Implications of experimental isospin mixing data on the spreading width