Displaying One Session

E-POSTER VIEWING (EXHIBITION HOURS)
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Channel
E-Poster Area
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Session Description
PLEASE NOTE: E-POSTER VIEWING IS DURING THE EXHIBITION HOURS OF EACH DAY.

SPURIOUS HIGH HBA1C: DO WE NEED TO REVIEW THE GUIDELINES?

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:30 - 09:31

Abstract

Background and Aims

Diabetes UK welcomes the 2011 decision by the WHO to accept the use of HbA1c testing in diagnosing diabetes.

Methods

A 51-year-old female was referred for diabetologist review of her newly diagnosed type 2 diabetes. The diagnosis was made in accordance with the NICE and WHO guidelines, based on two consecutive HbA1c results of 8.1% (65 mmol/mol). Her repeated fasting and postprandial plasma glucose were normal and she has no osmotic symptoms. Her fructosamine was normal. Her oral glucose tolerance test was normal as well. Her anti GAD and anti-islet cell antibodies are negative. The patient’s haematological indices were normal and she was unaware of any family history of hemoglobinopathy. She has family history of type 1 diabetes. After all these normal blood sugar reading, the diagnosis of diabetes was refuted and her treating GP was contacted to erase the diagnosis of type 2 diabetes from her medical record.. She was rejected in more than one occasion as a blood donor due to autoantibodies.

Results

We are not aware of any correlation between falsely high HbA1c and the presence of autoantibodies, currently, we are investigating that. We are sure during the conference of endo 2019; we will have a clear answer. We are aware of a case report from Canada of a similar scenario of false-positive HbA1c published in 2015(1).

Conclusions

We recommend that high HbA1c must be confirmed by another well-validated way of diabetes diagnosis such as osmotic symptoms or/and abnormal blood sugars readings to confirm the diagnosis of type 2.

Hide

TYPE 2 DIABETES: PATIENT EXPERIENCE OF CHRONIC ILLNESS CARE

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:31 - 09:32

Abstract

Background and Aims

We were to assess patient perception of chronic illness care in people with type 2 diabetes and to determine whether demographic variables, self-care behavior, as well as affective variables were related with perception of chronic illness care.

Methods

We conducted a secondary analysis of the previously published cross-sectional study in Iranian people with type 2 diabetes. Chronic illness care was assessed with the validated tool of patient assessment of chronic illness care (PACIC). Different aspects of care according with the chronic care model are measured on a scale of 1–5, with 5 being highest perception of care. The association between perception of chronic illness care and measured variables were tested using descriptive and bivariate statistics.

Results

Three hundred eighty participants completed the PACIC questionnaire (53.4% female, mean age: 54.73±8.0 years, mean PACIC score: 2.52±0.87). In univariate analysis considering PACIC score as the dependent variable, chronic illness care was inversely associated with level of education and distress, whereas, insulin treatment, wellbeing, family-social support and self-management were positively associated with chronic care (All p-value<0.05). In the multivariate forward stepwise logistic regression analysis, family-social support was positively related to chronic care while level of education, marital status, diabetes-related distress, and high density lipoprotein had significant negative relationship with PACIC score (All p-value<0.05).

Conclusions

Family-social support, level of education, marital status, and diabetes-related distress are the major determinants of patient experience of chronic illness care in people with type 2 diabetes.

Hide

REAL-WORLD USE OF IQCAST HYPOGLYCEMIA PREDICTION FEATURE IN THE GUARDIAN™ CONNECT SYSTEM AND ITS IMPACT ON CLINICAL OUTCOMES

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:32 - 09:33

Abstract

Background and Aims

Guardian™ Connect CGM system users can track the likelihood of experiencing low glucose episodes within a predictive window of 1 to 4 hours, using the IQCast feature in the Sugar.IQ™ diabetes assistant mobile application. The feature also notifies users of an impending low glucose episode. The frequency of low glucose episodes before and during four consecutive months of IQCast use were analyzed longitudinally.

Methods

Deidentified data from 259 individuals living with diabetes and who used the Sugar.IQ™ diabetes assistant prior to the introduction of the IQCast feature were selected for analysis. The mean (±SD) frequency of hypoglycemic (<70mg/dL) episodes was compared before IQCast use (15,639 days of data) and every month thereafter for four months (within-subjects ANOVA test, p<0.005 considered significant).

Results

Notifications of a predicted low glucose episode averaged 3.7/day. The overall and nighttime episodes/month observed for the 4 months before IQCast use was 17.8 (±16.2) and 5.1 (±5.9), respectively. For each month afterward, the frequency was 12.3 (±13.4), 12.0 (±12.9), 12.5 (±13.0) and 12.3 (±14.0) episodes/month and 3.5 (±4.4), 3.5 (±5.0), 3.6 (±4.5) and 3.4 (±5.7) episodes/month, respectively. Compared to the period before IQCast use, the frequency of hypoglycemic episodes post-use was significantly reduced (F[258,4]=15.63, p<0.001). This lowered frequency was maintained across the four-month period of IQCast use (F[258,3]=0.04, p=0.9867).

Conclusions

Results from this analysis demonstrate that using the Guardian™ Connect IQCast technology can result in a 31% decrease in the overall and nighttime hypoglycemic episode frequency.

Hide

EFFICACY AND SAFETY OF THE PATIENT EMPOWERMENT THROUGH PREDICTIVE PERSONALISED DECISION SUPPORT (PEPPER) SYSTEM: AN OPEN-LABEL RANDOMISED CONTROLLED TRIAL

Abstract

Background and Aims

The Patient Empowerment through Predictive Personalised Decision Support (PEPPER) system provides personalised insulin bolus advice for people with Type 1 diabetes (T1D) by means of an artificial intelligence-enhanced bolus calculator based on case-based reasoning. The system also incorporates a safety layer which includes predictive glucose alarms, low-glucose suspend for insulin pump users, and personalised carbohydrate recommendations. Here, we evaluate the safety, feasibility and usability of the PEPPER system compared to a standard bolus calculator.

Methods

Randomized controlled cross-over study. After a 4-week run-in, participants were randomized to PEPPER/Control or Control/PEPPER in a 1:1 ratio for 12-weeks. Participants then crossed over after a 3-week wash-out period. The primary endpoint is percentage time in range (3.9mmol/L – 10.0mmol/L) between the two groups. Secondary endpoints include percentage times in glycaemic ranges (hypo- and hyperglycaemia), glycaemic variability, HbA1c, quality of life questionnaires and safety system outcomes.

Results

58 participants (on multiple daily injections and insulin pump) were recruited at two clinical sites; Imperial College London in the UK and the Institut d'Investigació Biomèdica de Girona in Spain. Participants were (median(interquartile range)) aged 40.5 (30.8-50.3) years, with a diabetes duration of 21.0(11.5-26.3) years and HbA1c 61(59-66) mmol/mol. 51% used an insulin pump. Study completion is anticipated in November 2019 with a complete dataset of primary and secondary outcomes available for presentation in February 2020.

Conclusions

The results of this randomized controlled trial will help establish whether an adaptive bolus calculator with integrated safety system benefits individuals with T1D more than standard bolus advisors.

Hide

REAL-WORLD OUTCOMES VERSUS CLINICAL TRIALS RESULTS WITH DIPEPTIDYL PEPTIDASE-4 INHIBITORS: A SYSTEMATIC REVIEW PROTOCOL

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:34 - 09:35

Abstract

Background and Aims

Clinical guidance on the use of drugs is mainly supported by data from premarketing trials. However, results from clinical trials are not necessarily translated into real-world outcomes. The main objective of this study is to assess whether the outcomes of patients with type 2 diabetes mellitus treated under real-world conditions with dipeptidyl peptidase-4 (DPP-4) inhibitors reflect the results obtained from clinical trials.

Methods

This protocol conforms to the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) 2015 statement. This systematic review will assess the effects of DPP-4 inhibitors in pre-marketing studies (phase III randomized controlled trials [RCTs]) and in post-marketing real-world outcomes studies (phase IV RCTs and observational studies). A literature search will be performed at Medline, Embase, Cochrane Controlled Register of Trials (CENTRAL) and ClinicalTrials.gov. The following outcomes will be considered: i) efficacy endpoints: mean changes from baseline in hemoglobin A1C (HbA1c), fasting plasma glucose, and body weight, and patients achieving HbA1c <7%; ii) effectiveness endpoints: all-cause mortality, cardiovascular-related mortality, acute myocardial infarction, stroke, hospitalizations, emergency department visits, amputations, nephropathy and retinopathy. The risk of bias will be independently assessed according to the checklist proposed by Downs and Black. Data will be analyzed using descriptive statistics and meta-analysis when applicable.

Results

This review is ongoing and will be finished at the end of 2019.

Conclusions

The most relevant contribution of this study is providing evidence on both external validity of the results of clinical trials and on the extent to which diabetic patient benefit from DPP-4 inhibitors in clinical practice.

Hide

ASSESSMENT OF THE RELEVANCE OF PREDICTED BLOOD GLUCOSE CURVES AS A SUPPORT TO DECISION MAKING IN PATIENTS WITH TYPE 1 DIABETES

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:35 - 09:36

Abstract

Background and Aims

Patients with type 1 diabetes (T1D) make their decisions for insulin delivery from available blood glucose (BG) data and the expected effects on BG of forthcoming meals and activities according to education rules and their own experience. Enriched information on predicted BG glucose evolution could help them in better tuning insulin therapy. CDDIAB study’s objective was to evaluate the relevance of predicted BG trends in the decision-making process of the patient.

Methods

Eight patients (8F/6M, age: 51+/-15, T1D duration: 26+/-17 years, HbA1c: 7.09+/-0.82%) volunteered to track BG using a CGM device, meal intakes and insulin doses in real life conditions. The study ran over 30 days, and no specific intervention on the usual treatment was undertaken. Collected data has been used to train predictive models for each patient, in order to estimate future BG fluctuations up to 90 minutes. For each patient, low and high BG events were extracted, preditions were computed and patients were asked to make therapeutic decisions.

patient form.png

Results

Results were analysed by diabetologists of Montpellier Hospital, in order to evaluate the relevance of patients’ therapeutic decisions with prediction versus without:

results.png

Conclusions

Results show that in 84% of the cases presented to patients, prediction data drives better decision-making, and give insight into the benefits of implementing this technology in an open-loop system: predicted BG curve is a relevant and easy-to-read information to support decision making process.

The next step will be to test a decision support system, based on our prediction algorithms, which will provide therapeutic advices directly to the patient.

Hide

STOCHASTIC SEASONAL MODELS FOR GLUCOSE PREDICTION IN TYPE 1 DIABETES UNDER FREE-LIVING CONDITIONS

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:36 - 09:37

Abstract

Background and Aims

Accurate predictions of blood glucose (BG) concentration for large prediction horizons (PH) might improve type-1 diabetes therapy by allowing patients to adjust the therapy based on BG future values. Identification of seasonal local models after clustering BG data enhances other BG prediction approaches when used offline on available data sets. However, a methodology for the use under free-living conditions is needed.

Methods

Long-term full days BG historical data including post-prandial and nocturnal periods is partitioned into a set of event-to-event time subseries, driven by meals and night periods, and seasonality is enforced. Preprocessed data are then clustered into similar glycemic profiles and a seasonal model is identified for each cluster. The online BG prediction is obtained by local model integration through real-time membership-to-cluster estimation. As a proof of concept, the framework is tested over 6 months data of UVA/Padova simulator extended with several variability sources. Additionally to the BG prediction, an online monitoring system informs about prediction confidence and abnormal behavior detection.

Results

The framework exhibits high prediction accuracy for large PHs: a MAPE of 4.10%, 5.95%, 8.43%, 11.32%, 13.65%, and 13.97% has been achieved for 15-, 30-, 60-, 120-, 180-, and 240-min PHs, respectively.

Conclusions

The proposed system allows the use of seasonal models for BG prediction under free-living conditions, and therefore allows diabetic patients to anticipate therapeutic decisions and detect abnormal states.

Hide

A QUALITY OF CARE STUDY ON THE MANAGEMENT OF PATIENTS WITH DIABETIC FOOT ULCERS IN A TERTIARY HOSPITAL IN THE PHILIPPINES FROM 2013 TO 2017

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:37 - 09:38

Abstract

Background and Aims

Diabetic foot ulcers are among the leading causes of morbidity and mortality in diabetics. This study aims to assess the management of diabetic foot ulcers in East Avenue Medical Center in the last five years and compare it with the standard guideline of care.

Methods

A total of 267 charts were reviewed from 2013 to 2017 of patients with diabetic foot ulcers.

Results

The mean age of the patients was 57.31 years, while their mean HbA1c was 10.39%. The average duration of diabetes among the patients was 7.54 years. 41.95% of all patients received surgical intervention. The average number of days of hospital stay is 18.96 days. 14.61% of the total admission had adverse clinical outcome during their hospital stay. The most common of which were hospital acquired pneumonia and acute coronary syndrome. The mortality rate in this study is 13.11% The most common causes of death were acute coronary syndrome, septic shock secondary to infected wound and septic shock secondary to hospital acquired pneumonia.

Conclusions

The results of this study revealed the gravity of foot ulcers among diabetic patients. Improvement in the management of diabetic foot ulcers should be continued.

Hide

DISCREPANCY OF GLYCAEMIC RANGES IN REGARD TO CGM METRICS FOR CLINICAL CARE VERSUS GUIDANCE FOR GLYCAEMIC TARGETS WITHIN THE RECENT CONSENSUS ON TIR

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:38 - 09:39

Abstract

Background and Aims

The international consensus regarding clinical targets for continuous glucose monitoring defined percentages of total time people with type 1 diabetes (PWT1D) should spend in specified glycaemic ranges. According to the statement TBR-level 1 is defined as glucose concentrations 54-69mg/dL and TAR-level 1 181-250mg/dL, which is contrary to their recommendations regarding the guidance for the assessment of glycaemic control (TBR-level 1 <70mg/dL, TAR-level 2 >180mg/dL). The aim of this retrospective observational study was to assess if this discrepancy leads to different results for time spent in glycaemic ranges.

Methods

PWT1D (n=100, 49 female vs. 51 male, HbA1c 7.4±0.8%; age 42±14 years, 19 CSII, 81 MDI) used a flash glucose monitoring (FlashGM) system for 3 months from which >80% of the sensor data were available. FlashGM data were compared depending on the aforementioned recommendations via paired t-tests (p≤0.05) (Table 1).

Results

Table 1

Standardized CGM metrics for clinical care

Guidance for glycaemic targets

p-value

TBR-level 2

<54 mg/dL

2.0±2.2%

TBR-level 2 <54 mg/dL

2.0±2.2%

n/a

TBR-level 1

54-69 mg/dL

3.6±2.3%

TBR-level 1 <70 mg/dL

5.6±4.3%

p<0.0001

TIR

70-180 mg/dL

55.1±15.8%

TIR

70-180 mg/dL

55.1±15.8%

n/a

TAR-level 1 181-250 mg/dL

24.0±7.5%

TAR-level 1 >180 mg/dL

39.3±17.1%

p<0.0001

TAR-level 2 >250 mg/dL

15.3±11.3%

TAR-level 2 >250 mg/dL

15.3±11.3%

n/a

100±0%

117.3±11.1%

p<0.0001

Conclusions

Our results showed a difference between the standardized CGM metrics for clinical care versus guidance for glycaemic targets hence leading to different percentages in pre-specified glycaemic ranges.

Hide

THE PEPPER SYSTEM APPLICATION PROGRAM INTERFACE

REDUCTION OF POSTPRANDIAL HYPOGLYCEMIA IN INDIVIDUALS WITH TYPE 1 DIABETES USING AN INSULIN SENSITIVITY-INFORMED BOLUS CALCULATOR AFTER AN AEROBIC EXERCISE SESSION

Abstract

Background and Aims

Insulin sensitivity (SI) fluctuations, e.g. driven by physical activity and exercise, complicate insulin dosing and worsen glycemic control in individuals with type 1 diabetes (T1D). At last-year ATTD, we presented in-silico results demonstrating the benefit of using an SI-informed bolus calculator to account for exercise-induced SI changes. This year, we present the results from the first clinical deployment of the SI-informed system.

Methods

Fifteen subjects with T1D (male/female:10/5, age:45.1±12.6years, HbA1c:6.9±0.9%) using continuous glucose monitor (CGM) and insulin pump completed a 4-week at-home data collection, followed by two 24-hour hotel admissions. During the admissions, participants engaged into a 45-minute early-afternoon aerobic exercise session, after which they received a standardized dinner meal. The dinner bolus was computed using a standard or SI-informed bolus calculator; the latter modulates the insulin dose according to the deviation between usual SI estimated from historical data and real-time SI. Postprandial glycemic control was assessed by CGM-based low/high blood glucose indices (LBGI/HBGI) and percent time outside 70-180mg/dL, and compared between the two admissions.

Results

A 31% exercise-induced SI increase was observed (p=0.0002). The corresponding bolus modulation allowed to reduce postprandial hypoglycemia (ΔLBGI=(–)1.3, p=0.006; ΔPERC<70=(–)6.7%, p=0.049), without significant increase in hyperglycemia (ΔHBGI=1, p=0.075; ΔPERC>180=2.8%, p=0.5) [see figure]; the total number of administered hypoglycemia treatments was reduced from 12 to seven. Sensor glucose at dinnertime did not differ between the admissions (ΔCGM,DINN=(–)0.7mg/dL, p=0.925).

figure.png

Conclusions

This pilot study shows the safety and efficacy of using the proposed SI-informed bolus calculator in individuals with T1D. Future studies will be devoted to further testing the method.

Hide

DIABETES AND DEPRESSIVE SYMPTOMS: A CANADIAN REPRESENTATIVE SURVEY

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:41 - 09:42
Presenter

Abstract

Background and Aims

Depression prevalence is 2-3 times higher in people with versus without diabetes. Among Canadians (45-85 years) who participated in the Canadian Longitudinal Study on Aging (2010-2015), we evaluated associations across age groups between diabetes and (1) depression at baseline, (2) depressive symptoms at 18 months and (3) having sought medical care for these symptoms in the prior month.

Methods

Depression at baseline was identified by CES-D 10 (score ≥10). At 18 months, Kessler 10 scale (score ≥19) defined depressive symptoms and those who self-reported having sought care for these symptoms were identified.

Results

Among 29,933 individuals (mean age ± standard deviation 63 ± 10.4 years; 49% men), at baseline, 22.3% of those with diabetes had depression versus 15.2% of those without diabetes. In multivariate logistic regression models, individuals with (versus without) diabetes had higher risks of depression in all age groups (diabetes vs. no diabetes 45-60 years old: odds ratio, OR 2.00, 95% confidence interval, CI 1.70-2.39; 61-70 years: 1.39, 1.16-1.66 and 71-85 years: 1.37, 1.14-1.64). Among those with diabetes younger (versus older) individuals had higher risks of depression (diabetes 45-60 vs. 61-70 years old: OR 1.66, 95% CI 1.32-2.10 and vs. 71-85; 1.48, 1.17-1.87). Among those without diabetes, individuals 45-60 had a 13% increased risk of depression versus those 61-70 years old and similar risk versus those 71-85 years old. Depressive symptoms and seeking medical care for these symptoms at 18 months did not difer.

Conclusions

Younger individuals with diabetes had higher risks of depression compared to older individuals.

Hide

NEXT GENERATION HEALTH TECHNOLOGY ASSESSMENT TO SUPPORT PATIENT-CENTRED, SOCIETALLY ORIENTED, REAL-TIME DECISION-MAKING IN DIABETES

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:42 - 09:43

Abstract

Background and Aims

The current use of Health Technology Assessment (HTA) as a supporting tool for decision-making still varies considerably across different healthcare systems in Europe thus resulting in inefficiency of actual reporting of HTA. The new H2020 project, HTx (htx-h2020.eu) aims to develop a framework for next generation HTA that supports patient-centred, societally-oriented, real-time decision-making for integrated healthcare. Four different case studies will be performed to cover different types of disease areas, technologies and treatment strategies, one of them is focused on T1DM and T2DM.

Methods

HTx proposes the use of methods to bring together data from different sources such as Randomized Controlled Trials (RCT) and real-world data (RWD) with classical prediction modelling and artificial intelligence algorithms to integrate existing evidence and estimate relative clinical effectiveness and cost-effectiveness in complicated treatment and monitoring pathways.

Results

In T1DM and T2DM, the expected results are to assess the impact of using medical devices such as insulin infusion pumps, continuous glucose monitoring and glucose meters, e-health technologies such as tele-monitoring complemented by data visualization and decision support; and/or life-style interventions. We will link data from RCTs on these technologies to large population-based diabetes registries and health care claims data.

Conclusions

HTx should be able to provide more accurate estimations of the differential health impact of the technologies in specific subgroups of patients with Diabetes; predict which treatments combinations are most beneficial and cost-effective and facilitate tools that support patients and their healthcare providers in making personal decisions on the best treatment.

Hide

EFFICACY OF INSULCLOCK IN PATIENTS WITH POORLY CONTROLLED TYPE 1 DIABETES MELLITUS: A PILOT, RANDOMIZED CLINICAL TRIAL

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:43 - 09:44

Abstract

Background and Aims

Insulclock is an electronic device designed to improve treatment adherence and insulin injection tracking.We aimed, for the first time, to asses the impact of Insulclock on disease outcomes

Methods

Randomized, single-center, pilot study. We evaluated glycemic control, number of missed and mistimed insulin doses and quality of life after four weeks of Insulclock use in patients with uncontrolled T1DM. We also compared these outcomes between patients with or without receiving reminders and device alerts (Active or Blinded groups, respectively).

Results

Twenty one participants were recruited. Sixteen participants completed the study: 10 in the Active group and six in the Blinded group. The use of Insulclock was associated with a significant decrease in mean glucose levels (-27.0 mg/mL , p = 0.0126), a non-significant decrease (-2.8%, p = 0.6523) in the coefficient of variation (CV), and a significant increase in time in range (TIR) in the overall population (+7%, p = 0.038). No significant differences were observed in the change in HbA1C levels (-0.27%, p = 0.4098). TIR was more reduced in the Active group (-8%, p = 0.026). The number of missed and mistimed insulin doses decreased (-3.9; p = 0.1352, and -5.4; p = 0.0323, per month, respectively) in the overall population. Most of the items of The Insulin Treatment Satisfaction Questionnaire (ITSQ) improved after four weeks of Insulclock use.

Conclusions

This pilot study showed that Insulclock contributes to improving glycemic indices, decreasing missed and mistimed insulin doses, and improves treatment satisfaction in DM1 patients with persistent uncontrolled glycemic levels.

Hide

PROFESSIONAL CONTINUOUS GLUCOSE MONITORING IN HAEMODIALYZED DIABETIC PATIENTS.

Abstract

Background and Aims

It is inaccurate to assess blood glucose with glycated haemoglobin (HbA1c) in patients with chronic kidney disease and diabetes, also glycated albumin is ineffective to assess Time in Range (TIR) and Glycaemic Variability, emerging cardiovascular risks.

Methods

To evaluate glycaemic variability in Haemodialyzed Diabetic Patients we have used Glycaemic Holter (Medtronic Guardian Connect) in 81 Haemodialyzed Diabetic Patients, M 59, F 22, age 66,1 ± 9,5 ys, Diabetes Duration 21,7 ± 11,3 ys, T1DM 3 pts, T2DM 80 pts, Haemodialysis Duration 4,1 ± 3,4 ys, Haemodialysis 43,8%, Haemodialysis lfiltration 53,4%, Acetate-free biofiltration 2,7%, BMI 28,3 ±5,5 Kg/m2, HbA1c 7,2 ±1,36%, Hypertension 89,9%, Dyslipidaemia 60,7%, Heart Disease 29,1%, Stroke 7,59%, Diabetic Foot 26,5%, with amputation 6,5%, Peripheral Arterial Disease 35,4%, Retinopathy 52%. Insulin-Treated Patients 97%.

Results

Median data during 6 registration days were TIR 63%, Time Above Range TAR 33%, Time Below Range TBR 3%, calculated HbA1c 7±0.0%, first dialysis day TIR 82±20%, dialysis day interstitial glucose average 149±35mg/dl, non dialysis day interstitial glucose average 131mg/dl, pre dialysis interstitial glucose average 150±47 mg/dl, post dialysis interstitial glucose average126±38 mg/dl.

Conclusions

Our data demonstrate a moderate/good glucose control and confirm the burnt out of diabetes in dialysis. The observed trend was contrary to expectations with low TBR and higher TAR, without difference between dialysis days e non dialysis days, with the exception of lower interstitial glucose after dialysis. The work is in progress and we hope that new data will be able to help us to improve global treatment of these frail patients.

Hide

ILLNESS PERCEPTION AND ITS EFFECTS ON COPING, SELF-MANAGEMENT AND COMPLIANCE, AMONG TYPE 2 DIABETES PATIENTS

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:45 - 09:46

Abstract

Background and Aims

Type 2 diabetes is a chronic disease and requires an intensive and continuous daily monitoring and treatment in order to maintain a balanced and normal blood sugar levels, and to prevent complications in the various body systems. Diabetes can affect many areas of life, including various mental reactions, such as depression, and the patient's perception of illness has an impact on coping with this disease. Hence, the aim of this study was to examine the relationship between the components of illness perception and the coping of type 2 diabetes patients.

Methods

In this cohort study, 112 patients with type 2 diabetes were participated and completed the Illness Perception Questionnaire (IPQ-R), applied to diabetes. Demographic characteristics, such as details of diabetes status (duration of diabetes, treatments and complications) and glycosylated haemoglobin (HbA1c) were recorded. It is also included question about the coping with the disease and compliance to the treatment.

Results

Low compliance to the treatment was correlated with greater perceived symptom load (r=0.38, P<0.01), worse anticipated consequences (r=−0.45, P<0.01) and perceived lack of control of the disease (r=0.31, P<0.01). Linear regression revealed that perceived of high sense of control and less worse anticipated consequences predicted an effective coping with diabetes.

Conclusions

The importance of the study is to understand the factors that contribute to an effective diabetes coping and management and to increase responsiveness to treatment, balance and prevention of complications. The findings indicate the importance of diabetes patient's illness perception as contributing to coping with it.

Hide

HOW TO PREVENT AND REVERT DIABETES MELLITUS TYPE 2 (T2DM)

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:46 - 09:47

Abstract

Background and Aims

HOW TO PREVENT AND REVERT DIABETES MELLITUS TYPE 2 (T2DM)

Dora Mendoza, MD, PHD, MSC, MPH

Cellular 3115923210. Bogotá, Colombia.

ABSTRACT

This research has been taken more than 15 years in proven the accuracy

On what was learned from Professor Don Gregorio Marañón in his Institute

of Metabolism, Nutrition and Endocrinology, Madrid , Spain:

1-T2DM is dominantly transmitted to all generations

2- the only useful test to diagnose it is the Oral Glucose Tolerance Test (OGTT),

3- fasting test are not useful

4- the treatment is normal life style if not wanted the only medication useful is

insulin’s (basal or fast action)

Methods

More than thousand volunteers were studied using the OGTT

Results

6% of the studied were normal and had no family members with T2DM

The other 94% all had T2DM family members with T2DM. From these

13% had high levels of blood insulin

48% had Intolerance to Glucose

32% had Prediabetes

2% had asymptomatic hypoglycemia

Conclusions

The Dr. Gregorio Maranon’ teaching were proven scientific accurate

Methods

THE ORAL GLUCOSE TOLERANCE TEST WAS USED IN MORE THAN THOUSAND VOLUNTEERS

Results

THE DR. GREGORIO MARAÑÒN TEACHINGS WERE PROVEN SCIENTIFICALLY ACCURATE

Conclusions

1- T2DM IS A DOMINANTLY TRANSMITTED DISEASE

2- THE ONLY TEST USEFUL TO DIAGNOSE T2DM IS THE ORAL GLUCOSE TOLERANCE TESTS (OGTT) MEASUSIRN BLOOD INSULIN GLUCOSE AND GLUCAGON

3- FASTING TESTS ARE NOT USEFUL

4-THE TREATMENT IS HEALTY LIFE STYLE OR INSULINS INJECTIONS

5- HEMOBLOBIN A1C ONLY SHOWS FASTING HYPERGLYCEMIA

5-THE USE OF GLUCOMETERS IS A MUST

Hide

DECREASE IN HYPOGLYCEMIA EVENTS OVER TWO YEARS IN PATIENTS MONITORING WITH DIGITAL DIABETES MANAGEMENT SYSTEM

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:47 - 09:48

Abstract

Background and Aims

Hypoglycemia is a key risk factor and a major barrier in Diabetes management. Hypoglycemia events is a burden on health care systems due to the high cost of hypoglycemia-related emergency visits and hospitalizations. Dario, a digital Diabetes management system, may assist patients to reduce hypoglycemia events and glycemic control in users under insulin treatment.

Methods

A retrospective data analysis was performed on the Dario real-world database.

A population of 1,481 users under insulin therapy with type 1 and type 2 Diabetes that were using Dario over two years was evaluated.

Average numbers of level 1 hypoglycemia (<70mg/dL) and level 2 hypoglycemia (<54 mg/dL) events were calculated monthly and compared to baseline (first month).

Results

Continuous reduction in hypoglycemia events was observed throughout the 2 years period.
Average level 1 hypoglycemia events was reduced by 24% and 50% from baseline (1.39, 0.91 vs. 1.82 average events/month) after 6 months and after 2 years, respectively. Average level 2 hypoglycemia events were reduced by 16% and 56% from baseline (0.35, 0.18 vs. 0.42) after 6 months and after 2 years, respectively. Subgroup analyses of users with type 1 Diabetes (n=363) revealed substantial reduction of level 1 hypoglycemia events of 50% (2.5 vs. 5.0) and level 2 hypoglycemia of 55% (0.62 vs. 1.39) after 2 years. Moreover, a high reduction in hyper events was recorded as well.

Conclusions

To conclude: Patients using a digital Diabetes management platform have the potential to

improve their glycemic outcome and reduce emergency events and hospitalization.

Hide

UTILITY OF A GLUCOSE METER WITH BLUETOOTH AND WEB CONNECTIVITY IN AIDING PHYSICIANS AND PATIENTS ACHIEVE BETTER GLYCEMIC LEVELS

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:48 - 09:49

Abstract

Background and Aims

The One Touch Reveal (OTR) web application allows patients and HCPs to track progress and changes in glycemic control to support both self management and therapy decisions by summarizing BG or insulin data, displaying color coded trends and low or high BG patterns. The primary objective of our study was to assess changes in glycemic control and overall experiences of patients using the app in conjunction with bluetooth enabled OneTouch Verio Flex blood glucose meter(OTV).

Methods

Subjects were >18 years of age, diagnosed with T1DM(n=13) or T2DM(n=47) with an A1c < 11%. Glucose readings were instantly captured in the smartphone and web applications for the patient and physicians, thereby reducing the time frame required for a clinical decision making.

Results

Subjects had a mean age of 45.3±14.6 years and duration of diabetes of 12.9 ±10.6 years. At the end of 12 weeks, mean A1c decreased by 0.5%(p≤0.05). 78% of the subjects expressed visualisation of the colour coded trends and graphs as a motivation to continue SMBG. 67% of the patients attributed better adherence to the fact that the glycemic trends were visible to their HCPs in real time. 23% of the patients found the OTR app helpful in titrating the dosage of insulin.

Conclusions

OTR ecosystem via its bluetooth connectivity, mobile and web apps provides a broader data and actionable information for the patients and physicians. If an extra time is spent for interpreting the patterns and color coded information, it results in significant reduction in glycemic burden at no extra cost.

Hide

EFFECTIVE, SAFE AND USER-ACCEPTED BASAL-INSULIN THERAPY IN ELDERLY PATIENTS WITH TYPE 2 DIABETES RECEIVING DIGITALY ASSISTED DOMICILIARY NURSING CARE – THE GLUCOTAB@MOBILECARE STUDY

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:49 - 09:50

Abstract

Background and Aims

In elderly persons with type 2 diabetes (T2D), basal-insulin is recommended by international guidelines to simplify therapy regimens. A basal- and basal+ -insulin algorithm was developed and incorporated into a telemedical, digital workflow and decision support system (GlucoTab@MobileCare). The aim of this single-centre, non-controlled, proof-of-principle study was to investigate efficacy, safety and user-acceptance of GlucoTab@MobileCare in patients with T2D who received domiciliary nursing care.

Methods

Nine participants (5 females, age 77±10 years) received basal- or basal+ insulin therapy according to the suggestions of the GlucoTab@MobileCare algorithm during a three months study period.

Results

During the three months, HbA1c decreased from 60±13 mmol/mol to 57±12 mmol/mol. Mean morning BG was 171±68 mg/dl in the first study month vs. 145±35 mg/dl in the last study month. The glycaemic variability represented as standard deviation could be reduced by 50%. From 720 morning BG values, 60% were within, 37% above and 3% below BG target range. No severe hypoglycaemia <54 mg/dl occurred, three BG values (0.3%) were <70 mg/dl. The daily insulin dose was 24±13 IU at study start vs. 38±31 IU at study end. More than 95% of all requested actions (BG measurements, insulin dose calculations, insulin injections) were performed and documented in GlucoTab@MobileCare. Most nurses reported after study end, that the system supports the nursing staff to avoid errors (67%) and improves glycaemic control (67%).

Conclusions

The GlucoTab@MobileCare system supported an effective and safe treatment and was accepted by nurses. Further studies including larger sample sizes are required to confirm these proof-of-principle results.

Hide

A DIGITAL WORKFLOW AND DECISION SUPPORT SYSTEM TO PREVENT DIABETES-RELATED ACUTE HOSPITAL ADMISSIONS AND INPATIENT STAYS (THE GLUCOTAB@MOBILECARE STUDY)

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:50 - 09:51

Abstract

Background and Aims

International guidelines recommend basal-insulin for elderly persons with type 2 diabetes for safe glycaemic control. An algorithm-based (basal, basal+ insulin) digital workflow and decision support system (GlucoTab@MobileCare) was developed and tested in a proof-of-principle study. The system enables independent insulin adjustments based on algorithm suggestions for daily visiting of domiciliary nursing staff. The aim of this retrospective analysis was to investigate diabetes-related acute hospital admissions six months before, during and after the study.

Methods

Nine patients (5 females, age 77±10 years, HbA1c: start 60±13, end 57±12 mmol/mol) were treated for three months with GlucoTab@MobileCare. Data from electronical health records were analyzed regarding acute hospital admissions due to diabetes-related reasons six months prior to study, during the study period and six months after study termination.

Results

By using GlucoTab@MobileCare no acute diabetes-related hospital admissions occurred. Under routine care, in total six patients were admitted to the hospital. During six months prior to GlucoTab@MobileCare treatment, five emergency room (ER) visits occurred in three patients, resulting in four inpatient stays (length of stay (LoS): 11.3±3.4 days). During six months after GlucoTab@MobileCare treatment three ER visits occurred in two patients, resulting in two inpatient stays (LoS: 7.5±0.7 days). One patient was acutely admitted to the diabetes outpatient clinic. In eight cases, the cause for admission was hyperglycaemia; in one hypoglycaemia.

Conclusions

GlucoTab@MobileCare was able to prevent acute hospital admissions and subsequent inpatient stays due to blood glucose derailments. Further studies in a larger study population are needed to confirm these findings.

Hide

LOGISTIC REGRESSION FOR EARLY DETECTION OF HYPOGLYCEMIA

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:51 - 09:52

Abstract

Background and Aims

Detection of a potential hypoglycemia early enough is crucial to prevent its adverse effects ahead in time. Blood glucose prediction has been typically based on an ARX time series model or machine learning methods by training of neural networks, for example. Our goal is to investigate how logistic regression, where the probability of the future hypoglycemic event is modelled directly, performs in the task. The effect of variables other than CGM readings or insulin have been less studied with respect to predictive accuracy. Our aim is to also study how different predictor variables, particularly in relation to exercise, affect the predictions.

Methods

We use logistic regression to model the log-odds of the hypoglycemic event. The model is estimated from an individual T1DM patient’s data, and is thus personalized. We approach the early detection of hypoglycemia as a binary classification problem. At each time point, the classifier predicts, based on the values of predictor variables, that blood glucose will be either lower or higher than the hypoglycemic threshold at the end of the prediction time horizon. The classification is based on hard thresholding the estimated probability of hypoglycemia.

Results

We estimated our models on the OhioT1DM dataset, which contains six patients' data over a period of 8 weeks. We considered several hard threshold values for the classifier to compute the models' ROC curves, and compared predictive accuracy using different sets of predictor variables found in the dataset.

Conclusions

The results show that logistic regression is a viable alternative for the early detection of hypoglycemia.

Hide

DIGITAL DECISION SUPPORT FOR BASAL INSULIN THERAPY: RCT CONFIRMED READINESS FOR IMPLEMENTATION IN ACUTE-GERIATRIC CARE PATIENTS WITH TYPE 2 DIABETES

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:52 - 09:53

Abstract

Background and Aims

An algorithm for basal-insulin-therapy (including basal-plus) was developed based on the patients’ individual health status and incorporated into GlucoTab®. This workflow and decision support system provides suggestions for insulin dose and adjustment, blood glucose (BG) frequency and offers workflow support by visualization of open tasks. The aim was to investigate efficacy, safety and healthcare professional’s perception of the GlucoTab® system in patients at an acute-geriatric-hospital.

Methods

This randomized controlled trial was performed in 58 patients with type-2-diabetes aged ≥65 years. Patients in the intervention group (IG) (n=31, 65% females, 78±6 years, BMI: 29±6 kg/m2, HbA1c 56±10 mmol/mol) were treated according to GlucoTab® basal-insulin algorithm, patients in the control group (CG) (n=27, 63% females, 76±6 years, BMI 30±6 kg/m2, HbA1c 61±13 mmol/mol) received diabetes therapy (94% basal, basal-plus) according to physician’s prescription documented in GlucoTab®.

Results

The mean percentage of fasting blood glucose (FBG) values in the individual FBG target range was 59±33% (IG) vs. 51±31% (CG). BG values below, within, and above health dependent target range occurred in 4%, 75% and 22% vs. 1%, 75% and 25%, respectively. In both groups no severe hypoglycaemic event <40 mg/dl was detected. Healthcare-professionals stated that the decision support for basal-insulin is easy to handle (76%), is useful for routine care (53%) and assists in reduced consultations of physicians (65%).

Conclusions

The GlucoTab® basal-insulin algorithm provided an efficacious, safe and user-accepted glycaemic management at an acute-geriatric-hospital. Based on these results, the system will be further developed for routine use in geriatric patients (nursing home, domiciliary nursing).

Hide

IMPLEMENTATION OF A DIGITAL WORKFLOW AND DECISION SUPPORT SYSTEM CONFIRMS EFFICACY AND USER-SATISFACTION OF ROUTINE DIABETES CARE IN INPATIENTS WITH TYPE 2 DIABETES

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:53 - 09:54

Abstract

Background and Aims

The workflow and decision support system GlucoTab® was used for diabetes management in a regional hospital as replacement for the paper-based diabetes chart. Aim: to investigate the efficacy and user satisfaction of GlucoTab in hospitalized patients with type-2-diabetes in routine use. The use of decision support for basal-bolus-insulin was not focus of this implementation.

Methods

A retrospective comparison was conducted to analyze patient data before vs. after GlucoTab implementation (n=106 vs. n=241, age 72±14 vs. 75±11 years, female 49% vs. 51%, BMI 30±8 vs. 31±9 kg/m2, HbA1c 73±22 vs. 69±25 mmol/mol, respectively).

Results

Mean blood glucose (BG) was 174±32 (before) vs. 170±37 mg/dl (after) per patient. BG values in the ranges ≤54, ≤69, 70-179, 180-299 and ≥300 mg/dl occurred in 0.2% vs. 0.5%, 0.9% vs. 1.3%, 58.2% vs. 57.5%, 35.3% vs. 36.1% and 5.7% vs. 5.1% of all BG-values, respectively. Premixed insulin was prescribed in 51% vs. 30% during inpatient stay. Basal insulin was prescribed in 23% vs. 40%, respectively. Basal-bolus decision support was used in 7 patients. Healthcare professionals stated in a questionnaire that by using GlucoTab errors could be prevented (74%), workflows are completely digital supported (77%), interdisciplinary consultation is reduced (73%) and 10-15 minutes per patient-day could be saved.

Conclusions

GlucoTab completely replaced the paper-based insulin chart, was endorsed as support by healthcare professionals and is suitable for routine diabetes care in the hospital. It can be assumed that better glycaemic control can be achieved if the basal-bolus-insulin algorithm provided by GlucoTab is used in a larger proportion of inpatients.

Hide

THE ROLE OF TECHNOLOGY IN REAL-LIFE AND THE MANAGEMENT OF GLUCOSE VARIABILITY IN CHILDREN AND ADOLESCENTS WITH TYPE 1 DIABETES AND CELIAC DISEASE

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:54 - 09:55

Abstract

Background and Aims

Celiac disease (CD) can be associated to type 1 diabetes (T1D). Its treatment consists of a gluten-free diet. Some gluten-free foods have a high glycemic index, hypercaloric/low-fiber content and may affect glycaemic excursions. The aim of this study was to compare glucose monitoring (GM) profile of the last 14 days of subjects with T1D+CD on gluten-free diet to that with T1D alone.

Methods

All the subjects wore a glucose sensor (flash or continuous GM) and were on insulin pump (CSII) or multidaily injection (MDI) regimen. Insulin delivery sistems with predictive algorithms were excluded.

Results

155 subjects were studied, 43.5% males, mean age at T1D onset 6.3yrs±3.8 (T1D+CD 5.3yrs± 3.8vsT1D 7.0±3.7, p=0.006), mean age at study visit 12yrs ±3.9, mean % time sensor active 87,5%±13.1. Comparing patients with T1D+CD (n= 65) to those with T1D (n= 90), there were no significant differences in HbA1c (7,41%±0,8vs7,26±0,9 respectively), Time in range 70-180 mg/dl (53,0%±17,1vs53,3%±17,9), Time<70 mg/dl (3,3±3,0vs4,6±4,3), time>180 mg/dl (43,6±17,9vs42,2±18,9) and Coefficient of Variation (36,8±6,6vs37,5±7,0). Comparing patients with T1D+CD on CSII (n=33) to MDI (n=32), no significant differences in HbA1c (7,53% ±0,8vs7,3±0,9 respectively), time<70 mg/dl (2.9%±2.5vs3.7%±3.4), Coefficient of variation (35.9%±6.7vs37.8%±6.5), but higher time in range 70-180 mg/dl (58.6%±15.0vs47.5%±17.4, p=0.009), lower time>180 mg/dl (38.3%±16.0vs48.8%±18.4, p=0.025).

Conclusions

The use of GM allows patients on gluten-free diet to obtain a metabolic control comparable to that of patients without CD. The use of an integrated therapy with CSII has proved to be superior in improving metabolic control increasing the time in range 70-180 mg/dl.

Hide

DISCOVERING BLOOD GLUCOSE REGULATION PROCESSES WITH PROCESS MINING

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:55 - 09:56

Abstract

Background and Aims

Process mining (PM) is a knowledge extraction technique that aims to let us understand how processes are performing. PM is useful for gaining insights into how patients are self-managing and identifying the actions they perform effectively, as well as the actions that should be avoided. We aim to discover and compare blood glucose regulation processes using the indirect problem approach and process mining.

Methods

We first defined the in-day process of blood glucose (BG) regulation according to lifestyle observable variables (carbohydrate intake, insulin administration, physical activity and blood glucose levels). We mined the processes in a dataset containing recordings from Continuous Glucose Monitoring systems and self-recorded data. The observed time span was restricted to one month of observations. Main endpoints are to discover networks of events that affect the Time In Range (TIR) and the effect of insulin/food intake and physical activity on BG.

Results

The application of process mining allowed discovering significant differences in the observable BG regulation processes and thus helping to estimate and quantify the effect of insulin, food intake and physical activity over time. The process oriented view allowed to cluster different behaviors based on the TIR and transitions from BG ranges after insulin, carbohydrate and physical activity, providing more insights than time-oriented views and first order descriptors.

Conclusions

This is the first step towards discovering blood glucose regulation processes using the indirect problem approach and process mining techniques. Our results will produce new knowledge towards the improvement of closed-loop systems and optimization of self-management strategies

Hide

DYNAMIC RISK MODELS SUPPORTING PERSONALISED DIABETES HEALTHCARE WITH PROCESS MINING

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:56 - 09:57

Abstract

Background and Aims

Background and aims: Glycated hemoglobin A1c (HbA1c) has been recently shown as a weak indicator both short- and long-term blood glucose control. The popularization of continuous glucose monitoring (CGM) has fostered the upraise of the time-in-range (TIR) as a more robust and accurate metric for blood glucose control. Recent studies have shown a good correlation between HbA1c and TIR that may permit the transition to TIR as the preferred metric for determining the outcome of clinical.

This work aimed to examine HbA1c measures from a dynamic perspective by applying process mining tools, in order to obtain dynamic risk models of blood glucose control in general population.

Methods

Methods: We propose a method based on the use of process mining to discover and identify HbA1c changes during a period, and apply Clustering Algorithms to discover dynamic risk models for diabetes using HbA1c test results and other variables available in Health Electronic Records, such as Body Mass Index, age and co-morbidities.

We applied this methodology to real data from 50,169 patients followed-up for seven years (2012-2018) in the outpatient clinics of a tertiary hospital.

Results

Results: Results showed a population stratification and characterisation based on their dynamic evolution of HbA1c results over the given period, that let us inferring a Dynamic Diabetes Risk Model in an understandable way for health professionals.

Conclusions

Conclusion: This information will support health professionals to translate one-fits-all current approach of treatments and care, to a personalised one, fitting treatment strategy based on patients’ unique behaviour thanks to dynamic modelling.

Hide

A MACHINE LEARNING APPROACH FOR DETECTING INSULIN PUMP FAULTS

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:58 - 09:59

Abstract

Background and Aims

Recent advancements of closed-loop insulin delivery for T1D therapy aim to achieve better control while simultaneously reducing the need of patient interventions. As patients become increasingly confident in new technologies, it is important to develop automated methods that can detect possible malfunctioning in real-time. This work analyzes an innovative approach to detect insulin pump faults (IPF) based on advanced machine learning algorithms.

Methods

Using the latest version of the T1D Padova/UVA simulator, we generated a 30 days dataset in which we simulated IPF occurring at night, during fasting, and during the day, in correspondence of meals. From the data, we extracted a large pool of numerical attributes (features) capable of describing the patient status over time and highlighting suspicious portions. Then, a selection procedure was performed to determine the most effective feature set for detecting IPFs. Finally, we compared the performance of several state of art unsupervised anomaly detection algorithms on the data obtained.

Results

The best performance overall is obtained by the Histogram Based Outlier Score algorithm, which detected 87% of the IPF with 0.08 False Positives per day on average. Considering only the overnight period, we obtained a recall of 0.79 with 0.03 FP/day. In the diurnal portion, we obtained a higher recall of 0.95 with 0.05 FP/day.

Conclusions

In-silico data show that the proposed method can detect IPF and improve the safety of CL systems. In the future, we aim to test it inside dedicated clinical trials.

Hide

RETROSPECTIVE ADHERENCE DETECTION OF SIMULATED T2D PATIENTS ON BASAL INSULIN TREATMENT USING MACHINE LEARNING

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
09:59 - 10:00

Abstract

Background and Aims

Poor adherence of T2D patients on basal insulin treatment can lead to life-threatening health complications. In this context, CGM technology may be a future key to improve diabetes management by positively impacting treatment adherence. The aim of this study was to explore the use of Machine Learning (ML) for automated detection of basal insulin injection adherence in a retrospective view based on CGM data.

Methods

A cohort of subjects with simulated CGM data was generated using a T2D modified Medtronic Virtual Patient model with adherence annotation based on patient-specific once-daily basal insulin injection. Simple feature-engineered (CGM consensus inspired) ML classification models (logistic regression) were compared to more advanced deep learning (DL) models based on automatic feature-extraction (convolutional neural networks). Additionally, a fusion of both expert-dependent features and automatically extracted features (acquired from raw CGM data) were investigated. All models were reported as ensemble accuracies from a leave-one-patient-out cross-validation step.

Results

The results indicate that the simulated CGM data from the classification day and the consecutive day provide close to all information on whether a CGM day is considered adherent or not with ~ 80% accuracy (50% baseline), i.e. whether the basal insulin was injected or not. In this context, the fused DL models performed similarly to the simple feature-dependent classification models.

Conclusions

Given the simulated T2D based CGM data, simple feature-engineered ML models can be used for retrospective adherence detection based on the classificarion day and the consecutive day. The study should be followed up by analysis of real CGM data.

Hide

THE EFFECT OF GLYCEMIC LOAD ON POSTPRANDIAL BLOOD GLUCOSE CHARACTERISTICS IN PATIENTS WITH GESTATIONAL DIABETES

Abstract

Background and Aims

Despite there is a broad discussion concerning the importance of glycemic index and glycemic load in diabetes treatment, there is no clear evidence of how glycemic index affects postprandial glycemic response (PPGR). The aim of the study is to evaluate this relation in gestational diabetes patients.

Methods

The glycemic index was assigned to every item in the database, which patients used to record data on meals with a mobile app developed for the study. The GI values were sourced from University of Sydney’s database. The CGM and meal data were collected during a week in free living conditions for patients with gestational diabetes and in healthy pregnant women.

Results

The information on 2054 food intakes and postprandial blood glucose curves from 125 participants were collected. The Pearson correlation between amount of carbohydrates and glycemic load with the following PPGR characteristics for all collected data was: incremental area under the curve 120 minutes after the meal r=0.430 and r=0.418, the rise of blood glucose level from meal start to peak value r=0.420 and r=0.418, peak blood glucose value r=0.342 and r=0.336 respectively. P-value for all estimations was below 0.001.figure_1.png

Conclusions

Current study has shown no evidence of better correlation between glycemic load and PPGR characteristics in comparison to the amount of carbohydrates in a meal when used in remote health monitoring. Neither meal carbohydrates, nor glycemic load alone cannot be used effectively for PPGR prediction, while there are many other factors necessary to consider.

The study was funded by Russian Science Foundation (project №18-75-10042).

Hide

MISSED PRANDIAL INSULIN BOLUSES REDUCE TIME IN RANGE (TIR) AND INCREASE TIME ABOVE RANGE (TAR): NEW METHODS OF ANALYSIS

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
10:01 - 10:02

Abstract

Background and Aims

Insulin nonadherence leads to poor glycaemic control. Several methods have been proposed for identifying missed and suboptimal meal-related insulin bolus. To improve the analysis of results from clinical studies, we compared a new method based on the Mean Amplitude of Glycaemic Excursion (MAGE) concept (Method 1) and Kovatchev-Breton method (Method 2).

Methods

Method 1 uses Baghurst’s algorithm for MAGE to identify prandial glucose peaks utilising upstrokes which exceed a prespecified amplitude rather than the standard deviation of continuous glucose monitoring (CGM). Method 2 uses a >70 mg/dL increase within 2 hours of glucose increase. Missed bolus dose was defined as no injection within 2 hours prior to onset of the glucose excursion. The methods were compared using data from a 12-week study with two 21-day periods (period 1: masked CGM, 68 subjects; period 2: real-time CGM, 65 subjects; using Dexcom G5) in subjects with type 1 diabetes or type 2 diabetes using a basal-bolus regimen with prandial insulin lispro U-100 injected using a connected pen.

Results

In both periods, days with no missed bolus had lower mean glucose, higher TIR and lower TAR compared to days with ≥1 missed bolus (Table).

Conclusions

Methods 1 and 2 have excellent concurrence; both identify the effects of missed bolus on TIR and TAR. This study highlights the utility of connected insulin pens to identify suboptimal diabetes self-management and points to the need for development of consensus in the clinical community about the reporting of insulin dosing metrics.

table final 3 oct.jpg

Hide

CALCIFICATION BIOMARKERS AND VASCULAR DYSFUNCTION IN OBESITY AND TYPE 2 DIABETES: INFLUENCE OF ORAL HYPOGLYCEMIC AGENTS

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
10:02 - 10:03

Abstract

Background and Aims

Vascular aging in obesity and type 2 diabetes (T2D) is associated with progressive vascular calcification, an independent predictor of morbidity

and mortality. Pathways for vascular calcification modulate bone matrix deposition, regulating calcium deposits. We investigated

the association between biomarkers of vascular calcification and vasodilator function in obesity or T2D, and whether antidiabetic

therapies impact those markers.

.

Methods

Circulating levels of proteins involved in vascular calcification, as osteopontin (OPN), osteoprotegerin (OPG)), regulated on activation, normal T cell expressed and secreted (RANTES), fetuin-A were measured in lean subjects, individuals with metabolically healthy obesity (MHO), patients with metabolically unhealthy obesity (MUO) or T2D. Vasodilator function was assessed by infusion of ACh and sodium nitroprusside (SNP).

Results

Circulating levels of OPN were higher in the MUO/T2D group than in lean subjects (P 0.05); OPG and RANTES were

higher in MUO/T2D group than in the other groups (both P 0.001); fetuin-A was not different between groups (P 0.05); vasodilator

responses to either ACh or SNP were impaired in both MUO/T2D and MHO compared with lean subjects (all P 0.001). In patients with

T2D who were enrolled in the intervention trial, antidiabetic treatment with glyburide, metformin, or pioglitazone resulted in a significant

reduction of circulating OPG (P 0.001), without changes in the other biomarkers and vasodilator responses (all P 0.05).

Conclusions

In conclusion,obese patients with MUO/T2D have elevated circulating OPN, OPG, and RANTES; in these patients, antidiabetic treatment reduces

only circulating OPG. Further study is needed to better understand the mechanisms of vascular calcifications in obesity and diabetes.

Hide

USE OF AN AUTOMATED SMARTPHONE-BASED CARBOHYDRATE ESTIMATOR IN TYPE 1 DIABETES THERAPY: CLINICAL IMPACT ASSESSED BY COMPUTER SIMULATION

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
10:03 - 10:04

Abstract

Background and Aims

Carbohydrate (CHO) counting is a burdensome task for people with type 1 diabetes (T1D). Wrong estimation of meal CHO content increases risks of post-prandial hypo-/hyper-glycemia. An automated CHO estimator (ACE) using food pictures taken from a smartphone can potentially help subjects with T1D improving their glycemic control while reducing the burden of diabetes self-management. Here, we aim to in silico compare the clinical impact of the ACE vs. patient CHO estimation (PCE) in T1D therapy.

Methods

One hundred and twenty-eight food items, for which true CHO amounts were available, were scanned with an early ACE prototype. For each item, the Relative Signed Error (RSE) of CHO estimation was calculated and the resulting distribution was fitted to a Student's t probability function model. This model was incorporated into the UVA/Padova T1D simulator, already equipped with a statistical model describing PCE error (Vettoretti et al., 2018). Thus, a 3-day 3-meal/day in silico head-to-head trial was performed to compare the effects of ACE vs. PCE in terms of relevant glucose control metrics.

Results

Our data showed that ACE slightly overestimated true CHO amount (median [25th,75th] percentile RSEACE=0.9% [-7.5%,19.3%]), while patients usually tended to underestimate it (RSEPCE=-4.9% [-21.3%,7.5%]). Simulated glucose control metrics are reported in Table 1.

table.png

Conclusions

Despite statistically significant, the differences in outcome metrics are marginal from a clinical point of view. We conclude that, to date, the use of the ACE prototype likely would not improve glucose control but might ease T1D management. These results warrant confirmation in a head-to-head clinical trial.

Hide

ASSESSMENT OF GLYCAEMIA BY FINGERSTICK BLOOD GLUCOSE MONITORING MAY UNDERESTIMATE THE REQUIREMENT FOR INSULIN TO ADDRESS ELEVATED NOCTURNAL GLUCOSE LEVELS IN WOMEN WITH GDM

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
10:04 - 10:05

Abstract

Background and Aims

The commencement of insulin in women diagnosed with GDM is determined by health-care professional’s (HCP) perception of the patient’s glycaemia based on self-monitoring blood glucose (SMBG) and HbA1c. Macrosomia occurs in GDM patients despite optimal care. One reason may be that SMBG fails to correctly identify all patients requiring insulin.

Methods

We studied women attending two obstetric centres post-OGTT diagnosis of GDM. Within 2-weeks of GDM diagnosis, participants were provided dietary advice, taught SMBG, and 7-day CGM was initiated before insulin introduction with data masked to HCP and patients.

Results

Ninety women, mean age±SD 31±4 years; gestation 27±1 weeks were studied. Fasting OGTT glucose (5.2 vs 4.8mM; P=0.0004) and mean CGM glucose (5.7 vs 5.3mM; P<0.0001) was higher in those prescribed insulin (n=34). Figure 1 provides CGM profiles (median and IQR) and time >5.0mM overnight (0:00-03:00 and 03:00-06:00) and >6.5mM during the daytime. During the daytime, mean±SD time spent >6.5mM was 212±135 vs 123±98min for insulin prescribed vs non-insulin prescribed patients, respectively. Overnight (0:00-03:00 and 03:00-06:00) mean±SD time >5.0mM was 133±31 vs 96±29min and 118±42 vs 70±53min for insulin prescribed vs non-insulin prescribed patients. HCP time correlated with prescription of insulin (185 vs 108min in insulin vs non-insulin prescribed GDM; P=0.0001) but not with CGM and OGTT parameters.

Conclusions

More than 50% of women diagnosed with GDM not subsequently prescribed insulin during pregnancy had glucose levels above 5.0mM for >10% time between 0:00-03:00. CGM may facilitate triage of patients and an effective intervention by providing a better assessment of nocturnal hyperglycaemia vs SMBG.

attd figure 1_zaharieva.jpg

Hide

PERSONALIZED MEAL INSULIN BOLUS FOR TYPE 1 DIABETES USING DEEP REINFORCEMENT LEARNING

Session Name
CLINICAL DECISION SUPPORT SYSTEMS/ADVISORS
Session Type
E-POSTER VIEWING (EXHIBITION HOURS)
Date
20.02.2020, Thursday
Session Time
09:30 - 15:30
Channel
E-Poster Area
Lecture Time
10:05 - 10:06

Abstract

Background and Aims

Due to the high inter- and intra-population variability, it is often difficult for basal-bolus insulin therapy to effectively prevent postprandial hyperglycaemia and hypoglycaemia. Deep reinforcement learning (DRL) has recently achieved great success in different areas and has the potential to meet this challenge.

Methods

We propose a new DRL algorithm using an actor-critic architecture and fully connected neural networks, following the approach of deep deterministic policy gradient. We employ the UVa/Padova Type 1 Diabetes (T1D) Simulator as a testing platform and let the agent interact with the environment at each incidence of a meal intake. The agent’s states consist of real-time samples from continuous glucose monitoring, carbohydrate estimation, and meal ingestion time. Its action is to estimate the dose of premeal boluses. The percentage time in range (TIR) (70–180mg/dl) is set as the reward. We first trained the agent in a long-duration average scenario through long-term self-exploring to obtain a generalized model. Then, personalized tuning was performed for each T1D subject within a shorter scenario (two months). Finally, we evaluated the model on 10 virtual adult subjects over 12 months.

Results

Compared to baseline methods, i.e. standard bolus calculator, the overall mean TIR increased from 81.9±8.6% to 88.3±5.1% (p<0.005) with a non-significant decrease in hypoglycaemia.

Conclusions

This work presents a new meal insulin bolus recommender for T1D using DRL which has been proven to achieve, in silico, a significant improvement in glycaemic control.

Hide