Potential errors in POCT

Overview – errors in POCT

The influential Institute of Medicine report, To Err is Human – Building a Safer Health System, caused significant concern in the USA when it was release in 1999.¹³⁰ This report, which describes and categorises errors in the American health system, claims “that at least 44,000 people, and perhaps as many as 98,000 people, die in hospitals each year as a result of medical errors which could have been prevented.” The report defines medical errors as “the failure of a planned action to be completed as intended or the use of the wrong plan to achieve an aim.” Medical errors in laboratory testing, including POCT, are mainly failures in testing procedures or failure to complete planned actions and are generally described as pre-analytical errors, analytical errors or post-analytical errors. In order to achieve appropriate test results which are fit for purpose, all aspects of the testing process must be addressed in an efficient and reliable manner.¹³¹ To achieve this goal, operational standards, rules and guidelines must be applied. These same rules apply whether the objective is central laboratory testing or POCT.

The advantages of POCT largely depend on proof of acceptable analytical performance. This concept dominates policy statements from government agencies of many jurisdictions, articles published in peer reviewed journals, and statements made by international commentators on POCT. Regardless of the particular assay or method, POCT, in common with all laboratory testing requires regular QC and EQA to assure acceptable performance. Tirimacco provides a good summary of these concepts “Although POCT appears to be deceptively simple, if incorrectly performed it may present a risk to patient care and, if used inappropriately or overused can lead to significant increases in the cost of patient care. To ensure results obtained are comparable to the traditional pathology laboratory, POCT should be implemented within a framework of quality standards.” … “This quality framework should include: operator education, training and competency, quality control, proficiency testing and accreditation.”¹³²

There is now a growing literature describing the potential and actual errors arising from poorly controlled POCT. In the USA, home glucose-testing devices account for the largest number of complaints filed with the FDA for any medical device. According to information provided by Plebani¹³³ and Nichols¹³⁴, over 3200 incidents have been filed with the FDA in relation to blood glucose-testing devices with 16 deaths. In addition, poorly maintained blood gas analysers (and other POCT devices) that are carried from one patient to another, can act as reservoirs for anti-biotic resistant organisms and provide an easy path for the transfer of infection between patients. Thompson and Perz describe the investigation of 18 outbreaks of hepatitis B (HBV), with 15 (83%) in the past ten years (up to 2009) that were associated with improper use of blood glucose monitoring equipment. “At least 147 persons acquired HBV infection during these outbreaks, 6 (4.1%) of whom died from complications of acute HBV infection”.¹³⁵ Nichols also describes nine patients at two nursing facilities in Southern California who were diagnosed with hepatitis B following infection transmitted in association with blood glucose monitoring and linked to failure of staff to change gloves and adequately clean the testing device between patients.¹³⁴

According to Sharp etal, inspection and accreditation processes have been very successful in achieving the goal of reducing POC testing error.¹³⁶ In reviewing the impact of periodic inspection by the POCT manager, these authors “ …found that inspection correlated directly with laboratory error.” That is, fewer errors were produced as the number of inspections by the POCT manager increased:

• “During the 3-month period before the POCT staff was decreased 50%, the error rate averaged 2.0%.”
• “During a second period 5 months later, following suspension of POCT inspections by the POCT manager, the error rate averaged 2.7%.”
• “Finally, during a third period 6 months later, after the reinstitution of POCT inspections, the error rate averaged 1.1%.”
• “For the most recent 3-month period for which data are available, following 11 more months of POCT inspections, the error rate averaged 0.5%.”

Sharp etal suggest that there results “emphasize that the challenges to management of POCT are the same as those for clinical laboratory testing, and indicate the need for measures for continual and methodical surveillance of all laboratory testing.”¹³⁶

In a study of “errors in laboratory medicine”, Bonini etal found that the principal reasons for pre-analytical errors were: haemolysed specimen (54% of all pre-analytical errors), insufficient specimen (21% of all pre-analytical errors) and incorrect specimen (13% of all pre-analytical errors).¹³⁷ In a traditional laboratory these types of error are relatively easy to detect. In a POCT situation, this type of error may well pass undetected with a compromised patient result being produced. In addition to pre-analytical errors, POCT in particular is prone to “native interferences” which alter the reactivity of the analytical process to produce errors which may largely go undetected.¹³³ Haemolysis, which is undetected in poorly collected whole blood samples produces incorrect results in many circumstances. Heterophilic antibodies, ascorbic acid (vitamin C), INR tests due to inconsistent ISI values, patient haematocrit values higher or lower than the accepted POCT device range, provide a few examples.

By their very nature, errors in POCT are often difficult to identify and little information exists which describes error rates associated with POCT itself. A recent study by O’Kane etal, evaluated a variety of POC tests over a 14-month period and concluded “The quality error rate for POCT is variable and may be considerably higher than that reported previously for central laboratory testing.”¹³⁸ An interesting feature of these results, is the wide variability in reported error rate between test types; very low for “dip stick” tests, 0.52% for blood gas and electrolyte tests and 0.65% for HbA1c testing. Although the overall impact of POCT errors on patient care was considered low, with 51.8% of all reported errors considered to have no impact, 48.2% of errors were considered to have some impact on patient care.¹³⁸

Surveys of waived testing laboratories in the US

Since the commencement of the CLIA regulations in the USA there have been a number of surveys of waived testing laboratories, to review quality procedures and compliance with CLIA regulations. As described previously, waived tests include test systems cleared by the FDA for home use and those tests approved for waiver under the CLIA criteria. The published reviews of waived testing laboratories are broadly of two forms: (a) those conducted by or on behalf of CMS with results provided in a summarized de-identified CMS report, and (b), publications by professional commentators based on CMS report data and / or proficiency testing survey data.

To encourage improved testing in waived laboratories, in 2002 CMS initiated on-site visits to approximately 2% of laboratories that had been issued a certificate of waiver (COW) under CLIA. This was the first time that CMS had conducted inspections in all 50 states, although two previous surveys had included visits to a smaller number of states. An overview of the survey process is provided on the CMS internet site.¹³⁹ Results from a pilot study completed in 2001 formed the basis for concerns regarding waived laboratories.¹⁴⁰ Data from this pilot survey has been reviewed by Meier and Jones who conclude that POCT “errors are relatively common, their frequency is amplified by incoherent regulation, and their likelihood of affecting patient care is amplified by rapid availability of POCT results and immediate therapeutic implications.”¹⁴¹

Meier and Jones summarise the results of the CMS pilot report as follows:¹⁴¹

• “CMS examiners found that 19% of testing personnel had been neither trained nor evaluated in the performance of the assays they carried out”
• “32% of the observed test operators could not locate test instructions when asked to refer to them”
• “25% of test operators failed to follow manufacturer’s directions”
• “7% of test operators did not perform required calibrations”
• “32% of test operators failed to perform quality control (QC)” (as specified)
• “20% (of test operators) physically separated (against manufacturer’s directions)internal QC test fields from patient test fields in card test format tests”
• “6% (of test operators) use expired reagent kits, whose integrity manufacturers would” no longer guarantee.

In a similar manner, Howerton etal on behalf of the CDC, have summarised quality failures identified in waived testing sites on the basis of CMMS surveys conducted during 1999 – 2003, and studies of waived testing practices undertaken by CDC during 1999 – 2003.⁶²
This report, Good laboratory practices for waived testing sites described a similar catalogue of errors with the principal deficiencies being:

• 12% of sites did not have current manufacturer’s instructions
• 21% of sites did not routinely check new product inserts for changes
• 21% of site did not perform the minimum QC as specified by the manufacturer
• 18% of sited did not report test results with appropriate terminology or units
• 6% of sites did not comply with expiration dates
• 5% of sites did not perform the required function or calibration checks
• 3% of sites did not adhere to storage and handling instructions
• 3% of sites did not perform instrument maintenance
• 2% of sites did not use an appropriate specimen for the test

POCT, clinical trials and evidence-based practice

Actual clinical trials or large scale reviews to assess the effectiveness of POCT are much less numerous than publications which describe the potential risks associated with POCT, issues with specific tests or with specific devices. The evidence to support POCT has been recently reviewed by St John, with specific discussion including patient self-monitoring, POCT in the community (particularly POCT conducted within pharmacies), POCT in general practice and POCT in critical care areas including the emergency department.⁸⁵

The NACB publication Laboratory Medicine Practice Guidelines - Evidence-based practice for point-of-care testing reviews much of the evidence in support POCT and provides informed discussion on many aspects of the subject.⁶¹ The document provides specific recommendations based on evidence obtained from an extensive literature review. Most of the tests which would usually be considered in the POCT environment are covered in detail. Chapter 1, Management, provides insight into the topic of quality management (QM), quality assurance (QA) and quality control (QC).

A selection of clinical trials or literature reviews whose principal aim was to compare outcomes with central laboratory testing or to evaluate clinical effectiveness are summarised in appendix 6 and include:

• Australian Government point-of-care testing in general practice trial.¹⁴
• Kendall etal; Point-of-care testing, randomised controlled trial of clinical outcomes.¹⁴²
• Blattner etal; Changes in clinical practice and patient disposition following the introduction of point-of-care testing in a rural hospital.¹⁴³
• Hobbs etal; A review of near patient testing in primary care.¹⁴⁴
• Grieve etal; Near patient testing in diabetes clinics: appraising the costs and outcomes.¹⁴⁵
• Collinson etal; A prospective randomised controlled trial of point-of-care testing on the coronary care unit.¹⁴⁶
• Ryan etal; A multicenter randomised controlled trial comparing central laboratory testing and POCT cardiac marker testing strategies: the disposition impacted by serial POCT markers in acute coronary syndromes.¹⁴⁷
• Fitzmaurice etal; A randomised controlled trial of patient self management of oral anticoagulation treatment compared with primary care management.¹⁴⁸
• Fitzmaurice etal; Self management of oral anticoagulation: Randomised trial.¹⁴⁹
• Parry etal; Anticoagulation management in primary care: a trial-based economic evaluation.¹⁵⁰
• Wilson etal; Comparing the quality of oral anticoagulant management by anticoagulation clinics and by family physicians: a randomised controlled trial.¹⁵¹
• Khunti etal; Randomised controlled trial of near-patient testing for glycated haemoglobin in people with type 2 diabetes mellitus.¹⁵²
• Al-Ansary etal; Point-of-care testing for HbA1c in the management of diabetes: A systematic review and metaanalysis.⁷⁷