Do preferences to reduce health risks related to air pollution depend on illness type? Evidence from a choice experiment in Beijing, China