Messaging patterns for a plurality of subscribers are obtained and analyzed to determine a “trust score” that is an indication of the likelihood that a given message for a particular subscriber is of interest, as opposed to unwanted, e.g., spam or telemarketing phone calls. Subscribers establish or set trust score thresholds and call processing actions to take based on the thresholds and the trust score for a given incoming communication. For example, if a subscriber establishes a processing action of “block call” for calls with a “low” threshold trust score, and an incoming message (email, SMS, instant message, or phone call) has a trust score of “low” or “very low”, the message is blocked. Conversely, if the trust score for the message is “high” and the user has specified to pass through “high” trust score messages, the message is passed through. Various algorithms can be used compute trust scores based on messaging patterns.