Design, implementation and results of the quality control program for ...

Original Article Design, implementation and results of the quality control program for the Australian government’s point of care testing in general practice trial Mark Shephard1, Anne Shephard1, Les Watkinson1, Beryl Mazzachi1 and Paul Worley2 1 Community Point-of-Care Services, Flinders University Rural Clinical School; 2School of Medicine, Flinders University, GPO Box 2100, Adelaide, South Australia, Australia Corresponding author: Dr Mark Shephard, Director and Senior Research Fellow, Community Point-of-Care Services, Flinders University Rural Clinical School, Flinders University, GPO Box 2100, Adelaide, South Australia, Australia. Email: [email protected]

Abstract Background: From 2005 to 2007 the Australian Government funded a multicentre, clustered randomized controlled trial to determine the clinical effectiveness, cost-effectiveness, satisfaction and safety of point of care testing (PoCT) in general practice (GP). PoC tests measured (and devices used) in the trial were haemoglobin A1c and urine albumin:creatinine ratio (DCA 2000), lipids (Cholestech LDX) and international normalized ratio (CoaguChek S). Methods: An internal quality control (QC) program was developed as part of a quality management framework for the trial. PoCT device operators were provided with a colour-coded QC Result Sheet and QC Action Sheet for on-site recording and interpreting of their results. Within-practice imprecision for QC testing was calculated and compared with the analytical goals for imprecision set prior to the trial. Results: The average participation rate for QC testing was 91% or greater. Median within-practice imprecision met the analytical goals for all PoC tests, except for high-density lipoprotein-cholesterol (HDL-C) where observed performance was outside the minimum goal for one level and one lot number of QC. Most practices achieved the imprecision goals for all analytes, with the principal exception of HDL-C. Conclusions: Results from QC testing indicate that PoCT in the GP trial met the analytical goals set for the trial, with the exception of HDL-C. Ann Clin Biochem 2009; 46: 413– 419. DOI: 10.1258/acb.2009.009045

Introduction Point of care testing (PoCT) is being increasingly applied in primary care settings across the world. In Australia, there is now good evidence for the effectiveness of PoCT in the indigenous setting;1 – 4 however, a review of the role and value of PoCT in general practice (GP) in Australia found that further work was needed to verify the clinical and economic benefits of PoCT in this setting.5 As a result, the Australian Government funded a multicentre, clustered randomized controlled trial to determine the safety, clinical effectiveness, cost-effectiveness and satisfaction of PoCT in GP.6 In Australia, GP refers to the provision of primary continuing comprehensive whole-patient medical care to individuals, families and their communities; GP represents the first point of contact for the majority of people seeking health care, and often therefore the first point of referral. Fifty-three practices from urban, rural and remote locations across three states of Australia (New South Wales, Victoria

and South Australia) participated in the PoCT in GP Trial, which ran for 18 months across 2005 – 2007. Twenty-three practices were recruited to the control group and 30 to the intervention group. In Phase I of the trial, which lasted six months, patients in practices in the intervention group had pathology testing performed both by their local pathology laboratory and by PoCT in the practice. Patients in the control group had pathology testing performed by their local laboratory only. In Phase II of the trial, which lasted 12 months, patients in the intervention group were tested using PoCT in the practice only, while control patients continued to be tested by their local laboratory. Patients enrolled in the study (1958 control and 3010 intervention) must have had a pre-existing diagnosis of diabetes, hyperlipidaemia or been on warfarin therapy.6 Pathology tests measured as part of the trial included haemoglobin A1c (HbA1c), urine albumin:creatinine ratio (ACR) and international normalized ratio (INR). The PoCT devices used in the trial were the DCA 2000 (Siemens Annals of Clinical Biochemistry 2009; 46: 413– 419

414

Annals of Clinical Biochemistry

Volume 46

September 2009

................................................................................................................................................

HealthCare Diagnostics [formerly Bayer Australia], Melbourne, Australia) for measuring HbA1c and urine ACR, the Cholestech LDX analyser (Point of Care Diagnostics, Sydney, Australia) for lipids and the CoaguChek S (Roche Diagnostics Australia Pty Ltd, Sydney, Australia) for INR. Three lead organizations were contracted by the Australian Government to deliver the trial, working collaboratively from an Adelaide base. They were the University of Adelaide (responsible for overall trial management and evaluation), the Community Point-of-Care Services (CPS) unit at Flinders University and the RCPA Quality Assurance Programs Pty Ltd. A key aspect of the trial was to develop and implement a quality management system for GP that (i) ensured PoCT device operators had the competency skill set required to conduct PoCT safely and effectively; (ii) continually assessed the analytical quality of the PoCT devices used in the trial; and (iii) determined whether the observed performance met the analytical goals established for this trial.7 PoCT device management, training and competency assessment and an internal quality control (QC) program were delivered by a PoCT Device Group comprising scientists from the CPS unit from Flinders University, working collaboratively with industry partners from Bayer Australia, Point of Care Diagnostics Australia and Roche Diagnostics Australia. The RCPA Quality Assurance Program was responsible for the delivery of an external quality assurance (EQA) program. This paper comments on aspects of device management and specifically describes the design, implementation and results of the QC program for the trial.

Methods Device management Each GP participating in the intervention (PoCT) arm of the trial was provided with a DCA 2000, Cholestech LDX and CoaguChek S device by the PoCT Device Group. A Device Assets Register, which contained a list of all PoCT devices, their make, model, serial number, practice ID number, name and address of practice, acquisition date, general comments on device history (repair/service/movement), retrieval date (at the end of the trial) and condition at retrieval, was maintained by the PoCT Device Group during the trial. Bayer Australia’s Therapeutic Drugs Administration (TGA)-licensed Central Warehouse acted as the single distribution point for all PoCT reagents, QC materials and consumables from all three industry partners required for the trial, through a collaborative arrangement brokered between the partners. This Central Warehouse was officially licensed by the TGA, Australia’s peak regulatory agency for medical drugs and devices. Training for PoCT device operators An education and training resource package was developed by the PoCT Device Group for the trial; this package comprised a training manual, a set of posters, which provided a step-by-step guide on how to conduct patient, QC and

quality assurance testing on each device, and a CD-ROM. Initial training workshops were held in each geographic region to train practice nursing staff as device operators. At the end of training, nurses underwent a written and practical competency assessment and were awarded a Certificate of Competency. Refresher training workshops were held for all active device operators 12 months into the trial. A competency register was maintained during the life of the trial. A total of 80 device operators were trained. Full details on the training program will be reported elsewhere. QC materials and testing regimen Once the live phase of the trial commenced, device operators were required to test one set of QC materials (comprising two levels of analyte for HbA1c, urine ACR and lipids and one level for INR) for each test every fortnight for the first three months of the trial (September– November 2005) and monthly thereafter (from December 2005 to February 2007). Three different lot numbers of QC material were used for lipid testing, two for HbA1c and INR testing, and a single lot number for urine ACR testing. The participation rate for QC testing (defined as the total number of QC results returned by practices for each test divided by the total number of QC results expected to be returned according to the required protocol, expressed as a percentage) was monitored monthly throughout the trial. On-site recording and interpretation of QC results by practices Device operators were provided with a colour-coded QC Result Sheet on which to record their results. The Result Sheet documented:

† The target value for each sample (as assigned by the manufacturer);

† Limits for acceptable performance (as determined by the PoCT Device Group) (coloured green);

† Limits where analytical performance required close monitoring (warning limits [orange]) and;

† Limits where analytical performance was considered unacceptable and required the cessation of patient testing until the reason for unacceptable performance had been investigated and corrected (action limits [red]). PoCT operators were also required to record on their QC Result Sheet the date of testing, their initials and their decision as to whether to accept or reject their QC results (using their QC Action Sheet described below). The following QC limits for acceptable performance were used in the trial and were set by the PoCT Device Group: HbA1c + 10%, urine albumin + 12.5%, urine creatinine + 7.5%, urine ACR + 15%, total cholesterol + 10%, triglyceride + 15%, high-density lipoprotein-cholesterol (HDL-C) + 15%, INR + 15%. QC limits were designed to provide a crude guide to overall analytical performance at the practice level and allowed for potential discrepancies between the manufacturer-assigned target values and the mean of participant results. The QC limits equated with between-practice

415

Shephard et al. Quality control results from PoCT Trial

................................................................................................................................................

Quality control action sheet General oractice tests 2 QCs

Both QCs acceptable

No

1 GREEN 1 ORANGE

Continue to test patients until further notice

2 ORANGE

2 ORANGE = RED Do NOT test patients repeat both ORANGE phone help desk

Yes

Continue to test patients 1 GREEN 1 RED

GO

1 ORANGE 1 RED

Processed with caution

STOP

Do NOT test patients repeat all ORANGE and REDS phone help desk

2 RED Help desk support hotline 1800 085 474

Figure 1

Quality control action sheet

goals for total allowable error; that is, they encompassed allowable error due to both the inaccuracy and imprecision of test results across sites (in this case GPs). A laminated QC Action Sheet (Figure 1) was provided for each practice to enable the device operator to interpret the QC results on site and make an informed decision as to whether to:

† Accept the results of QC testing and proceed with patient †

testing (where QC results were green/green or green/ orange); or Reject the results of QC testing, stop testing patients and telephone the PoCT Device Group (through the telephone help desk) support service described later in this paper to determine the next course of action (where QC results were orange/orange or any QC result was red).

The colour codes used for deciding whether to accept or reject the results of QC testing were designed to mimic a ‘traffic light’ system (that is, green means ‘go’, orange means ‘proceed with caution’ and red means ‘stop’) and provide non-laboratory trained device operators with a practical and user-friendly method of working with QC data.

Management and analysis of QC testing by the PoCT Device Group Upon completion of testing a set of QC samples, the device operator was required to immediately fax the QC Result Sheet to the PoCT Device Group, who then entered the results into an Excel spreadsheet. This spreadsheet was used to calculate the within-practice imprecision (coefficient of variation [CV%]) for QC testing for each individual practice, following receipt of a minimum of five QC results for each level tested. The within-practice imprecision for each test was compared with the analytical goals for imprecision set prior to the commencement of the trial by an expert Government

Working Party.7,8 Minimal, desirable and/or optimal goals were set for the trial (Table 1), depending on how well (or otherwise) the test can be measured in practice by current laboratory methods.9 – 11 The approach to goal setting used by the expert group was to examine data from the internationally accepted five-tiered hierarchical approach to analytical goal setting promulgated by the Stockholm Conference on strategies to set global quality specifications in laboratory medicine in 1999;12 these approaches included data from clinical studies, biological variation, professional recommendations from expert groups, quality specifications (for imprecision) used in external proficiency programs and published data on state of the art.9 – 10,12 The analytical goals for imprecision set for POCT in the GP trial represented a middle-ground consensus position that was considered by the expert group to be practical and achievable in Australian GP. Regular monthly meetings between scientific representatives of the PoCT Device Group and the RCPA Quality Assurance Programs group were held to discuss and review all QC and EQA results returned by practices for that month. Extracts from meeting minutes were taken and sent to the Chair of the Safety Subcommittee (under the direction of the Trial Management Committee). These QC/EQA Review Meetings enabled poor analytical performance and/or non-compliance with QC and/or EQA testing schedules to be actioned and addressed in a timely fashion during the trial. A QC Feedback Report was sent to each practice at sixmonthly intervals during the trial. The Feedback Report documented the practice’s QC results in both graphic and tabular form, and recorded the practice’s imprecision for QC testing and the median imprecision achieved by all practices to provide a peer-reviewed comparative assessment of their performance.

Table 1 Analytical goals for within-practice imprecision set for the PoCT Trial by an expert Government Working Party Goal for imprecision Analyte HbA1c Urine albumin Urine creatinine Urine ACR Total cholesterol Triglyceride HDL cholesterol INR

Minimum 4%

Desirable

Optimal

3% 10% 6% 12%

5% 6% 10%

3% 7.5% 4%

5%

PoCT, point of care testing; HbAlc, haemoglobin A1c; ACR, albumin:creatinine ratio; HDL, high-density lipoprotein; INR, international normalized ratio An optimal analytical goal provides a measure of the standard of analytical performance that the very best methods for a particular test can achieve. A desirable analytical goal provides a measure of the standard of analytical performance that should be generally achievable by most methods for a particular test. A minimum analytical goal provides a measure of the widest allowable standard of analytical performance that methods for a particular test should be able to achieve Although an imprecision goal for INR was not specifically set for this trial, a coefficient of variation % of 10% was considered to represent acceptable performance for patient care

416


Volume 46

September 2009

................................................................................................................................................

Other quality management practices In addition to the lodging of QC Result Sheets, each practice was required to complete and return the following sheets to the PoCT Device Group:

† Test Error Log Sheet (to record specific error codes displayed on PoCT devices when such an event occurred);

† Optical Test Result Sheet (as part of general maintenance,

†

an optical test was required to be performed monthly on the Bayer DCA 2000 device and weekly on the Cholestech LDX device); Device Maintenance Schedule (detailing a set time frame for basic maintenance procedures performed on all devices).

The telephone helpdesk hotline was manned by a member of the PoCT Device Group from 8 am to 8 pm Monday to Friday and from 8 am to 1 pm Saturday mornings. A telephone help desk register logged all telephone communications conducted between practices and the PoCT Device Group. Incoming calls were categorized according to whether they related to a device, reagent, QC, consumable or training issue. A record of the summary of the problem, action taken to address the problem and follow-up action required to resolve the problem was also made. For situations where the PoCT Device Group could not readily resolve a help desk enquiry or where a common problem with a specific device occurred across several practices in a short time frame, the relevant industry partner was notified immediately and the problem was addressed and resolved with their support.

Results Participation rate The total number of QC tests expected to be performed over the period September 2005 to February 2007 was 1102 for HbA1c, 3303 for urine ACR (1101 each for urine albumin, urine creatinine and urine ACR), 3303 for lipids (1101 each for total cholesterol, triglyceride and HDL-C) and 610 for INR. The actual participation rate for QC testing across the live phase of the trial averaged 93% for HbA1c (1026/ 1102), 93% for urine ACR (3072/3303), 93% for lipids (3063/3303) and 91% for INR (553/610).

Key analytical performance indicators The within-practice imprecision was calculated for each practice that returned a minimum of five QC results for each level and lot number of QC tested. The within-practice imprecision for each level and lot number was then ranked from lowest to highest and the median (50th percentile) CV% was calculated. The median within-practice imprecision was selected as a key performance indicator for QC testing because the RCPA Quality Assurance Program (the EQA provider for the trial) routinely uses the median imprecision to report on imprecision for EQA testing. This strategy thus ensured that QC and EQA performance

could be compared with the same performance indicator. (EQA results for the trial will be reported elsewhere.) Table 2 summarizes the median within-practice imprecision for each QC level and PoC test and also compares the observed performance with the within-practice analytical goals for imprecision set for this trial. For HbA1c, the median within-practice imprecision met the minimum analytical goal of 4% for all levels and both lot numbers of QC tested. For LN 27 and LN 28 (Level 1), the observed imprecision for HbA1c also met the desirable analytical goal of 3%. For urine albumin, urine creatinine and urine ACR, the median within-practice imprecision readily met the optimal analytical goals of 10%, 6% and 12%, respectively. For total cholesterol, the median within-practice imprecision met the minimum analytical goal of 5% for all levels and both lot numbers of QC tested. For LN 5230 and LN 5082 (Level 2), the observed imprecision for total cholesterol also met the desirable analytical goal of 3%. For triglyceride, the median within-practice imprecision met the desirable analytical goal of 7.5% for all levels and both lot numbers of QC tested. For LN 5082 and LN 5230 (Level 1), the observed imprecision for triglyceride also met the optimal analytical goal of 5%. The median within-practice imprecision for HDL-C met the minimum analytical goal of 6% for LN 5082 and LN 5230 (Level 2), but did not meet the minimum goal for LN 5082 (Level 1). For LN 5082 (Level 2), imprecision also met the desirable analytical goal of 4%. An imprecision goal for INR was not set for this trial, but the PoCT Device Group considered that the observed within-practice imprecision for INR of less than 10% was acceptable. Table 3 displays the percentage of practices that met the within-practice imprecision goals. Overall, greater than 70% of practices achieved the imprecision goals for all analytes and lot numbers of QC material used, except for

Table 2 Comparison of median within-practice imprecision and the analytical goals for imprecision set for the trial

Test HbA1c Urine albumin Urine creatinine Urine ACR Total cholesterol Triglyceride HDL cholesterol INR

QC lot number 27 28 28 28 28 5082 5230 5082 5230 5082 5230 800,042 800,049

Imprecision goal (CV%) set for the trial

Within-practice median CV% QC level 1

QC level 2

4 4 10 6 12 5

2.5 2.9 5.0 3.5 5.0 3.1

2.8 3.3 4.1 3.5 4.0 3.0

5 7.5 7.5 6 6 10 10

2.3 3.9 5.0 5.5 6.7 7.0 6.4

2.9 3.7 5.4 3.9 4.8 n/a n/a

QC, quality control; CV, coefficient of variation; HbAlc, haemoglobin A1c; ACR, albumin:creatinine ratio; HDL, high-density lipoprotein; INR, international normalized ratio


417

................................................................................................................................................

Table 3 Percentage of practices which met the imprecision goals set for the PoCT Trial

Tests HbA1c HbA1c Urine albumin Urine creatinine Urine ACR Total cholesterol Total cholesterol Triglyceride Triglyceride HDL cholesterol HDL cholesterol INR INR

Percentage of practices which met the imprecision goals

Imprecision goal (CV%) set for the trial

QC level 1

QC level 2

27 28 28 28 28 5082

4 4 10 6 12 5

96 89 100 96 100 96

73 67 100 100 100 96

5230

5

100

96

7.5 7.5 6 6 10 10

100 100 58 33 89 88

100 79 96 63 n/a n/a

QC lot numbers

5082 5230 5082 5230 800,042 800,049

QC, quality control; CV, coefficient of variation; HbAlc, haemoglobin A1c; ACR, albumin:creatinine ratio; HDL, high-density lipoprotein; INR, international normalized ratio, n/a, not applicable

HbA1c (Level 2 LN 28 [67%]) and HDL-C (Level 1 LN 5082 [58%] and Level 1 LN 5230 [33%], and Level 2 LN 5230 [63%]). Eighty-nine percent of all practices achieved an imprecision of less than 10% for INR. Other quality management practices Telephone hotline A total of 302 calls were received by the PoCT Device Group via the telephone help desk support line. Eighteen calls (6%) were received out of hours. Thirty-five calls (12%) required additional help from industry partners to resolve the issue. Seventy-six calls originated from urban practices (average 9.6 calls per practice), 84 calls from rural practices (average 9.3 calls per practice) and 142 calls from remote practices (average 10.9 calls per practice). Most calls (106, 34%) related to QC issues, notably processes concerning the appropriate interpretation and recording of QC results. Seventy-eight calls were received regarding device issues. Fifty-two calls were received in relation to reagents, particularly delivery of start-up reagents at the beginning of the trial. Device events and test errors The PoCT Device Group were notified of 15 test error codes for the DCA 2000, 20 for the Cholestech LDX and 23 for the CoaguChek S. The DCA 2000 errors related to cartridges being exposed to too much heat or used before they had warmed to room temperature after being taken out the fridge, and the capillary blood sample being allowed to dry out prior to analysis or excess blood being applied to the sample holder. The Cholestech LDX errors mainly related to insufficient sample being collected or the analyser being used outside its specified temperature range (maximum 308C) within the practice itself. Prolonged heat wave

conditions were experienced in rural/remote Australia when these ‘temp out of range’ errors were reported. The CoaguChek S errors mainly concerned insufficient sample being collected or difficulty in applying the sample to the device’s application area, which was a significant issue for some operators particularly early in the trial. Across the trial, one DCA 2000 was replaced due to mechanical failure and three Cholestech LDX devices were replaced due to persistent ‘mag read errors’, which related to the device not being able to successfully read the calibration information encoded on the magnetic strip of the lipid reagent cassette. Four CoaguChek S devices were replaced; two due to code chip errors, one due to a power supply problem and one due to a dirty electrical connection.

Discussion Interest in the discipline of PoCT in Australia has blossomed over the past decade. Large-scale national models such as the Quality Assurance for Aboriginal and Torres Strait Islander Medical Services (QAAMS) PoCT program for diabetes management has provided a sound evidence base for the analytical, clinical and cultural effectiveness of PoCT in this most challenging of primary health-care environments.1,2 While local PoCT models have been developed for use in the Australian rural GP environment,13 the evidence base for the usefulness of PoCT in GP both within Australia and globally has thus far been limited.14 – 16 The PoCT in GP Trial represents one of the largest and most comprehensive studies of PoCT ever conducted in this primary care setting, and the results of the study should enable an objective assessment of the effectiveness of PoCT to be made in this clinical setting.6 A multifaceted approach to quality management that included training of device operators, QC and EQA testing, and device management was developed for the trial to enable PoCT to be conducted safely for patient care and within the standards of analytical quality set for the trial. This paper reports principally on the framework developed, and the results obtained, for QC testing for the trial. The use of a colour-coded QC reporting system based on a traffic light system to interpret and action QC results is unique to PoCT models managed by the Flinders CPS unit. The system aims to translate a laboratory-oriented quality concept into a readily understandable, user-friendly form that has practical application for non-laboratory health professionals conducting PoCT in community-based environments. Feedback from participants during the trial indicated widespread acceptance of this system, which engendered operator ownership and confidence. In terms of results, PoCT device operators generally met the analytical performance standards set for the trial for QC testing, as evidenced by (i) the median within-practice imprecision meeting the analytical goals for imprecision set for the trial for all PoC tests, except for HDL-C where observed performance was outside the minimum required goal for one level and one lot number of QC, and

418


Volume 46

September 2009

................................................................................................................................................

(ii) greater than 70% of practices achieving the imprecision goals for all analytes with the principal exception of HDL-C. For lipid PoCT in general, two further observations warrant comment. The short expiry of QC material necessitated three changes of lot number across the 18-month life of the trial. In terms of technical handling, a different method was recommended by the manufacturer for loading QC and patient samples into the sample well of the reagent cassette. With QC, a transfer pipette was used to transfer the QC sample onto the sample well whereas, for a patient sample, a glass capillary was used. The use of the pipette was more technically demanding for device operators (who were generally not experienced in laboratory techniques) compared with using a glass capillary; this may have led to a slightly higher degree of uncertainty of measurement for QC testing as compared with patient testing. The PoCT devices selected for the trial were user-friendly and generally robust. Nonetheless, the ‘mag read error’ problem on the Cholestech LDX became a minor issue in the latter stages of the trial while sample application was an area of concern for several operators with the CoaguChek S device. The temperature limits under which PoCT devices operated were well defined prior to the trial (15 –308C) but, in everyday practice under Australian conditions, practice staff were confronted with the very real issue of what to do when their device would not operate under temperature extremes of heat (for the Cholestech LDX) and cold (for the CoaguChek S). This is an issue that device manufacturers should be conscious of when introducing PoCT technology into this country. Were the analytical goals used for the trial appropriate? As a general principle, analytical goals for point-of-care (and laboratory) tests should be flexible and continually reviewed and refined as more clinical outcome and state-of-the-art analytical data become available.9 – 11 In general, we believe that the analytical goals set for this trial were appropriate, with the possible exception of urine ACR. While there have been no calls in the international literature to tighten the goals for this test, it is clear from the results of this trial that the goals set were readily achieved, with 100% of practices achieving the optimal goals set. In the national QAAMS Program, the median imprecision achieved by Aboriginal Medical Services for urine ACR testing has averaged 4.4% over recent years.2 It is therefore the view of the PoCT Device Group that the analytical goal for imprecision for urine ACR should be tightened to at least 8%. In conclusion, the methods established for the implementation and delivery of the internal QC program were appropriate for the PoCT in GP Trial. Results from QC testing indicate that PoCT in GP trial met the analytical goals set for the trial, with the exception of HDL-C. DECLARATIONS

Competing interests: None. Funding: Australian Government’s Department of Health and Ageing through the Pathology Section, Diagnostics Services Branch.

Guarantor: MS. Ethical approval: The PoCT Trial was approved by five relevant independent Australian Human Research Ethics Committees. The trial is registered with the Australian Clinical Trial Registry, Number 12612605000272695. Contributorship: MS wrote the manuscript and was PoCT Device Manager for the trial. AS commented on drafts of the manuscript and was QC Manager for the PoCT Device Group for the trial. LW commented on drafts of the manuscript and was scientist for the PoCT Device Group for the trial. BM commented on drafts of the manuscript and was scientist for the PoCT Device Group for the trial. PW commented on drafts of the manuscript and was Chair of the PoCT Device Management Committee for the trial. Acknowledgements: The authors wish to acknowledge the significant support received throughout the life of the Trial from members of the Pathology Section, Diagnostic Services Branch, Department of Health and Ageing, in particular Fifine Cahill, Debbie Stanford, Pamela McKittrick, Helen Blackall, Rob Lillie and Robert Walsh. We acknowledge members of the other two lead organizations responsible for delivery of the PoCT in the General Practice Trial, namely the University of Adelaide (Trial Management and Evaluation Group) led by Professor Justin Beilby, Dr Briony Glastonbury and Dr Caroline Laurence, and the RCPA Quality Assurance Programs Pty Ltd (Quality Assurance Program) led by Janice Gill and Meredith Liddy. Dr Andrew St John is thanked for his work and incisive input as Consultant to the PoCT Device Group. We also thank and recognize the outstanding efforts of our industry partners (Dr Dean Whiting, Alison Halfnights, Jan Gilbert and Jonathan Trengove from Siemens Health Care Diagnostics [formerly Bayer Australia]; Rupert Haines and Peter Merilees from Point of Care Diagnostics Australia; and Dr George Koumantakis, Adrienne Ripley, Kevin Falzon and Agnes Godfrey-Roberts from Roche Diagnostics). Julie Gardner, Ross Forbes, Andy Dallison, Ben Flink and Deb Kelly from Flinders Consulting organized all administrative matters for the PoCT Device Group and managed finances for the trial with great efficiency. REFERENCES 1 Shephard M. Clinical and cultural effectiveness of the ‘QAAMS’ point-of-care testing model for diabetes management in Australian Aboriginal medical services. Clin Biochem Rev 2006;27:161 –70 2 Shephard MDS, Gill JP. The analytical quality of point-of-care testing in the ‘QAAMS’ model for diabetes management in Australian Aboriginal medical services. Clin Biochem Rev 2006;27:185 –90 3 Shephard M, Mazzachi B, Shephard A, et al. Point-of-care testing in Aboriginal hands – a model for chronic disease prevention and management in Indigenous Australia. Point Care 2006;5:168 –76 4 Shephard MDS. Point-of-care testing in the indigenous rural environment – the Australasian Experience. In: Price C, Hicks J, St John A, eds. Point-of-Care Testing. Washington, DC: AACC Press, 2004:293– 301 5 Guibert R, Schattner P, Sikaris K, et al. Review of the Role and Value of Near Patient Testing in General Practice. Canberra: Commonwealth Department of Health and Aged Care, 2001 6 Laurence C, Gialamas A, Yelland L, et al. A pragmatic cluster randomised controlled trial to evaluate the safety, clinical effectiveness, cost effectiveness and satisfaction with point of care testing in a general


419

................................................................................................................................................

7

8

9

10

practice setting – rationale, design and baseline characteristics. Trials 2008;9:50 Shephard M. Analytical performance criteria for point-of-care testing instruments. In: Standards for point of care testing in general practice: incorporating POCT trial guidelines. Appendix 3, Pages 65 –69. Commonwealth Department of Health and Ageing, Canberra. 2004. Available at: http://www.health.gov.au/internet/main/publishing.nsf/Content/ D5EF81600AF9726DCA256F180046A4B2/$File/InterimStandardsPoCT. pdf Shephard M. Analytical goals for point-of-care testing used for diabetes management in Australian health care settings outside the laboratory. Point Care 2006;5:177 – 85 Fraser CG. General strategies to set quality specifications for reliability performance characteristics. Scand J Clin Lab Invest 1999;59:487 –90 Ricos C, Alvarez V, Cava F, et al. Current databases on biological variation: pros, cons and progress. Scand J Clin Lab Invest 1999;59: 491– 500

11 Fraser CG. Optimal analytical performance for point of care testing. Clin Chim Acta 2001;307:37– 43 12 Kenny D, Fraser C, Hyloft Petersen P, et al. Consensus agreement. Scand J Clin Lab Invest 1999;59:565 13 Shephard M, Mazzachi B, Shephard A, et al. The impact of point of care testing on diabetes services along Victoria’s Mallee Track. Results of a community-based diabetes risk assessment and management program. Rural Remote Health [online] 2005;5:371 14 Hobbs F, Delaney B, Fitzmaurice D, et al. A review of near patient testing in primary care. Health Tech Assess 1997;1:1– 129 15 Delaney BC, Hyde CJ, McManus RJ, et al. Systematic review of near patient test evaluations in primary care. BMJ 1999;319:824 – 7 16 Cohen J, Piterman L, McCall LM, Segal L. Near-patient testing for serum cholesterol: Attitudes of general practitioners and patients, appropriateness, and costs. Med J Aust 1998;168:605 – 9

(Accepted 14 May 2009)