Comparator – clinical claim for the proposed medical service
Please provide details of how the proposed service is expected to be used, for example is it to replace or substitute a current practice; in addition to, or to augment current practice.
The proposed medical service will be a new MBS item for a PoC test conducted at or near the site of patient care (within consulting rooms and excluding the laboratory setting) using a PoC instrument that meets defined performance criteria for the quantification of HbA1c in human whole blood.
It is proposed that the appropriate comparator for HbA1c PoC testing will be HbA1c testing performed in a laboratory (MBS items 66841, 66551 and 66554). Currently, HbA1c tests are predominantly performed in NATA accredited laboratories. Point of care testing has been funded in a limited capacity under the Quality Assurance for the Aboriginal and Torres Strait Islander Medical Services (QAAMS).
The current clinical pathways for diagnosis and management of diabetes are based on recommendations in the NHMRC case detection guidelines (5) and general practice guidelines (2).
The three new proposed MBS items will be similar to the currently listed diagnostic tests for HbA1c currently used to diagnose diabetes in asymptomatic patients at high risk (see MBS item 66841) and to monitor the effectiveness of diabetes treatment and long-term blood glucose in people with established diabetes (see MBS items 66551 and 66554 in pregnant patients).
For the proposed medical service, practitioners would perform the test under the same circumstances as they would order the existing HbA1c tests.
As with current HbA1c testing arrangements the medical practitioner may request alternative tests such as fasting blood glucose (FBG) or OGTT if they are clinically indicated.
PASC agreed with the comparator being HbA1c test as analysed in an accredited pathology laboratory. PASC also agreed that provision of HbA1c PoC testing via QAAMS is not an appropriate comparator.
PASC agreed that the reference standard is the HbA1c test analysed in an accredited pathology laboratory at the diagnostic cut-off of 48 mmol/mol (6.5%).
Expected health outcomes relating to the medical service
Identify the expected patient-relevant health outcomes if the service is recommended for public funding, including primary effectiveness (improvement in function, relief of pain) and secondary effectiveness (length of hospital stays, time to return to daily activities).
Improvements in therapeutic control, even slight, can prevent or delay the onset of diabetes-related complications, including kidney failure, heart disease and diabetic retinopathy, lowering associated healthcare costs and resource use.
Immediate availability of HbA1c results at the time of consultation can increase the frequency of intensification of therapy (7), lowers HbA1c levels in type 2 diabetes (8;14), positively impact on medication adherence (10) and assist in the better management of chronic conditions (7-14). The avoidance of diabetes-related diseases and complications leads to both an increase in quality and quantity of life for patients, and a consequent reduction in the healthcare burden associated with diabetes-related death, disability and decline in quality of life when compared to no intervention.
The RCT performed in Australian general practice (15) reported the following patient-relevant health outcomes:
GP visits: The number of GP visits per person-year in the diabetes sub-analyses were similar between both treatment groups. The total number of tests per person-year for HbA1c in Phase ll was slightly higher for the HbA1c PoC test group (1.8) compared to the pathology laboratory group (1.5) (15).
Process of care actions: Across both treatment groups, a total of 3014 (55.03%) HbA1c tests were within target range and 2463 (44.96%) tests were outside the target range (15).
For HbA1c test results within the target range:
the most common process of care actions for patients with diabetes, which were not part of the diabetes annual cycle of care, were review of the results (94.01%), a GP consultation (91.10%) and a review of all medications (59.32%).
the least common action was change or cessation of medication (8.31%).
when viewed by treatment group, the proportion of actions in all areas was greater in the HbA1c PoC test group than in the pathology laboratory group. More tests in the HbA1c PoC test group were associated with a GP consultation (94.48%) compared with the pathology laboratory group (83.39%) as well as review of all medications (61.22%) and other actions (45.62%).
For HbA1c tests outside the target range:
the most common actions were similar to those reported for the tests within range being review of test results (93.61%), GP consultation (85.06%) and review of all medications (61.18%)
when comparing treatment groups the HbA1c PoC test group had a larger proportion of actions in terms of GP consultations (90.83% versus 78.56%), review of all medications (64.18% versus 57.81%) and other actions (42.49% versus 27.94%). However, the pathology laboratory group reported a slightly higher proportion of changes in medication (19.73% versus 17.01%).
In terms of impact on processes of care:
More GPs changed or ceased diabetes medication if the test was outside the target range (18.3%) than within range (8.3%).
More GP consultations were noted (90.1%) within range than outside the range (85.1%). This was an interesting finding, as it would be expected that those patients outside the range would require additional GP follow-up visits.
Larger number of follow-up HbA1c test for those outside the range than within range (20.2% v 15.0%). Interestingly, GPs in the pathology laboratory group were more likely to order an HbA1c follow-up test than the HbA1c PoC test group GPs despite having access to a PoC testing device.
Medication compliance: Based on all Trial participants, using the adjusted analysis, the percentage of MARS-5 questionnaire responses indicating compliance with disease management (use of medicines), was higher in the HbA1c PoC test group (39.27%) compared to the pathology laboratory group (36.98%), with a difference of 2.29% and a 90% confidence interval for the difference of (-0.06, 4.64) (10;15). The lower limit of the 90% confidence interval is greater than the non-inferiority limit of -3.70%, indicating PoC testing is non-inferior to pathology laboratory testing in relation to the proportion of MARS-5 questionnaire responses indicating compliance with disease management (p<0.0001) (10;15). The results of the unadjusted analysis confirm these findings. Over 50% of diabetes patients reported forgetting to take their medicines with approximately 10-15% of both treatment groups altering the dose, missing out a dose or stopping taking them for a while (10;15).
Patient satisfaction: No subgroup analysis for diabetes patients was provided. However, the Trial findings overall found strong statistical evidence that PoC testing patients on average were more satisfied than patients in the pathology laboratory group in regards to the PoC testing process, the convenience of not travelling to an outside laboratory and to disease management (12;15).
PASC advised that the proposed outcomes should include the following:
Analytical performance (% coefficient of variation, trueness, imprecision);
Clinical performance (diagnostic accuracy, sensitivity, specificity in proposed setting (GP));
Clinical effectiveness (safety and effectiveness);
Health resources (number of HbA1c tests, total tests per year, number of GP consultations, and other healthcare resource use).
A 2014 study found that four HbA1c PoC test systems met generally accepted performance criteria (16). This is supported by other studies reporting that PoC testing met the desired goals for imprecision (9-12), although the study of Schwartz et al (17) showed the mean HbA1c result of the PoC testing device to be significantly higher compared to the laboratory.
The RCT performed in Australian general practice (15) conducted in 2006 reported weighted estimates of the number of serious adverse events (SAEs), the number and percentage of patients experiencing one or more SAE and the number of SAEs per 10,000 person-years were calculated both overall and by treatment group. Descriptive analysis of Trial incidents was also undertaken. Issues of under-reporting were noted where reports were made direct by Trial staff or patients/carers using the appropriate SAE form.
Of patients with diabetes, there were 372 SAEs, with the most common being inpatient hospitalisations (64.54%), followed by other important medical events (17.55%) (15). A higher percentage of SAEs within the pathology laboratory group were inpatient hospitalisations (80.59%) compared with the HbA1c PoC test group (53.765) (15). The HbA1c PoC test group had a higher percentage of events classified as ‘other important medical events, compared to the pathology laboratory group (25.19% vs 6.18%) (15). Table presents the type of SAE for the diabetes treatment group (weighted estimates).
Table Type of SAE for the diabetes treatment group (weighted estimates)
Type of SAE
HbA1c PoC test
Inpatient hospitalisation or prolongation of existing hospitalisation
Reference: Australian Government PoCT report (2009)
Specify the type of economic evaluation.
A cost-effectiveness/ cost-utility analysis will be conducted to extrapolate costs associated with patients who receive the proposed medical service (HbA1c PoC testing) and those who receive the comparator (HbA1c testing via pathology lab), and associated gains in patient outcomes ie responder (controlled diabetes) and QALYs. This analysis is based on the following clinical claim: HbA1c PoC testing is non-inferior (same or better) in terms of clinical effectiveness and clinical utility for the diagnosis and management of patients with diabetes when compared with pathology laboratory testing. In areas with limited access to pathology laboratory services HbA1c testing is likely to be superior. HbA1c PoC testing is non-inferior and comparable in terms of diagnostic performance, clinical validity and analytical performance compared to pathology laboratory testing.