Terence Stephenson

Exploring Variation in Glycemic Control Across and Within Eight High-Income Countries: A Cross-Sectional Analysis of 64,666 Children and Adolescents With Type 1 Diabetes

OBJECTIVE International studies on childhood type 1 diabetes (T1D) have focused on whole-country mean HbA1c levels, thereby concealing potential variations within countries. We aimed to explore the variations in HbA1c across and within eight high-income countries to best inform international benchmarking and policy recommendations. RESEARCH DESIGN AND METHODS Data were collected between 2013 and 2014 from 64,666 children with T1D who were <18 years of age across 528 centers in Germany, Austria, England, Wales, U.S., Sweden, Denmark, and Norway. We used fixed- and random-effects models adjusted for age, sex, diabetes duration, and minority status to describe differences between center means and to calculate the proportion of total variation in HbA1c levels that is attributable to between-center differences (intraclass correlation [ICC]). We also explored the association between within-center variation and children’s glycemic control. RESULTS Sweden had the lowest mean HbA1c (59 mmol/mol [7.6%]) and together with Norway and Denmark showed the lowest between-center variations (ICC ≤4%). Germany and Austria had the next lowest mean HbA1c (61–62 mmol/mol [7.7–7.8%]) but showed the largest center variations (ICC ∼15%). Centers in England, Wales, and the U.S. showed low-to-moderate variation around high mean values. In pooled analysis, differences between counties remained significant after adjustment for children characteristics and center effects (P value <0.001). Across all countries, children attending centers with more variable glycemic results had higher HbA1c levels (5.6 mmol/mol [0.5%] per 5 mmol/mol [0.5%] increase in center SD of HbA1c values of all children attending a specific center). CONCLUSIONS At similar average levels of HbA1c, countries display different levels of center variation. The distribution of glycemic achievement within countries should be considered in developing informed policies that drive quality improvement.

Psycho-educational interventions for children and young people with Type 1 Diabetes in the UK: How effective are they? A systematic review and meta-analysis.

Aims To synthesise evidence from UK-based randomised trials of psycho-educational interventions in children and young people (CYP) with Type 1 Diabetes (T1D) to inform the evidence-base for adoption of such interventions into the NHS. Methods We searched Medline, Embase, Cochrane, PsycINFO, CINAHL, and Web of Science up to March 2016. Two reviewers independently selected UK-based randomised trials comparing psycho-educational interventions for improving management of T1D for CYP with a control group of usual care or attention control. The main outcome was glycaemic control measured by percentage of glycated haemoglobin (HbA1c); secondary outcomes included psychosocial functioning, diabetes knowledge, adverse and other clinical outcomes. A narrative synthesis and meta-analysis were conducted. Pooled effect sizes of standardised mean difference (SMD) were calculated. Results Ten eligible trials of three educational and seven psycho-educational interventions were identified. Most interventions were delivered by non-psychologists and targeted adolescents with more than one year duration of diabetes. Meta-analysis of nine of these trials (N = 1,838 participants) showed a non-significant reduction in HbA1c attributable to the intervention (pooled SMD = -0.06, 95% CI: -0.21 to 0.09). Psycho-educational interventions aiming to increase children’s self-efficacy had a moderate, beneficial effect (SMD = 0.50, 95% CI: 0.13 to 0.87). No benefits on diabetes knowledge and other indicators of psychosocial functioning were identified. Conclusions There is insufficient evidence to recommend the use of particular psycho-educational programme for CYP with T1D in the UK. Further trials with sufficient power and reporting standards are needed. Future trials could consider active involvement of psychological specialists in the delivery of psychologically informed interventions and implementation of psycho-educational interventions earlier in the course of the disease. Systematic review registration PROSPERO CRD42015010701

Ethnic differences in early glycemic control in childhood-onset type 1 diabetes

Some ethnic minorities with type 1 diabetes (T1D) have worse glycemic control (higher glycated hemoglobin (HbA1c)) and increased risk for vascular complications. There is limited evidence on the impact of ethnicity on early glycemic control when most patients experience transient remission postdiagnosis. We examined associations between ethnicity and longitudinal HbA1c trajectories during the first 6 months postdiagnosis in a multiethnic cohort in East London. Research design and methods Data on 443 (50% female) children <19 years of age, with T1D and attending one of three clinics in East London between January 2005 and December 2015 were included. Linear mixed-effects modeling was used to assess ethnic differences in longitudinal HbA1c trajectories during the first 6 months postdiagnosis (1,028 HbA1c data points), adjusting for sex, age at diagnosis, socioeconomic status and pH at diagnosis. Growth curve modeling was used to plot discrete HbA1c trajectories by ethnicity. Results Longitudinal modeling revealed that all ethnic minorities had higher mean HbA1c at diagnosis compared with White children and highest in Bangladeshi (9.7 mmol/mol, 95% CI 5.1 to 14.3), Asian-Other (5.8 mmol/mol, 95% CI 2.2 to 9.3) and Somali (5.2 mmol/mol, 95% CI 0.1 to 10.2) children, and these differences persisted over the 6-month period after diagnosis. During the first month, HbA1c decreased on average by 19.6 mmol/mol (95% CI −21 to −18) for all children. Population averaged HbA1c decreased between diagnosis and 4 months, followed by a gradual increase in HbA1c levels (mean difference of −30 mmol/mol between diagnosis and 6 months). Conclusions Ethnic minorities present with higher HbA1c at diagnosis, with the largest mean differences observed in Bangladeshi, Asian-Other and Somali children. These higher levels (indicating poorer glycemic control) track into the first 6 months postdiagnosis.

Clinic variation in glycaemic control for children with Type 1 diabetes in England and Wales: a population-based, multilevel analysis

To understand the scope for improving childrens glycaemic outcomes by reducing variation between clinics and examine the role of insulin regimen and clinic characteristics.

Predictors of glycemic control in the first year of diagnosis of childhood onset type 1 diabetes: A systematic review of quantitative evidence

Early glycemic control is associated with reduced future vascular complications risk in type 1 diabetes (T1D). The aim of this study was to systematically review evidence on the predictors of glycemic control within 12 months of diagnosis of childhood onset T1D. Inclusion criteria for the electronic search were: interventional and observational studies that assessed and quantified an association between the predictor and glycemic control within 12 months of diagnosis of childhood onset T1D. A total of 17 915 articles were identified from 6 databases and 20 studies were finally included in the analysis. Harvest plots and narrative synthesis were used to summarize data from intervention (n = 0), prospective/retrospective cohort (n = 15), and cross‐sectional (n = 5) studies. Significant predictors of poorer glycemic control 0 to 3 months after diagnosis were older age and female gender. Non‐white ethnicity, diabetes autoantibody positivity, measures of deprivation, and non‐private health insurance were potential predictors. Predictors of poorer glycemic control 4 to 12 months after diagnosis were: older age, non‐white ethnicity, a single parent family, high hemoglobin A1c (HbA1c) levels at diagnosis, longer T1D duration, and non‐intensive insulin therapy. Potential predictors included: family with health issues, clinical factors, and comorbidities at diagnosis. Most significant predictors of poor glycemic control within 12 months of diagnosis of childhood onset T1D are non‐modifiable. These factors need to be recognized and addressed through individualized and multidisciplinary diabetes care. Further research is required to confirm the association of potential predictors with early glycemic control.

A survey of staffing levels in paediatric diabetes services throughout the UK

To assess staffing levels of healthcare professionals involved in the care of children and young people with diabetes in the UK.

G224 Development of cardiovascular risk factors in children with type 1 diabetes – a longitudinal cohort study

Aims Cardiovascular disease is the leading cause of death in adult diabetes. Atherosclerosis is related to modifiable risk factors beginning in childhood but can be reversed. Screening for complications is recommended from 12 years in the UK. This study investigated the prevalence and evolvement from diagnosis of cardiovascular risk factors (hypertension, obesity (BMI), dyslipidaemia, albuminuria) in an ethnically diverse population of children with Type 1 diabetes (T1D). Methods The study included all children diagnosed with T1D attending three paediatric diabetes clinics in East London between 2005–2015. Clinical and demographic information was collected prospectively during routine check-ups. Linear longitudinal mixed effects modelling (growth curve analyses) were used to analyse the development of cardiovascular risk factors from diagnosis. Models were adjusted for age, gender, ethnicity, diabetes clinic, glycated haemoglobin (HbA1c). Results Of 565 children 60% were non-white, of whom 8% were Bangladeshi and 8% from Somali backgrounds. Mean age at diagnosis was 8.5 years (0.9–19.4). Mean length of follow-up was 4.3 years (0–10.8). There was variation between clinics in commencing screening, range 8.8–13.8 years (p<0.001). 33% of all measures demonstrated an unhealthy BMI, 15% revealed microalbuminuria, 66% were above lipid targets with 5% reaching treatment thresholds. The frequency of abnormalities before or after 12 years of age was unchanged. Longitudinal modelling revealed mean increases in BMI (0.6 kg/m2, 95% CI: 0.6 to 0.7), blood pressure (1.8 mmHg, 1.5–2.1), total cholesterol (0.05 mmol/mol, 0.03–0.06) and low density lipoprotein (LDL) (0.2 mmol/mol, 0.01–0.04) per year from diagnosis for the entire cohort. HbA1c directly affected all variables other than albuminuria. Results from subgroup analysis in children over 12 years old were similar but not statistically significant. There was significant ethnic variation: Annual increments were greater for Bangladeshi children’s cholesterol and LDL (0.38 mmol/mol, 0.1–0.6 and 0.32 mmol/mol, 0.09–0.55 respectively) and Somali children’s BMI (1.6 kg/m2, 0.7–2.5) compared to White. Conclusions Levels of abnormalities were above regional and national averages in children of all ages. Somali and Bangladeshi children had less favourable profiles, which supports findings of increased cardiovascular risk in these ethnicities. Commencing screening in younger children and introducing tailored interventions for ethnic minorities with T1D may help reduce cardiovascular events in adulthood.

Diabetic Ketoacidosis Severity at Diagnosis and Glycaemic Control in the First Year of Childhood Onset Type 1 Diabetes—A Longitudinal Cohort Study

It is unclear whether diabetic ketoacidosis (DKA) severity at diagnosis affects the natural history of type 1 diabetes (T1D). We analysed associations between DKA severity at diagnosis and glycaemic control during the first year post-diagnosis. We followed 341 children with T1D, <19 years (64% non-white) attending paediatric diabetes clinics in East London. Data were extracted from routine medical registers. Subjects were categorized with normal, mild, moderate, or severe DKA. Linear mixed-effects modelling was used to assess differences in longitudinal HbA1c trajectories (glycaemic control) during 12 months post-diagnosis (1288 HbA1c data-points) based on DKA, adjusting for sex, age, ethnicity, SES (Socioeconomic Status) and treatment type. Females (OR 1.6, 95% CI 1.1–2.4) and younger age, 0–6 vs. 13–18 years (OR 2.9, 95% CI 1.5–5.6) had increased risk for DKA at diagnosis. Moderate or severe DKA was associated with higher HbA1c at diagnosis (adjusted estimates 8 mmol/mol, 2–14, and 10 mmol/mol, 4–15, respectively, compared to normal DKA). Differences in HbA1c trajectories by DKA were no longer apparent at six months post-diagnosis. All subjects experienced a steep decrease in HbA1c during the first three months followed by a gradual increase. While, DKA severity was not associated with glycaemic control at 12 months post-diagnosis, age at diagnosis, ethnicity, gender, and treatment type were significantly associated. For example, Black and mixed ethnicity children had increased risk for poor glycaemic control compared to White children (adjusted RRR 5.4, 95% CI 1.7–17.3 and RRR 2.5, 95% CI 1.2–6.0, respectively). DKA severity at diagnosis is associated with higher initial HbA1c but not glycaemic control from six months post-diagnosis. Age at diagnosis, ethnicity, gender, and insulin pump are associated with glycaemic control at one year post-diagnosis.

G443 Greater prevalence of type 2 diabetes and poorer glycaemic control among ethnic minority children in England and Wales

Background Ethnic minority children are at greater risk for type 2 diabetes. The current prevalence of type 2 diabetes in children in England and Wales is not known. Additionally, very little is known on glycaemic control in paediatric type 2 diabetes globally. Methods Using data from the National Paediatric Diabetes Audit (NPDA) for 2012–13, we estimated A. The overall, gender- and ethnic-specific prevalence of type 2 diabetes in children <16 years and B. Whether ethnicity predicts glycaemic control (mean HbA1c) in children <19 years. Ethnicity was self-identified and categorised into White, Asian, Black, Mixed, Other and ’Not-stated’. Multivariable linear regression was used to estimate differences in glycaemic control by ethnicity adjusting for socioeconomic status, age, diabetes duration and gender. Results 307 children <16 years were identified with type 2 diabetes in 2012–13. The table shows prevalence estimates by gender and ethnicity. Overall prevalence of type 2 diabetes was 2.9/100,000 with females having a higher prevalence than males (4.3 vs. 1.5/100,000). Asians had greater than five-fold increased prevalence of type 2 diabetes compared to White children (8 vs. 1.4/100,000). The highest prevalence of type 2 diabetes was found in Asian (12.2/100,000) followed by Mixed-ethnicity (4.4/100,000) females. White males had the lowest prevalence (0.6/100,000). Mixed-ethnicity children had the highest mean HbA1c (83mmol/mol) compared to other groups; White (62 mmol/mol), Asian (65 mmol/mol), Black (63 mmol/mol), Other (66 mmol/mol) and Not-stated (61 mmol/mol). In regression analysis, Mixed-ethnicity children had the highest HbA1c levels (adjusted mean difference with the White group was 22.3 mmol/mol, 95% CI 10.9–33.6), with no significant differences observed in other ethnic groups.Abstract G443 Table 1 Prevalence of type 2 diabetes by gender and ethnicity among children under 16 years of age in England and Wales Conclusion Children of all ethnic-minorities have an increased prevalence of type 2 diabetes compared to White children, with Asian females being particularly affected. Results indicate a significant increase in paediatric type 2 diabetes over the past decade. Those belonging to mixed-ethnic backgrounds had the poorest glycaemic control.

G474 Variation in patient experience across paediatric diabetes units (pdus) in england and wales and association with glycaemic outcomes

Background/Aims Patient Reported Experience Measures (PREMs) are increasingly recognised as an important indicator of healthcare quality. We investigated variation in patient experience across Paediatric Diabetes Units (PDUs) in England and Wales and explored links between PREMs and glycaemic outcomes. Methods An aggregate analysis of linked PREM and glycaemic data for each PDU collected as part of the 2012–2013 National Paediatric Diabetes Audit (NPDA) was conducted. The percentage of patients/carers who reported a high level of satisfaction (scores 9 or 10/10) in the friends and family recommendation test was used as a quality indicator of diabetes care. A funnel plot of the above indicator was constructed and the 95% and 99.8% control limits around the mean were used to distinguish between common–cause (units lying within the control limits) and special-cause variation (units lying outside the control limits). A linear regression model was also fitted for the relationship between mean satisfaction score and mean HbA1c at PDU level. Results Percentage of patients who reported high satisfaction showed a 2.9-fold variation (range 35% to 100%) across PDUs with 60/170 (35.3%) and 33/170 (19.4%) units falling outside the 95% and the 99.8% control limits respectively. Of the 170 PDUs, 21 (12.4%) lay below the lower 99.8% control limit (Figure 1). Linear regression showed a significant negative association between mean HbA1c (mmol/mol) and average satisfaction score for each PDU (regression coefficient= –2.9, 95% CI: –4.7 to –1.1, p = 0.002) (Figure 2). Abstract G474(P) Figure 1 Funnel plot of the% of children and young people/carers reporting high satisfaction (scores 9–10/10) in the friends and family recommendation test across 170 PDUs in England and Wales,2012–201 Abstract G474(P) Figure 2 Scatterplot with regression line for the association between mean satisfaction and mean HbA1c at PDU level (2012–2013 NPDA data for England and Wales) Conclusion Patient satisfaction varied considerably across PDUs with one in three units exhibiting more variation than expected by chance. At aggregate level, clinics with higher mean satisfaction had better average glycaemic control. Individual-level analyses allowing for nested sources of variation (patients within clinics) are needed to confirm these clinic-level analyses. Further work should also adjust for “case-mix variation” i.e. correct for differences in the sociodemographic and disease profile of children served by different clinics.