Table 1 Quality assessment criteria

A specific objective has been clearly formulated.

The response rate is at least 80%.

The response rate is at least 80% after at least one year or the non-responders are not selective.

Cases and controls originate from the same population and there is a clear definition of cases and controls.

The original ERI questionnaire was used, including the ascertainment of over-commitment.

Data on physical work-related stress was collected and taken into account in the analysis.

Prospective enrolment was used (identification of new cases and selection of controls occurs at the same time).

Follow-up completed at least one year later.

Risk predictors including confidence intervals or p-values were calculated and confounding was controlled for.