Laboratory data mining: monitoring of the local applicability of prolactin reference interval

This study aimed to explore the feasibility and methodology of long-term monitoring the local applicability of reference intervals (RIs) by statistically analyzing the positive rates derived from Laboratory Information System (LIS) data. The stability of the coefficient of variation (CV) of 2018–2020 positive rates was evaluated using 1/2 total allowable error (TEA). The ±1/2 TEA range around their mean was calculated to verify if the 2022–2024 positive rates were consistently outside this range. If so, prolactin results of first-visit patients with irregular menstruation were analyzed to exclude age and sample collection time effects, with discrepancies attributed to inappropriate RIs. Re -evaluation was performed using the local RI upper limit, with a reverse correlation analysis between the two periods' data. The 2018–2020 CV was 11.41% (within 1/2 TEA). The 2022–2024 positive rates exceeded the ±1/2 TEA range but fell within the acceptable scope after applying the new RI. Age and sample collection time were ruled out as discrepancy causes. Both periods' positive rates mutually conformed to each other's ± 1/2 TEA ranges. The range of ±1/2 TEA around the mean positive rate, derived from LIS data analysis, is applicable for the long-term evaluation of the local applicability of reference intervals after detection system transition. Can the reference interval be subjected to long-term monitoring in the same way as internal quality control? Study on the Value of Positive Rate in Laboratory Information System Data. • Addressing RI applicability monitoring post-transition via LIS data • Verifies pre-transition rate mean ± 2SD serves as a valid acceptable ranges.