Prospective, observational study comparing automated and visual point-of-care urinalysis in general practice

Machine learning-assisted image label-free smartphone platform for rapid segmentation and robust multi-urinalysis

This study presents a groundbreaking approach for the early detection of chronic kidney disease (CKD) and other urological disorders through an image-label-free, multi-dipstick identification method, eliminating the need for complex machinery, label libraries, or preset coordinates. Our research successfully identified reaction pads on 187 multi-dipsticks, each with 11 pads, leveraging machine learning algorithms trained on human urine data. This technique aims to surpass traditional colourimetric methods and concentration-colour curve fitting, offering more robust and precise community screening and home monitoring capabilities. The developed algorithms enhance the generalizability of machine learning models by extracting primary colours and correcting urine colours on each reaction pad. This method's cost-effectiveness and portability are significant, as it requires no additional equipment beyond a standard smartphone. The system's performance rivals professional medical equipment without auxiliary lighting or flash under regular indoor light conditions, effectively managing false positives and negatives across various categories with remarkable accuracy. In a controlled experimental setting, we found that random forest algorithms, based on a Bagging strategy and applied in the HSV colour space, showed optimal results in smartphone-assisted urinalysis. This study also introduces a novel urine colour correction method, significantly improving machine learning model performance. Additionally, ISO parameters were identified as crucial factors influencing the accuracy of smartphone-based urinalysis in the absence of additional lighting or optical configurations, highlighting the potential of this technology in low-resource settings.

Assessment of inter-observer agreement among experienced and inexperienced observers and an automated device for dipstick urinalysis in dogs and cats

Standard visual urine dipstick analysis (UDA) is performed routinely in veterinary medicine; results can be influenced by both the operator and the method. We evaluated the agreement of results for canine and feline urine samples analyzed using a 10-patch dipstick (Multistix10SG; Siemens), both visually under double-anonymized conditions by students and a laboratory technician, and with an automated device (AD; Clinitek Status, Siemens). The mean concordance for semiquantitative urinalysis results between students and the technician and between students and the AD was fair (κ0.21-0.40) in dogs and cats; concordance was moderate between the technician and the AD (κ0.41-0.60) in dogs and good (κ0.61-0.80) in cats. For pH, the mean concordance between students and the technician and between the technician and the AD was good (ρ0.80-0.92) in dogs and cats; concordance was good between students and the AD (ρ0.80-0.92) in dogs and moderate (ρ0.59-0.79) in cats. Repeatability was higher (p < 0.001) for the technician and the AD than for a student. We found good agreement between UDA performed by an experienced operator and an AD in dogs and cats but found low reproducibility and low repeatability for urinalysis performed by an inexperienced operator.

Comparative Evaluation between Visual and Automated Dipstick Urinalyses in Dogs

Urine test strips are commercially available and can be assessed with semi-automated analyzers or by visual assessment. This study aimed to compare the visual and automated evaluations of dipstick variables in canine urine samples. One hundred and nineteen urine samples were evaluated. Automated analysis was performed on a veterinary urine analyzer URIT-50Vet (URIT Medical Electronic) with UC VET13 Plus strips. Multistix 10 SG dipsticks (Siemens Healthcare GmbH, Erlangen, Germany) were used for visual evaluation, along with a refractometer (Clinical Refractometer Atago T2-Ne, Atago Co., Tokyo, Japan) for urine specific gravity measurements. A linear relationship was observed between the pH measurements (p = 0.2) of the two methods; the Passing-Bablok procedure was valid since neither proportional nor systematic significant errors were observed. Comparing the two methods, the correlation for urine specific gravity was poor (p = 0.01, CI 0.667-1.000). Moderate agreement was demonstrated for proteins (κ = 0.431), bilirubin (κ = 0.434) and glucose (κ = 0.450). Agreement was substantial for blood (κ = 0.620) and poor for leukocytes (κ = 0.100). Poor agreement was observed for ketones (κ = -0.006). Apart from the pH analysis, visual and automated dipstick urinalyses should not be used interchangeably. Multiple urine samples obtained from the same dog during the day should be evaluated using the same method to overcome erroneous results.

A dataset examining the impact of direct electronic medical record interfacing on the accuracy of point-of-care urinalysis results

Point-of-care testing is widely used in a variety of clinical settings. While this testing provides immediate and actionable clinical information, it is prone to error in both the interpretation and reporting of results. Point-of-care urinalysis presents unique opportunities for errors, ranging from variation in visual interpretation to input of results. The data included here represent the results from 63,279 urinalyses from 36,780 unique patients performed over a span of three years at an academic medical center and its associated clinics. The data include the patient age/legal sex, methodology (instrument and test strip used), and the available test results (color, clarity, glucose, bilirubin, ketones, specific gravity, blood, pH, protein, urobilinogen, nitrite, and leukocyte esterase). Additionally, we include the method of interface between the testing instrumentation and our electronic medical record (EMR). These fell into one of three broad categories: "Interfaced" (results directly transmitted from the urinalysis instrument to the EMR via specialized data interface), "Manual" (results input by selecting from a drop-down menu in the laboratory information system), and "Enter/Edit" (results typed freely into a text field in the EMR). Analysis of this data was primarily a direct comparison of detectable errors (typos, uninterpretable results, and results outside the reportable range) as a function of the method of entry into the EMR. Secondary analysis comparing the impact of restricting drop-down menu options for urine color and clarity was also performed. These data are of use to others as they are diverse in terms of the test performed and the method of interface. Others may wish to analyze these data when making decisions as to how to perform and report these tests and when estimating risks of error with various methods of data entry.

Bacteriuria resistance patterns and the suitability of urinalysis as an initial diagnostic tool in a post-antibiotic era

AIMS

Urinalysis is used as a first-line investigation throughout healthcare to indicate bacteriuria and guide treatment of potential urinary tract infections. In light of rising bacterial multi-resistance, we aim to analyse its diagnostic accuracy, determine its usefulness in a present-day setting and evaluate current antibiotic resistance patterns across a Trust population.

METHODS

A retrospective case series of 712 paired urinalysis and urine culture results was obtained over a 1-month period. Sensitivity, specificity and diagnostic accuracy were calculated, and resistance profiles of commonly used Trust antibiotics assessed using statistical analysis.

RESULTS

A high false negative rate of nitrites on urinalysis, with sensitivity of 38.4%, was found. Leucocyte sensitivity was 87.6% and specificity 39.7%, with no improvement in diagnostic accuracy seen when combining both. Positive urine culture growth demonstrated a substantial resistance pattern to trimethoprim of 48%, compounded by a statistically significant correlation with gentamicin resistance (p < 0.0001).

CONCLUSION

Our study has highlighted a reduced accuracy of urinalysis compared with previous literature, questioning its usefulness in the real world. We have consolidated growing published trends doubting the efficacy of trimethoprim, revealing co-existing resistance patterns between commonly used antibiotics. This will have implications for future antibiotic-prescribing protocols and requires further research to ensure guidelines are progressive in consciously managing this growing concern in modern-day healthcare.

Optimizing the u411 automated urinalysis instrument for veterinary use

BACKGROUND

The Cobas u411 Analyzer (Roche Diagnostics) is an automated, reflectance photometry-based urinalysis instrument designed for use with Roche's CHEMSTRIP 10UA technology and human urine samples.

OBJECTIVE

We aimed to optimize and validate the Cobas u411 Analyzer for use in canine and feline urinalysis.

METHODS

Patient urine samples presenting to the Clinical Pathology Laboratory at the Colorado State University Veterinary Teaching Hospital were analyzed with the Cobas u411 and by manual readings in parallel. Initially, 223 canine and 83 feline urine samples were run using the u411 factory settings. Following comparisons with manual results, and evaluation for directional bias, adjustments to the reflectance values were made in the instrument's programming. An additional 183 canine and 95 feline samples were run using the adjusted settings. Total urine protein concentrations were measured in 48 samples and used to generate receiver operating characteristic curves for the protein test pad.

RESULTS

Following adjustments in reflectance programming, concordance between u411 and manual results was increased by 17.7% for protein, 11.7% for ketones, and 4.5% for bilirubin. Concordances for pH, glucose, and blood were not substantially changed. Discordance for all analytes was ≤3%. Canine and feline samples had similar levels of discordance, though marginal concordance was higher in dogs for ketones, bilirubin, and blood.

CONCLUSIONS

Adjustments to the reflectance programming of the Cobas u411 Analyzer improved concordance with manual results for canine and feline samples. This instrument has the potential to greatly increase both efficiency and consistency of urinalysis procedures in higher throughput veterinary diagnostic laboratories.

Dipstick analysis of urine chemistry: benefits and limitations of dry chemistry-based assays

Urinalysis is a commonly utilized laboratory test, and analysis of urine has been studied and used since ancient times. Urine contains a wide array of metabolites that can provide information regarding the current physiologic state of the body and clinical manifestations of disease. In this review, we discuss the mechanics of the dry chemistry component of the urine dipstick such as the reaction principles underlying various assays and potential effects of collection and storage on results. Additionally, we discuss the benefits and limitations of the urine dipstick as it pertains to its use as a low-cost tool in point-of-care settings and the reasoning for a lack of its use as a broad screening tool.

Evaluation of the Clinitek®, a point-of-care urinalysis system for the measurement of clinically significant urinary metabolites and detection of haematuria in Schistosoma haematobium infected children in southern Côte d'Ivoire

BACKGROUND

Urinary schistosomiasis, caused by Schistosoma haematobium, remains a significant public health problem worldwide, despite years of efforts to control it. Haematuria is one of the notable indirect indicators of S. haematobium infection and is commonly assessed along with other routine screens using a urinary dipstick test. A portable "field friendly" electronic analyser would offer an automated and thus more objective read-out compared to visual-read dipstick methods.

METHODS

Within the framework of a Phase 2 praziquantel dose finding study in preschool- and school-aged children infected with S. haematobium, in southern Côte d'Ivoire, we compared a visual-read of the urine dipstick strips (Multistix PRO, Siemens Healthcare Diagnostics) to an automated reader (CLINITEK Status+ analyser™ Siemens Healthcare Diagnostics). Urine samples were collected from 148 pre-school aged and 152 school-aged children for urinalysis. Values were compared using a linear weighted kappa statistic and Bland-Altman analysis.

RESULTS

A very good correlation between the two methods for nitrites and haematuria was observed (κ coefficient of 0.88 and 0.82, respectively), while a good correlation was observed for leukocytes (κ coefficient of 0.63) A moderate to fair correlation was calculated (κ coefficient ≤ 0.6) for all other parameters. When the results were stratified according to infection intensity, the agreements were stronger from the high infection intensity sample measurements, for most of the parameters.

CONCLUSION

Our results demonstrate the device's utility in detecting haematuria and nitrites but underline the need for further development of this tool in order to improve its performance in the field.