Global Burden of Disease estimates of depression--how reliable is the epidemiological evidence?