The study is clearly explained but suffers from a few methodological problems:
1) the quality of life of patients at hospices compared to those at hospital was ignored, particularly with regard to location of death, and pain endured in final months;
2) there was no randomisation of patients to the treatment regimes, although this is partly compensated for by the matching of cases to controls;
3) of greater concern is that there appeared to be no difference in costs when subgroups of similar patients were considered. However, the authors justified this on the basis of the relatively small number in the study subgroups.