Accuracy of daily fluid intake measurements using a “smart” water bottle

Michael S. Borofsky, Casey A. Dauw, Nadya York, Colin Terry, James E. Lingeman

High fluid intake is an effective preventative strategy against recurrent kidney stones but is known to be challenging to achieve. Recently, a smart water bottle (Hidrate Spark™, Minneapolis, MN) was developed as a non-invasive fluid intake monitoring system. This device could help patients who form stones from low urine volume achieve sustainable improvements in hydration, but has yet to be validated in a clinical setting. Hidrate Spark™ uses capacitive touch sensing via an internal sensor. It calculates volume measurements by detecting changes in water level and sends data wirelessly to users' smartphones through an application. A pilot study was conducted to assess accuracy of measured fluid intake over 24 h periods when used in a real life setting. Subjects were provided smart bottles and given short tutorials on their use. Accuracy was determined by comparing 24-h fluid intake measurements calculated through the smart bottle via sensor to standard volume measurements calculated by the patient from hand over the same 24 h period. Eight subjects performed sixty-two 24-h measurements (range 4-14). Mean hand measurement was 57.2 oz/1692 mL (21-96 oz/621-2839 mL). Corresponding mean smart bottle measurement underestimated true fluid intake by 0.5 ozs. (95% CI -1.9, 0.9). Percent difference between hand and smart bottle measurements was 0.0% (95% CI - 3%, 3%). Intraclass correlation coefficient (ICC), calculated to assess consistency between hand measures and bottle measures, was 0.97 (0.95, 0.98) indicating an extremely high consistency between measures. 24-h fluid intake measurements from a novel fluid monitoring system (Hidrate Spark™) are accurate to within 3%. Such technology may be useful as a behavioral aide and/or research tool particularly among recurrent stone formers with low urinary volume.