1431 Final Protocol

Final Protocol 1431 – HbA1c PoC testing in diabetes

MEDICAL SERVICES ADVISORY COMMITTEE

HbA1c point of care testing for the diagnosis and management of diabetes mellitus

Protocol

December 2016

1. Title of application

HbA1c point of care testing for the diagnosis and management of diabetes mellitus.

2. Purpose of application

Please indicate the rationale for the application and provide one abstract or systematic review that will provide background.

Diabetes is a common chronic disease in Australia, increasing at a faster rate than other chronic diseases such as heart disease and cancer. It is estimated that around 1.7 million Australians have diabetes, with around 280 people developing it every day (1;2). This includes all types of known and registered diabetes, as well as silent, undiagnosed type 2 diabetes which is estimated to affect up to 500,000 Australians (1;2). The total annual cost impact of diabetes in Australia is estimated at $14.6 billion (1).

It is estimated that for every 100 people with the diagnosis of diabetes there are between 25 (3) and 80 (4) people living with undiagnosed diabetes. Many patients do not seek help until they have developed complications which may be many years after diabetes actually began. The high prevalence of undiagnosed diabetes demonstrates the importance of early detection of the disease.

Diabetes is a serious complex condition resulting from defective insulin production and/or action. If left undiagnosed or poorly controlled, the chronic hyperglycaemia associated with diabetes can affect the entire body and lead to coronary artery disease (CAD), stroke, kidney failure, limb amputations and blindness (5).

The diagnosis of diabetes typically involves one or more biochemical analyses: blood glucose levels (fasting and random); the oral glucose tolerance test (OGTT); and glycated haemoglobin (HbA1c) (2). HbA1c, a marker of long-term glycaemic control, reflects the average blood glucose concentration over the preceding 3 month period.

Current guidelines recommend HbA1c tests every 3 to 6 months to assess glycaemic control as a routine component of disease management for patients with established diabetes (2). HbA1c has been endorsed as a diagnostic test for diabetes by the World Health Organization (WHO), and recently the Australian Diabetes Society, the Royal College of Pathologists of Australasia, and the Australasian Association of Clinical Biochemists confirmed its use to establish the diagnosis of diabetes (2). HbA1c testing performed in NATA accredited pathology laboratories is currently MBS listed for both the diagnosis and the management of patients with established diabetes (refer to MBS items 66551, 66554, 66841 for further details). Point of care (PoC) testing is funded by the Federal Government in a limited capacity under the Quality Assurance for the Aboriginal and Torres Strait Islander Medical Services (QAAMS) program (MBS Item 73840).

Early identification and optimal glycaemic control can slow the onset and progression of diabetes-related complications (2). Innovative HbA1c PoC tests have been developed to allow diagnostic testing at or near the site of patient care, providing clinicians and patients an alternative to laboratory testing (6). The addition of HbA1c PoC testing to usual patient management leads to improved patient outcomes (ie improved and quicker clinical decisions), and allows prevention of unnecessary healthcare resource use (ie more rational prescribing, fewer laboratory or doctor visits) (7-14).

During the recent MSAC application for HbA1c for the diagnosis of diabetes (1267) there was strong public support for the consideration of the PoC testing however this was ultimately not included in the final assessment report.

This application seeks MBS listing for the quantification of HbA1c by PoC for the diagnosis and usual management of patients with diabetes.

The proposed MSAC application will be undertaken by Optum on behalf of IVD Australia.

Please refer to Appendix for background information from the Australian Government Point of Care Trial in General Practice (2009).

3. Population and medical condition eligible for the proposed medical services

Provide a description of the medical condition (or disease) relevant to the service.

Diabetes mellitus is a group of metabolic disorders characterised by hyperglycaemia resulting from defects in insulin secretion and/or action, and is categorised into three main types: type 1, type 2 and gestational. Several pathogenic processes are involved in its development ranging from autoimmune destruction of the β-cells of the pancreas to abnormalities resulting in insulin resistance. For patients with diabetes, chronic hyperglycaemia is associated with long-term damage, dysfunction and failure of various organs including eyes, kidneys, nerves, heart and blood vessels. The symptoms of hyperglycaemia include polyuria, polydipsia, weight loss and blurred vision. Long-term complications involve loss of vision due to retinopathy, renal failure, peripheral and autonomic neuropathy. The incidence of atherosclerotic cardiovascular, peripheral arterial and cerebrovascular disease, as well as hypertension and lipoprotein abnormalities in patients with diabetes is increased. Early detection and effective therapy providing good metabolic control, can delay the onset and progression of diabetes late complications, resulting in better outcomes for patients.

Define the proposed patient population that would benefit from the use of this service. This could include issues such as patient characteristics and /or specific circumstances that patients would have to satisfy in order to access the service.

The two proposed patient populations are the same as those currently funded for HbA1c testing for the diagnosis and management of diabetes (MBS items 66841, 66551 and 66554) for rural, remote and urban areas.

NHMRC guidelines (5) suggest case detection should be conducted on an opportunistic basis (eg in GP consulting rooms) in individuals who are judged to be at risk, either through a score of ≥12 on the AUSDRISK assessment tool, or are in one of the following population groups with a known higher risk:

• 
  people with impaired glucose tolerance or impaired fasting glucose;
  
• 
  women with a history of gestational diabetes mellitus;
  
• 
  women with a history of polycystic ovary syndrome;
  
• 
  people presenting with a history of cardiovascular disease event (ie myocardial infarction, stroke); and
  
• 
  people on antipsychotic medication.
  

The proposed HbA1c PoC test will be used for the detection and diagnosis in this same at risk population.

The HbA1c POC test will also be used for ongoing management of patients with established diabetes on an as needed basis every 3-6 months to assess blood glucose control.

The HbA1c PoC test will be an alternative to the HbA1c test currently performed by a laboratory.

PASC noted that the potential benefits of the HbA1c POC test versus laboratory testing may be greater in rural, remote, disadvantaged and aged populations. Therefore these groups are listed as sub-populations of interest in the protocol. These populations will be assessed if HbA1c POC test is not cost-effective in the general population.

Indicate if there is evidence for the population who would benefit from this service i.e. international evidence including inclusion / exclusion criteria. If appropriate provide a table summarising the population considered in the evidence.

Two clinical trials were identified that recruited patients with similar criteria described above. These prospective, randomised controlled trials (RCTs) assessed the clinical effectiveness of HbA1c PoC testing in patients with established diabetes. Additionally, an RCT funded by the Australian Government investigated PoC testing in general practice. It assessed patients across three conditions (anticoagulant therapy, diabetes, hyperlipidaemia) and reported the results separately. Results from the diabetes subgroup are presented in this protocol. The eligibility criteria used to select and randomise patients in the RCTs are presented in Table .

Table Patient inclusion and exclusion criteria

Author and year	N	Inclusion criteria	Exclusion criteria
Khunti 2006	681	Patients with type 2 diabetes who attended participating general practices for review of their diabetes care.	Patients who were unable to attend the practice. Patients who were exclusively under hospital care.
Miller 2003	597	Patients with type 2 diabetes for at least 6 months.
Australian Government 2009	1967	Patients with established diabetes. Patients with a fasting plasma glucose of ≥ 7.0 mmol/L or a two-hour post glucose load of ≥ 11.1 mmol/L. (This equates to 53 mmol/mol.)	Patients who were < 18 years of age. Patients whose condition was not stabilised. Patients who had dementia. Patients deemed by their GP as unable to comply with the requirements of the Trial. Patients who were unable to understand the instructions written in English.

In both trials by Khunti et al (2006) and Miller et al (2003) trials patients were randomised to receive HbA1c results either during (“rapid”) or after (“routine”) the patient visit. Khunti et al (2006) (14) found that the proportion of patients with HbA1c < 7.0% was not statistically different between either group at 12 months. Miller et al (2003) (7) reported more intensification of therapy in patients with HbA1c ≥ 7.0% at baseline occurred in the PoC testing group (51 versus 32, p=0.01), particularly when HbA1c ≥ 8.0%. Additionally, a statistically significant reduction was reported in HbA1c in the PoC testing group (8.4% to 8.1%, p=0.04), whilst no reduction was reported in the routine care group (8.1% to 8.0%, p=0.31) (7). Overall no statistically significant changes in HbA1c between groups were reported, however, the trial was not powered to determine a statistically significant difference in routine care provided over the 4 month trial duration (7).

The multi-centre, cluster RCT evaluated the safety and clinical effectiveness of HbA1c PoC testing in Australian general practice (15). The main clinical outcomes of the PoC test trial findings in relation to HbA1c were:

• 
  The median number of HbA1c tests performed was the same across treatment groups and the median HbA1c test result was similar between groups, though slightly lower in the PoC testing group (15).
  
• 
  PoC testing was non-inferior compared to pathology laboratory testing in relation to the proportion of diabetes patients who have shown an improvement from baseline and are within the target range (p<0.0001). Based on the adjusted analysis, the percentage of patients within target range was higher in the PoC testing group (65.48%) compared to the pathology laboratory testing group (56.18%) with a difference of 9.31%. The results of the unadjusted analysis confirmed these findings (15).
  
• 
  PoC testing was non-inferior compared to pathology laboratory testing (p<0.0001). Based on the adjusted analysis, the percentage of patients with a reduction in their HbA1c test from baseline was higher in the PoC testing group (57.33%) compared with the pathology laboratory group (44.91%) with a difference of 12.42% (15).
  
• 
  PoC testing was non-inferior compared to pathology laboratory testing in relation to the proportion of HbA1c tests within the target range (p<0.0001). Based on the adjusted analysis, the percentage of HbA1c tests within the target range was higher in the PoC testing group (64.11%) compared with the pathology laboratory group (54.74%), with a difference of 9.36%. The results of the unadjusted analysis confirm these findings (15).
  

This trial conducted an analysis of the comparison of PoC testing and pathology laboratory test results for HbA1c and found the mean difference in results and the 95% limits of agreement were clinically acceptable. In addition, no serious adverse events (AEs) were attributable to PoC testing (15).

Additional data from non-randomised trials were located. Kennedy et al (2006) investigated the impact of active versus usual monitoring of algorithmic insulin titration and PoC versus laboratory HbA1c measurement on glycaemic control (11). 7,893 adults with type 2 diabetes were enrolled and assigned to treatment with insulin, with either:

• 
  Usual titration (no contact between visits) using a simple algorithm with laboratory HbA1c testing;
  
• 
  Usual titration with HbA1c PoC testing;
  
• 
  Active (weekly monitoring) titration with laboratory HbA1c testing; or
  
• 
  Active titration with HbA1c PoC testing.
  

The HbA1c PoC testing arm was associated with an increase in the proportion of patients receiving active insulin titration achieving HbA1c<7.0% (41% for PoC testing vs 36% for laboratory testing, p<0.0001) (11).

Cagliero et al (1999) tested the hypothesis that immediately available HbA1c results could improve glycaemic control by changing physician or patient behaviour, or both (9). In the PoC testing group, HbA1c decreased significantly at 6 and 12 months (-0.57±1.44,and -0.40 ±1.65%, respectively), whilst HbA1c levels did not change in the laboratory testing arm (-0.11±0.79, and -0.19±1.16%, respectively, NS) (9). The difference between the two groups was statistically significant at 6 months (p=0.029) but not at study end (p=0.346).

Provide details on the expected utilisation, if the service is to be publicly funded.

Table below shows the number of services for HbA1c MBS item numbers between March 2015 and February 2016. HbA1c pathology for diagnosis testing was made available on the MBS in November 2014. The most recent 12 months of data between March 2015 and February 2016 shows approximately 1.4 million HbA1c tests were provided. Of these the majority of tests have been in patients with established diabetes (1.1 million services).

For the diagnosis of diabetes, it is expected that POC testing would be additional testing in situations where current laboratory testing is not practical. This is based on the experience of the introduction of laboratory HbA1c testing for the diagnosis of diabetes which has results in additional testing rather than the replacement of other diagnostic tests such as OGTT. To verify this assumption the number of services provided from the OGTT were compared with the HbA1c test (MBS items 66542 and 66841). This information is shown in Table .

For the management of diabetes, it is estimated that PoC testing would be a combination of new testing (5%) where current laboratory testing is not practical, and replacement of laboratory testing (5%).

Table Expected utilisation

MBS item description	MBS item number	Total number of services (between March 2015 and February 2016)	Uptake of HbA1c PoC testing in the first year
OGTT for the diagnosis of diabetes mellitus	66542	263,872	26,387
Quantitation of HbA1c (glycated haemoglobin) performed for the diagnosis of diabetes in asymptomatic patients at high risk. (Item is subject to rule 25)	66841	230,209	23,021
Quantitation of glycated haemoglobin performed in the management of established diabetes – (Item is subject to rule 25)	66551	1, 161,359	58,068 – New testing 58,068 – Replacement testing
Quantitation of glycated haemoglobin performed in the management of pre-existing diabetes where the patient is pregnant – including a service in item 66551 (if performed) – (Item is subject to rule 25)	66554	11,353	590 – New testing 590 – Replacement testing
Total HbA1c services		1,402,921	140,337

Ultimately, the level of uptake is uncertain. Uptake is expected to grow in with line with number of patients at risk and diagnosed with diabetes. It is anticipated that uptake will be greatest in rural and remote areas where patients are required to travel long distances for pathology services. The level of uptake would likely be affected by the implementation of the service when it is funded.

As noted by PASC, due to ‘episode coning’ which limits the payment of pathology services within a patient episode, it is difficult to determine the total number of HbA1c tests that were performed based on MBS data. HbA1c tests that are not paid for by Medicare due to episode coning are not captured in MBS data. The Sponsor notes this coning effect and will attempt to account for this in the economic and budget impact models within the SBA based on available data.

4.