Female sex is a risk factor for painful diabetic peripheral neuropathy: the EURODIAB prospective diabetes complications study.

AIMS/HYPOTHESIS: While the risk factors for diabetic peripheral neuropathy (DPN) are now well recognised, the risk factors for painful DPN remain unknown. We performed analysis of the EURODIAB Prospective Complications Study data to elucidate the incidence and risk factors of painful DPN. METHODS: The EURODIAB Prospective Complications Study recruited 3250 participants with type 1 diabetes who were followed up for 7.3±0.6 (mean ± SD) years. To evaluate DPN, a standardised protocol was used, including clinical assessment, quantitative sensory testing and autonomic function tests. Painful DPN (defined as painful neuropathic symptoms in the legs in participants with confirmed DPN) was assessed at baseline and follow-up. RESULTS: At baseline, 234 (25.2%) out of 927 participants with DPN had painful DPN. At follow-up, incident DPN developed in 276 (23.5%) of 1172 participants. Of these, 41 (14.9%) had incident painful DPN. Most of the participants who developed incident painful DPN were female (73% vs 48% painless DPN p=0.003) and this remained significant after adjustment for duration of diabetes and HbA1c (OR 2.69 [95% CI 1.41, 6.23], p=0.004). The proportion of participants with macro- or microalbuminuria was lower in those with painful DPN compared with painless DPN (15% vs 34%, p=0.02), and this association remained after adjusting for HbA1c, diabetes duration and sex (p=0.03). CONCLUSIONS/INTERPRETATION: In this first prospective study to investigate the risk factors for painful DPN, we definitively demonstrate that female sex is a risk factor for painful DPN. Additionally, there is less evidence of diabetic nephropathy in incident painful, compared with painless, DPN. Thus, painful DPN is not driven by cardiometabolic factors traditionally associated with microvascular disease. Sex differences may therefore play an important role in the pathophysiology of neuropathic pain in diabetes. Future studies need to look at psychosocial, genetic and other factors in the development of painful DPN.

Sustaining the benefits of structured education: Participants' experiences of receiving structured individual support during a programme (DAFNEplus) informed by behavioural science.

AIMS: The DAFNEplus programme seeks to promote sustained improvements in glycaemic management by incorporating techniques from behavioural science. It includes five sessions of structured individual support delivered over 12 months following group education. As part of a broader evaluation, and to inform decision-making about roll-out in routine care, we explored participants' experiences of, and engagement with, that individual support. METHODS: We interviewed DAFNEplus participants (n = 28) about their experiences of receiving individual support and the impact they perceived it as having on their self management practices. We analysed data thematically. RESULTS: Participants described several important ways individual support had helped strengthen their self management, including: consolidating and expanding their understandings of flexible intensive insulin therapy; promoting ongoing review and refinement of behaviour; encouraging continued and effective use of data; and facilitating access to help from healthcare professionals to pre-empt or resolve emergent difficulties. Participants characterised themselves as moving towards independence in self management over the time they received individual support, with their accounts suggesting three key stages in that journey: 'Working with healthcare professionals'; 'Growing sense of responsibility'; and, 'Taking control'. Whilst all portrayed themselves as changed, participants' progress through those stages varied; a few continued to depend heavily on DAFNEplus facilitators for advice and/or direction at 12 months. CONCLUSIONS: While all participants benefited from individual support, our findings suggest that some may need, or gain further benefit from, longer-term, tailored support. This has important implications for decision-making about roll-out of DAFNEplus post-trial and for the development of future programmes seeking to bring about sustainable changes in self management practices.

Analysis of real-world capillary blood glucose data to help reduce HbA1c and hypoglycaemia in type 1 diabetes: Evidence in favour of using the percentage of readings in target and coefficient of variation.

AIMS: To examine real-world capillary blood glucose (CBG) data according to HbA1c to define proportions of CBG readings at different HbA1c levels, and evaluate patterns in CBG measurements to suggest areas to focus on with regard to self-management. METHODS: A retrospective analysis stratified 682 adults with type 1 diabetes split into quartiles based on their HbA1c . The proportions of results in different CBG ranges and associations with HbA1c were evaluated. Patterns in readings following episodes of hyperglycaemia and hypoglycaemia were examined, using glucose to next glucose reading table (G2G). RESULTS: CBG readings in the target range (3.9-10 mmol/L) increase by ~10% across each CBG quartile (31% in the highest versus 63% in the lowest quartile, p < 0.05). The novel G2G table helps the treatment-based interpretation of data. Hypoglycaemia is often preceded by hyperglycaemia, and vice-versa, and is twice as likely in the highest HbA1c quartile. Re-testing within 30 min of hypoglycaemia is associated with less hypoglycaemia, 1.6% versus 7.2%, p < 0.001, and also reduces subsequent hyperglycaemia and further hypoglycaemia in the proceeding 24 h. The coefficient of variation, but not standard deviation, is highly associated with hypoglycaemia, r = 0.71, and a CV ≤ 36% equates to 3.3% of CBG readings in the hypoglycaemic range. CONCLUSIONS: HbA1c <58 mmol/mol (7.5%) is achievable even when only ~60% of CBG readings are between 3.9-10 mmol/L. Examining readings subsequent to out-of-range readings suggests useful behaviours which people with type 1 diabetes could be supported to adhere to, both in a clinic and structured education programmes, thereby decreasing the risk of hypoglycaemia whilst also reducing hyperglycaemia and improving HbA1c .

Interpretable Machine Learning for Inpatient COVID-19 Mortality Risk Assessments: Diabetes Mellitus Exclusive Interplay.

People with diabetes mellitus (DM) are at elevated risk of in-hospital mortality from coronavirus disease-2019 (COVID-19). This vulnerability has spurred efforts to pinpoint distinctive characteristics of COVID-19 patients with DM. In this context, the present article develops ML models equipped with interpretation modules for inpatient mortality risk assessments of COVID-19 patients with DM. To this end, a cohort of 156 hospitalised COVID-19 patients with pre-existing DM is studied. For creating risk assessment platforms, this work explores a pool of historical, on-admission, and during-admission data that are DM-related or, according to preliminary investigations, are exclusively attributed to the COVID-19 susceptibility of DM patients. First, a set of careful pre-modelling steps are executed on the clinical data, including cleaning, pre-processing, subdivision, and feature elimination. Subsequently, standard machine learning (ML) modelling analysis is performed on the cured data. Initially, a classifier is tasked with forecasting COVID-19 fatality from selected features. The model undergoes thorough evaluation analysis. The results achieved substantiate the efficacy of the undertaken data curation and modelling steps. Afterwards, SHapley Additive exPlanations (SHAP) technique is assigned to interpret the generated mortality risk prediction model by rating the predictors' global and local influence on the model's outputs. These interpretations advance the comprehensibility of the analysis by explaining the formation of outcomes and, in this way, foster the adoption of the proposed methodologies. Next, a clustering algorithm demarcates patients into four separate groups based on their SHAP values, providing a practical risk stratification method. Finally, a re-evaluation analysis is performed to verify the robustness of the proposed framework.

Assessment of the psychometric properties and refinement of the Health and Self-Management in Diabetes Questionnaire (HASMID).

BACKGROUND: The Health And Self-Management In Diabetes (HASMIDv1) questionnaire consists of 8 attributes, 4 about quality of life, and 4 about self-management. The overall aim of this study was to rigorously examine the psychometric properties of the HASMIDv1 questionnaire. METHODS: The study comprised two phases. Phase 1 identified items of the HASMIDv1 questionnaire that potentially required rewording through consultation with a patient involvement panel and two focus groups of people with diabetes. Phase 2 involved a cross-sectional longitudinal survey where HASMID, EQ-5D-5L, health, treatment and sociodemographic questions were administered using both paper and online versions to people with diabetes. Participants were asked to complete the survey again approximately 3 months later. Psychometric analyses were undertaken to examine floor and ceiling effects, item distributions, known group differences and internal consistency. Rasch analysis was undertaken to assess differential item functioning and disordered thresholds. RESULTS: Phase 1 derived five alternative wordings to items: Irritable, Affects Mealtimes, Daily Routine, Social Activities and Problem. Phase 2 achieved 2835 responses at time point 1 (n = 1944 online, n = 891 paper version) and 1243 at time point 2 (n = 533 online, n = 710 paper version). Overall the HASMID items performed well, though two alternative worded items (Irritable and Social Activities) provided additional information not fully captured by the original HASMID items. CONCLUSION: Psychometric evaluation and Rasch analysis were used in conjunction with expert opinion to determine the final questionnaire. The application of psychometric analyses or Rasch analysis alone to inform item selection would have resulted in different items being selected for the final instrument. The benefit of a combined approach has produced an instrument which has a broader evaluation of self-management. The final validated HASMID-10 is a short self-report PRO that can be used to evaluate the impact of self-management for people living with diabetes. HASMID-10 can be scored using total summative scores, with utility and monetary values also available for use in cost-utility and cost-benefit analyses.

Painful and Painless Diabetic Neuropathies: What Is the Difference?

PURPOSE OF REVIEW: The prevalence of diabetes mellitus and its chronic complications are increasing to epidemic proportions. This will unfortunately result in massive increases in diabetic distal symmetrical polyneuropathy (DPN) and its troublesome sequelae, including disabling neuropathic pain (painful-DPN), which affects around 25% of patients with diabetes. Why these patients develop neuropathic pain, while others with a similar degree of neuropathy do not, is not clearly understood. This review will look at recent advances that may shed some light on the differences between painful and painless-DPN. RECENT FINDINGS: Gender, clinical pain phenotyping, serum biomarkers, brain imaging, genetics, and skin biopsy findings have been reported to differentiate painful- from painless-DPN. Painful-DPN seems to be associated with female gender and small fiber dysfunction. Moreover, recent brain imaging studies have found neuropathic pain signatures within the central nervous system; however, whether this is the cause or effect of the pain is yet to be determined. Further research is urgently required to develop our understanding of the pathogenesis of pain in DPN in order to develop new and effective mechanistic treatments for painful-DPN.

PROM Validation Using Paper-Based or Online Surveys: Data Collection Methods Affect the Sociodemographic and Health Profile of the Sample.

OBJECTIVE: This study examines the impact of data collection method on the sociodemographic and health profile of samples of people with diabetes who complete either an online or postal patient-reported outcome measure (PROM) validation survey. METHODS: A longitudinal survey of people with diabetes was conducted using online and postal survey versions. The survey consisted of sociodemographic and health questions, a health and self-management PROM (Health and Self-Management in Diabetes [HASMID]), and 5-level version of EQ-5D. Dose adjustment for normal eating Online, Diabetes UK, and social media were used to recruit online survey participants. A panel of patients at a local National Health Service Trust was randomly allocated to participate in either survey version (two-thirds to postal version). Participants were asked to complete the survey again approximately 3 months later. RESULTS: A total of 2784 participants completed the survey (1908 online, 876 postal). The samples (online versus postal) differed; the online sample was younger, with a larger proportion of women and respondents with type 1 diabetes. There were significant differences in sociodemographic characteristics by type of diabetes across data collection mode. The proportion of respondents who responded again at point 2 was higher in the postal sample (525 postal, 698 online). CONCLUSION: The sociodemographic and health profile of samples of people with diabetes differed depending on whether they completed the online or postal survey. Differences are likely due to different recruitment methods and differences in those choosing to respond to different survey versions. Future PROM validation surveys should select data collection methods carefully because these can affect sample characteristics and results.

Disruptive illness contexts and liminality in the accounts of young people with type 1 diabetes.

We utilise Bury's (1982) biographical disruption to examine young people's experiences of type 1 diabetes. Our findings show that young adults adopted various 'subject positions' across different illness contexts. The subject positions deployed are intended to produce a particular kind of normal embodied identity unaffected by diabetes. First, participants concealed their illness in public spaces and challenged cultural stereotypes of diabetes to maintain a normal illness biography. Disruption was ever present and required careful negotiation to avoid exposure of illness in public. Young adults upheld a 'normal public presentation'. Second, they resisted the medical system's pressure to adhere to glucose targets asserting and maintaining a subject position of 'independent and autonomous young adults'. Here, disruption was transient and temporary, present in the clinic but not always beyond. It remained in the background for much of the time until it was reinforced by parents or at meal times. Third, young adults acquired a 'pragmatic subject position' with diabetes viewed as complex but manageable, no longer a target for resistance. Frank's (1995) 'narrative restitution' is adopted to describe the transition to life with 'normal' illness. We argue that illness experience was 'liminal' and reflected the subject positions adopted by young adults.

Using a Discrete-Choice Experiment Involving Cost to Value a Classification System Measuring the Quality-of-Life Impact of Self-Management for Diabetes.

OBJECTIVES: To describe the use of a novel approach in health valuation of a discrete-choice experiment (DCE) including a cost attribute to value a recently developed classification system for measuring the quality-of-life impact (both health and treatment experience) of self-management for diabetes. METHODS: A large online survey was conducted using DCE with cost on UK respondents from the general population (n = 1497) and individuals with diabetes (n = 405). The data were modeled using a conditional logit model with robust standard errors. The marginal rate of substitution was used to generate willingness-to-pay (WTP) estimates for every state defined by the classification system. Robustness of results was assessed by including interaction effects for household income. RESULTS: There were some logical inconsistencies and insignificant coefficients for the milder levels of some attributes. There were some differences in the rank ordering of different attributes for the general population and diabetic patients. The WTP to avoid the most severe state was £1118.53 per month for the general population and £2356.02 per month for the diabetic patient population. The results were largely robust. CONCLUSIONS: Health and self-management can be valued in a single classification system using DCE with cost. The marginal rate of substitution for key attributes can be used to inform cost-benefit analysis of self-management interventions in diabetes using results from clinical studies in which this new classification system has been applied. The method shows promise, but found large WTP estimates exceeding the cost levels used in the survey.

Follow-Up Support for Effective type 1 Diabetes self-management (The FUSED Model): A systematic review and meta-ethnography of the barriers, facilitators and recommendations for sustaining self-management skills after attending a structured education programme.

BACKGROUND: People with type 1 diabetes who attend structured education training in self-management using flexible intensive therapy achieve improved blood glucose control and experience fewer episodes of severe hypoglycaemia. However, many struggle to sustain these improvements over time. To inform the design of more effective follow-up support we undertook a review of qualitative studies which have identified factors that influence and inform participants' self-management behaviours after attending structured education and their need for support to sustain improvements in glycaemic control. METHODS: We undertook a meta-ethnography of relevant qualitative studies, identified using systematic search methods. Studies were included which focused on participants' experiences of self-managing type 1 diabetes after attending structured education which incorporated training in flexible intensive insulin therapy. A line of argument approach was used to synthesise the findings. RESULTS: The search identified 18 papers from six studies. The studies included were judged to be of high methodological quality. The line of argument synthesis developed the Follow-Up Support for Effective type 1 Diabetes self-management (FUSED) model. This model outlines the challenges participants encounter in maintaining diabetes self-management practices after attending structured education, and describes how participants try to address these barriers by adapting, simplifying or personalising the self-management approaches they have learned. To help participants maintain the skills taught during courses, the FUSED model presents ten recommendations abstracted from the included papers to provide a logic model for a programme of individualised and responsive follow-up support. CONCLUSIONS: This meta-ethnography highlights how providing skills training using structured education to people with type 1 diabetes does not necessarily result in participants adopting and sustaining recommended changes in behaviour. To help people sustain diabetes self-management skills after attending structured education, it is recommended that support be provided over the longer-term by appropriately trained healthcare professionals which is responsive to individuals' needs. Although developed to inform support for people with type 1 diabetes, the FUSED model provides a framework that could also be applied to support individuals with other long term conditions which require complex self-management skills to be learned and sustained over time. TRIAL REGISTRATION: PROSPERO registration: CRD42017067961 .

Feasibility study of portable technology for weight loss and HbA1c control in type 2 diabetes.

BACKGROUND: The study investigated the feasibility of conducting a future Randomised Controlled Trial (RCT) of a mobile health (mHealth) intervention for weight loss and HbA1c reduction in Type 2 Diabetes Mellitus (T2DM). METHODS: The intervention was a small wearable mHealth device used over 12 weeks by overweight people with T2DM with the intent to lose weight and reduce their HbA1c level. A 4 week maintenance period using the device followed. The device records physical activity level and information about food consumption, and provides motivational feedback based on energy balance. Twenty-seven participants were randomised to receive no intervention; intervention alone; or intervention plus weekly motivational support. All participants received advice on diet and exercise at the start of the study. Weight and HbA1c levels were recorded at baseline and weeks 6, 12, and 16. Qualitative interviews were conducted with participants who received the intervention to explore their experiences of using the device and involvement in the study including the training received. RESULTS: Overall the device was perceived to be well-liked, acceptable, motivational and easy to use by participants. Some logistical changes were required during the feasibility study, including shortening of the study duration and relaxation of participant inclusion criteria. Descriptive statistics of weight and HbA1c data showed promising trends of weight loss and HbA1c reduction in both intervention groups, although this should be interpreted with caution. CONCLUSIONS: A number of methodological recommendations for a future RCT emerged from the current feasibility study. The mHealth device was acceptable and promising for helping individuals with T2DM to reduce their HbA1c and lose weight. Devices with similar features should be tested further in larger studies which follow these methodological recommendations.

Type 1 diabetes patients' experiences of, and need for, social support after attending a structured education programme: a qualitative longitudinal investigation.

AIMS AND OBJECTIVES: To explore patients' experiences of, views about and need for, social support after attending a structured education programme for type 1 diabetes. BACKGROUND: Patients who attend structured education programmes attain short-term improvements in biomedical and quality-of-life measures but require support to sustain self-management principles over the longer term. Social support can influence patients' self-management practices; however, little is known about how programme graduates use other people's help. DESIGN: This study was informed by the principles of grounded theory and involved concurrent data collection and analysis. Data were analysed using an inductive, thematic approach. METHODS: In-depth interviews were undertaken postcourse, six and 12 months later, with 30 adult patients with type 1 diabetes recruited from Dose Adjustment for Normal Eating courses in the United Kingdom. RESULTS: Patients' preferences for social support from other people ranged from wanting minimal involvement, to benefiting from auxiliary forms of assistance, to regular monitoring and policing. New self-management skills learnt on their courses prompted and facilitated patients to seek and obtain more social support. Support received/expected from parents varied according to when patients were diagnosed, but parents' use of outdated knowledge could act as a barrier to effective support. Support sought from others, including friends/colleagues, was informed by patients' domestic/employment circumstances. CONCLUSION: This study responds to calls for deeper understanding of the social context in which chronic illness self-management occurs. It highlights how patients can solicit and receive more social support from family members and friends after implementing self-care practices taught on education programmes. RELEVANCE TO CLINICAL PRACTICE: Health professionals including diabetes specialist nurses and dietitians should explore: patients' access to and preferences for social support; how patients might be encouraged to capitalise on social support postcourse; and new ways to inform/educate people within patients' social networks.

Data-driven blood glucose level prediction in type 1 diabetes: a comprehensive comparative analysis.

Accurate prediction of blood glucose level (BGL) has proven to be an effective way to help in type 1 diabetes management. The choice of input, along with the fundamental choice of model structure, is an existing challenge in BGL prediction. Investigating the performance of different data-driven time series forecasting approaches with different inputs for BGL prediction is beneficial in advancing BGL prediction performance. Limited work has been made in this regard, which has resulted in different conclusions. This paper performs a comprehensive investigation of different data-driven time series forecasting approaches using different inputs. To do so, BGL prediction is comparatively investigated from two perspectives; the model's approach and the model's input. First, we compare the performance of BGL prediction using different data-driven time series forecasting approaches, including classical time series forecasting, traditional machine learning, and deep neural networks. Secondly, for each prediction approach, univariate input, using BGL data only, is compared to a multivariate input, using data on carbohydrate intake, injected bolus insulin, and physical activity in addition to BGL data. The investigation is performed on two publicly available Ohio datasets. Regression-based and clinical-based metrics along with statistical analyses are performed for evaluation and comparison purposes. The outcomes show that the traditional machine learning model is the fastest model to train and has the best BGL prediction performance especially when using multivariate input. Also, results show that simply adding extra variables does not necessarily improve BGL prediction performance significantly, and data fusion approaches may be required to effectively leverage other variables' information.

Physical Activity Integration in Blood Glucose Level Prediction: Different Levels of Data Fusion.

Blood glucose level (BGL) prediction contributes to more effective management of type 1 diabetes. Physical activity (PA) is a crucial factor in diabetes management. It affects BGL, and it is imperative to effectively deploy PA in BGL prediction to support diabetes management systems by incorporating this crucial factor. Due to the erratic nature of PA's impact on BGL inter- and intra-patients and insufficient knowledge, deploying PA in BGL prediction is challenging. Hence, optimal approaches for PA fusion with BGL are demanded to improve the performance of BGL prediction. To address this gap, we propose novel methodologies for extracting and integrating information from PA data into BGL prediction. This paper proposes several novel PA-informed prediction models by developing different approaches for extracting information from PA data and fusing this information with BGL data in signal, feature, and decision levels to find the optimal approach for deploying PA in BGL prediction models. For signal-level fusion, different automatically-recorded PA data are fused with BGL data. Also, three feature engineering approaches are developed for feature-level fusion: subjective assessments of PA, objective assessments of PA, and statistics of PA. Furthermore, in decision-level fusion, ensemble learning is used to combine predictions from models trained with different inputs. Then, a comparative investigation is performed between the developed PA-informed approaches and the no-fusion approach, as well as between themselves. The analyses are performed on the publicly available Ohio dataset with rigorous evaluation. The results show that deploying PA can statistically significantly improve BGL prediction performance. The results show that deploying PA can statistically significantly improve BGL prediction performance. Also, among the developed approaches to leveraging PA in BGL prediction, fusing heart rate data at the signal-level and PA intensity categories at the feature-level with BGL data are the most effective ways. Our developed methodologies contribute to determining optimal approaches, including the kind of PA information and fusion method, to improve the performance of BGL prediction effectively.

In Vitro Glucose Measurement from NIR and MIR Spectroscopy: Comprehensive Benchmark of Machine Learning and Filtering Chemometrics.

The quantitative analysis of glucose using spectroscopy is a topic of great significance and interest in science and industry. One conundrum in this area is deploying appropriate preprocessing and regression tools. To contribute to addressing this challenge, in this study, we conducted a comprehensive and novel comparative analysis of various machine learning and preprocessing filtering techniques applied to near-infrared, mid-infrared, and a combination of near-infrared and mid-infrared spectroscopy for glucose assay. Our objective was to evaluate the effectiveness of these techniques in accurately predicting glucose levels and to determine which approach was most optimal. Our investigation involved the acquisition of spectral data from samples of glucose solutions using the three aforementioned spectroscopy techniques. The data was subjected to several preprocessing filtering methods, including convolutional moving average, Savitzky-Golay, multiplicative scatter correction, and normalisation. We then applied representative machine learning algorithms from three categories: linear modelling, traditional nonlinear modelling, and artificial neural networks. The evaluation results revealed that linear models exhibited higher predictive accuracy than nonlinear models, whereas artificial neural network models demonstrated comparable performance. Additionally, the comparative analysis of various filtering methods demonstrated that the convolutional moving average and Savitzky-Golay filters yielded the most precise outcomes overall. In conclusion, our study provides valuable insights into the efficacy of different machine learning techniques for glucose measurement and highlights the importance of applying appropriate filtering methods in enhancing predictive accuracy. These findings have important implications for the development of new and improved glucose quantification technologies.

Is consulting patients about their health service preferences a useful exercise?

As part of the shift toward patient-centered care, patients are increasingly being consulted about their preferences for health services and interventions, including those explored during randomized controlled trials (RCTs), to ensure that service recommendations are aligned to their own circumstances and needs. Hence, we interviewed patients (N = 40) who participated in a randomized control trial comparing diabetes education courses delivered using two different formats to establish whether, and why, they preferred one format to the other, to inform recommendations for future course delivery. Not only did patients report changing their preferences, and the reasons underlying these preferences, over time, but all patients also claimed to prefer the particular course they had attended. We use our findings and experiences to problematize the notion of a patient preference and to raise questions about what we can really learn from consulting patients about the care they receive within the context of an RCT.

Causality analysis in type 1 diabetes mellitus with application to blood glucose level prediction.

Effective control of blood glucose level (BGL) is the key factor in the management of type 1 diabetes mellitus (T1D). BGL prediction is an important tool to help maximise the time BGL is in the target range and thus minimise both acute and chronic diabetes-related complications. To predict future BGL, histories of variables known to affect BGL, such as carbohydrate intake, injected bolus insulin, and physical activity, are utilised. Due to these identified cause and effect relationships, T1D management can be examined via the causality context. In this respect, this work initially investigates these relations and quantifies the causality strengths of each variable with BGL using the convergent cross mapping method (CCM). Then, considering the extended CCM, the causality strengths of each variable for different lags are quantified. After that, the optimal time lag for each variable is determined according to the quantified causality effects. Subsequently, the feasibility of leveraging causality information as prior knowledge for BGL prediction is investigated by proposing two approaches. In the first approach, causality strengths are used as weights for relevant affecting variables. In the second approach, the optimal causal lags and the corresponding causality strengths are considered the shifts and weights for the variables, respectively. Overall, the evaluation criteria and statistical analysis used for comparing results show the effectiveness of using causality analysis in T1D management.

Blood Glucose Level Time Series Forecasting: Nested Deep Ensemble Learning Lag Fusion.

Blood glucose level prediction is a critical aspect of diabetes management. It enables individuals to make informed decisions about their insulin dosing, diet, and physical activity. This, in turn, improves their quality of life and reduces the risk of chronic and acute complications. One conundrum in developing time-series forecasting models for blood glucose level prediction is to determine an appropriate length for look-back windows. On the one hand, studying short histories foists the risk of information incompletion. On the other hand, analysing long histories might induce information redundancy due to the data shift phenomenon. Additionally, optimal lag lengths are inconsistent across individuals because of the domain shift occurrence. Therefore, in bespoke analysis, either optimal lag values should be found for each individual separately or a globally suboptimal lag value should be used for all. The former approach degenerates the analysis's congruency and imposes extra perplexity. With the latter, the fine-tunned lag is not necessarily the optimum option for all individuals. To cope with this challenge, this work suggests an interconnected lag fusion framework based on nested meta-learning analysis that improves the accuracy and precision of predictions for personalised blood glucose level forecasting. The proposed framework is leveraged to generate blood glucose prediction models for patients with type 1 diabetes by scrutinising two well-established publicly available Ohio type 1 diabetes datasets. The models developed undergo vigorous evaluation and statistical analysis from mathematical and clinical perspectives. The results achieved underpin the efficacy of the proposed method in blood glucose level time-series prediction analysis.

National Trends in Hyperglycemia and Diabetic Ketoacidosis in Children, Adolescents, and Young Adults With Type 1 Diabetes: A Challenge Due to Age or Stage of Development, or Is New Thinking About Service Provision Needed?

OBJECTIVE: Adolescence is associated with high-risk hyperglycemia. This study examines the phenomenon in a life course context. RESEARCH DESIGN AND METHODS: A total of 93,125 people with type 1 diabetes aged 5 to 30 years were identified from the National Diabetes Audit and/or the National Paediatric Diabetes Audit for England and Wales for 2017/2018-2019/2020. For each audit year, the latest HbA1c and hospital admissions for diabetic ketoacidosis (DKA) were identified. Data were analyzed in sequential cohorts by year of age. RESULTS: In childhood, unreported HbA1c measurement is uncommon; however, for 19-year-olds, it increases to 22.3% for men and 17.3% for women, and then reduces to 17.9% and 13.1%, respectively, for 30-year-olds. Median HbA1c for 9-year-olds is 7.6% (60 mmol/mol) (interquartile range 7.1-8.4%, 54-68 mmol/mol) in boys and 7.7% (61 mmol/mol) (8.0-8.4%, 64-68 mmol/mol) in girls, increasing to 8.7% (72 mmol/mol) (7.5-10.3%, 59-89 mmol/mol) and 8.9% (74 mmol/mol) (7.7-10.6%, 61-92 mmol/mol), respectively, for 19-year-olds before falling to 8.4% (68 mmol/mol) (7.4-9.7%, 57-83 mmol/mol) and 8.2% (66 mmol/mol) (7.3-9.7%, 56-82 mmol/mol), respectively, for 30-year-olds. Annual hospitalization for DKA rose steadily in age from 6 years (2.0% for boys, 1.4% for girls) and peaked at 19 years for men (7.9%) and 18 years for women (12.7%), reducing to 4.3% for men and 5.4% for women at age 30 years. For all ages over 9 years, the prevalence of DKA was higher in female individuals. CONCLUSIONS: HbA1c and the prevalence of DKA increase through adolescence and then decline. Measurement of HbA1c, a marker of clinical review, falls abruptly in the late teenage years. Age-appropriate services are needed to overcome these issues.

Automatic inference of hypoglycemia causes in type 1 diabetes: a feasibility study.

Background: Hypoglycemia is the most common adverse consequence of treating diabetes, and is often due to suboptimal patient self-care. Behavioral interventions by health professionals and self-care education helps avoid recurrent hypoglycemic episodes by targeting problematic patient behaviors. This relies on time-consuming investigation of reasons behind the observed episodes, which involves manual interpretation of personal diabetes diaries and communication with patients. Therefore, there is a clear motivation to automate this process using a supervised machine learning paradigm. This manuscript presents a feasibility study of automatic identification of hypoglycemia causes. Methods: Reasons for 1885 hypoglycemia events were labeled by 54 participants with type 1 diabetes over a 21 months period. A broad range of possible predictors were extracted describing a hypoglycemic episode and the subject's general self-care from participants' routinely collected data on the Glucollector, their diabetes management platform. Thereafter, the possible hypoglycemia reasons were categorized for two major analysis sections - statistical analysis of relationships between the data features of self-care and hypoglycemia reasons, and classification analysis investigating the design of an automated system to determine the reason for hypoglycemia. Results: Physical activity contributed to 45% of hypoglycemia reasons on the real world collected data. The statistical analysis provided a number of interpretable predictors of different hypoglycemia reasons based on self-care behaviors. The classification analysis showed the performance of a reasoning system in practical settings with different objectives under F1-score, recall and precision metrics. Conclusion: The data acquisition characterized the incidence distribution of the various hypoglycemia reasons. The analyses highlighted many interpretable predictors of the various hypoglycemia types. Also, the feasibility study presented a number of concerns valuable in the design of the decision support system for automatic hypoglycemia reason classification. Therefore, automating the identification of the causes of hypoglycemia may help objectively to target behavioral and therapeutic changes in patients' care.