Reliability and minimal detectable change of the MX3 hydration testing system

Abstract

Assessing hydration status outside of laboratories can be challenging given that most hydration measures are invasive, stationary, costly, or have questionable validity. This study investigated the within-day, test-retest reliability, and minimal detectable change (MDC) of the MX3 Hydration Testing System (MX3 Diagnostics), a relatively low cost, noninvasive, and portable method to measure saliva osmolality. Seventy-five adults (44 men, 31 women; 29.6±10.8 yr, 171.1±9.2 cm, 79.1±15.4 kg) presented two saliva samples approximately 3 to 5 minutes apart. Fluid intake was avoided for at least 5 minutes prior to sample collections. For each sample collection, a researcher used the MX3 to tap the tip of a test strip to saliva on a participant's tongue. Intra-class correlation coefficients (ICCs) and MDC were calculated as measures of reliability. Means for the two measurements were 64.3±19.0 mOsm and 65.5±17.9 mOsm (mean difference of -1.2±13.0 mOsm; t = -0.806, p = 0.423). Further analysis with a two-one-sided test procedure revealed equivalence between the two SOSM measurements (p < 0.001 for upper and lower bounds). Within-day repeat measures yielded an ICC of 0.75 and an MDC at the 90% confidence level of 21.3 mOsm. With moderate-to-good reliability, the MX3 appears to be a practical choice for reliably measuring moderate-sized changes (>20 mOsm) in saliva osmolality outside of laboratory constraints.

Fig 1. Bland-Altman plot showing the individual differences in SOSM measurements 1 and 2 by the average of the two measurements.

The mean difference (black), 95% limits of agreement (dashed), and regression (grey) lines are shown.

Fig 2. Scatterplot of the two SOSM measurements with a regression line (black) and a line of identity (grey).