Do you know how to recover a backup? Have you tested your ability to recover a backup? Have you tested that a recovered backup is valid?
If no? Being able to answer yes to all three of those questions is a much more productive use of time than thinking about a version upgrade without a concrete reason to consider.
If yes… I have a middling-strong recommendation against it. OpenEMR runs in the Dockerized environment, and won’t enjoy any benefits from the upgrade, and the AWS tooling is tested only on and specifically against Ubuntu 16.04; while I don’t see what would break it may well be that a backup created on a Ubuntu 18 system will fail to recover on a Ubuntu 16 instance.
Proving this true or false won’t be hard… as long as you have a parallel instance you can test your backup and recovery process on, and are familiar enough with the process to be able to test it on terms of your own choosing – changing the variables, seeing what happens, and knowing enough about the process to catch a non-obvious fault, and recover from one if a fault is discovered.
In truth, the odds are very good that the upgrade will go flawlessly, but if it /does/ introduce a problem I’m dreadfully certain it’ll be in the interaction between the host system and the container, instead of something within the application container, and the pool of people able to help you diagnose that is… thinner than you might prefer.