Is a threshold-based model a superior method to the relative percent concept for establishing individual exercise intensity? a randomized controlled trial