Work-related diseases--some problems in study design.