diabetes mellitus using intensive insulin therapy. The glycaemic profiles resulting from high protein loads differ significantly from the excursion from glucose in terms of time to peak glucose and duration of the glycaemic excursion. This research supports recommendations for insulin dosing for large amounts of protein.

8th Annual Symposium on Self-Monitoring of Blood Glucose (SMBG): April 16-18, 2015, Republic of Malta

International experts in the fields of diabetes, diabetes technology, endocrinology, mobile health, sport science, and regulatory issues gathered for the 8(th) Annual Symposium on Self-Monitoring of Blood Glucose (SMBG) with a focus on personalized diabetes management. The aim of this meeting was to facilitate new collaborations and research projects to improve the lives of people with diabetes. The 2015 meeting comprised a comprehensive scientific program, parallel interactive workshops, and two keynote lectures.

The relationship between carbohydrate and the mealtime insulin dose in type 1 diabetes

A primary focus of the nutritional management of type 1 diabetes has been on matching prandial insulin therapy with carbohydrate amount consumed. Different methods exist to quantify carbohydrate including counting in one gram increments, 10g portions or 15g exchanges. Clinicians have assumed that counting in one gram increments is necessary to precisely dose insulin and optimize postprandial control. Carbohydrate estimations in portions or exchanges have been thought of as inadequate because they may result in less precise matching of insulin dose to carbohydrate amount. However, studies examining the impact of errors in carbohydrate quantification on postprandial glycemia challenge this commonly held view. In addition it has been found that a single mealtime bolus of insulin can cover a range of carbohydrate intake without deterioration in postprandial control. Furthermore, limitations exist in the accuracy of the nutrition information panel on a food label. This article reviews the relationship between carbohydrate quantity and insulin dose, highlighting limitations in the evidence for a linear association. These insights have significant implications for patient education and mealtime insulin dose calculations.

Computer vision-based carbohydrate estimation for type 1 patients with diabetes using smartphones

BACKGROUND

Individuals with type 1 diabetes (T1D) have to count the carbohydrates (CHOs) of their meal to estimate the prandial insulin dose needed to compensate for the meal's effect on blood glucose levels. CHO counting is very challenging but also crucial, since an error of 20 grams can substantially impair postprandial control.

METHOD

The GoCARB system is a smartphone application designed to support T1D patients with CHO counting of nonpacked foods. In a typical scenario, the user places a reference card next to the dish and acquires 2 images with his/her smartphone. From these images, the plate is detected and the different food items on the plate are automatically segmented and recognized, while their 3D shape is reconstructed. Finally, the food volumes are calculated and the CHO content is estimated by combining the previous results and using the USDA nutritional database.

RESULTS

To evaluate the proposed system, a set of 24 multi-food dishes was used. For each dish, 3 pairs of images were taken and for each pair, the system was applied 4 times. The mean absolute percentage error in CHO estimation was 10 ± 12%, which led to a mean absolute error of 6 ± 8 CHO grams for normal-sized dishes.

CONCLUSION

The laboratory experiments demonstrated the feasibility of the GoCARB prototype system since the error was below the initial goal of 20 grams. However, further improvements and evaluation are needed prior launching a system able to meet the inter- and intracultural eating habits.

A food recognition system for diabetic patients based on an optimized bag-of-features model

Computer vision-based food recognition could be used to estimate a meal's carbohydrate content for diabetic patients. This study proposes a methodology for automatic food recognition, based on the bag-of-features (BoF) model. An extensive technical investigation was conducted for the identification and optimization of the best performing components involved in the BoF architecture, as well as the estimation of the corresponding parameters. For the design and evaluation of the prototype system, a visual dataset with nearly 5000 food images was created and organized into 11 classes. The optimized system computes dense local features, using the scale-invariant feature transform on the HSV color space, builds a visual dictionary of 10000 visual words by using the hierarchical k-means clustering and finally classifies the food images with a linear support vector machine classifier. The system achieved classification accuracy of the order of 78%, thus proving the feasibility of the proposed approach in a very challenging image dataset.

Both dietary protein and fat increase postprandial glucose excursions in children with type 1 diabetes, and the effect is additive

OBJECTIVE

To determine the separate and combined effects of high-protein (HP) and high-fat (HF) meals, with the same carbohydrate content, on postprandial glycemia in children using intensive insulin therapy (IIT).

RESEARCH DESIGN AND METHODS

Thirty-three subjects aged 8-17 years were given 4 test breakfasts with the same carbohydrate amount but varying protein and fat quantities: low fat (LF)/low protein (LP), LF/HP, HF/LP, and HF/HP. LF and HF meals contained 4 g and 35 g fat. LP and HP meals contained 5 g and 40 g protein. An individually standardized insulin dose was given for each meal. Postprandial glycemia was assessed by 5-h continuous glucose monitoring.

RESULTS

Compared with the LF/LP meal, mean glucose excursions were greater from 180 min after the LF/HP meal (2.4 mmol/L [95% CI 1.1-3.7] vs. 0.5 mmol/L [-0.8 to 1.8]; P = 0.02) and from 210 min after the HF/LP meal (1.8 mmol/L [0.3-3.2] vs. -0.5 mmol/L [-1.9 to 0.8]; P = 0.01). The HF/HP meal resulted in higher glucose excursions from 180 min to 300 min (P < 0.04) compared with all other meals. There was a reduction in the risk of hypoglycemia after the HP meals (odds ratio 0.16 [95% CI 0.06-0.41]; P < 0.001).

CONCLUSIONS

Meals high in protein or fat increase glucose excursions in youth using IIT from 3 h to 5 h postmeal. Protein and fat have an additive impact on the delayed postprandial glycemic rise. Protein had a protective effect on the development of hypoglycemia.