Comparison of Patient-based Real Time Quality Control Programme with Conventional Quality Control Program for Serum Electrolytes.

Abstract

Background: Patient-based real time quality control (PBRTQC) programme uses patient data to monitor the quality of test results in a clinical laboratory. It can be used as a cost-effective method to detect systematic errors in the analytical phase supplementing the conventional internal quality control (IQC) programme. The research data is sparse for PBRTQC in Sri Lanka.

Objective: To implement and compare PBRTQC program with conventional IQC program for serum electrolytes (Na+ and K+), at Teaching Hospital Karapitiya

Methods: Patient data within the reference ranges were extracted from the laboratory information system (LIS) for a period of one year retrospectively (76 549 data for Na+ and 109048 for K+) and used to calculate the target values; mean (M) and standard deviation (SD) for the two analytes. PBRTQC plots were constructed for serum Na+ and K+ separately, using reference ranges as truncation limits. Two methods were selected to set the control limits of the plots; maximum and minimum of Moving Average (MA) values and 3SD limits. Patient data (6281) were reviewed prospectively for two weeks utilizing the PBRTQC system for both analytes. The block size of the MA was taken as 80 based on sample size calculation. Number of rejected test results and bias of each analyte were compared with the error-detection rate of the conventional IQC. Calculations were done using SPSS version 22.0. All data were descriptively analyzed using constructed PBRTQC plots.

Results: The calculated target value (population M) for serum sodium and potassium were 139.5 (3SD 5.1) and 4.2 mmol/L (3SD 0.78), respectively. Maximum and minimum MA values for serum sodium and potassium were (145-136) mmol/L and (5.2-3.5) mmol/L, respectively. The number of rejected test results for serum Na+ in MA-QC and SD-based PBRTQC were 2032 (31.6%) and 273 (4.3%), respectively. A bias of 2.15% was observed for serum Na+ results. Analysis of serum K+ results using PBRTQC program neither gave any bias nor alarms. There were no conventional IQC failures reported for both analytes within the two weeks.

Conclusions: Although conventional IQC did not detect any errors, PBRTQC raised QC alarms for serum Na+ based on MA and 3SD limits. The QC limits and alarms being affected by patient specific factors are a limitation. The causes for QC alarms by PBRTQC need further scrutiny to refine this tool for optimization.