We use the Kotliar and Ruckenstein slave-boson mean-field representation for the periodic Anderson model and the Hartree-Fock approximation for the Kondo lattice model to determine the phase boundary between antiferromagnetic and paramagnetic states. A comparison is made with the quantum Monte Carlo and spin-density-wave mean-field results. We find that at the mean-field level the Kotliar and Ruckenstein slave-boson formalism is better than the spin-density-wave approach, but both deviate from the quantum Monte Carlo result at large hybridization V, where the Hartree-Fock approximation for the Kondo lattice gives a better description.