The Effort-Reward Imbalance model of psychosocial stress at the workplace—a comparison of ERI exposure assessment using two estimation methods

It is not unusual for old data to be used in epidemiological studies. Recently developed instruments for measuring work-related stress did not exist when the data collection was carried out. Therefore, approximate questions are sometimes used. An apparent problem is the lack of validation of proxy questions. The aim of this study was to compare the original questions for measuring Effort-Reward Imbalance (ERI) with approximate questions. The study population corresponded with a subgroup in the WOLF-n (WOrk, Lipids, Fibrinogen-north) cohort study of cardiovascular risk in a working population in the north of Sweden: 655 men and 178 women. The agreement in exposure between the original and the approximate ERI single questions was relatively low throughout, whereas the correlation between the ERI subscales and the ERI ratios was reasonable. The latter agreement between the original and the approximate ERI ratio indicates that the approximate measures might have been useful in the past. Yet, whenever possible the implementation of the original questionnaire in study protocols is recommended since a complete measurement of ERI might help to further improve the internal consistency and the predictive validity of this exposure to effort-reward imbalance.