Evaluation of sigma metrics in a Medical Biochemistry lab

Sigma metrics is calculated in our lab and evaluated from December 2013 to November 2014. It is observed that Triglycerides, Lactate, Uric acid, AST, Urea, Creatine kinase(CK), Phosphate, Total Bilirubin are the best performers and the sigma value is more than 6.0 in both in normal and abnormal levels. Iron and Creatinine are best performers in normal level and Prolactin and Vitamin B12 are best performers in abnormal level. Amylase and LDH are the poor performers at level 1 (normal), though they show sigma between 3 to 6 at level 2 which is clinically acceptable.


1.Introduction
It is the time to improve the quality accompanied by reduction of cost in healthcare system of both public and private sectors.This pressurises to implement Total Quality Management which includes Quality planning, Quality Laboratory Process, Quality Control, Quality assessment, Quality Improvement.Quality refers to satisfaction of the needs and expectations of the users or customers.Fundamental requirements for all objective quality control systems are clearly defined quality goals.Laboratories must define their service goals and establish clinical analytical requirements for testing processes.Without such quality goals, there is no objective way to determine whether acceptable quality is being achieved.
Six sigma is an evolution in quality management that is being widely implemented in business and industry in the new millennium.The principles of Six sigma was adopted by Motorola in early 1990s and won the award of Malcolm Baldridge Quality Award.The application of sigma metrics for assessing analytical performance depends on measuring the process variation and determining process capability in sigma units.Sigma(σ) is the mathematical symbol for standard deviation(SD).
Any process can be evaluated in terms of a sigma metric that describes how many sigma's fit within the tolerance limits.Two methods can be used to assess the process performance in terms of a sigma metric.One approach is to measure outcomes by inspection.The other approach is to measure variation and predict process performance.Measurement of outcome is done by calculating defects per million(DPM) and converting it into sigma metric.A defect rate of 0.033% would be considered excellent in any healthcare organization, where error rates from 1 to 5% are often considered acceptable.A 5% error rate corresponds to a 3.15 sigma performance, and a 1% error rate corresponds to 3.85 sigma.Quality is assessed on the sigma scale with a criterion of 3 σ as the minimum allowable sigma for routine performance and a sigma of 6 σ being the goal for world-class quality.[1] To achieve Six sigma goal, the error rate should be 0.1% (4.6

Materials and Methods
Six Sigma is unique in its rigorous approach to outlining the details that are necessary to achieve significant improvement in process quality and efficiency.The process begins with developing a clear understanding of required performance.It then applies a variety of statistical tools to analyze process measures, which facilitates proving the root cause(s) for problem(s).The task then becomes revising the process in order to eliminate the causative factor(s).
We determined the sigma values for various parameters and evaluated sigma metrics from December 2013 to November 2014.Sigma (σ) value is calculated with the formula Sigma metrics (σ) = (TEa % -Bias %) / CV% where TEa% is Total allowable error percentage and CV% is Coefficient of Variation.

Precision
Precision has been defined as the closeness of agreement between independent results of measurements obtained under stipulated conditions.The degree of precision is usually expressed on the basis of statistical measures of imprecision, such as CV%.CV% is calculated from Internal Quality Control (IQC) data with the formula CV% = (SD/Mean)* 100.Monthly CV% of all the analytes from Dec 2013 to Nov 2014 is shown in Table 1 and Table 2.

Trueness
Trueness is defined as closeness of agreement between the average value obtained from a large series of results of measurements and the true value.The difference between the average value and the true value is the bias, which is expressed numerically and so is inversely related to the trueness.Bias% is calculated from External Quality Assurance Scheme (EQAS) with the formula: Bias% = [(Our lab result -Peer group mean) / (Peer group mean)]*100.Bias% of all the analytes from Dec 2013 to Nov 2014 is shown in

Discussion
In a routine accredited clinical biochemistry laboratory it is conventional practice to run the Internal and External quality controls to assess precision and accuracy.In this practice, laboratory personnel usually follow the Westgard rules like 12S, 13S, R4S, and 10X for internal quality assurance.
Rule 12S indicates one control observation exceeding the mean ± 2 standard deviations is used as a warning rule that intiates testing of the control data by the other control rules.
Rule 13S indicates one control observation exceeding the mean ± 3 standard deviations is rejection rule that is primarily sensitive to a random error, www.ssjournals.com

Rule 22S
indicates two control observations exceeding the same mean plus 2 standard deviations or minus 2 standard deviations limit is rejection rule that is sensitive to systematic error.
Rule R4S indicates one observation exceeding the mean plus 2 standard deviation and the other exceeding the mean minus 2 standard deviation is a rejection rule that is sensitive to random error.
Rule 10 x indicates ten consecutive control observations falling one side of the mean is a rejection rule that is sensitive to systematic error.[2] Common checklist followed when a systemic error is noted includes: 1. Change in the Reagent/Control/ numbers.Total error (TE) = 1.96 * CV% + Bias %.CV% can be calculated from the internal quality control and Bias percentages can be known from EQAS results.So, it is clear that calculation of total error takes into the consideration of both precision and accuracy.From this it appears that if Total error (TE) is less than Total allowable error (TEa), we can cosider the process satisfactory.But it is not so because other factors like sample carry over, non linear bias and sample matrix effects can also affect the Total analytical error.[4] Total analytical error (TE) for observed analytes from Dec 2013 to Nov 2014 is showed in Table 6 and Table 7.
Total allowable errors (TEa) for different analytes are published by different groups like Clinical Laboratory Improvement Amendments (CLIA), Royal College of Australasian Pathologists (RCPA), as well as the Guidelines of the German Medical Association (RiliBäk).Dr. Carmen Ricos and her colleagues have provided a continuously updated database of biologic variation since 2000.For over 300 different analytes, they have tabulated desirable specifications for imprecision, inaccuracy, and total allowable error.[5] The Six Sigma scale typically runs from zero to six, but a process can actually exceed Six Sigma, if variability is sufficiently low as to decrease the defect rate.In industries outside of healthcare, 3 Sigma is considered the minimal acceptable performance for a process.When performance falls below 3 Sigma, the process is considered to be essentially unstable and unacceptable.[5] In contrast to other industries, healthcare and clinical laboratories appear to be operating in a 2 to 3 Sigma environment.The routine use of "2s" (i.e., 2 standard deviations or 2 SD) control limits is indicative of a complacent tradition in quality control practices.Despite the well-known problems of 2s limitsthey can generate false rejection rates of up to 10 to 20%, depending on the number of controls run-many laboratories use them for all testing processes.The misuse of 2s limits in laboratory testing frequently results in erroneously-repeated controls, excessive trouble-shooting, or worse still, workarounds that artificially widen control limits to the point that laboratories can no longer detect critical analytical errors.
Six sigma scale has the power to provide a universal bench mark.It allows the comparison between different instruments, different labs and different methods all over the world.Nevalainen's data on Sigma assessment in preanalytic, analytic and post analytic phases of the clinical lab showed that many were not adequate.[6] Nanda et al data showed that out of the 13 analytes evaluated for sigma assessment in their laboratory, 5 analytes showed the performance of above 6 sigma, 4 analytes showed the performance between 3 to 6 sigma metrics and remaining 4 showed the performance of below 3 sigma metrics.[7] Singh et al data showed that among the 15 analytes observed for sigma assessment three analytes showed the performance of below 3 sigma metrics.[8] In present study, 11 of 23 analytes showed above six sigma performances, 10 analytes showed 3 to 6 sigma performance, 2 analytes showed less than 3 sigma performances in normal level (level 1).Sigma metrics of abnormal level (level 2) showed 11 of 23 analytes showed above six sigma performances, 12 analytes showed the performance between 3 and 6.
Though many laboratories are following the ISO 15189 guidelines and participating in the Internal and external quality control programmes, unable to achieve the six sigma performance.Six sigma being the goal for world-class quality, there is a need to implement the sigma metrics in the laboratories.Sigma metrics in combination with a rational QC design for each analyte can improve the quality there by reducing the wastage.[5] Schoenmaker et al described the importance of application of sigma metrics and preparation of rational QC design based on the sigma values with the help of westgard operational specifications chart (OPSpecs chart) in clinical biochemistry laboratories.[9] An example of QC design is shown in the Table 8.

checklist for corrective action when a random error is noted includes
1. Weekly / monthly maintenance due. 2. Date and Time of current calibration.3. Calibrator lot change.4. Any maintenance / Service done since the shift /trend is noted.5. Any lamp change / lamp deterioration.6. Reagent shelf / on board stability.7. Date of reconstitution of the control.8. Status after repetition of control.9. Correlation of lab Mean and SD with peer group Mean and SD.Common : 1. Proper mixing of the control.2. Any bubbles noted in aliquot.3. Shelf life of reconstituted control.