Practical considerations for continuous glucose monitoring in elite athletes with type 1 diabetes mellitus: A narrative review

Type 1 diabetes mellitus (T1DM) refers to a metabolic condition where a lack of insulin impairs the usual homeostatic mechanisms to control blood glucose levels. Historically, participation in competitive sport has posed a challenge for those with T1DM, where the dynamic changes in blood glucose during exercise can result in dangerously high (hyperglycaemia) or low blood glucoses (hypoglycaemia) levels. Over the last decade, research and technological development has enhanced the methods of monitoring and managing blood glucose levels, thus reducing the chances of experiencing hyper- or hypoglycaemia during exercise. The introduction of continuous glucose monitoring (CGM) systems means that glucose can be monitored conveniently, without the need for frequent fingerpick glucose checks. CGM devices include a fine sensor inserted under the skin, measuring levels of glucose in the interstitial fluid. Readings can be synchronized to a reader or mobile phone app as often as every 1-5 min. Use of CGM devices is associated with lower HbA1c and a reduction in hypoglycaemic events, promoting overall health and athletic performance. However, there are limitations to CGM, which must be considered when being used by an athlete with T1DM. These limitations can be addressed by individualized education plans, using protective equipment to prevent sensor dislodgement, as well as further research aiming to: (i) account for disparities between CGM and true blood glucose levels during vigorous exercise; (ii) investigate the effects of temperature and altitude on CGM accuracy, and (iii) explore of the sociological impact of CGM use amongst sportspeople without diabetes on those with T1DM.

Developing and exploring the validity of a patient reported experience measure for adult inpatient diabetes care

AIM

To develop and explore the validity of a Patient Reported Experience Measure (PREM) for adult inpatient diabetes care.

METHOD

27 in-depth interviews were conducted to inform the development of the 42-item PREM which was cognitively tested with 10 people. A refined 38-item PREM was piloted with 228 respondents completing a paper (n = 198) or online (n = 30) version. The performance of the PREM was evaluated by exploring (i) uptake/number of responses and (ii) survey validity by investigating whether the PREM data were of adequate quality and delivered useful information.

RESULTS

The PREM had low drop-out or missing data rates suggesting it was appropriately constructed. Analysis of item frequencies and variances, and problem score calculations concluded that questions provided sufficient score differentiation.

CONCLUSIONS

This new PREM allows for experiences of inpatient diabetes care to be measured, understood and reported on to help identify priority areas for improving care quality.

Real-world outcomes of Omnipod DASH system use in people with type 1 diabetes: Evidence from the Association of British Clinical Diabetologists (ABCD) study

AIMS

To evaluate real-world outcomes in people with Type 1 Diabetes (PwT1D) initiated on Omnipod DASH® Insulin Management System.

METHODS

Anonymized clinical data were submitted to a secure web-based tool within the National Health Service network. Hemoglobin A1c (HbA1c), sensor-derived glucometrics, total daily dose of insulin (TDD), and patient-reported outcome changes between baseline and follow-up were assessed. Individuals were classified to "new-to-pump" (switched from multiple daily injections) and "established-on-pump" (switched from a tethered insulin pump) groups.

RESULTS

276 individuals from 11 centers [66.7 % female; 92 % White British; median age 41 years (IQR 20-50); diabetes duration 20 years (IQR 11-31); 49.3 % within "new-to-pump" group] were included. Baseline HbA1c was 8.0 ± 1.3 % (64 ± 14 mmol/mol). At follow-up [3 years (IQR 1.5-3.2)], HbA1c reduced by 0.3 % [(3 mmol/mol); p = 0.002] across the total population, 0.4 % [(5 mmol/mol); p = 0.001] in those "new-to-pump" and remained unchanged in those "established-on-pump". TDD decreased in the "new-to-pump" cohort (baseline:44.9 ± 21.0units vs follow-up:38.1 ± 15.4units, p = 0.002). Of those asked, 141/143 (98.6 %) stated Omnipod DASH had a positive impact on quality of life.

CONCLUSIONS

Omnipod DASH was associated with improvements in HbA1c in PwT1D "new-to-pump" and maintained previous HbA1c levels in those "established-on-pump". User satisfaction in all groups and TDD reduction in those "new-to-pump" were reported.

Hybrid Closed-Loop Therapy in Adults With Type 1 Diabetes and Above-Target HbA1c: A Real-world Observational Study

OBJECTIVE

We explored longitudinal changes associated with switching to hybrid closed-loop (HCL) insulin delivery systems in adults with type 1 diabetes and elevated HbA1c levels despite the use of intermittently scanned continuous glucose monitoring (isCGM) and insulin pump therapy.

RESEARCH DESIGN AND METHODS

We undertook a pragmatic, preplanned observational study of participants included in the National Health Service England closed-loop pilot. Adults using isCGM and insulin pump across 31 diabetes centers in England with an HbA1c ≥8.5% who were willing to commence HCL therapy were included. Outcomes included change in HbA1c, sensor glucometrics, diabetes distress score, Gold score (hypoglycemia awareness), acute event rates, and user opinion of HCL.

RESULTS

In total, 570 HCL users were included (median age 40 [IQR 29-50] years, 67% female, and 85% White). Mean baseline HbA1c was 9.4 ± 0.9% (78.9 ± 9.1 mmol/mol) with a median follow-up of 5.1 (IQR 3.9-6.6) months. Of 520 users continuing HCL at follow-up, mean adjusted HbA1c reduced by 1.7% (95% CI 1.5, 1.8; P < 0.0001) (18.1 mmol/mol [95% CI 16.6, 19.6]; P < 0.0001). Time in range (70-180 mg/dL) increased from 34.2 to 61.9% (P < 0.001). Individuals with HbA1c of ≤58 mmol/mol rose from 0 to 39.4% (P < 0.0001), and those achieving ≥70% glucose time in range and <4% time below range increased from 0.8 to 28.2% (P < 0.0001). Almost all participants rated HCL therapy as having a positive impact on quality of life (94.7% [540 of 570]).

CONCLUSIONS

Use of HCL is associated with improvements in HbA1c, time in range, hypoglycemia, and diabetes-related distress and quality of life in people with type 1 diabetes in the real world.

UK's Association of British Clinical Diabetologist's Diabetes Technology Network (ABCD-DTN): Best practice guide for hybrid closed-loop therapy

This best practice guide is written with the aim of providing an overview of current hybrid closed-loop (HCL) systems in use within the United Kingdom's (UK) National Health Service (NHS) and to provide education and advice for their management on both an individual and clinical service level. The environment of diabetes technology, and particularly HCL systems, is rapidly evolving. The past decade has seen unprecedented advances in the development of HCL systems. These systems improve glycaemic outcomes and reduce the burden of treatment for people with type 1 diabetes (pwT1D). It is anticipated that access to these systems will increase in England as a result of updates in National Institute of Health and Care Excellence (NICE) guidance providing broader support for the use of real-time continuous glucose monitoring (CGM) for pwT1D. NICE is currently undertaking multiple-technology appraisal into HCL systems. Based on experience from centres involved in supporting advanced technologies as well as from the recent NHS England HCL pilot, this guide is intended to provide healthcare professionals with UK expert consensus on the best practice for initiation, optimisation and ongoing management of HCL therapy.

Variation in the Current Use of Technology to Support Diabetes Management in UK Hospitals: Results of a Survey of Health Care Professionals

BACKGROUND

There has been a significant increase in the use of wearable diabetes technologies in the outpatient setting over recent years, but this has not consistently translated into inpatient use.

METHODS

An online survey was undertaken to understand the current use of technology to support inpatient diabetes care in the United Kingdom.

RESULTS

Responses were received from 42 different organizations representing 104 hospitals across the United Kingdom. Significant variation was found between organizations in the use of technology to support safe, effective inpatient diabetes care. Benefits of the use of technology were reported, and areas of good practice identified.

CONCLUSION

Technology supports good inpatient diabetes care, but there is currently variation in its use. Guidance has been developed which should drive improvements in the use of technology and hence improvements in the safety and effectiveness of inpatient diabetes care. Key recommendations include implementation of this guidance (especially for continuous glucose monitoring), ensuring specialist support is available for the use of wearable diabetes technology in hospital, optimizing information sharing across the health care system, and making full use of data from networked glucose and ketone meters.

Using Technology to Improve Diabetes Care in Hospital: The Challenge and the Opportunity

The past 10 years have seen a revolution in technology improving the lives of people with diabetes. This has implications for diabetes care in hospitalized inpatients. These technological developments have the potential to significantly improve the care of people with diabetes in hospital. Combining point of care glucose monitoring, electronic prescribing, electronic observations with electronic referral, and electronic health records allow teams to daily oversee the whole hospital population. To make the most of these tools as well as developing the use of pumps and glucose sensors in hospital, the diabetes team needs to work in new ways. To date, very little work has described how these should be combined. We describe how this technology can be combined to improve diabetes care in hospital.

How Can Point-of-Care Technologies Support In-Hospital Diabetes Care?

People with diabetes admitted to hospital are at risk of diabetes related complications including hypoglycaemia and diabetic ketoacidosis. Point-of-care (POC) tests undertaken at the patient bedside, for glucose, ketones, and other analytes, are a key component of monitoring people with diabetes, to ensure safety. POC tests implemented with a quality framework are critical to ensuring accuracy and veracity of results and preventing erroneous clinical decision making. POC results can be used for self-management of glucose levels in those well-enough and/or by healthcare professionals to identify unsafe levels. Connectivity of POC results to electronic health records further offers the possibility of utilising these results proactively to identify patients 'at risk' in real-time and for audit purposes. In this article, the key considerations when implementing POC tests for diabetes in-patient management are reviewed and potential to drive improvements using networked glucose and ketone measurements are discussed. In summary, new advances in POC technology should allow people with diabetes and the teams looking after them whilst in hospital to integrate to provide safe and effective care.

Insulin Pumps and Hybrid Close Loop Systems Within Hospital: A Scoping Review and Practical Guidance From the Joint British Diabetes Societies for Inpatient Care

This article is the second of a two-part series providing a scoping review and summary of the Joint British Diabetes Societies for Inpatient Care (JBDS-IP) guidelines on the use of diabetes technology in people with diabetes admitted to hospital. The first part reviewed the use of continuous glucose monitoring (CGM) in hospital. In this article, we focus on the use of continuous subcutaneous insulin infusion (CSII; insulin pumps) and hybrid closed-loop systems in hospital. JBDS-IP advocates enabling people who can self-manage and are willing and capable of using CSII to continue doing so as they would do out of hospital. CSII should be discontinued if the individual is critically ill or hemodynamically unstable. For individuals on hybrid closed-loop systems, the system should be discontinued from auto-mode, and may be used individually (as CGM only or CSII only, if criteria are met). Continuing in closed-loop mode may only be done so under specialist guidance from the Diabetes Team, where the diabetes teams are comfortable and knowledgeable about the specific devices used. Health care organizations need to have clear local policies and guidance to support individuals using these wearable technologies, and ensure the relevant workforce is capable and skilled enough to ensure their safe use within the hospital setting.

Continuous Glucose Monitoring Within Hospital: A Scoping Review and Summary of Guidelines From the Joint British Diabetes Societies for Inpatient Care

Increasing numbers of people, particularly with type 1 diabetes (T1D), are using wearable technologies. That is, continuous subcutaneous insulin infusion (CSII) pumps, continuous glucose monitoring (CGM) systems, and hybrid closed-loop systems, which combine both these elements. Given over a quarter of all people admitted to hospital have diabetes, there is a need for clinical guidelines for when people using them are admitted to hospital. The Joint British Diabetes Societies for Inpatient Care (JBDS-IP) provide a scoping review and summary of guidelines on the use of diabetes technology in people with diabetes admitted to hospital.JBDS-IP advocates enabling people who can self-manage and use their own diabetes technology to continue doing so as they would do out of hospital. Whilst people with diabetes are recommended to achieve a target of 70% time within range (3.9-10.0 mmol/L [70-180 mg/dL]), this can be very difficult to achieve whilst unwell. We therefore recommend targeting hypoglycemia prevention as a priority, keeping time below 3.9 mmol/L (70 mg/dL) at < 1%, being aware of looming hypoglycemia if glucose is between 4.0 and 5.9 mmol/L (72-106 mg/dL), and consider intervening, particularly if there is a downward CGM trend arrow.Health care organizations need clear local policies and guidance to support individuals using diabetes technologies, and ensure the relevant workforce is capable and skilled enough to ensure their safe use within the hospital setting. The current set of guidelines is divided into two parts. Part 1, which follows below, outlines the guidance for use of CGM in hospital. The second part outlines guidance for use of CSII and hybrid closed-loop in hospital.

Time in range following flash glucose monitoring: Relationship with glycaemic control, diabetes-related distress and resource utilisation in the Association of British Clinical Diabetologists national audit

AIMS

The aim of this study was to understand the relationship between time in range (TIR) achieved using the isCGM with changes in glycaemic control, diabetes-related distress (DRD) and resource utilisation in people living with diabetes.

METHODS

Clinicians from 106 National Health System (NHS) UK hospitals submitted isCGM user baseline and follow-up data in a web-based tool held within the UK NHS network. Linear regression analysis was used to identify the relationship between follow-up glucose TIR (3.9-10 mmol/L) categories (TIR% 50-70 and TIR% >70) with change in haemoglobin A1c (HbA1c), DRD and Gold score (measure of hypoglycaemia unawareness, where a score ≥4 suggests impaired awareness of hypoglycaemia).

RESULTS

Of 16,427 participants, 1241 had TIR follow-up data available. In this cohort, the mean TIR was 44.8% (±22.5). With the use of isCGM, at 7.9 months mean follow-up, improvements were observed in HbA1c (-6.9 [13.5] mmol/mol, p < 0.001), Gold score (-0.35 [1.5], p < 0.001) and Diabetes Distress Screening (-0.73 [1.23], p < 0.001). In the regression analysis restricted to people living with type 1 diabetes, TIR% 50-70 was associated with a -8.9 mmol/mol (±0.6, p < 0.001) reduction in HbA1c; TIR% >70 with a -14 mmol/mol (±0.8, p < 0.001) reduction in HbA1c. Incremental improvement in TIR% was also associated with significant improvements in Gold score and DRD. TIR% >70 was associated with no hospital admissions due to hypoglycaemia, hyperglycaemia/diabetic ketoacidosis, and a 60% reduction in the paramedic callouts and 77% reduction in the incidence of severe hypoglycaemia.

CONCLUSION

In a large cohort of UK isCGM users, we demonstrate a significant association of higher TIR% with improvement in HbA1c, hypoglycaemia awareness, DRD and resource utilisation.

Current provision and HCP experiences of remote care delivery and diabetes technology training for people with type 1 diabetes in the UK during the COVID-19 pandemic

BACKGROUND

The COVID-19 pandemic has led to the rapid implementation of remote care delivery in type 1 diabetes. We studied current modes of care delivery, healthcare professional experiences and impact on insulin pump training in type 1 diabetes care in the United Kingdom (UK).

METHODS

The UK Diabetes Technology Network designed a 48-question survey aimed at healthcare professionals providing care in type 1 diabetes.

RESULTS

One hundred and forty-three healthcare professionals (48% diabetes physicians, 52% diabetes educators and 88% working in adult services) from approximately 75 UK centres (52% university hospitals, 46% general and community hospitals), responded to the survey. Telephone consultations were the main modality of care delivery. There was a higher reported time taken for video consultations versus telephone (p < 0.001). Common barriers to remote consultations were patient familiarity with technology (72%) and access to patient device data (67%). We assessed the impact on insulin pump training. A reduction in total new pump starts (73%) and renewals (61%) was highlighted. Common barriers included patient digital literacy (61%), limited healthcare professional experience (46%) and time required per patient (44%). When grouped according to size of insulin pump service, pump starts and renewals in larger services were less impacted by the pandemic compared to smaller services.

CONCLUSION

This survey highlights UK healthcare professional experiences of remote care delivery. While supportive of virtual care models, a number of factors highlighted, especially patient digital literacy, need to be addressed to improve virtual care delivery and device training.

Delivering evidence-based interventions for type 1 diabetes in the virtual world - A review of UK practice during the SARS-CoV-2 pandemic

AIMS

This review considers the impact of the SARS-CoV-2 pandemic on access to interventions for those living with type 1 diabetes and discusses the solutions which have been considered and actioned to ensure ongoing access care.

METHODS

We performed a focussed review of the published literature, and the guidelines for changes that have been effected during the pandemic. We also drew from expert recommendations and information about local practice changes for areas where formal data have not been published.

RESULTS

Evidence based interventions which support the achievement of improved glucose levels and/or reduction in hypoglycaemia include group structured education to support self-management, insulin pump therapy and continuous glucose monitoring. The SARS-CoV-2 pandemic had impacted the ability of diabetes services to deliver these intervention. Multiple adaptations have been put in place - transition to online delivery of education and care, and usage of diabetes technology.

CONCLUSIONS

Although various adaptations have been made during the pandemic that have positively influenced uptake of services, there are many areas of delivery that need immediate improvement in the UK. We recommend a proactive approach in recognising the digital divide and inequity in distribution of these changes and we recommend introducing measures to reduce them.

Association of British Clinical Diabetologists, Diabetes Technology Network UK and Association of Children's Diabetes Clinicians Survey of UK Healthcare Professional Attitudes Towards Open-Source Automated Insulin Delivery Systems

INTRODUCTION

Automated insulin delivery (AID) systems can enable improved glycaemic outcomes with reduced mental burden. Open-source AID (OS-AID) systems overcome some of the developmental and access barriers enabling a wider use of these systems. Limited data are available on healthcare professional (HCP) opinions and current practice regarding these systems. The aim of this survey was to gain insight into HCP perceptions and practices around OS-AID.

METHODS

This survey was developed collaboratively with OS-AID users and distributed to adult and children's teams, using an online survey tool. Results were received between February and April 2019. Responses were assessed using simple descriptive statistics with analyses stratified by respondent characteristics.

RESULTS

317 responses were obtained from a range of HCPs in both adult and paediatric services. Key results include: HCP perception of OS-AID as "risky in the wrong hands" (43%); 91% felt uncomfortable initiating discussions around OS-AID because of lack of regulation (67%) and/or their own lack of knowledge (63%). Half of HCPs (47%) reported that they would choose OS-AID if they themselves had type 1 diabetes.

CONCLUSIONS

HCPs are generally supportive of OS-AID users but many feel uncomfortable with the technicalities of the systems given the lack of approval. Knowledge around the use of these systems was limited. Re-assessment of HCP perceptions should be performed in the future given the evolving landscape of diabetes technology, recent consensus statements and emerging ethical and legal perspectives.

The impact of islet mass, number of transplants, and time between transplants on graft function in a national islet transplant program

The UK islet allotransplant program is nationally funded to deliver one or two transplants over 12 months to individuals with type 1 diabetes and recurrent severe hypoglycemia. Analyses were undertaken 10 years after program inception to evaluate associations between transplanted mass; single versus two transplants; time between two transplants and graft survival (stimulated C-peptide >50 pmol/L) and function. In total, 84 islet transplant recipients were studied. Uninterrupted graft survival over 12 months was attained in 23 (68%) single and 47 (94%) (p = .002) two transplant recipients (separated by [median (IQR)] 6 (3-8) months). 64% recipients of one or two transplants with uninterrupted function at 12 months sustained graft function at 6 years. Total transplanted mass was associated with Mixed Meal Tolerance Test stimulated C-peptide at 12 months (p < .01). Despite 1.9-fold greater transplanted mass in recipients of two versus one islet infusion (12 218 [9291-15 417] vs. 6442 [5156-7639] IEQ/kg; p < .0001), stimulated C-peptide was not significantly higher. Shorter time between transplants was associated with greater insulin dose reduction at 12 months (beta -0.35; p = .02). Graft survival over the first 12 months was greater in recipients of two versus one islet transplant in the UK program, although function at 1 and 6 years was comparable. Minimizing the interval between 2 islet infusions may maximize cumulative impact on graft function.

Hybrid closed-loop therapy: Where are we in 2021?

Hybrid closed-loop systems are characterized by the coexistence of algorithm-driven automated insulin delivery combined with manual mealtime boluses. Used correctly, these insulin delivery systems offer better glucose control and reduced risk of hypoglycaemia and represent the most advanced form of insulin delivery available for people with type 1 diabetes. The aim of this paper was to compare the currently available commercial hybrid closed-loop systems in the UK: the Medtronic 670G/780G, Tandem t:slim X2 Control IQ and CamAPS FX systems. The Medtronic 670G/780G systems use Guardian 3 sensor (7-day use, two to four calibrations per day), while Tandem and CamAPS systems use the calibration-free Dexcom G6 sensor (10 days). The CamAPS system is available as an android app, whereas the other two systems have the algorithm embedded in the insulin pump. During pivotal studies, depending on the study population and baseline glycated haemoglobin level, these systems achieve a time spent in the target range 3.9 to 10 mmol/L (70 to 180 mg/dL) of 65% to 76% with low burden of hypoglycaemia. All three systems allow a higher glucose target for announced exercise, while the Tandem system offers an additional night-time tighter target. The CamAPS system offers fully customizable glucose targets and is the only system licensed for use during pregnancy. Additional education is required for both users and healthcare professionals to harness the best performance from these systems as well as to troubleshoot when "automode exits" occur. We provide consensus recommendations to develop pragmatic pathways to guide patients, clinicians and commissioners in making informed decisions on the appropriate use of the diabetes technology.

Inpatient hypoglycaemia in older people is associated with a doubling in the increased length of stay compared with the younger population

BACKGROUND

Hypoglycaemia during hospital admission is associated with poor outcomes including increased length of stay. In this study, we compared the incidence of inpatient hypoglycaemia and length of stays among people of three age groups: ≤65 years, 65-80 years and >80 years old.

METHODS

The study was conducted using a 4-year electronic patient record dataset from Oxford University Hospitals NHS Foundation Trust. The dataset contains hospital admission data for people with diabetes. We analysed the blood glucose (BG) measurements and identified all level 1 (BG <4 mmol/l) and level 2 (BG <3 mmol/l) hypoglycaemic episodes. We compared the length of stays between different age groups and with different levels of hypoglycaemia.

RESULTS

We analysed data obtained from 17,658 inpatients with diabetes who underwent 32,758 hospital admissions. The length of stays for admissions with no hypoglycaemia were 3[1,6], 3[1,8] and 4[2,11] (median[interquartile range]) days for age groups ≤65 years, 65-80 years and >80 years, respectively. These were statistically significantly lower (P < 0.01 for all pairwise comparisons) than the length of stays for admissions with level 1 hypoglycaemia, which were 6[3,13], 10[5,20] and 12[6,22] days, and level 2 hypoglycaemia, which were 7[3,14], 11[5,24] and 13[6,24] days.

CONCLUSIONS

In all age groups, admissions with either level 1 or level 2 hypoglycaemia were associated with an increased length of stay. However, in both the older groups, the length of stay increments were much higher (double) than the younger counterparts. The clinical consequences of hypoglycaemia were more severe in older people compared with the younger population.

Time in range: A best practice guide for UK diabetes healthcare professionals in the context of the COVID-19 global pandemic

The emergence of continuous glucose monitoring has driven improvements in glycaemic control and quality of life for people with diabetes. Recent changes in access to continuous glucose monitoring systems within UK health services have increased the number of people able to benefit from these technologies. The COVID-19 pandemic has created an opportunity for diabetes healthcare professionals to use continuous glucose monitoring technology to remotely deliver diabetes services to support people with diabetes. This opportunity can be maximized with improved application and interpretation of continuous glucose monitoring-generated data. Amongst the diverse measures of glycaemic control, time in range is considered to be of high value in routine clinical care because it is actionable and is visibly responsive to changes in diabetes management. Importantly, it is also been linked to the risk of developing complications associated with diabetes and can be understood by people with diabetes and healthcare professionals alike. The 2019 International Consensus on Time in Range has established a series of target glucose ranges and recommendations for time spent within these ranges that is consistent with optimal glycaemic control. The recommendations cover people with type 1 or type 2 diabetes, with separate targets indicated for elderly people or those at higher risk from hypoglycaemia, as well as for women with type 1 diabetes during pregnancy. The aim of this best practice guide was to clarify the intent and purpose of these international consensus recommendations and to provide practical insights into their implementation in UK diabetes care.

Guidelines for the management of diabetes services and patients during the COVID-19 pandemic

The UK National Diabetes Inpatient COVID Response Group was formed at the end of March 2020 to support the provision of diabetes inpatient care during the COVID pandemic. It was formed in response to two emerging needs. First to ensure that basic diabetes services are secured and maintained at a time when there was a call for re-deployment to support the need for general medical expertise across secondary care services. The second was to provide simple safe diabetes guidelines for use by specialists and non-specialists treating inpatients with or suspected of COVID-19 infection. To date the group, comprising UK-based specialists in diabetes, pharmacy and psychology, have produced two sets of guidelines which will be continually revised as new evidence emerges. It is supported by Diabetes UK, the Association of British Clinical Diabetologists and NHS England.

Inpatient hypoglycaemia: understanding who is at risk

AIMS/HYPOTHESIS

We analysed data obtained from the electronic patient records of inpatients with diabetes admitted to a large university hospital to understand the prevalence and distribution of inpatient hypoglycaemia.

METHODS

The study was conducted using electronic patient record data from Oxford University Hospitals NHS Foundation Trust. The dataset contains hospital admission data for patients coded for diabetes. We used the recently agreed definition for a level 1 hypoglycaemia episode as any blood glucose measurement <4 mmol/l and a level 2 hypoglycaemia episode as any blood glucose measurement <3 mmol/l. Any two or more consecutive low blood glucose measurements within a 2 h time window were considered as one single hypoglycaemic episode.

RESULTS

We analysed data obtained from 17,658 inpatients with diabetes (1696 with type 1 diabetes, 14,006 with type 2 diabetes, and 1956 with other forms of diabetes; 9277 men; mean ± SD age, 66 ± 18 years) who underwent 32,758 hospital admissions between July 2014 and August 2018. The incidence of level 1 hypoglycaemia was 21.5% and the incidence of level 2 hypoglycaemia was 9.6%. Recurrent level 1 and level 2 hypoglycaemia occurred, respectively, in 51% and 39% of hospital admissions in people with type 2 diabetes with at least one hypoglycaemic episode, and in 55% and 45% in those with type 1 diabetes. The incidence of level 2 hypoglycaemia in people with type 2 diabetes, when corrected for the number of people who remained in hospital, remained constant for the first 100 h at approximately 0.15 events per h per admission. With regards to the hypoglycaemia distribution during the day, after correcting for the number of blood glucose tests per h, there were two clear spikes in the rate of hypoglycaemia approximately 3 h after lunch and after dinner. The highest rate of hypoglycaemia per glucose test was seen between 01:00 hours and 05:00 hours. Medication had a significant impact on the incidence of level 2 hypoglycaemia, ranging from 1.5% in people with type 2 diabetes on metformin alone to 33% in people treated with a combination of rapid-acting insulin analogue, long-acting insulin analogue and i.v.-administered insulin.

CONCLUSIONS/INTERPRETATION

Retrospective analysis of data from electronic patient records enables clinicians to gain a greater understanding of the incidence and distribution of inpatient hypoglycaemia. This information should be used to drive evidence-based improvements in the glycaemic control of inpatients through targeted medication adjustment for specific populations at high risk of hypoglycaemia.

Predicting the Risk of Inpatient Hypoglycemia With Machine Learning Using Electronic Health Records

OBJECTIVE

We analyzed data from inpatients with diabetes admitted to a large university hospital to predict the risk of hypoglycemia through the use of machine learning algorithms.

RESEARCH DESIGN AND METHODS

Four years of data were extracted from a hospital electronic health record system. This included laboratory and point-of-care blood glucose (BG) values to identify biochemical and clinically significant hypoglycemic episodes (BG ≤3.9 and ≤2.9 mmol/L, respectively). We used patient demographics, administered medications, vital signs, laboratory results, and procedures performed during the hospital stays to inform the model. Two iterations of the data set included the doses of insulin administered and the past history of inpatient hypoglycemia. Eighteen different prediction models were compared using the area under the receiver operating characteristic curve (AUROC) through a 10-fold cross validation.

RESULTS

We analyzed data obtained from 17,658 inpatients with diabetes who underwent 32,758 admissions between July 2014 and August 2018. The predictive factors from the logistic regression model included people undergoing procedures, weight, type of diabetes, oxygen saturation level, use of medications (insulin, sulfonylurea, and metformin), and albumin levels. The machine learning model with the best performance was the XGBoost model (AUROC 0.96). This outperformed the logistic regression model, which had an AUROC of 0.75 for the estimation of the risk of clinically significant hypoglycemia.

CONCLUSIONS

Advanced machine learning models are superior to logistic regression models in predicting the risk of hypoglycemia in inpatients with diabetes. Trials of such models should be conducted in real time to evaluate their utility to reduce inpatient hypoglycemia.

Importance of inpatient hypoglycaemia: impact, prediction and prevention

Hypoglycaemia is a key barrier to achieving euglycaemic control in people who are hospitalized. Inpatient hypoglycaemia has been linked to adverse clinical outcomes, including mortality and longer stay in hospital. A number of studies have applied mathematical tools and statistical models to predict inpatient hypoglycaemia and identify factors that may result in hypoglycaemic events. Several different approaches have been tested to prevent inpatient hypoglycaemia. These can be categorized as human intervention, computerized methods or application of medical devices. In this review we provide an overview of the epidemiology of inpatient hypoglycaemia and its impact on patients and hospitals. We also discuss the existing methodology used to predict inpatient hypoglycaemia and the limited number of trials performed to prevent inpatient hypoglycaemia. The review highlights the urgent need for evidence-based methods to reduce inpatient hypoglycaemia.

Barriers and facilitators to integrating primary and specialist healthcare in the United Kingdom: a narrative literature review

Many national policies propose integration between primary and specialist care to improve the care of people with long-term conditions. There is an increasing need to understand how to practically implement such service redesign. This paper reviews the literature on the barriers to, and facilitators of, integrating primary and specialist healthcare for people with long-term conditions in the UK, with the aim of informing the development and implementation of similar initiatives in integration. MEDLINE and CINAHL databases were searched and 14 articles discussing factors hindering or enabling integration were identified. The factors were extracted and synthesised and key lessons were tabulated. Successful integration of care requires synchronised changes on different levels, a well-resourced team, a well-defined and evidence-based service, agreed and articulated new roles and responsibilities, and a willingness among healthcare professionals to co-work and co-learn. Barriers to successful implementation of integrated care include a lack of commitment across organisations, limited resources, poorly functioning information technology (IT), poor coordination of finances and care pathways, conflicting objectives, and conflict within teams. The examples of integrated working provide insights into problems and solutions around interorganisational and interprofessional working that will guide those planning integration in the future.

Diabetes and exercise

Exercise has a beneficial effect on metabolic parameters affecting cardiovascular risk, such as lipids and blood glucose, and is a key component in both the prevention and the management of type 2 diabetes. Glycaemic control improves with both aerobic and resistance exercise in type 2 diabetes, but no glycaemic benefit is seen in type 1 diabetes. This probably results from glucose fluctuations commonly seen with exercise. Low and moderate intensity exercise are generally associated with a fall in blood glucose, and high intensity exercise can be associated with a rise in blood glucose. Trial evidence is suggestive of a reduction in cardiovascular risk with exercise, although evidence from prospective, randomised controlled trials is certainly not conclusive.

Fuelling the athlete with type 1 diabetes

People with type 1 diabetes (T1DM) want to enjoy the benefits of sport and exercise, but management of diabetes in this context is complex. An understanding of the physiology of exercise in health, and particularly the control of fuel mobilization and metabolism, gives an idea of problems which may arise in managing diabetes for sport and exercise. Athletes with diabetes need to be advised on appropriate diet to maximize performance and reduce fatigue. Exercise in diabetes is complicated both by hypoglycaemia and hyperglycaemia in particular circumstances and explanations are advanced which can provide a theoretical underpinning for possible management strategies. Management strategies are proposed to improve glycaemic control and performance.

Expanding boundaries of endocrinology

There have been huge advances in endocrine care as a consequence of improved biochemistry and diagnostic techniques as well as improved imaging. Specialist transethmoidal endoscopic surgery has improved results in pituitary tumour patients and minimally invasive parathyroid surgery has had the same consequence in patients with parathyroid disease. Multidisciplinary teams have improved outcomes in a number of areas and, as described above, endocrinologists are dealing with more in the way of endocrine disease to expand boundaries. Much work remains to be done particularly concerning the care of children and adults with late endocrine effects of cancer treatment and obesity.

'Seeing old people with real problems like leaving the gas on': students' and tutors' reflections after piloting the second phase of a new, community based course for second-year medical students

This short communication presents the results of an evaluation carried out with tutors and students at the end of the pilot phase of a new, community-based teaching course for second year students at St Bartholomew's and the Royal London School of Medicine and Dentistry. Students were placed with different community-based healthcare teams in East London. The aims of the course were to enable students to observe, and be part of, the work of the multidisciplinary team, and to experience the continuing care of patients with chronic illnesses within the community.

Monitoring long-term ecological changes through the Ecological Monitoring and Assessment Network: science-based and policy relevant

Ecological monitoring and its associated research programs have often provided answers to various environmental management issues. In the face of changing environmental conditions, ecological monitoring provides decision-makers with reliable information as they grapple with maintaining a sustainable economy and healthy environment. The Ecological Monitoring and Assessment Network (EMAN) is a national ecological monitoring network consisting of (1) about 100 case study sites across the country characterized by long-term multi-disciplinary environmental work conducted by a multitude of agencies (142 partners and counting); (2) a variety of less comprehensive yet more extensive monitoring sites; (3) a network where core monitoring variables of ecosystem change are measured; and (4) geo-referenced environmental observations. Environment Canada is the co-ordinating partner for the network through the EMAN Co-ordinating Office. EMAN's mission is to focus a scientifically-sound, policy-relevant ecosystem monitoring and research network based on (a) stabilizing a network of case-study sites operated by a variety of partners, and (b) developing a number of cooperative dispersed monitoring initiatives in order to deliver unique and needed goods and services. These goods and services include: (1) an efficient and cost-effective early warning system which detects, describes and reports on changes in Canadian ecosystems at a national or ecozone scale; and (2) cross-disciplinary and cross-jurisdictional assessments of ecosystem status, trends and processes. The early warning system and assessments of ecosystem status, trends and processes provide Environment Canada and partner organizations with timely information that facilitates increasingly adaptive policies and priority setting. Canadians are also informed of changes and trends occurring in Canadian ecosystems and, as a result, are better able to make decisions related to conservation and sustainability.

High-dose ondansetron regimen vs droperidol for morphine patient-controlled analgesia

We have performed a randomized, double-blind study comparing droperidol and high-dose ondansetron mixed with morphine for patient-controlled analgesia (PCA). To detect a reduction in the incidence of postoperative nausea and vomiting from 55% to 20% with a power of 80% at the P < 0.05 level, 29 patients per group were required. We studied 60 healthy women undergoing abdominal hysterectomy, anaesthetized using a standard technique. Group D received a bolus dose of droperidol 1.25 mg at induction followed by droperidol 0.1 mg per 1 mg of morphine from the PCA system. Group O received a bolus dose of ondansetron 4 mg at induction followed by ondansetron 0.32 mg per 1 mg of morphine. This dose of ondansetron is more than double that studied previously. Mean nausea and vomiting scores at 4, 8, 12 and 24 h, mean time to first vomit, sedation scores, incidence of side effects, and doses of prochlorperazine did not differ between the groups. In group D, 24 patients did not vomit compared with 23 in group O. The only significant difference between the groups was increased morphine consumption in the ondansetron group up until 12 h after operation (P < 0.05), but by 24 h this difference was not significant. The ondansetron regimen was more expensive (at local prices) by a factor of 27, and our results suggested no clinical advantage over droperidol.