I'm not sure about re-writing the Galaxy formation model, just yet.
Their base data is derived from looking at low metallicity regions in only three z∼3 redshifted galaxies, capable of sustaining high star formation rates without frequent mergers with other galaxies.