The biggest glaring issue with it (eg in the form (square^2+m^2)\psi=0) is that it is a manifestly Lorentz invariant equation in which particle number is conserved (which is highly unlikely for any relativistic interaction). I know that you can extend it into a scalar field theory proper, quantize it, and sidestep around those issues (and use it as a model in cmp!), but the bigger problem I think is that you really need spin -- and ideally all other interactions...