This project aims to evaluate the performance of a Roche Cobas B221 blood gas analyser in terms of accuracy, precision and by means of a method comparison with a reference method. The cobas b221 has been purchased for the provision of point of care blood gas analysis in the neonatal intensive care unit, and delivery suite of a busy maternity hospital. Measurement of Sodium, Potassium, Bilirubin, Haematocrit and Haemoglobin on the analyser will be compared to central laboratory testing with an aim to ascertain if Point of Care testing for same is comparable to central laboratory testing, in order to allow important clinical decisions to be made based on point of care results for above parameters. This would allow the introduction of testing for above analytes at the point of care, in the neonatal intensive care unit of this hospital. It is hoped to address the issue of costing in order to determine which method is the most cost effective.
Hypothesis :
The Roche Cobas B221 measured blood gases will correlate to statistically acceptable limits against its predecessor, the Radiometer ABL720, and measured analytes such as haemoglobin, haematocrit, bilirubin and electrolytes will correlate with measurement by the central laboratory analyser. Results generated by the B221 will be precise, and accurate, to allow for key clinical decision making.
Aims: (1) To evaluate performance of the Roche Cobas b221 blood gas analyser in order to verify its accuracy, precision and uncertainty of measurement for the following assays:
(2) To verify an acceptable comparable performance between an established platform, a Radiometer ABL 725, for the measurement of pH, PO2 and PCO2 in the framework of predefined statistical limits of acceptability.
(3) To show that there is no significant bias, in terms of a clinical decision making and statistical context, in results obtained for the measurement of sodium, potassium, bilirubin, haematocrit and haemoglobin between the central laboratory analyser and the blood gas analyser at the point of care, in order to allow the introduction of testing for the above parameters at the POC in the hospital in which this study is taking place.
(4) To assess if there is a cost benefit to testing at the point of care versus central laboratory for above parameters.
Background:
Quality, equity, people focused and accountability are the cornerstones of the Irish healthcare system. When we refer to quality we mean best standards and best practices.
The role of quality in Medical laboratories nationwide is a fast emerging concept with most laboratories engaging in both internal quality audits and peer review. Currently, the Rotunda Hospital department of Clinical Laboratory Medicine is accredited under INAB, ISO15125, and CPA.
Validation and acceptance testing for new analysers are of paramount importance in determining whether new methods/analysers perform to an acceptable level of quality and cost effectiveness. It is the quality policy of the Rotunda Laboratory to subject every analyser to the following;
Installation Qualification assessment; this phase includes engineer installation, on site testing and system design, calibration etc.
Operational Qualification assessment; this phase includes system set up, method/assay calibrations, and demonstration of basic functions required for effective performance.
Performance qualification assessment; this phase involves assessment of the analyser in terms of a method comparison with a reference method, precision, accuracy, and uncertainty of measurement for all analytes generated.
This project aims to evaluate the performance qualification assessment of a newly purchased Roche Cobas B221 blood gas analyser to ensure it meets such acceptable standards and is comparable to the existing platform, a Radiometer ABL 725 blood gas analyser.
Today, Point of Care testing is crucial element in the area of critical care medicine, providing almost instant results at the patient’s bedside, using minimal volumes of blood, and enabling clinicians to make rapid decisions with regard to patient care. The blood gas analyser is one such point of care device which provides immediate information on patients’ status allowing for necessary medical intervention to take place as quickly as possible.
Perhaps one such are where blood gas analysers have become invaluable is in the field of obstetrics and neonatal care.
In the area of obstetrics, monitoring of fetal well being can be challenging for the obstetrician. The use of fetal heart rate monitoring alone has been shown to be unreliable, associated with a high false positive rate, meaning that following observation of concerning fetal heart rate patterns, many foetuses are diagnosed as asphyxiated and delivered by Caesarean section, only to have normal cord blood gases(Henderson & Ecker, 2003). However, when fetal scalp blood gas analysis is used in conjunction with electronic fetal heart rate monitoring, the rate of Caesarean section has been found to be reduced(Haverkamp, Orleans, & Langendoerfer, 1979). In the days following birth, serious difficulties may develop in the adaptation of the cardiorespiratory system of the newborn. Blood gas measurements play a crucial role in the care of these ill newborns in the neonatal intensive care unit, providing essential information regarding oxygenation (Brouilette & Waxman, 1997).
Blood gas analysers have advanced over the years and can now provide an array of tests at the point of care. In critical care areas such as the neonatal intensive care, medical decisions have to be made rapidly, often based on parameters such as electrolytes, bilirubin, haemoglobin measured in the central laboratory, which can now be measured at the point of care. However, this is still an emerging area subject to a degree of uncertainty by both users and medical laboratory scientists practicing in central laboratories. Historically reliance on POC testing as a sole source for key clinical decision making in terms of patient diagnosis and management has always been supplemented by confirmatory testing within the central laboratory. One of the downsides of central laboratory testing is the turnaround time involved in processing of specimens. By the time results become available from the central laboratory, the infants condition may have changed and not be reflective of its present condition. By bringing testing closer to the patient and hence reducing the amount of steps involved in process of central laboratory testing, results are obtained much more rapidly to bring about faster medical intervention if required, leading to improved outcomes (St. Louis, 2000). In addition, phlebotomy associated blood loss may be may reduced, which is paramount in the neonatal setting.
However, if medical decisions are to be made on results provided from a blood gas analyser for above parameters, the performance of the analyser must be acceptable when compared to the central laboratory analyser. Studies by Patel 2007, and Jain, Subhan and Joshi, 2009, report poor correlation for haemoglobin and sodium measurement between the central laboratory analysers and point of care device. In one study, increased imprecision is reported for measurement of bilirubin by a blood gas analyser versus a central laboratory analyser (Meilsch, Zimmerman, Wagner, Matthes, Schlebusch, & Luppa, 2010), while another concludes that bilirubin measurements exceeding 250umol/l from a blood gas analyser photometric method should be confirmed with central laboratory methods (Grohmann, Roser, & Rolinski, 2006).
In summary, it is the aim of this project to validate the analytical performance of a Roche Cobas b221 blood gas analyser and hence introduce testing for sodium, potassium, bilirubin, haemoglobin, haematocrit and lactate at the point of care in the neonatal intensive care unit of this maternity hospital. It is aimed to evaluate whether there is a difference in results following measurement of electrolytes, bilirubin, haemoglobin and haematocrit between the blood gas analyser and the central laboratory analyser, in order to determine if medical decisions can be made based on the results generated from the blood gas analyser at the point of care.
The benefits and cons of the analyser in the point of care setting will be discussed against the pros and cons of measurement of the same parameters in the central laboratory.
Workplan
Part one of this project involves the validation of the performance of the Roche Cobas B221 analyser. This will involve the assessment of the precision of the analyser in the measurement of its platform of assays: pH, PCO2, PO2, sodium, potassium, bilirubin, lactate, haematocrit haemoglobin, tHb, SO2, O2Hb, COHb, HHb, MethHb, and ionised calcium. Within run precision will be determined via 40 consecutive measurements of QC specimens at two different levels for the above parameters. Between run precision will be determined via the measurement of two QC specimens at two different levels twice daily for 20 days. This is in accordance with Westgard recommendations (Westgard, 2011) and NCCLs guidelines (NCCLS document EP5A, 1999) for precision studies which suggest to run at least two different QC materials that reflect low and high medical decision concentrations for the test of interest, and is acceptable for accreditation purposes by INAB and CPA. Precision will be expressed using the coefficient of variation, CV, and from the within run and between run precison, uncertainty of measurement will be determined.
Accuracy of the method for the measurement of the above parameters will be examined by the analysis of external QA material from WEQAS. Six specimens from two distributions will be measured to cover the physiological range and examine accuracy. This will provide an objective analysis for any potential random or systematic bias of the instrument relative to a large peer group.
Part two of the project will involve comparison of the Roche Cobas B221 analyser with the current analyser in use in the hospital, a Radiometer ABl 700. Electrolytes, bilirubin, lactate, and ionised calcium are not measured on the radiometer instrument, so comparison will be confined to pH, PCO2, PO2, In order to be statistically valid and in accordance with NCCLS (NCCLS document EP9A, 1995)guidelines and Westgard recommendations (Westgard, 2011), it is aimed to analyse 40 specimens for the above parameters on both analysers. In order to determine if there is a difference between methods, statistical analysis will be performed via linear regression analysis and the determination of the Pearson Correlation Coefficient, and Bland Altman analysis. Ethical approval will not be required as no additional testing will be performed on patients; an infant will only undergo blood gas analysis testing if or when clinically indicated as determined by a paediatrician. Following processing on the ABL 725 for pH, PCO2, PO2, then that specimen will also be analysed for same parameters on Roche Cobas B221. Only results from the Radiometer ABL 725 will be interpreted and acted on medically.
Part three of the project will involve the assessment of the performance of the Roche Cobas B221 point of care device for the measurement of electrolytes, bilirubin, haemoglobin and haematocrit versus the central laboratory analyser, a Roche Cobas 6000 and an Abbott Cell Dyn 3200, a haematology analyser specifically equipped with the methodology for measurement of haemoglobin in a neonatal subject population. This is being performed in order to determine if POC testing for the above parameters is same is comparable to central laboratory testing, in order to allow important clinical decisions to be made based on point of care results for above parameters. This would allow the introduction of testing for above analytes at the point of care, in the neonatal intensive care unit of the hospital in which this study is taking place. Again, in order to be statistically valid and in accordance with NCCLS guidelines (NCCLS document EP9A, 1995) and Westgard recommendations (Westgard, 2011), it is aimed to analyse 40 specimens, covering the physiological range, for above parameters by both methods. To determine if there is a difference between methods, statistical analysis using linear regression and the Pearson Correlation coefficient, and Bland Altman analysis will be utilised. Ethical approval will not be required as only specimens received in the central laboratory already requesting testing for electrolytes, bilirubin, and haemoglobin will be used. These specimens will be analysed on the roche blood gas analyser for said parameters and then following centrifugation, the plasma from such specimens will be analysed on the central laboratory analyser, the Roche Cobas 6000 (electrolytes and bilirubin) and the Abbott Cell Dyn 3200 (Haemoglobin and haematocrit, whole blood). Only results obtained from the central laboratory analysers will be issued and acted on. No additional blood will be required from patients.
