Six sigma matrix and Quality Goal Index ratio in improving the quality of analytical phase in a clinical laboratory

INTRODUCTION

The clinical laboratory is the core of the healthcare system. Providing an accurate report is quite challenging as the diagnosis, treatment, and prognosis of the disease by a physician are in accordance with the report.^[1] Here comes the importance of a quality management system (QMS) for releasing accurate reports. There are recommended guidelines provided by ISO15189 to assess and monitor the QMS.^[2]

Quality control (QC) is one of the components of QMS, which monitors and evaluates the analysis phase of the clinical laboratory. QC is the statistical analysis of internal QC (IQC) and external quality assessment service (EQUAS) programs.^[3] IQC is evaluated daily on Levey–Jenning (LJ) charts using west guard rules.^[4] EQUAS analyses unknown concentrations of controls monthly, provided by an external agency. EQUAS is interpreted by Z SCORE or standard deviation index.^[5] A Z-score is a calculated value that indicates how many standard deviations a control result has shifted from the mean value that is expected for that material. IQC checks the precision, and EQUAS checks the accuracy of the parameter to the mean value.^[6]

Neither IQC nor EQUAS can give the exact number of errors occurring in the analytical phase of the clinical laboratory.^[7] Six Sigma, which integrates accurate evaluation and process improvement, has come to light.^[8] Six Sigma metrics evaluate the errors in the QC system and quantify the performance as “defects per million.”^[9] The power of six sigma is measured on the “Sigma Scale.” It typically runs from 0 to 6, but a process can exceed six sigma if variability is sufficiently low to decrease the defect rate.

These values indicate the chance of false test results by the clinical laboratory, a value <3 Indicates poor performance, while >3 is considered as good and >6 is world-class performance.^[10] Sigma metrics are a calculation in which sigma is a matrix that quantifies the performance as “defects per million” and EQUAS data along with total allowable error (TEa). Mao et al., in their study, stated that “Sigma metrics is a self-assessment method in guiding clinical laboratories to make QC strategy and plan QC frequency”^[11]. Previous studies were done to elicit the individual laboratory performances but very few studies have reviewed and corrected the parameters with poor sigma scale.^[4]

An accurate report is needed for proper clinical diagnosis. This study aims to assess the performance of individual biochemistry parameters on the sigma scale. In addition, we will conduct a root cause analysis and implement corrective actions for parameters that perform poorly on the sigma scale to enhance the overall performance of our clinical laboratory.

MATERIALS AND METHODS

This is a retrospective-prospective study conducted in the central laboratory of a tertiary care hospital from January 2023 to September 2023. Retrospective data of daily IQC and monthly EQUAS from January 2022 to December 2022 were collected. Subsequently, daily IQC and monthly EQUAS data were collected prospectively from April 2023 to September 2023. The performance of the following ten biochemistry parameters was studied: Serum creatinine, total bilirubin, total protein, albumin, calcium, triglycerides, high-density lipoprotein cholesterol (HDL-C), aspartate transaminase (AST), alkaline phosphatase (ALP), and amylase.

Data collection and calculations

The IQC data for the 10 parameters serum creatinine, total bilirubin, total protein, albumin, calcium, triglycerides, HDL-C, AST, ALP, and amylase were collected retrospectively. The performance of these 10 biochemical parameters was done on a Sigma Scale by calculating the sigma metrics for each parameter.

Sigma metrics were calculated with the following formula:^[12]

$Sigma=\frac{\left[TEa-Bias\right]}{CV}$

Where TEa is the total allowable error and Bias and coefficient of variation (CV) are the indicators of systematic and random errors, respectively.

TEa for each parameter was obtained from Clinical Laboratory Improvement Amendment proficiency testing criteria for acceptable analytical performance printed in the federal register.^[13]

Bias is the systematic difference between the expected results obtained by the laboratory test method and the results obtained from an accepted reference method. The bias percentage for each parameter was calculated from the EQAS report using the formula.^[14]

$Bias\%=\frac{\left[OurEQASresult-Peergroupmeanusingthesameinstrumentandmethod\right]}{Peergroupmeanusingthesameinstrumentandmethod}\times 100$

CV was calculated for the Bio-Rad IQC each month. CV is a measure of the variability of an assay and is expressed as a percentage.

$CV\%=\frac{StandardDeviation\left[SD\right]}{Mean}*100$

The Sigma value was calculated for all the 10 parameters.

The standardized sigma values were categorized into six categories, i.e., world-class (σ ≥6), excellent (6-5), good (<5-4), marginal (<4-3), poor (<3-2), and unacceptable (<2).^[10] For the parameters with sigma values 3, Quality goal index (QGI) was calculated. Based on the values of QGI, the reason for the poor sigma performance of the parameters was interpreted. A value of <0.8 indicates imprecision, 0.8–1.2 indicates imprecision and inaccuracy, and >1.2 indicates inaccuracy.^[15,16] The formula used for QGI was:

$QGI=\frac{\left[Bias\%*CV\%\right]}{1.5}$

With reference from QGI, root cause analysis (RCA) was performed, and corrective actions were taken using a cause-effect chart^[17] (fishbone diagram), as represented in Figure 1. After corrective actions, the sigma values for the parameters were calculated from April 2023 to September 2023, that is, for 6 months.

Close

Figure 1:

Standardized QC sigma chart for 10 analytes (level-1). The slope of the five lines is the negative value of sigma. The colored circles represent the sigma value of the analytes. The X-axis is the percentage of CV normalized to TEa, and the Y-axis is the percentage of bias normalized to TEa. QC: Quality control, TEa: Total allowable error, CV: Coefficient of variation. HDL: High-density lipoprotein, AST: Aspartate transaminase, ALP: Alkaline phosphatase.

Export to PPT

RESULTS

Tables 1 and 2 represent the monthly and average CV of levels 1 and 2 IQC from January 2022 to December 2022, respectively. It was observed that CV % of both levels of IQC for the parameters creatinine, total bilirubin, total protein, albumin, calcium, triglycerides, L1 of HDL, and L2 of amylase was <5%; both L1 and L2 of AST and amylase were <10%.

Table 3 represents the Bias percentage of EQUAS for all the parameters. It was observed that all parameters had bias <10% except amylase, which was 13.5%.

Table 4 represents Sigma METRICS for all the parameters. It was observed that sigma values for both levels of IQC for parameters albumin, calcium, triglycerides, HDL, L1 of total bilirubin, L2 of creatinine, AST, and amylase were >3. Whereas both the levels of total protein and ALP, L1 of creatinine, AST, and amylase, and L2 of total bilirubin were <3.

Figures 1 and 2 represent operational process specifications (OP) spec graphs. The slope of the five lines is the negative value of sigma. The colored circles represent the sigma value of the parameters; the X-axis is the percentage of CV normalized to TEa and shows imprecision, and the Y-axis is the percentage of bias normalized to TEa and shows inaccuracy.

Close

Figure 2:

Standardized QC sigma chart for 10 analytes (level-2). The slope of the five lines is the negative value of sigma. The colored circles represent the sigma value of the analytes. X-axis is the percentage of CV normalized to TEa, and the Y-axis is the percentage of Bias normalized to TEa. QC: Quality control, TEa: Total allowable error, CV: Coefficient of variation. HDL: High-density lipoprotein, AST: Aspartate transaminase, ALP: Alkaline phosphatase.

Export to PPT

Table 5 represents QGI parameters whose level 1, level 2, or both sigma values were <3. It was observed that the cause for the poor sigma values for L1 and L2 of total protein and total bilirubin were inaccuracy. The poor sigma values for L1 and L2 of AST and ALP was imprecision.

Table 6 represents RCA done as per the cause-effect chart and appropriate the corrective actions are taken. Figure 3 represents the cause effect chart (Fish bone analysis) for identifying cause and effect on parameters with poor sigma performance.

Close

Figure 3:

Cause-effect chart (Fish-bone diagram) for the potential cause and effect for parameters with low sigma levels. QC: Quality Control

Export to PPT

Tables 7 and 8 represent the CV of levels 1 and 2 IQC from April 2023 to September 2023, respectively. Table 9 represents the sigma values for the 6 parameters after corrective action. Sigma values for L1 of creatinine and amylase; L1 and 2 of total bilirubin, total protein, AST, and ALP were observed to be >3.

DISCUSSION

In this era with the advancement of medical sciences, clinicians are depending on clinical laboratories for proper diagnosis and treatment. Automation has become an integral part of laboratories to meet the increasing workload and decrease TAT. The working of automation is checked by Good Laboratory Practice as per the National Accreditation Board for Testing and Calibration Laboratories guidelines. Most of the clinical laboratories follow IQC and EQUAS, monitor LJ charts, follow Westgard rules, and do corrective actions for outliers to maintain the quality of reports.^[18] However by introducing a QMS tool, sigma metrics, the quality of the report at the analytical phase of testing can be improved.^[11]

A sigma level of more than 3 Standard Deviations (SD) is always desirable.^[16] The Sigma model pursues a Plan, Do, Check, Act cycle for QMS. The QMS is dominated by defining, measuring, analyzing, improving, and controlling, which are salient features of six sigma metrics.^[19]

In the present study, we obtained values of six sigma for 10 biochemical parameters of both L1 and L2 levels of IQC. In our study, we observed L1 IQC of total bilirubin, level 2 of amylase, AST, and creatinine, and both levels of albumin, calcium, TG, and HDL-C were between 3 and 6 sigma values. It was also observed that both IQC levels of total protein and ALP failed to meet minimum sigma quality performance with a value <3.

Similar studies were done by Kumar and Mohan,^[14] Verma et al.,^[20] Pradhan et al.,^[21] and Maheshwari et al.^[6] Variations in sigma values between our study and others can be due to the difference in the methodology, traceability calibrators, instrument, QC material, and other pre-analytical and analytical conditions.^[14,22]

In our study, QGI was calculated for parameters whose Sigma values were poor (ie <3) parameters like total protein, ALP, creatinine , AST, Amylase, total bilirubin. Creatinine and amylase had problems of imprecision, and in accuracy (QGI 0.8–1.2), AST and ALP were imprecision (QGI <0.8), and total protein was in accuracy (QGI >1.2).

The cause-effect chart (Fish-bone diagram) for RCA was carried out to determine the cause of low sigma values and Westgard rules were used for corrective actions.^[17,23]

Incubation temperature fluctuations were found to be cause for enzymatic reagents such as AST and ALT, as in the study conducted by Goel et al.^[23] For creatinine, sampling, and reagent dispense issues were resolved by major instrument maintenance. For total bilirubin and total protein, SD was narrowed using 60 points of IQC data and setting up lab mean and SD. This is similar to the study done by Pradhan et al.^[21] Other factors that can affect the quality are working conditions in the laboratory, instrument proficiency, frequent calibration, training of the staff about quality management, and troubleshooting.^[24] Proper designing and implementing of

QC strategies can be done by evaluating the sigma matrix. Any QC procedure in a clinical laboratory should work for a high probability of error detection and a low probability of false rejection.^[14] Six sigma metrics help in providing a strategy and plan for QC frequency.^[5] Guidelines proposed by Westgard for QC planning as per the Six Sigma matrix are given in Table 10.^[17] Charuruks, in their study, stated the sigma scale can be applied as a universal benchmark for the comparative evaluation of performance between tests, methods, equipment, and laboratories.^[25] The errors encountered in the clinical laboratory can be decreased by prioritizing the quality improvement plan through the evaluation of low sigma value analytes and monitoring of daily quality indicators.^[26]

CONCLUSIONS

Our study states that sigma metrics are a reliable quality tool for assessing the analytical performance of a clinical laboratory. However, a few parameters had sigma values <3, with RCA and corrective actions performed based on personnel, equipment, materials, method, and environment, sigma values raised above 3. Sigma metric analysis is a tool to determine the performance of QC design. This gives the laboratory a select the right QC strategy. This will help to save time, effort, unnecessary runs, calibration, and reagent waste, which affect the outcome of turnaround time.