highly significantP < 0.001
The log-likelihood ratio test comparing this two-piecewise model to a single-slope linear model was highly significant ( P < 0.001), confirming that the two-piecewise model provided a substantially better fit to the data than a one-piece linear model.