1.6 Data Assumptions / Decisions / Data 'Manipulation'

In analysing the data, as in previous years, the following data assumptions, decisions and data ‘manipulations’ were made using the following ‘rules’:

Missing data were also calculated conditionally where relevant:

• (Pregnancy Yes/No, for female only), (Erectile Dysfunction Yes/No for male only);
• (On Insulin since-Year, only if on Insulin);
• (Fasting Yes/No, only if Lipids recorded);
• (HbA1c/Microalbumin range, only if HbA1c/Microalbumin value recorded);
• (Microalbumin units, only if Microalbumin value recorded);
• (Paediatric/Adolescent fields based on Form Level [ie fields only on that form]). [N/A in 2011]

Clearly invalid data were excluded:

• (Date Of Birth > today, if not corrected after site questioned);
• (On insulin since year, if not on insulin);
• (Pregnancy for males or females age <15 or >50), (Erectile Dysfunction for females).

Data ‘manipulations’ were necessary:

• Visual Acuity was rounded up to the next valid level [6/4 6/5; 6/30 6/36];
• Visual Acuity [a]<=6/60 in both eyes and Blindness = Yes {New &/or Past} Yes was/were changed to No and [b] >6/60 in either eye and Blindness New and Past = No, Past Blindness was changed to Yes [in line with the Blindness Definition] if not corrected after site questioned;
• In instances where Retina reports contained both Diabetes Abnormality and Non Diabetes Abnormality, given the ability to record (and report) one only, the Diabetes Abnormality was recorded and reported;
• Both ‘Year of Diagnosis’ and ‘Insulin Since’ results were deleted if < Date Of Birth;
• The ‘On Anti Lipid Rx’ result was changed from Yes to No, if no further Lipid Rx details were provided;
• If Age >= 20 and Site = Paediatric, then we changed the record to ‘Adult’; [N/A in 2011]
• Changed ‘Diet Only’ from Yes to No if other Management Methods details were provided;
• HbA1c and Microalbuminuria/Proteinuria were only considered in relation to range calculations when range [and units (microalbuminuria/proteinuria)] were stated;
• If Type 1, but insulin was commenced >3 years after diagnosis and age at diagnosis was ≥40 and BMI was ≥ 25, they were changed to Type 2;
• HDL recorded as > 4.0 after site checks, and HbA1c values < 3.5 after site checks were excluded from analyses.
  • previous page
  • next page