Number of pain sites – a simple measure of population risk?