Should data on gender and ethnicity inform ergonomics interventions? Lessons from four case studies