Considerations on the calculation of fractions of cardiovascular disease attributable to psychosocial work factors