• occupational health
• public health surveillance
• transients and migrants

Surveillance data are essential to identify and target prevention for all public health activity, including occupational safety and health. Accurate and timely surveillance data are needed to identify new hazards, monitor prevention activity, plan interventions and to evaluate the efficacy of these interventions. All occupational health specialists need to understand the limitations of the surveillance data available to them in order to make the best use of the data in their work.

The inadequacies of the surveillance system for nonfatal work-related injuries and illnesses in the USA have been well documented.1 On the other hand, some European countries with their national compensation programmes (eg, the Nordic countries) and labour force surveys (eg, UK) have been looked to as models to follow to improve non-fatal work-related injury and illness surveillance.1 The paper by Kreshpaj et al in this current issue2 and the previous recent publication from the same authors3 show the Swedish system has deficiencies in providing accurate surveillance data that, not surprisingly, are even greater in precarious workers without job security. The Swedish studies used capture–recapture methodology to look at overlap between two databases: Information System on Occupational Injuries (ISA), that has data on work-related injuries and illnesses reported by employers; and data from the AFA, a mutual insurance company, which mainly insures employees covered by collective …