Inappropriate prescription of antibiotics for abnormal urine tests increases the risk of antimicrobial resistance in the elderly patient population.
Aims: To compare the performance of manual urine dipsticks versus an automated/ validated machine.
Methods: A prospective audit of urinalyses taken from a cohort of 64 adult patients suspected of a urinary tract infection in the Emergency Department (ED) over a 2-month period in May 2022. Urine samples were analysed, comparing manual versus automated urine dipstick outcomes. Leukocyte/ nitrite/ blood parameters were measured on manual and then on the automated machine for each patient. The automated machine which was validated was the reference standard for this study.
Results:
|
|
Nitrite |
Leukocytes |
Blood parameters |
|
Sensitivity1 |
1.00 |
0.93 |
1.00 |
|
Specificity1 |
0.9 |
0.6 |
0.8 |
|
FPR2 |
8% |
39% |
23% |
|
FNR3 |
0% |
7% |
0% |
|
Accuracy |
94% |
81% |
89% |
|
Agreement4 |
0.8 |
0.6 |
0.8 |
1 The results are based on the categorising all machine and dipstick method outputs to dichotomous positive and negative.
2 Where dipstick gave a positive of any magnitude, but machine gave negative
3 Where Dipstick gave negative, but machine gave positive of any magnitude
4 Cohen’s kappa
Conclusion: Our results show that the manual urine stick method is equivalent to the automated method for nitrites but its specificity for leucocytes and blood is low, false positive rate for nitrates was 8% and for leucocytes it was 39%. This suggests shortcomings in the manual dipsticks which could potentially over diagnose and over treat infections thereby contributing to poor antimicrobial stewardship and the rise of antimicrobial resistance.