Interdisciplinary research is being promoted in many quarters as the way forward, but "research islands" still persist. Taking computer-assisted learning (CAL) within health sciences as a case in point, this paper describes a detailed study of the references to source material within papers published in general medical, specific nursing and general information and communications technology journals as pointing to papers published either in the same category of journal or journals of other categories. The results show that research within this area exhibits a disquieting inbred approach. Furthermore, there is clear evidence of a tendency amongst the papers examined to cite research published in journals with broadly the same impact factors as the journal in which the papers themselves are published. A discussion of the value of journal impact factors, which are a crude but useful index of the quality of journal, and their possible effect on the future of research in CAL within health sciences is embedded in the paper.