Australasian Diabetes Data Network Research Articles

Semantic privacy-preserving framework for electronic health record linkage

The combination of digitized health information and web-based technologies offers many possibilities for data analysis and business intelligence. In the healthcare and biomedical research domain, applications depending on electronic health records (EHRs) identify privacy preservation as a major concern. Existing solutions cannot always satisfy the evolving research demands such as linking patient records across organizational boundaries due to the potential for patient re-identification. In this work, we show how semantic methods can be applied to support the formulation and enforcement of access control policy whilst ensuring that privacy leakage can be detected and prevented. The work is illustrated through a case study associated with the Australasian Diabetes Data Network (ADDN – www.addn.org.au), the national paediatric type-1 diabetes data registry, and the Australian Urban Research Infrastructure Network (AURIN – www.aurin.org.au) platform that supports Australia-wide access to urban and built environment data sets. We demonstrate that through extending the eXtensible Access Control Markup Language (XACML) with semantic capabilities, finer-grained access control encompassing data risk disclosure mechanisms can be supported. We discuss the contributions that can be made using this approach to socio-economic development and political management within business systems, and especially those situations where secure data access and data linkage is required.

Prevalence of Celiac Disease in 52,721 Youth With Type 1 Diabetes: International Comparison Across Three Continents.

Celiac disease (CD) has a recognized association with type 1 diabetes. We examined international differences in CD prevalence and clinical characteristics of youth with coexisting type 1 diabetes and CD versus type 1 diabetes only. Data sources were as follows: the Prospective Diabetes Follow-up Registry (DPV) (Germany/Austria); the T1D Exchange Clinic Network (T1DX) (U.S.); the National Paediatric Diabetes Audit (NPDA) (U.K. [England/Wales]); and the Australasian Diabetes Data Network (ADDN) (Australia). The analysis included 52,721 youths <18 years of age with a clinic visit between April 2013 and March 2014. Multivariable linear and logistic regression models were constructed to analyze the relationship between outcomes (HbA1c, height SD score [SDS], overweight/obesity) and type 1 diabetes/CD versus type 1 diabetes, adjusting for sex, age, and diabetes duration. Biopsy-confirmed CD was present in 1,835 youths (3.5%) and was diagnosed at a median age of 8.1 years (interquartile range 5.3-11.2 years). Diabetes duration at CD diagnosis was <1 year in 37% of youths, >1-2 years in 18% of youths, >3-5 years in 23% of youths, and >5 years in 17% of youths. CD prevalence ranged from 1.9% in the T1DX to 7.7% in the ADDN and was higher in girls than boys (4.3% vs. 2.7%, P < 0.001). Children with coexisting CD were younger at diabetes diagnosis compared with those with type 1 diabetes only (5.4 vs. 7.0 years of age, P < 0.001) and fewer were nonwhite (15 vs. 18%, P < 0.001). Height SDS was lower in those with CD (0.36 vs. 0.48, adjusted P < 0.001) and fewer were overweight/obese (34 vs. 37%, adjusted P < 0.001), whereas mean HbA1c values were comparable: 8.3 ± 1.5% (67 ± 17 mmol/mol) versus 8.4 ± 1.6% (68 ± 17 mmol/mol). CD is a common comorbidity in youth with type 1 diabetes. Differences in CD prevalence may reflect international variation in screening and diagnostic practices, and/or CD risk. Although glycemic control was not different, the lower height SDS supports close monitoring of growth and nutrition in this population.

The Australasian Diabetes Data Network: first national audit of children and adolescents with type 1 diabetes.

To assess glycaemic control, anthropometry and insulin regimens in a national sample of Australian children and adolescents with type 1 diabetes. Cross-sectional analysis of de-identified, prospectively collected data from the Australasian Diabetes Data Network (ADDN) registry. Five paediatric diabetes centres in New South Wales, Queensland, South Australia, Victoria and Western Australia. Children and adolescents (aged 18 years or under) with type 1 diabetes of at least 12 months' duration for whom data were added to the ADDN registry during 2015. Glycaemic control was assessed by measuring haemoglobin A1c (HbA1c) levels. Body mass index standard deviation scores (BMI-SDS) were calculated according to the CDC-2000 reference; overweight and obesity were defined by International Obesity Task Force guidelines. Insulin regimens were classified as twice-daily injections (BD), multiple daily injections (MDI; at least three injection times per day), or continuous subcutaneous insulin infusion (CSII). The mean age of the 3279 participants was 12.8 years (SD, 3.7), mean diabetes duration was 5.7 years (SD, 3.7), and mean HbA1c level 67 mmol/mol (SD, 15); only 27% achieved the national HbA1c target of less than 58 mmol/mol. The mean HbA1c level was lower in children under 6 (63 mmol/mol) than in adolescents (14-18 years; 69 mmol/mol). Mean BMI-SDS for all participants was 0.6 (SD, 0.9); 33% of the participants were overweight or obese. 44% were treated with CSII, 38% with MDI, 18% with BD. Most Australian children and adolescents with type 1 diabetes are not meeting the recognised HbA1c target. The prevalence of overweight and obesity is high. There is an urgent need to identify barriers to achieving optimal glycaemic control in this population.

Australasian Diabetes Data Network: Building a Collaborative Resource.

Australasia is a region with a high incidence of type 1 diabetes (T1D). There are approximately 140 000 individuals with T1D, and of these 10 000 are children. Although the region covers a huge geographical area, most children with T1D are managed by tertiary academic centers in the major capital cities. Local longitudinal data collection has been in place for several decades in most of these centers, however ongoing national data collection had not been attempted. In 2012, with funding from the Juvenile Diabetes Research Foundation (JDRF) Australian Type 1 Clinical Research Network, a national collaboration was formed to provide ongoing longitudinal collection of T1D patient characteristics and outcomes. The initial phase of this collaboration, known as the Australasian Diabetes Data Network or ADDN, was led by the Australasian Paediatric Endocrine Group (APEG) and thus included only children and adolescents. The next phase, commenced in 2016, will see adult sites added through collaboration with the Australian Diabetes Society (ADS). As most of the initial centers had longitudinal data collection in place the model employed was to establish the transfer and collation of data already collected into a central database. This required the definition of a common data dictionary, ethics and governance procedures and the employment of technology to enable efficient and accurate information transfer and accessibility. The ADDN project received widespread support from the diabetes research community with study investigators representing 20 pediatric centers across the region. The first phase focused on the 5 largest centers and at the end of 2015 these centers were uploading patient data to the ADDN database on a quarterly basis resulting in 5271 patients with 83 506 diabetes visits.

The Australasian diabetes data network (ADDN): first steps towards a national database resource

The Australasian Diabetes Data Network (ADDN) is a collaboration that aims to collect and centrally collate a suite of patient data across Australia and New Zealand that will be available to all diabetes researchers. ADDN will collect information prospectively, facilitating data sharing and building capacity in clinical service delivery and investigation. The central database will provide a research resource, facilitate study recruitment and be an ongoing source of data to enable benchmarking against national outcomes. By April 2015 ADDN will be populated with over 4,000 participants presenting an opportunity to analyse national data for the first time. To date ADDN holds the data of 1136 children and adolescents. Of the T1D participants (n=1064), males and females were equally represented, with 3% (35) aged 20 years of age. A total of 10% were treated with twice daily injections, 42% with Multiple Daily Injections and 47% with continuous subcutaneous insulin infusion (CSII). The TID group had a median HbA1c of 8% (64 nmol/mol) with 31% (330) 9.0% (74.9 nmol/mol). From these preliminary results it is clear that a large proportion of participants are not meeting recommended glycaemic targets[1] despite the high uptake of intensified insulin therapy. Further analysis of the ADDN dataset is needed to understand the influence of different practices and therapies for T1D used in Australian youth, and the clinical and demographic predictors of glycaemic outcomes. This research and resource will have major implications for clinical investigation in Australasia and will provide an evidence base to inform health policy. On behalf of the ADDN Study Group