NATA


Verification and Statistical Comparison of Blood Gas Analysers: 3 September 2008

Laboratories performing blood gas analysis or who are responsible for blood gas analysis performed outside of their laboratory are advised of the Association's approach to assessing this area.

Initial Comparison

(i) If the blood gas analyser (BGA) is replacing an existing analyser

a) Patient comparison studies must be performed at the time of installation. Often the old and new analysers will be co-located during this time thereby allowing active comparison and limiting concerns about sample stability.

b) For branch laboratories installing a new analyser, the comparison studies may be performed in the central laboratory before relocation of the analyser to the branch site.

c) Where multiple new instruments are being installed within a multi-site organisation, the studies may involve the comparison of one new analyser against one old analyser and this data then taken as representative of the overall comparison and is not replicated for every installation.

The findings of any statistical comparison studies must be available for review at the assessment of the branch laboratory and these studies must show evidence of review and authorisation.

(ii) If the BGA is new, and there is no previous instrument or central laboratory for comparison

Exchange of patient samples with another facility is difficult due to the lability of blood gas samples. There may be data in the literature or the analyser's manufacturer may be able to provide some information.

The findings of this research must  be available for review at the assessment and this information must show evidence of review and acceptance.

Alternatively, where this is not available the laboratory must establish appropriate internal QC limits and demonstrate satisfactory performance in internal QC and QAP.  The appropriateness of these measures will be reviewed at assessment.

On-going Comparison of blood gas analysers

(i) A laboratory having a number of blood gas analysers at one geographical site

Where analysers are closely located (e.g. multiple analysers at one hospital site) a single sample should be analysed on a number of analysers within a short time, allowing comparison of results on patient material. The use of patient samples will eliminate any matrix issues with internal QC and QAP material. These comparisons are a useful adjunct to daily QC testing and are recommended as part of good laboratory practice.

(ii) For blood gas analysers that are not located at one geographical site

Where analysers are not closely located (e.g. at different laboratory sites) the lability of patient samples makes on-going comparison difficult. Where internal QC and QAP results suggest there may be performance problems with a particular analyser, patient comparisons must be investigated as part of the laboratory's assessment of the possible impact on patient results. Where internal QC and QAP results demonstrate acceptable performance on-going comparisons are recommended as part of good laboratory practice but are not mandatory.

Comparison of blood gas analyser results against main chemical pathology analyser results

(i) Where the blood gas analysers and main chemistry analysers are located at one geographical site

Where a laboratory performs serum/plasma electrolytes on the main chemistry analyser and blood electrolytes on a blood gas analyser, the laboratory should assess ongoing agreement between the platforms.  It is acknowledged that the two platforms may not measure the same thing (e.g. direct ISE versus indirect ISE, different sample types etc) and that there may be a bias, however this bias should be relatively small and stable. The comparison can be performed by measuring a serum/plasma sample on each instrument at defined intervals and assessing the level of agreement. This is useful information and will reflect the laboratory users' experience.

(ii) Where the blood gas analysers and main chemistry analysers are not located at one geographical site

On-going comparison studies would be good practice, however it is appreciated that logistics may complicate or preclude the performance of such studies.

Provided adequate agreement is demonstrated at the time of analyser implementation and QC and QAP performance is satisfactory, these comparisons are not mandatory.

In both scenarios (i) and (ii) described above, on-going comparisons are recommended as part of good laboratory practice but are not mandatory.

If you have any questions please do not hesitate to contact NATA in Sydney on (02) 9736 8222 or in Melbourne on (03) 9329 1633.