In short, what you’re asking for is most likely close to impossible; patch relevance only considers if the patch is installed or not, not how it was installed. It could have been deployed manually through running a downloaded MSU, it could have been installed by Windows Update or SCCM, or it could have been installed by BigFix; in all of these cases, the patch relevance would be considered False if the patch successfully installed.
Action reports in BigFix are useful in the short term (we use them to track success rate and pinpoint possible errors through Exit Codes), but for long term tracking they will be useless if, like us, you stop and delete your obsolete patching actions.
The only surefire way I’m aware of is to parse the Windows Update logs, as these will contain the source application and the source location of the installer. Considering Windows 10 no longer has a live Windows Update log, and even if it did parsing thousands and thousands of lines of logs is not exactly efficient, I consider this approach a non-starter.
Basically, I think it would be wise to re-evaluate the reasons why you believe you need this level of reporting and examine other options. For example, we track when Windows Update was last used to check for, download, and install updates; at least this way we can tell if someone used WU to patch their system.