A comparative study was conducted at Komfo Anokye Teaching Hospital in Kumasi, Ghana, to evaluate the concordance between fully automated urinalysis (Sysmex UN series) and manual methods. Sixty-seven fresh urine samples were analyzed using both methods. Kappa and Bland-Altman plot analyses assessed agreement and correlation.
Assessing Agreement and Correlation in Urinalysis Methods
Substantial agreement (κ = 0.711, p < 0.01) was observed for urine color, while slight agreement was found for appearance (κ = 0.193, p = 0.004) and pH (κ = 0.109, p < 0.001). Specific gravity showed a strong, significant correlation (r = 0.593, p < 0.001). Red blood cell, white blood cell, and epithelial cell counts exhibited strong positive linear correlations (r = 0.951, R2 = 0.904, p < 0.001; r = 0.907, R2 = 0.822, p < 0.001; and r = 0.729, R2 = 0.532, p < 0.001, respectively).
Nitrite demonstrated perfect agreement (κ = 1.000, p < 0.001), while protein showed fair agreement (κ = 0.395, p < 0.001). Cast presence exhibited strong agreement (κ = 0.734, p < 0.001). However, no concordance was observed for crystals (κ = 0.115, p = 0.326) or yeast-like cells (κ = 0.171, p = 0.116).
The Role of Manual Microscopy in Urinalysis
Automated and manual urinalysis methods demonstrated similar performance and good correlation, particularly for physical and chemical examination. However, manual microscopy remains essential for classifying urine sediments, especially bacteria and yeast-like cells. This highlights the importance of manual microscopy in identifying specific microscopic elements that automated analysis may not accurately differentiate. Automated systems excel in providing rapid, quantitative data for routine urinalysis parameters.
Future research with larger sample sizes could further validate automated urinalysis for broader clinical application and pinpoint areas needing improved automated detection capabilities. This comparative analysis underscores the ongoing need for both automated and manual methods in comprehensive urinalysis. While automated platforms offer efficiency in high-throughput settings, manual expertise is crucial for accurate interpretation and identification of certain urinary components.
