replication lag versus consistency in distributed databases

heey everyon, i've been having a bit of a time with replication lag in our distributed database lately. it's a primary-replica setup, and reads are super quick, but we're noticing some pretty clear delays before writes actually show up on all the replicas. this is starting to mess with features that need data to be pretty much the same everywhere, right away. how do you guys handle wanting strong consistency and dealing with the lag that just comes with distributed systems? are there specific ways you set things up or patterns you like when this gets to be a big deal