Using HbA1c to simplify diabetes screening in remote Aboriginal communities

Type 2 diabetes and its associated complications are a significant health problem facing Aboriginal and Torres Strait Islander peoples in Australia. While early diagnosis and treatment of diabetes leads to improved outcomes this condition remains undiagnosed in up to 50% of people. Current screening methods rely on fasting glucose measurements, and are failing to detect people with diabetes in northern Australia.

In December 2009 the American Diabetes Association updated their criteria for diagnosing diabetes to include glycated haemoglobin A (HbA1c) testing. We wanted to see if a combination of point-of-care (POC) and laboratory (lab) HbA1c testing, in real world settings, is an effective method of screening for diabetes in Kimberley Aboriginal people compared to the existing glucose screening model.

How was this study done?

  • From September 2011 to November 2013, 255 Aboriginal and Torres Strait Islander patients over 17 years old without confirmed diabetes and due for screening were enrolled in the study. This was done by clinic staff on an opportunistic basis depending on clinic resources.
  • We compared POC capillary HbA1c with venous HbA1c levels measured in a reference lab (n=241). A train-the-trainer model was used to train staff in how to operate POC analysers.
  • At the same time clinic staff screened patients for diabetes using:
    1. existing glucose screening model; and
    2. HbA1c screening model
  • Participants were classified independently using both diabetes screening models for comparison.

What did we find?

POC HbA1c testing is sufficiently accurate to be used in the HbA1c screening model:

  • Concordance (agreement) between POC and lab results was good (ρ=0.88).
  • The POC value for screening for diabetes or a high risk of developing diabetes was ≥ 5.7% (sensitivity, 91%; specificity, 76.7% for lab measurements ≥ 6.0%).
  • Staff trained by other clinic staff ‘on the job’ performed as well as people with formal accredited training.
  • Staff reported difficulty in maintaining formal accreditation.

The HbA1c model was better for screening and diagnosing diabetes than the glucose model:

  • Participants were more likely to have a definitive result using the HbA1c than the glucose model (250 v 214; p < 0.001).
  • HbA1c results were also much more timely with only 6 (2.4%) participants not having a result within seven days compared to 56 (22.0%) participants using the glucose model (p < 0.001).
  • Participants were more likely to be diagnosed with diabetes using HbA1c than the glucose model (15 v 4; p = 0.003):
    • For everyone with diabetes that we found using the glucose model we missed two: 8 participants who were classified as normal with the glucose model had diabetes.
    • Normal POC HbA1c measurements (<5.7%) did not miss anyone with diabetes.

What happens now?

  • In late 2014 an MBS rebate for diagnosing diabetes using a single lab HbA1c test was introduced in Australia – this is different to the rebate for HbA1c tests for managing diabetes.
  • As a result of this we have dropped the requirement for an additional lab HbA1c test to confirm diabetes and have included the simplified Kimberley HbA1c diabetes screening and detection model in the recently revised Kimberley Chronic Disease Protocol for Type II Diabetes.
  • Adoption of the Kimberley HbA1c diabetes screening and detection model should simplify the testing process in those previously undiagnosed and provide a timelier and more accurate diagnosis of diabetes for Aboriginal people and other high-risk remote populations.

It is important to note that these result are not applicable for screening for gestational diabetes later in pregnancy. The use of HbA1c is limited by physiological changes that result in generally lower HbA1c levels in third trimester. We have started a new project (the ORCHID Study – Optimisation of Rural Clinical and Haematological Indicators of Diabetes) to determine if we can use HbA1c (or other glycation products) to screen for gestational diabetes, and reduce the number of women needing to take an oral glucose tolerance test.

Barriers and enablers to implementing HbA1c screening:

We want to see how well the new Kimberley HbA1c diabetes screening and detection model is implemented. Some of issues that may need to be addressed include:

  • FUNDING: the MBS rebate for is for lab not POC HbA1c – we need to show that using POC testing as an initial screening measure is more cost effective than only using a single lab test.
    • Awareness of the new screening model: lab HbA1c 6.5-6.9% is not normal.
    • Developing resources for clinicians to explain the new tests to patients.
  • TRAINING: adding POC HbA1c analyser training to existing regional training programs.
  • UPDATING PROTOCOLS: adding POC HbA1c to routine tests for people due for screening.
  • HARDWARE: ensuring clinics have POC HbA1c analysers.
  • MONITORING: developing a regional quality assurance program.

Many thanks to patients, staff & Council of participating services.  Without your help this research would not have been possible.


Julia V Marley, May S Oh, Nyssa T Hadgraft, Sally Singleton, K Isaacs, David Atkinson. Using glycated haemoglobin testing to simplify diabetes screening in remote Aboriginal Australian health care settings. Med J Aust 2015; 203:28-33.

Julia V Marley, May S Oh, Nyssa T Hadgraft, Sally Singleton,K Issacs, David Atkinson. Cross-sectional comparison of point-of-care with laboratory HbA1c in detecting diabetes in real-world remote Aboriginal settings. BMJ Open 2015; 5:e006277.


Download the plain language report for staff (111KB .pdf)

Download the plain language report for community members (82KB .pdf)