At the risk of repeating myself. Here's the rub. The disease DCM in dobermans which causes them to drop dead does actually have a genetic test, only quite recently is it available. However in Europe alone the prevalence of the disease is considered far greater than 50% of the population, particularly in showing dogs. At the present time there is no breeding programme to eliminate the gene, even given the test available.

You would think this horrifically irresponsible, and yes it is. But not for the reasons you imagine.

These diseased show dogs where the prevalence is highest make up so much of the population due to line breeding to the few winning sires that the showing doberman would go extinct if they removed all those dogs with the gene for DMC from the gene pool. It would reduce the gene pool so drastically the breed wouldn't have any where to turn. There just wouldn't be enough showing dobermans left to select to breed with. Unfortunately the entire population is not much better. By reducing the entire gene pool to such an extent by eliminating all the dogs carrying the gene would quite correctly inadvertently massively increase the frequency of other genetic diseases. A genetic bottle neck of disease with no where to go. The point of no return.

Such problems could be avoided by out-crossing to dogs of other breeds. Choosing related breeds without the gene for DCM would solve the problem entirely and certainly increase the gene pool. This is not "foolish" this is science. For it to work like it has with the Dalmatian which had a similar crippling gene, breeders have to open their minds enough to let the simple science in. It has to be done responsibly and not just any unrelated dog or breed will work, the solution is simple but putting it into practise is rather more complicated than your average breeder understands presently (see the link).

However far from being frozen into paralysis like you might expect given the prevalent thoughts in show breeding concerning purity, they're rather just continuing like nothing was wrong! This is Dodo with head so firmly in the ground it's head has grown roots and it's died of old age already, never having seen the light. Much like the breed will become. Extinct.

Rare breeds face much the same problems but get to the point of no return just that much quicker, some even start there. Health testing as in the above example (the doberman isn't even considered a rare breed as such) is not "responsible breeding" either. It's not "saving" the breed. For your breed I would be more concerned with the inbreeding coefficient than health testing simply because by definition of "rare-breed" you don't have the numbers to exclude any of the dogs in the gene pool even those with debilitating genetic diseases. You can't assume that dogs from different lines or even different countries are less related than your own dogs. All the health testing in the world is not going to increase that gene pool. You will just find more and more diseases appearing. In so far as a breed should be "saved" it can be. Its just a matter of will.

There is definitely a place for health testing but particularly DNA profiling, if nothing else just to find out how much of genetic bottle neck your breed is in, so you can discover if there are any unrelated dogs to breed to. However it's not the panacea most show breeders suddenly have jumped onto the band wagon for. It's simply not the solution to many many breeds problems. It's also thus not the automatic guarantee of a "responsible" breeder by any means. 

Short of waffling endlessly on I leave you with a link and a quote to a very readable Q and A on the subject.

http://www.instituteofcaninebiology.org/blog/coi-faqs-understanding-the-coefficient-of-inbreeding

"The key to managing recessive mutations in any population is keeping them rare, so adding animals to the population that share many of the same mutations is asking for trouble in a very short time. Also, incorporating new genetic material into the breed requires a well-designed strategy worked out for at least four to five generations. A single crossbreeding (to a different breed) followed by sequential backcrossing into the breed will remove most of the genetic diversity you were hoping to introduce. You definitely need to start with a carefully designed plan. Without the tools of a geneticists at your disposal and short of becoming one yourself you risk waisting what for most pedigree dogs is a critical amount of time."

So yes you are perfectly right "anything on four legs" won't necessarily work. But without something entirely different on those four legs you can't solve anything either.

And don't you just hate it when rare breeds are marketed as such as though this were a plus instead of a big red flag in your face.