The NADC Board was approached in late 2009 and a sub-committee was nominated and formed to meet by Teleconference and formulate the data set, ie the data items to be collected. Members were Jeff Flack [Chair], Gil Cremer, Trisha Dunning, Gloria Kilmartin and Pamela Jones.
Any data items selected would use current, agreed, preferably national definitions [if in existence], or original NDOW [National Diabetes Outcomes Workshop], (now METeOR5) Dataset definitions. Any data item collected would need to be able to provide relevant information for the Centre about its patients with a view to instituting change for identified ‘problems’. These data items and their definitions would be widely discussed and consensus reached where necessary before implementation. The ANDIAB2 2005 dataset was placed on the ADEA Website and feedback sought, and incorporated where possible. This was not done for ANDIAB2 2010.
A separate version of the ANDIAB Database was created and converted to accept the data items to be collected, and Teleform Software was utilised to create a scannable form for use in data collection [See Appendices > Data Collection Form]. Definitions for each data field were agreed and provided to participants [See Appendices > ANDIAB2 Data Definitions].
The NADC Secretariat coordinated the project, which was conducted in a ‘double blind’ fashion. Diabetes Centres were invited to complete the one-page data collection form during the months of April, May (or June) 2010. Centres who wished to participate, responded to a formal invitation distributed by the NADC Secretariat to all Diabetes Centre members of the NADC across Australia. All subsequent contact and correspondence with participants was conducted through the Secretariat.
The Secretariat allocated a unique code to each Diabetes Centre using the predetermined ANDIAB proforma, and holds the only copy of the code. Centres who had participated in ANDIAB previously used the code already allocated to them. The Central Data Analysis Site generated .pdf electronic ‘Master Copies’ of forms uniquely numbered for each Centre, and sent them to the NADC Secretariat where a copy was forwarded to each participating Centre with printing instructions. Using de-identified data via this methodology, it is not possible for the Benchmarking Centre – Central Analysis Site to identify individual Centres or patients.
Every effort was made to ensure data completeness and correctness. As in ANDIAB, specific ‘Validation Reports’ were generated for each Centre requesting missing data and correction of questionable or potentially invalid data [See Appendices > ANDIAB2 Data Definitions, for data definitions & validation rules]. Specific items that were altered if not done so by sites were:
- ‘Management Method’ not Insulin, but ‘Insulin Since’ not null - (Removed Insulin Since);
- Not ‘Current Smoker’ but ‘Tried to Stop Smoking’ Yes - (Tried to Stop Smoking made Null);
- ‘Know who to contact’ not Yes - ‘Have Telephone Number’ Yes (Telephone No. made Null).
- Process assessment:
- Missing Data [presented in both tabular and graphical formats];
- Outcome measurements such as:
- Frequency Counts Data, Mean Descriptive Results and Outcome Data [where relevant].
- A 2005 and 2010 Demographic Data Report;
- A 2005 and 2010 Missing Data Analysis Report;
- A 2005 and 2010 Frequency Counts Report.
The pooled data analysis in this Final Report provides a ‘snapshot’ of the status of patient
practices and Diabetes Centre services – as detailed in Project Milestones