Urine dry reagent strip “error” rates using different reading methods

Abstract The need for “quality” in near patient testing (NPT) has been acknowledged since the mid 1980s. The commonest biochemical NPT device is the dry reagent strip or “dipstick” for urinalysis. Dipsticks may be read in three ways, against the color chart printed along the side of the bottle, using a benchreader (the color chart printed on a flat card) or using an electronic reader. This report uses the results of a urinalysis quality assurance (QA) program, over 1998, to evaluate the “error” rates which occur using the three different reading methods. The QA samples are buffered aqueous solutions which are “spiked” to give concentrations midway between two color blocks for each analyte. Results are scored as ±1 if a color block adjacent to the target value, ±2 for results two color blocks (defined as “error”) and ±3 for results three color blocks (defined as “gross error”) from the target value. Analysis of the results show that the error rates are similar reading visually by either method, but greatly reduced when read electronically. Some persisting errors when using the electronic reader are explained by observation studies. The study highlights the value of a urinalysis QA program for NPT urinalysis in understanding the error rates of this simple but ubiquitous test.