Example of use of clock time-dependent targets for patient-based quality control.

Abstract

Running means for total calcium (Ca) results at our laboratory exhibit a stable time-of-day (TOD) periodic pattern. We examined use of TOD-dependent targets for running means in patient-based quality control (PBQC) for Ca. Primary data were Ca results over a 3 month interval, restricted to weekday data within the Ca reference interval (8.5-10.3mg/dL; 2.12-2.57mmol/L). Running means were evaluated as sliding averages of 20 samples (20-mers). Data comprised 39,629 consecutive Ca measurements (75.3% inpatient (IP)) for which Ca was 9.29±0.47mg/dL. The all data average for 20-mers was 9.29 ± 0.18mg/dL. When parsed in 1h TOD intervals, however, averages among 20-mers ranged from 9.1 to 9.5mg/dL, with blocs of contiguous results above (0800-2300 h; 53.3% of results; IP = 75.3%) and below (2300-0800 h; 46.7% of results; IP = 99.9%) the all-data mean. There was thus an inherent TOD-dependent pattern of deviation of means from target when using a fixed PBQC target. Using Fourier series analysis as an example approach, characterization of the pattern to produce TOD-dependent PBQC targets eliminated this inherent inaccuracy. In circumstances of periodic variation in running means, simple characterization of that variation can reduce the probability of both false positive and false negative flags in PBQC.