In my experience, the value of an ISR depends partially on the International Searching Authority (ISA) and partially on the style of the claims. Many applicants choose to file broad-scope claims at the international phase, aiming to restrict them to novel subject matter during the various national phases.
That being said, it is a fact that some authorities are better at scrutinizing than others. EPO examiners usually spot even the oldest prior art. Japanese examiners are very skilled searchers as well and as you can imagine they retrieve valuable prior art from old japanese patents, which others might miss for obvious reasons. I suppose this relates also to the availability of search databases in each patent office. Chinese examiners are particularly good in analyzing details, so their comments about dependent claims are extremely useful.
Regarding the usefulness of an ISR, I have asked this question myself a thousand times. I ended up thinking that, for big enterprises they are not particularly useful, they have large departments which already know all the available prior art. For middle or small businesses, an ISR assures you that you have done your homework well (or not!), because most likely they do not have all the resources available, to do an in-depth prior art search. And don't forget that an ISR is useful to a competitor monitoring patents related to his products: he quickly aqcuires knowledge of relevant prior art, for a given patent application. He is in a position of evaulating the situation with less effort.