The invoked long SMART test will simply stop in case of a found error; there is no destructive behaviour of the internal disk’s test routine.
But … if you run a long smart test it will most possible end in an “aborted by host” state after some minutes since the disk has to be kept busy before the kernel tries to hibernate the idle USB drive.
For that I let run a tiny script in a second session or background job till the SMART test finished
while true
do
sudo smartctl -d sat,12 -Hc /dev/sdd
sleep 60
done
(-Hc shows the health state and capabilities which includes the Self-test execution status)
After all at least you know whether the disk has any surface issue or not.