With occasional exceptions (every manufacturer has a bad run of drives now and then, and you will also get the occasional random one-off defective drive), hard drives will generally last until they are obsolete provided they are treated reasonably well. This means no wide variations in temperature/power/humidity, not subjecting them to mechanical shock, and not power cycling them any more than necessary. Using a non-enterprise rated drive in an application where it gets hammered 24x7 is probably not a good idea either.
MTBF numbers are statistical in nature, and generally meaningless for real world usage unless you're running a large data center. E.g., WD claims a MTBF of 1 million hours for the WD Red series. 1,000,000 hours is over 100 years! Nobody is actually expecting their hard drives to last 100 years; but if you've got a few hundred of them spinning in a server farm, then differences in MTBF could make a meaningful difference in how many of them fail over the lifetime of those servers.
Don't buy models with known hardware/firmware issues, don't expose the drives to environmental extremes, use RAID-1 or -6 if availability or uptime is critical (ECC RAM on the server too if data integrity matters), and -- above all -- back up your important data. Everything else is secondary.
(this space intentionally left blank)