A study of pre-analytical errors in the clinical biochemistry laboratory at Victoria hospital, BMCRI, Bangalore

To categorize and calculate the percentage error of pre-analytical variables in the clinical biochemistry laboratory. Prospective observational study conducted for two months with documenting the frequency and type of pre-analytical errors occurring in venous samples. The total errors recorded were 1.31%. Insufficient volume followed by haemolysis amounted to a major proportion of errors. Continuous pre-analytical phase evaluation and taking corrective measures to make this phase error-free, have to be done.

Evaluation of Quality Control Data of Clinical Chemistry Parameters using Six Sigma Metrics Tool in Clinical Laboratory

Background: Physicians rely on laboratory results for treating patients. So it is the duty of laboratories to assure quality of the results released. So laboratory performance should be validated to maintain the quality. Six sigma has now gained recent interest in monitoring the laboratory quality.This study was designed to gauge the clinical chemistry parameters based on six sigma metrics. Materials and Methods: In this retrospective study, both the internal and external quality control data of 26 clinical chemistry parameters were collected for a period of 6 months from June 2020 to November 2020 and the six sigma analysis was done at the Central clinical biochemistry laboratory of Chettinad Hospital and research institute. Results: AST, amylase, lipase, triglyceride, HDL, iron, magnesium, creatine kinase showed sigma values more than 6.Uric acid, total protein, ALT, direct bilirubin, GGT,cholesterol, cholesterol, calcium, TIBC and phosphorus shows sigma values between 3.5 to 6. Glucose, BUN, creatinine, albumin, Na, K, Chloride, showed sigma values less than 3.5. Conclusion: Six sigma metrics can help in improving the quality of laboratory performance and also helps to standardisethe actual amount of QC that is required by the laboratory for maintaining quality.

Machine learning predictive models of LDL-C in the population of eastern India and its comparison with directly measured and calculated LDL-C

Background LDL-C is a strong risk factor for cardiovascular disorders. The formulas used to calculate LDL-C showed varying performance in different populations. Machine learning models can study complex interactions between the variables and can be used to predict outcomes more accurately. The current study evaluated the predictive performance of three machine learning models—random forests, XGBoost, and support vector Rregression (SVR) to predict LDL-C from total cholesterol, triglyceride, and HDL-C in comparison to linear regression model and some existing formulas for LDL-C calculation, in eastern Indian population. Methods The lipid profiles performed in the clinical biochemistry laboratory of AIIMS Bhubaneswar during 2019–2021, a total of 13,391 samples were included in the study. Laboratory results were collected from the laboratory database. 70% of data were classified as train set and used to develop the three machine learning models and linear regression formula. These models were tested in the rest 30% of the data (test set) for validation. Performance of models was evaluated in comparison to best six existing LDL-C calculating formulas. Results LDL-C predicted by XGBoost and random forests models showed a strong correlation with directly estimated LDL-C (r = 0.98). Two machine learning models performed superior to the six existing and commonly used LDL-C calculating formulas like Friedewald in the study population. When compared in different triglycerides strata also, these two models outperformed the other methods used. Conclusion Machine learning models like XGBoost and random forests can be used to predict LDL-C with more accuracy comparing to conventional linear regression LDL-C formulas.

Quality improvement in clinical biochemistry laboratory using six sigma metrics and quality goal index

Introduction: One of the most popular quality management system tackle employed for process perfection is six sigma. When the process outcome is measurable, six sigma can be used to assess the quality. Aim: Present study was conducted with the objective to apply six sigma matrices and quality goal index for the assessment of quality assurance in a clinical biochemistry laboratory. Materials and methods: Present study is a retrospective study. Internal and external quality control data were analyzed retrospectively during July 2020 to December 2020. Descriptive statistics like laboratory mean ± standard deviation; bias and coefficient of variation (CV) were calculated for the parameters glucose, urea, creatinine, ALT (SGPT), AST (SGOT), cholesterol, triglyceride and HDL. Sigma value was calculated for both level I & level II of internal quality control (IQC). Results: Satisfactory sigma values (≥3) were elicited for blood glucose, cholesterol, triglyceride, HDL, urea and creatinine, while ALT (SGPT) and AST (SGOT) performed poorly (<3) on the sigma scale. The quality goal index (QGI) ratio was found (> 1.2) for only 2 parameters SGPT and SGOT (with sigma value <3) for both levels 1 and 2, indicating inaccuracy. Conclusion: Results of present study focuses on meticulous appraisal and execution of quality measures to improve sigma standards of all the analytical processes. Even though six sigma provides benefits over former approaches to quality assurance, it also opens newer challenges for laboratory practitioners. Therefore, sigma metric analysis provides a point of reference to design a protocol for IQC for the laboratory; address poor assess performance, and assess the existing laboratory process efficiency.

Serum lithium test requesting across three UK regions: an evaluation of adherence to monitoring guidelines

Background Bipolar disorder is the fourth most common mental health condition, affecting ~ 1% of UK adults. Lithium is an effective treatment for prevention of relapse and hospital admission, and is widely recommended as a first-line treatment. We previously showed in other areas that laboratory testing patterns are variable with sub-optimal conformity to guidance. We therefore examined lithium results and requesting patterns relative to monitoring recommendations. Methods Data on serum lithium levels and intervals between requests were extracted from Clinical Biochemistry laboratory information systems at the University Hospitals of North Midlands, Salford Royal Foundation Trust and Pennine Acute Hospitals from 2012 to 2018 (46,555 requests; 3371 individuals). Data were examined with respect to region/source of request, age and sex. Results Across all sites, lithium levels on many requests were outside the recommended UK therapeutic range (0.4–0.99 mmol/L); 19.2% below the range and 6.1% above the range (median [Li]: 0.60 mmol/L). A small percentage were found at the extremes (3.2% at < 0.1 mmol/L, 1.0% at ≥1.4 mmol/L). Most requests were from general practice (56.3%) or mental health units (34.4%), though those in the toxic range (≥1.4 mmol/L) were more likely to be from secondary care (63.9%). For requesting intervals, there was a distinct peak at 12 weeks, consistent with guidance for those stabilised on lithium therapy. There was no peak at 6 months, as recommended for those aged < 65 years on unchanging therapy, though re-test intervals in this age group were more likely to be longer. There was a peak at 0–7 days, reflecting those requiring closer monitoring (e.g. treatment initiation, toxicity). However, for those with initial lithium concentrations within the BNF range (0.4–0.99 mmol/L), 69.4% of tests were requested outside expected testing frequencies. Conclusions Our data showed: (a) lithium levels are often maintained at the lower end of the recommended therapeutic range, (b) patterns of lithium results and testing frequency were comparable across three UK sites with differing models of care and, (c) re-test intervals demonstrate a noticeable peak at the recommended 3-monthly, but not at 6-monthly intervals. Many tests were repeated outside expected frequencies, indicating the need for measures to minimise inappropriate testing.

Serum lithium test requesting across three UK regions: An evaluation of adherence to monitoring guidelines

Background: Bipolar disorder is the fourth most common mental health condition, affecting ~1% of UK adults. Lithium is an effective treatment for prevention of relapse and hospital admission, and is widely recommended as a first-line treatment. We previously showed in other areas that laboratory testing patterns are variable with sub-optimal conformity to guidance. We therefore examined lithium results and requesting patterns relative to monitoring recommendations.Methods: Data on serum lithium levels and intervals between requests were extracted from Clinical Biochemistry laboratory information systems at the University Hospitals of North Midlands, Salford Royal Foundation Trust and Pennine Acute Hospitals from 2012-2018 (46,555 requests; 3,371 individuals). Data were examined with respect to region/source of request, age and sex.Results: Across all sites, lithium levels on many requests were outside the recommended UK therapeutic range (0.4-0.99 mmol/L); 19.2% below the range and 6.1% above the range (median [Li]: 0.60 mmol/L). A small percentage were found at the extremes (3.2% at <0.1mmol/L, 1.0% at ≥1.4mmol/L). Most requests were from general practice (56.3%) or mental health units (34.4%), though those in the toxic range (≥1.4 mmol/L) were more likely to be from secondary care (63.9%). For requesting intervals, there was a distinct peak at 12 weeks, consistent with guidance for those stabilised on lithium therapy. There was no peak at 6 months, as recommended for those aged <65 years on unchanging therapy, though re-test intervals in this age group were more likely to be longer. There was a peak at 0-7 days, reflecting those requiring closer monitoring (e.g. treatment initiation, toxicity). However, for those with initial lithium concentrations within the BNF range (0.4-0.99 mmol/L), 69.4% of tests were requested outside expected testing frequencies.Conclusions: Our data showed: (a) lithium levels are often maintained at the lower end of the recommended therapeutic range, (b) patterns of lithium results and testing frequency were comparable across three UK sites with differing models of care and, (c) re-test intervals demonstrate a noticeable peak at the recommended 3-monthly, but not at 6-monthly intervals. Many tests were repeated outside expected frequencies, indicating the need for measures to minimise inappropriate testing.

Sigma Metrics: A Powerful Tool for Performance Evaluation and Quality Control Planning in a Clinical Biochemistry Laboratory

Introduction: Any error in the laboratory testing processes can affect the diagnosis and patient management. Six Sigma is a data driven quality management system for identifying and reducing errors and variations in clinical laboratory processes. Aim: This study was carried out to estimate Sigma metrics of various biochemical analytes in order to evaluate performance of quality control and implement optimum quality control strategy for each analyte. Quality goal index (QGI) was also calculated to identify the problems of inaccuracy and imprecision for parameters having lower sigma values. Materials and Methods: This retrospective, observational study was conducted at the Central Clinical Biochemistry Laboratory of Seth GS Medical College and KEM hospital in Mumbai for a period of six months (July 2019 to December 2019). Sigma metrics for 20 analytes was calculated by using internal quality control and external quality control data. Further, QGI was calculated for analytes having sigma value of <4 to identify imprecision or inaccuracy. Statistical analysis was performed using Microsoft office excel 2010 software. Results: Total protein, Glucose, Urea, Triglyceride (TAG), High Density Lipoprotein (HDL), and Low Density Lipoprotein (LDL) for normal (L1) and pathological (L2) controls achieved excellent performance (>6 sigma). Westgard rule (13s) with two control measurement (N2) per QC event and run size (R1000) i.e. 1000 patient samples between consecutive QC events was adopted for these analytes. For analytes with sigma value of 4-6, more rules (sigma 4-5: Westgardrules-13s/22s/R4s/41s, N4 and R200 and for sigma value 5-6: 13S/22s/R4s, N2 and R450) were adopted. The sigma values of six analytes (Creatinine, Sodium, Potassium, Calcium, Chloride, Inorganic phosphate) were <4 at one or more QC levels. For these analytes, strict QC procedures (Westgard rules-13s/22s/R4s/41s/6x, N4 and R45) were incorporated. QGI of these analytes was <0.8 which indicated the problem of imprecision. Staff training programs and review of standard operating procedures were done for these analytes to improve method performance. Conclusion: Sigma Metrics estimation helps in designing optimum QC protocols, minimising unnecessary QC runs and reducing the cost for analytes having high sigma metrics. Focused and effective QC strategy for analytes having low sigma helps in improving the performance of those analytes.

High preanalytical non-compliance and samples rejection rate in clinical biochemistry laboratory are decreased by nurse staff training in phlebotomy and sample handling

Background: Preanalytical phase of biomedical analysis remains an important source of diagnostic errors that deserves special attention. This study aims to evaluate the training in phlebotomy and sample handling impact on the preanalytical non-compliances. Material and Methods: we performed a prospective study before and after staff training in phlebotomy and sample handling by systematically recording all clinical samples non-compliances. First, we assessed and describe the non-compliance baseline rate from January to December 2017 in the clinical biochemistry laboratory of Hôpital Sominé DOLO de Mopti. After two sessions of one week staff training in January 2018, we performed the same study from January to December 2018. We compared the proportions of non-compliances between the two assessments. Data were collected on the case report forms, captured in Excel and analyzed by R software for (Mac) OS X version 4.0.3. Pearson Ch2 or Fisher exact tests were used for the comparison of proportions. The statistical significance was set at p < 5%. Results: a total of 27,810 venous blood samples were received during the study period; 48% was for biochemistry, 41% for immuno-serology, 9% for blood cell count and 2% for coagulation tests. There were 3,826 instances of preanalytical non-compliances (13.76%) identified that led to sample rejection. Out of the 11 types of non-compliances investigated, 5 (45.4%) accounted for nearly 91% of the problems: insufficient sample volume (28.9%), hemolyzed samples (20.5%), inappropriate collection time (17.8%), sample clot (12.9%), and inappropriate sample collection tube (10.8%). We observed a significant difference in rates of non-compliance between inpatients and outpatients samples (44.4% vs 7.3%; p < 0.001). The proportion of non-compliance have significatively decreased after the two training sessions of hospital staff in phlebotomy and sample handling 3,826/27,810 (13.8%) vs 3,009/32,476 (9.3%); p < 0.001. Conclusion: we report a significantly higher rate of non-compliance in inpatients. Hospital staff training in phlebotomy and sample handling reduce the proportion of preanalytical non-compliance and thereby improve patient management and safety.