Interest recently has extended to the problem of looking for HLA haplotype associations with particular diseases. Higher order interactions, that is to say of particular haplotypes with a disease, are much more complicated than individual antigen associations, with which most disease studies have so far been concerned. Our aim in this work is to explain the concept of higher order interactions and to discuss the difficulties involved when trying to measure such haplotype associations. The main emphasis is to point out that it is not easily possible to detect haplotype associations if an individual antigen association with the disease exists.