Is there a 'recent occupational paradox' where highly active physically active workers die early? Or are there failures in some study methods?