The tests and requirements set out in 60601-1 are, as you identify, Type Tests.
It is generally understood that some of these tests can be damaging to the unit under test. For instance, the impact test tends to dent panels and the marking durability tests can scuff labels. Neither necessarily constitutes test failure - but leave the unit in an unsaleable condition.
This accumulation of damage is why the sequence of tests (see Annex B) is defined. Your device should still be functional and safe after all he tests have been completed - but it may look a little worse for wear.
The dielectric strength test places significant stress upon electrical insulation. It may even permanently damage the insulation. And that damage may not be visible. So, it is not possible to guarantee that a test unit has undamaged insulation - so should never enter the market.
So what tests should be used on production?
Clause 5.1 describes the use of Type Tests. The 'Note 2' to that clause identifies IEC 62354 Annex K as guidance for 'Production Tests' and IEC62353 as guidance for 'Recurrent Tests'. Both have alternate methods for both the dielectric tests and various leakage current tests.
The guidance rationale for Clause 5.1 is useful:
Production tests need not be identical with TYPE TESTS, but can be adapted to manufacturing conditions and possibly invoking less RISK for the quality of the insulation or other characteristics...
One option for the dielectric tests that is quite commonly used, for instance by manufacturer's of transformers, is to perform the full voltage, full 1 minute, test on a sample to 'Type Test' it but then, for production, test at full voltage - but only for 5 seconds. This is, basically, the test we perform for production.
One thing to highlight - the dielectric strength tests only apply to solid insulations. If you are using 'creepage and clearance' insulations then you use a ruler as your test equipment - not the dielectric tester!
Also, when it comes to production dielectric tests, it is often a mistake to test from the 'back to the front'. By that I mean testing from the mains input to the front panel connectors. What happens then is that the test voltage stress is applied across more than one insulation and will be 'shared out' unevenly and unpredictably.
