From: Evidence for measurement bias of the short form health survey based on sex and metropolitan influence zone in a secondary care population