Australasian Diabetes Data Network

Abstract

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.

Australia is one of the high incidence countries for type 1 diabetes (T1D): there are approximately 10 000 children (aged <18 years) with T1D nationally and 1000 new onset cases per annum.¹ The majority of these cases receive clinical care in tertiary academic centers or from clinical teams linked to tertiary centers.² Incidence registries in Australia have documented an increased rate of T1D over 4 decades and provided insights into the epidemiology of the disease.^3-7 In addition to investigating disease epidemiology, the value of prospectively monitoring patient characteristics and outcomes was recognized. In parallel, the availability of databases in the clinical setting led to the establishment of local datasets in the late 1980s and early 1990s. These have facilitated and informed clinical care, service planning and research.^8-13

In 2012 the recognition of the benefits of national collaboration along with advances in information technology led to the formation of the Australasian Diabetes Data Network (ADDN; www.addn.org.au). This initiative, funded through the Juvenile Diabetes Research Foundation (JDRF) Australian Type 1 Clinical Research Network (T1DCRN, www.t1dcrn.org.au), sought to establish a national pediatric diabetes registry and was led in its first phase by the Australasian Paediatric Endocrine Group (APEG). Critical technological direction, database development and maintenance were provided by the University of Melbourne eResearch group. More recently the Australian Diabetes Society (ADS) has joined the collaboration, a development that will see adult cases included in the registry and will provide valuable data on outcomes across the life span.

This article describes the initial phase of ADDN that focused on major pediatric centers within Australia. A staged expansion to include regional sites, centers managing adults with T1D and New Zealand is planned for the next phase, which commenced in 2016. Although Australasia covers a huge geographical area, the majority of children with T1D are managed by tertiary academic centers in the major capital cities. As most of these centers had longitudinal data collection in place the model employed was to establish the collection and collation of these data into a central database on an on-going basis. 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.

Objective and Specific Aims

The overall objective of ADDN is to provide national data on people with diabetes—to enhance service planning and clinical care through audit and benchmarking, and to facilitate research and collaboration.

Specific aims include:

1. Long-term monitoring of diabetes outcomes in the Australasian population with T1D across the life span.

2. Linkage of biobank samples (DNA, plasma, serum) with a detailed registry of clinical phenotype in a population based sample of T1D.

3. Population based recruitment of young people for trials in new onset and established T1D.

4. Enrolment of patient cohorts for specific cross-sectional and longitudinal research.

Methods

Overview

The initial phase included the design and development of a central database with a web-based user interface and the ingestion of data from 5 tertiary pediatric centers across Australia—Princess Margaret Hospital for Children (PMH) in Western Australia, Children’s Hospital at Westmead (CHW) in New South Wales, Royal Children’s Hospital (RCH) in Victoria, Lady Cilento Children’s Hospital (LCCH) in Queensland, and Women’s and Children’s Hospital (WCH) in South Australia.

The project was overseen by a steering committee including representatives from each of the 5 phase 1 centers, APEG and JDRF. The group investigated potential database systems available at the time, prior to deciding to custom build a system to leverage a number of existing center-based clinical and research databases already in place.

University of Melbourne’s eResearch group was appointed following a competitive selection process to build the ADDN database and user interface. User requirements were gathered through a series of workshops and consultation.

A data dictionary was developed to define the 130+ data points to be captured in the ADDN database. Data points related to either a patient or to a patient visit and included demographic and clinical characteristics, therapeutic interventions, comorbidities and complications, documentation of biobanked samples and consent to be contacted for future research. Where possible, International Society for Pediatric and Adolescent Diabetes (ISPAD)/APEG/ADS definitions and standard classifications were used. The data dictionary was defined in close cooperation with the diabetes research community.

Implementation of ADDN

The specification for the ADDN registry included several key capabilities: an intuitive and user-friendly interface; data ingestion from existing center-based databases; a range of queries and reports; the overall performance of the system, the hosting and back-up of the data; user documentation and training; and, importantly, support for enhanced security including data integrity, data privacy, and data auditability.

It was critical to support extraction of data from existing center-based clinical databases to avoid double data entry. Electronic medical records and personally controlled electronic health records (PCEHR) are being progressively rolled out across Australasia but currently there is a wide range of IT systems being used for pediatric diabetes patient management. These systems are predominantly custom-built applications developed using various versions of SQL Server and Microsoft Access.

Data extracts are produced locally and uploaded at quarterly intervals with center-specific data sets mapped to the core data model as defined by the data dictionary. The core data model has been essential to underpin the ADDN work, with the software systems and processes focused on ensuring that heterogeneous data sets from the different centers are systematically aligned with this model. To support this, standardized data extraction routines were developed and aligned with local IT systems. The canonical data schema was represented in XML, with versions in other formats provided by centers where technical expertise was limited.

Guarantees of confidentiality and protection of access to data are fundamental to ADDN. Identifying data are not loaded to ADDN and only the contributing diabetes data centers are aware of the identity of individuals on the registry. To allow patients to be tracked across multiple centers as the project progresses, individual participants must be uniquely identifiable. This will be achieved by using a defined set of personally identifying information to create a unique data linkage key that is stored in the local data set and loaded to ADDN together with the deidentified patient data.

Architecture of ADDN

The ADDN architecture has been developed to be scalable. It has been realized by an n-tier service-oriented architecture adopting the Model-View-Controller (MVC) paradigm¹⁴ which provides a clean separation of the data layer (database/data) from the application logic and security and the front-end user interface. The underlying database technology is MySQL, which provides a robust, open source relational database. It provides direct support for a service–oriented architecture with Representational State Transfer–based services utilizing the Spring IO Platform for security and data persistence.¹⁵ The high-level structure of the ADDN architecture is shown in Figure 1. The system supports a range of data processing pipelines that enable Extraction, Transformation and Loading (ETL) capabilities. This includes support for data extraction from the heterogeneous diabetes data resources, which are transformed into the standardized/agreed ADDN data model for subsequent querying and analysis through the ADDN user interface.

Figure 1.

High level ADDN architecture. Local database DPMS (Diabetes Patient Management System) is used at PMH (WA), RCH (Vic), and WCH (SA); EPMS (Endocrine Patient Management System) is used at CHW (NSW).

User Interface, Reports, and Data Interrogation

The user interface to the ADDN system provides a range of capabilities reflecting the needs of the ADDN community. The front page includes a report of the patients loaded to the registry summarized by center and diabetes type and for each center, the date of the most recent upload. ADDN includes options for the interrogation and analysis of the data. The number of patients matching a range of primary and secondary characteristics can be determined using a Registry Query tool to assist with study planning (Figure 2). The tool shows the number of individuals resulting from the application of each filter, providing a richer understanding of the underlying data.

Figure 2.

Query builder interface.

The ADDN system also supports a range of reports targeted to specific centers and the functionality to search across all individual patients using any of the data points that are defined in the data model. Specific values and upper and lower bounds can be specified for data points where applicable. To support the validation of data the ADDN model includes a quality and completeness report, detailed reporting of a random selection of individual patients, and a set of predefined reports of summary information on a range of measures. The same reports can also be run for data aggregated from all centers, allowing benchmarking of individual centers against the total group, highlighting missing data and other data quality issues.

To support data export from the registry for approved research purposes, the front-end web information has been formatted so that the HTML on the page can be copied directly. Data can be saved in various formats such as Excel or Comma Separated Variable for analysis in a range of statistical packages including Stata and SPSS.

Security

ADDN users are authenticated using individual usernames and passwords. All users are assigned a security level and the ADDN records to which they have access are determined according to the ADDN Data Access Model (Figure 3):

Figure 3.

Model for data access.

• Administrator security allows access to all of the deidentified data within the ADDN database. Individuals with this role can run standard reports for individual centers and for aggregated data, perform data queries and extraction and have access to administration functions such as adding new users. This role is restricted to system administrators and lead investigators on ADDN.

• JDRF T1DCRN members are able to view the front page report of patients loaded to the registry summarized by center and diabetes type. They are also able to use the Registry Query tool to find the number of patients (across all centers) that match selected criteria. This group of users do not have access to individual patient level data.

• In addition to the access granted to JDRF T1DCRN members, center-based users will have detailed access to all of the data within their center. Patient records can be searched and viewed individually and standard reports can be run for the center.

The wider community, including researchers who do not belong to the JDRF T1DCRN, will be able to view a summary of aggregated data on an ADDN website to be developed in the next phase of the ADDN project.

Researchers apply to the ADDN Committee to use the data and if approved, they are provided with a data extract of the relevant fields. Protocols cover the application process, security, use and disposal of ADDN data.

Ethics

Institutional regulatory board (ethics) approval was obtained by each center according to center-specific requirements. Newly diagnosed patients were consented prospectively and existing patients were consented retrospectively as they attended routine clinic appointments.

Governance and Data Access

Governance, publication and data access protocols were developed based on input from US-based T1D Exchange Clinic Registry, BioGrid Australia and a range of Australian cohort studies unrelated to diabetes. The governance protocol includes sections on the phases of ADDN, scope, roles and responsibilities, risk management, performance assessment and ADDN committee details such as composition, term of office and member obligations. The publication and data access protocol include sections on the use of ADDN data, publication of findings and authorship.

Results

The ADDN project received widespread support from the diabetes research community with study investigators representing 20 pediatric centers across Australia (Appendix A). The first phase of ADDN was completed at the end of June 2015, reaching all agreed milestones within the set timeframe and within budget. Major achievements included the consent of >6000 patients, the development and refinement of a common data dictionary (Appendix B), and the establishment of a secure, centralized database populated with clinical and demographic data from 5 major pediatric centers across Australia. At the completion of the project these centers were uploading patient data to the ADDN database on a quarterly basis resulting in 5271 patients with 83 506 diabetes visits.

Within 6 months of the project start the ADDN Committee and project managers had been appointed, ethics applications were lodged with participating centers, a database prototype was demonstrated, a shared data dictionary was agreed and a requirements specification document was completed. At 18 months local ethics approvals were obtained by each of the 5 major diabetes centers and an additional 16 diabetes centers throughout NSW and Queensland, and 1 New Zealand diabetes center.

For the majority of diabetes centers, ethics approval was granted on the basis that written patient consent be obtained. It was usual practice to obtain written consent from patients, or in the case of minors their caregivers, at the time of diagnosis for the national incidence registry. An amendment to include the option to consent to ADDN on the same consent form as that used for the national incidence registry was submitted and approved by ethics. Diabetes centers that demonstrated significant strain on resources due to the burden of obtaining consent from prevalent cases received funding to enable the allocation of dedicated staff to consent patients in clinic. In 1 center the ethics board approved the transfer of deidentified data to ADDN without the need for written consent.

The number of patients consented to ADDN increased from 1500 at 12 months to 4800 at 2 years and 6000 by the end of the first phase. There have been no participants who have withdrawn from the study. The first data were uploaded to ADDN from 2 centers within 2 years of the project start with a third center uploading data 6 months later.

One of the major challenges for meeting the ADDN project milestones was communication with the participating centers and the wider diabetes community. A substantial proportion of participants were to be drawn from prevalent cases, largely recruited through busy diabetes clinics. The technical challenges of transferring data from numerous local centers to the central ADDN database also warranted clear communication between the local IT staff and the database developers. The project required actively engaged stakeholders to proceed in a timely manner. To achieve effective communication weekly project management meetings and monthly ADDN investigator meetings were held. Face-to-face investigator meetings were held every 6 to 12 months. ADDN Newsletters were utilized to update APEG members and the research community on the project’s progress. The ADDN project has resulted in poster and oral presentations at national scientific meetings and at international scientific meetings.^16-18 In a recent T1D Exchange meeting, “Global Data Collection and Outcomes” (Boston 2015), ADDN investigators were invited panelists, an important step to developing partnerships with other groups collecting diabetes data.

Conclusion

The development of prospective national datasets is increasing globally for T1D particularly in pediatrics: a phenomenon that has been driven by the recognition of the importance of monitoring outcomes and, at the same time, facilitated by the rapid evolution of information technology. As therapies change and the incidence of the disease rises it is increasingly important to obtain representative data to inform clinicians, patients and their carers, the institutions that fund treatments, policymakers and the broader community. Such data are a rich source of information that can initiate and inform research activity and facilitate study recruitment, and inform clinical practice. Although Australia has a long track record of data collection, this had predominantly been at an individual center level and the expansion to a national level was overdue. The existence of local databases meant that the most efficient approach was to integrate these systems into a central collection rather than to develop a new local and central platform. Although in concept this was straightforward, there were a number of challenges. These included the definition of a common acceptable data dictionary, the technical issues around having differing source databases, security of transfer of information and agreement on the governance of the resource. To date these have been addressed and over 50% of Australian children with T1D are now included in the ADDN prospective database that will provide important data over time.

A major challenge is integration with local databases and electronic medical records at the source centers. There are multiple systems in use and the systems used change. Addressing this challenge is complex and requires close collaboration with the IT teams at the local site with technical and clinical input from the ADDN team as well as coordination from the local clinicians.

The next phase of ADDN is to include sites managing adult patients and to expand to other pediatric centers across Australia and New Zealand. Potentially over 80% of Australasian children with T1D and a significant number of adult patients can be included in this phase. Differences in health care provision in adult patients may present a challenge and the first patients will be those who receive their clinical care in the public as opposed to private health care system. The development of nationally representative data will enable benchmarking at an international level and facilitate collaboration. The ongoing challenge is to maintain the integrity of the data, ensure its efficient use and sustainability. Whilst the resource needs to retain independence, ongoing support will be obtained not only from the research community but also from a range of agencies, including industry and health care funders, who may utilize the information to the benefit of patient care.

Australasian Diabetes Data Network (ADDN) Study Group

Name	Organization
Professor Maria Craig	The Children’s Hospital at Westmead, Sydney
Clinical Professor Tim Jones	Princess Margaret Hospital for Children, Perth
Clinical Professor Geoff Ambler	The Children’s Hospital at Westmead, Sydney
Professor Jenny Batch	Lady Cilento Children’s Hospital, Brisbane
Dr Phil Bergman	Monash Children’s Hospital, Melbourne
Professor Fergus Cameron	Royal Children’s Hospital, Melbourne
Professor Peter Colman	Royal Melbourne Hospital, Melbourne
A/Prof Louise Conwell	Lady Cilento Children’s Hospital, Brisbane
A/Prof Andrew Cotterill	Lady Cilento Children’s Hospital, Brisbane
Professor Jennifer Couper	Women’s and Children’s Hospital, Adelaide
A/Prof Elizabeth Davis	Princess Margaret Hospital for Children, Perth
Professor Kim Donaghue	The Children’s Hospital at Westmead, Sydney
Dr Jan Fairchild	Women’s and Children’s Hospital, Adelaide
Dr Leonie Gray	Mater Medical Centre, Rockhampton
Professor Paul Hofman	University of Auckland, New Zealand
Dr Neville Howard	The Children’s Hospital at Westmead, Sydney
Dr Michelle Jack	Royal North Shore Hospital, Sydney
Dr Craig Jefferies	Starship Children’s Hospital, New Zealand
A/Prof Bruce R King	John Hunter Children’s Hospital, Newcastle
Dr Antony Lafferty	The Canberra Hospital, Canberra
Dr Robert McCrossin	Gladstone Hospital, Gladstone
Dr Mark Pascoe	Royal Hobart Hospital, Hobart
Dr Alexia Peña	The University of Adelaide
A/Prof Darrell Price	Pacific Private Clinic, Gold Coast
A/Prof Christine Rodda	Monash Children’s Hospital, Melbourne
Professor Richard Sinnott	eResearch, University of Melbourne
Dr Alan Sive	Townsville Hospital, Townsville
Dr Carmel Smart	John Hunter Children’s Hospital, Newcastle
Dr Monique Stone	Royal Darwin Hospital, Darwin
Dr Elaine Tham	Women’s and Children’s Hospital, Adelaide
Dr Charles Verge	Sydney Children’s Hospital, Sydney
Professor Jerry Wales	Lady Cilento Children’s Hospital, Brisbane
Dr Tim Warnock	Cairns Diabetes Centre, Cairns
Dr Judy Williams	Bundaberg Diabetes Centre, Bundaberg
Dr Michael Williams	Mackay Base Hospital, Mackay
A/Prof Esko Wiltshire	Wellington Children’s Hospital, New Zealand
Dr Nick Woolfield	Caboolture Hospital, Caboolture
Professor Sophia Zoungas	Monash Medical Centre, Melbourne

Appendix

Appendix A.

ADDN Phase 1 Investigators and Centers (Pediatric).

Principal Investigators	Organization
Professor Maria Craig	The Children’s Hospital at Westmead, Sydney
Clinical Professor Tim Jones	Princess Margaret Hospital for Children, Perth
Co-investigators
Clinical Professor Geoff Ambler	The Children’s Hospital at Westmead, Sydney
Professor Jenny Batch	Lady Cilento Children’s Hospital, Brisbane
Dr Phil Bergman	Monash Children’s Hospital, Melbourne
Professor Fergus Cameron	Royal Children’s Hospital, Melbourne
A/Prof Louise Conwell	Lady Cilento Children’s Hospital, Brisbane
A/Prof Andrew Cotterill	Lady Cilento Children’s Hospital, Brisbane
Professor Jennifer Couper	Women’s and Children’s Hospital, Adelaide
A/Prof Elizabeth Davis	Princess Margaret Hospital for Children, Perth
Professor Kim Donaghue	The Children’s Hospital at Westmead, Sydney
Dr Jan Fairchild	Women’s and Children’s Hospital, Adelaide
Dr Leonie Gray	Mater Medical Centre, Rockhampton
Dr Neville Howard	The Children’s Hospital at Westmead, Sydney
Dr Michelle Jack	Royal North Shore Hospital, Sydney
Dr Bruce King	John Hunter Children’s Hospital, Newcastle
Dr Antony Lafferty	The Canberra Hospital, Canberra
Dr Robert McCrossin	Gladstone Hospital, Gladstone
Dr Mark Pascoe	Royal Hobart Hospital, Hobart
Dr Alexia Peña	Women’s and Children’s Hospital, Adelaide
A/Prof Darrell Price	Pacific Private Clinic, Gold Coast
A/Prof Christine Rodda	Monash Children’s Hospital, Melbourne
Dr Alan Sive	Townsville Hospital, Townsville
Dr Carmel Smart	John Hunter Children’s Hospital, Newcastle
Dr Monique Stone	Royal Darwin Hospital, Darwin
Dr Elaine Tham	Women’s and Children’s Hospital, Adelaide
Dr Charles Verge	Sydney Children’s Hospital, Sydney
Dr Tim Warnock	Cairns Diabetes Centre, Cairns
Dr Judy Williams	Bundaberg Diabetes Centre, Bundaberg
Dr Michael Williams	Mackay Base Hospital, Mackay
Dr Nick Woolfield	Caboolture Hospital, Caboolture
Phase 1 Clinical Centers
State/territory	Name and address
Western Australia	Department Endocrinology and Diabetes, Princess Margaret Hospital for Children Roberts Road, Subiaco, Perth, WA 6008
New South Wales	Institute of Endocrinology & Diabetes, The Children’s Hospital at Westmead Locked Bag 4001, Westmead, NSW 2145
	Department of Endocrinology, Sydney Children’s Hospital High St, Randwick, NSW 2031
	Department Paediatric Endocrinology, Royal North Shore Hospital Reserve Road, St Leonard’s, NSW 2065
	Department of Endocrinology, John Hunter Children’s Hospital Lookout Road, Newcastle, NSW 2310
Victoria	Department Endocrinology and Diabetes, Royal Children’s Hospital Flemington Road, Parkville, VIC 3052
	Paediatric Endocrinology and Diabetes, Monash Children’s Hospital 246 Clayton Road, Clayton, VIC 3168
South Australia	Endocrinology and Diabetes Centre, Women’s and Children’s Hospital 72 King William Road, North Adelaide, SA 5006
Queensland	Department Endocrinology and Diabetes, Lady Cilento Children’s Hospital 501 Stanley Street, South Brisbane, QLD 4101
	Child and Adolescent Health Service, Mackay Base Hospital PO 5580, Mackay Mail Centre, QLD 4741
	Cairns Diabetes Centre, North Cairns Community Health Centre 381 Sheridan St, Cairns, QLD 4870
	Diabetes Centre, Bundaberg Base Hospital PO Box 34, Bundaberg, QLD 4670
	Diabetes Centre, Townsville Hospital P O Box 670, Townsville, QLD 4810
	Gladstone Hospital Park Street, Gladstone, QLD 4680
	Pacific Private Clinic, Gold Coast 8/123 Nerang Street, Southport, QLD 4215
	Rockhampton Diabetes Clinic Suite 9, Mater Medical Centre, Jessie Street, Rockhampton, QLD 4700
	Diabetes Clinic, Caboolture Hospital Locked Mail Bag 3, Caboolture, QLD 4510
Tasmania	Royal Hobart Hospital Diabetes Centre 70 Collins Street, Hobart, TAS 7000
Australian Capital Territory	Paediatric and Adolescent Diabetes Unit, Canberra Hospital PO Box 11, Woden ACT 2606
Northern Territory	Royal Darwin Hospital 105 Rocklands Drive, Tiwi, NT 0810

Appendix B.

ADDN Data Dictionary (Phase 1—Pediatric Data).

Table: Patient
Field	Description	Expected values
ADDN ID	Unique identifier for patient
Center code	State-city-center	Valid ADDN center code
Local ID	Local identifier for patient not linked to identifying information (source system)
Active flag	Active flag (patient is no longer active if they have transferred to another center, died, or if it is >2 years since last visit)	Yes, no
Date of birth	Patient DOB	Valid date
Gender	Gender of patient	M, F, undetermined
Date of consent for ADDN	Date patient consented to ADDN	≥date of diagnosis
Consent to be contacted for future research	Indicates whether patient consents to be contacted re future research	Yes, no
Country of birth	Patient country of birth	SACC (Standard Australian Classification of Countries, ABS) lowest level (4 digits)—2nd revision, 2011
Ethnicity (primary)	Patient ethnicity—primary	ASCCEG (Australian Standard Classification of Cultural and Ethnic Groups, ABS) 2nd level (2 digits)
Ethnicity (secondary)	Patient ethnicity—secondary	As for primary
Indigenous status	Indigenous status for patients	For Australian centers—Aboriginal, Torres Strait Islander (TSI), both Aboriginal and TSI, neither Aboriginal nor TSI For NZ centers—Maori, non-Maori
Language spoken at home	Language spoken at home	ASCL (Australian Standard Classification of Languages, ABS) 2nd level (2 digits)
Birth weight	Birth weight (kg)	0.4-6.0
Birth weight SDS for gestational age	Calculated from tables	−5.0-5.0
Birth weight for gestational age percentile	Calculated from tables	0.0-100.0
Gestation at birth	Gestation at birth (weeks)	20.0-43.0
Mode of birth	Mode of birth	Caesarian, NVD, forceps/assisted
Date of menarche	Date of menarche (females only)	After date of birth + 8 years and no later than today’s date
Transfer date	Date patient transferred to another center	After date of diagnosis
Transfer to	Center patient transferred to	Tertiary, primary, private, unknown
ADDN withdrawal date	Date patient withdrew from ADDN	After date of diagnosis and no later than today’s date
Deceased date	Date of death	After date of diagnosis and no later than today’s date
Cause of death	Cause of death	DKA, hypoglycemia, dead in bed, sepsis, other diabetes related, not diabetes related
Cause of death other	Description of “other diabetes-related” cause of death	Free text
DNA stored	DNA stored flag	yes, no
HLA A-1	HLA A-1 genotype	15 digits
HLA A-2	HLA A-2 genotype
HLA B-1	HLA B-1 genotype
HLA B-2	HLA B-2 genotype
HLA C-1	HLA C-1 genotype
HLA C-2	HLA C-2 genotype
HLA DRB1-1	HLA DRB1-1 genotype
HLA DRB1-2	HLA DRB1-2 genotype
HLA DQA1-1	HLA DQA1-1 genotype
HLA DQA1-2	HLA DQA1-2 genotype
HLA DQB1-1	HLA DQB1-1 genotype
HLA DQB1-2	HLA DQB1-2 genotype
HLA DPA1-1	HLA DPA1-1 genotype
HLA DPA1-2	HLA DPA1-2 genotype
HLA DPB1-1	HLA DPB1-1 genotype
HLA DPB1-2	HLA DPB1-2 genotype
Diabetes type	Diabetes type	Type 1, type 2, gestational, monogenic, CFRD, neonatal, other
Diabetes type other	Description of “other” diabetes type	Free text
Date of diagnosis	Date of diagnosis	No later than today’s date
No days hospitalized at diagnosis	Number of days hospitalized at diagnosis (days)	≥0 and <15
Country at diagnosis	Country at diagnosis	SACC (Standard Australian Classification of Countries, ABS) lowest level (4 digits)-2nd revision, 2011
Postcode at diagnosis	Postcode at diagnosis	Valid postcode
DKA at diagnosis	DKA at diagnosis (ISPAD guidelines: Hyperglycemia (blood glucose >11 mmol/L), venous pH <7.3 or bicarbonate <15 mmol/L, ketosis)	Yes, no
Current postcode	Current postcode	Valid postcode
Table: Comorbidity
Field	Description	Expected values
Comorbidity	Comorbidity	Valid comorbiditya
Comorbidity other	Comorbidity other than those defined	Free text
Date of diagnosis	Date of diagnosis	No later than today’s date
Table: Family history
Field	Description	Expected values
Relationship to proband	Relationship of family member to the patient	Father, mother, sibling, half-sibling, child
Diabetes type	Diabetes type affecting family member	Type 1, type 2, gestational, monogenic, CFRD, neonatal, other, unspecified
Diabetes type other	Description of “other” diabetes type	Free text
Autoimmune disease	Autoimmune disease affecting family member	Thyroid, celiac, other
Autoimmune disease other	Description of “other” autoimmune disease	Free text
Table: Visit
Field	Description	Expected values
Date of visit	Date of patient visit to clinic	Valid date
Diagnosis visit	Diagnosis visit flag	Yes, no
Center code at visit	Center code at visit	Valid ADDN center code
HbA1c (NGSP)	HbA1c (NGSP) (%)	3.0-20.0
HbA1c (IFFC)	HbA1c (IFFC) (mmol/mol)	9-195
Height	Height (cm)	<200.0
Height SDS	Calculated from tables	−5.0-5.0
Height percentile	Calculated from tables	0.0-100.0
Weight	Weight (kg)	<200.0
Weight SDS	Calculated from tables	−5.0-5.0
Weight percentile	Calculated from tables	0.0-100.0
BMI	Calculated from height and weight (kg/m^2)
BMI SDS	Calculated from tables	−5.0-5.0
BMI percentile	Calculated from tables	0.0-100.0
Waist circumference	Waist circumference (cm)	<120.0
Waist to height ratio	Calculated field
Systolic blood pressure	Systolic blood pressure (mmHg)	50-250
Systolic BP SDS	Calculated from tables	−5.0-5.0
Systolic BP percentile	Calculated from tables	0.0-100.0
Diastolic blood pressure	Diastolic blood pressure (mmHg)	40-120
Diastolic BP SDS	Calculated from tables	−5.0-5.0
Diastolic BP percentile	Calculated from tables	0.0-100.0
IAA pos	IAA antibodies positive flag	Yes, no
IA-2 pos	IA-2 antibodies positive flag	Yes, no
GAD pos	GAD antibodies positive flag	Yes, no
ZnT8 pos	ZnT8 antibodies positive flag	Yes, no
ICA pos	ICA antibodies positive flag	Yes, no
Insulin regimen	Insulin regimen	CSII, BD/twice daily, MDI, Other
Pump make/model	Pump make/model for CSII patients	Valid pump make and model
Pump make/model other	Pump make/model other than those defined	Free text
ICR	Insulin to carb ratio used for MDI patients	Yes, no
ISF	Insulin sensitivity factor used for MDI patients	Yes, no
No injections per day	Number of injections per day for MDI patients	≤10
Insulin (1)	Insulin 1 name	Valid insulin name
Insulin daily dose (1)	Insulin 1 daily dose (units)	0-200
Insulin (2)	Insulin 2 name	Valid insulin name
Insulin daily dose (2)	Insulin 2 daily dose (units)	0-200
Insulin (3)	Insulin 3 name	Valid insulin name
Insulin daily dose (3)	Insulin 3 daily dose (units)	0-200
Insulin (4)	Insulin 4 name	Valid insulin name
Insulin daily dose (4)	Insulin 4 daily dose (units)	0-200
Total daily insulin dose	Total daily insulin dose (units)	0-300
Daily basal insulin dose	Total basal insulin dose (units)	0-200
Basal insulin %	Calculated field	0-100
SMBG frequency per day	Average frequency of self-monitoring of blood glucose per day since last visit. If available use 7 day average from glucometer	0-30
CGM	Continuous glucose monitoring since last visit or if last visit>6 m ago then in the last 6 m	Real-time, retrospective, no
CGM %	% of time CGM worn during real-time monitoring since last visit	None, <25%, 25-50%, 50-75%, >75%, 100%
Severe hypoglycemia episodes	# episodes of severe hypoglycemia (coma or convulsion) since last visit	Integer
Moderate hypoglycemia episodes	# episodes of moderate hypoglycemia (requiring assistance but not coma or convulsion) since last visit	Integer
DKA episodes	# episodes of DKA since last visit	Integer
Biological samples	Biological samples stored flag	Yes, no
Tanner stage (breast)	Tanner stage (breast) (females only)	1-5
Tanner stage (genitalia)	Tanner stage (genitalia) (males only)	1-5
Tanner stage (pubic hair)	Tanner stage (pubic hair)	1-5
Right testicular volume	Right testicular volume (ml) (males only)	≤30
Left testicular volume	Left testicular volume (ml) (males only)	≤30
Fasting Lipids	Flag to indicate whether lipid results fasting	yes, no
Total cholesterol	Total cholesterol (mmol/L)	1-20
TG	Triglycerides (mmol/L)	1-20
HDL	HDL (mmol/L)	1-20
LDL	LDL (mmol/L)	1-20
ACR	Albumin creatinine ratio	0.1-50.0
AER	Albumin excretion rate (mg/min)	0.1-100.0
Smoking	Smoking status	Current, past, never, unknown
Table: Medication
Field	Description	Expected values
Name	Name of medication other than insulin being taken by patient	Free text
Start date	Start date for other medication	Valid date
End date	End date for other medication	After start date

^aValid comorbidities are the following: hematological—thalassemia; endocrine/nutritional/metabolic—hypothyroidism, hyperthyroidism, thyroid autoimmunity, Wolfram syndrome, Addison disease, PCOS, obesity, dyslipidemia, hypercholesterolemia, cystic fibrosis; mental/behavioral—depression, needle phobia, anxiety, anorexia, bulimia, mental retardation, autism, ADHD, ADD; nervous system—Friedreich’s ataxia, epilepsy, TIA, mononeuritis, peripheral neuropathy, myotonic dystrophy, autonomic neuropathy; eye—cataracts, nonproliferative retinopathy, proliferative retinopathy; circulatory—hypertension, angina, AMI, angioplasty, bypass graft, heart failure, stroke, carotid artery disease, claudication, amputation; respiratory—asthma; digestive—gastroparesis, coeliac disease; skin and subcutaneous—ulceration, infection; musculoskeletal—Charcot joint, osteomyelitis; genitourinary—ESRF, albuminuria, microalbuminuria, erectile dysfunction; congenital/chromosomal—Prader-Willi syndrome, Laurence-Moon-Biedl syndrome, Down syndrome, Turner syndrome, Klinefelter syndrome.