Is Australia ready to use glycated haemoglobin for the diagnosis of diabetes?

To the Editor: I would like to add some detail to the article by Shaw and colleagues1 on the costs of screening for diabetes and glycated haemoglobin (HbA1c) testing. At face value, using the 85% Medicare Benefits Schedule (MBS) rebate, the item for HbA1c testing costs $1.85 less than the item for a glucose tolerance test (more than 10% cheaper), but it is a little more complicated than these simple figures suggest. In 2010, 295 023 glucose tolerance tests (item number 66542) were claimed on the MBS — up 61% on the number ordered in 2004 (183 090)2 — which reflects the increase in ordering by general practitioners, who I believe are more aware of the increased incidence and prevalence of one of the most common chronic diseases in Australia. I also believe that it reflects the use of the glucose tolerance test as the definitive test for diabetes (rather than relying on a single fasting glucose level) in private practice.

In contrast, 1 021 247 HbA1c tests (item number 66551) were claimed in 2010, versus 911 623 in 2004 — up by only 12% over the same 7-year period.2 This reflects “coning” of pathology items in the MBS. There are two types of cone that affect billing of HbA1c tests: the “grand cone”, which restricts billing to the three most expensive items ordered by a GP on a single occasion (regardless of the number of tests ordered), and the “temporal cone”, where only four HbA1c tests can be billed in any 1 year. The glucose tolerance test has no temporal restrictions and is usually performed on its own, so it avoids the grand cone, but the HbA1c test is often ordered with a bank of other tests (eg, as part of diabetes monitoring) and is thus not usually billed to Medicare. In my practice, only 30% of reported HbA1c tests can be billed to Medicare, hence the cost to Medicare per reportable test is about a third of the listed rebate of $14.40. Furthermore, as Shaw et al point out, HbA1c testing cannot currently be billed for the diagnosis of diabetes, although my personal observation is that many doctors are already using this as a screening test.

There are essentially three powerful drivers for HbA1c testing: the increased prevalence of diabetes, the (honest) push to test HbA1c levels every 4 months (through care plans etc) and the use of HbA1c tests to diagnose diabetes. I believe that Medicare currently pays for less than 40% of these tests and that this proportion will fall as more HbA1c tests are requested.

The majority of the costs for HbA1c testing are subsidised by pathology practices — which, philosophically, I find quite odd. These points need to be taken into account when undertaking a cost–benefit analysis of screening for diabetes in Australia.