A study in which the objective is to uncover cause and effect relationships but in which it is not feasible to use controlled experimentation, in the sense of being able to impose the procedure or treatments whose effects it is desired to discover, or to assign subjects at random to different procedures.