spam issues and spamassassin running

[IsaacT]

I have spamassassin setup on my plesk 10.4 server, and it seems to be running. The bayesian filter is setup to learn nightly for each user account, and I have verified that it is learning. But still spam seems to be a large issue. My account is fine, as it is not signed up anywhere so it hasn't gotten onto any spam lists. However other users are seeing large volumes. One person on average gets 100+ spam a week that is not blocked. Spam score limit is set to 5. However the majority of the email that they get shows up in the ~40% bayesian range so doesn't score very high and isn't marked as spam.

I think the problem is this one mailbox wasn't trained at first, so there are probably a lot of emails that were trained as ham that were in fact spam. Here are the stats for the bayes DB for this mail account.

0.000 0 3 0 non-token data: bayes db version
0.000 0 57565 0 non-token data: nspam
0.000 0 28247 0 non-token data: nham
0.000 0 204301 0 non-token data: ntokens
0.000 0 1366147191 0 non-token data: oldest atime
0.000 0 1366842727 0 non-token data: newest atime
0.000 0 1366826859 0 non-token data: last journal sync atime
0.000 0 1366817818 0 non-token data: last expiry atime
0.000 0 0 0 non-token data: last expire atime delta
0.000 0 0 0 non-token data: last expire reduction count

I'm not sure what might help to get bayes to identify spam better, any thoughts?

I have also noticed in the headers that I see this in a lot of emails.

0.0 URIBL_BLOCKED ADMINISTRATOR NOTICE: The query to URIBL was blocked.
* See http://wiki.apache.org/spamassassin/DnsBlocklists#dnsbl-block
* for more information.

I'm not sure how to go about dealing with this issue of being blocked. our mail server does not go through enough email to be blocked as far as I know, so any information about this would be greatly helpful.

If anyone has any other pointers for how to deal with spam I would greatly appreciate it.

thanks