Yes, we scan all messages leaving our network for SPAM and have partnered with a leading anti-SPAM provider, SpamExperts, in order to help ensure the system is as accurate as possible.  SpamExperts has been proven to be more than 99.9% accurate in our testing conducted over a period of more than 6 months but unfortunately it is not completely infallible and will sometimes have false positives and false negatives.

There are two generalized methods of controlling spam:
1. Proactive - scanning outbound messages for spam using industry trusted algorithms and signatures.
2. Reactive - allowing all messages out and taking action when spam is noticed or reported.

The largest benefit to proactive scanning is that when brought to our attention a failed message can be released as fast as within a minute or two.  Since *our scanner* is the one holding the message in quarantine we are able to release it immediately to its intended recipient.  The largest downsides to the scanning system are admin-facing such as maintaining and updating the systems which isn't really a large deal as we already have administrators handling our existing infrastructure.

The largest drawback to handling spam in a reactive fashion is that we generally won't be aware of an issue until damage do the IP reputation and email deliver-ability is damaged - possibly permanently.   In some cases these issues can be resolved as fast as in 24 to 48 hours or as long as never.  It is up to the organization(s) that are blocking the IP address as to whether they will unblock or not and when.  In some cases, such as with Yahoo Mail, contacting the postmaster is similar to talking to a brick wall.

The second largest result of reactive handling of SPAM is that any failures/blocks will affect everybody on the whole server and not an individual message from an individual account.  It takes longer to resolve and affects more users - it is not the ideal way to handle things but it is, admittedly, the cheapest way.  We do spend a fairly substantial amount of funds on both the infrastructure for our scanning servers as well as the licensing for the software that runs on it.

We do understand that you obviously want all of your legitimate outbound messages to pass without error and we want that as well.  In the event that you ever have a legitimate message that you've sent that you got a bounce for let us know immediately.  It is helpful if you can provide the following details [as much as possible so we can find it quickly and release it]:
1. The sender email address [usually yours].
2. The recipient email address(es).
3. The exact subject line of the message.

We can likely find any stuck/quarantined messages with just one piece of data such as sender or subject but the more information you provide on the message the easier it is to narrow down the logs from millions of messages per hour to your one individual failed message.