Automated in-toilet hydration sensor for urinalysis on participants during a four-day prolonged walking exercise event: Prototype validation study

Abstract

Underhydration is a common problem in elderly, which can lead to serious health complications if left untreated. However, currently, there is no sufficient, non-obtrusive method to monitor underhydration status automatically. Therefore, a prototype was developed that automatically measures urine concentration during a toilet visit. This study validates this prototype for its capability to repeatedly and accurately measure the urine of 106 participants participating in a 4-day prolonged walking exercise event. The prototype contains 4 sensors that measure the color, conductivity, pH, and temperature of the urine sample while the participant uses the toilet. In total, the prototype analyzed 514 urine samples. In addition, the urine was automatically collected to measure the gold standard for urine concentration, urine specific gravity (USG), and osmolality. With a linear regression classification model and the USG reference, the data collected with the prototype was classified. The measured reference USG values range between 1.0025 and 1.0345. The prototype measurements strongly correlated to the USG reference, with an R-squared of 0.85 and a mean absolute error of 0.00215. As such, this study reports on a method that allows automated, repeated, and accurate urinalysis during a toilet visit. When used daily, this prototype offers potential for the non-obtrusive and time-effective underhydration estimation by monitoring urine.