Power guy here. Nothing, and I mean not even the wildest circuit breaker, can act faster than a properly selected fast blow fuse - even the grid still uses them to avert catastrophic damage to the high tension lines.
Each output ideally would have its own protection that was tailored to the load characteristics for which it was being used. In any event, the input to the power supply should be fused, making your protection scheme downstream more flexible.
Thermal devices are based on current over time, and as such the delay should be predictable - should you be able to find a datasheet. But the delay, in most cases, will probably work out such that a 10 amp thermal element will blow soon enough that a 10 amp output channel wouldn't fail, provided the channel's absolute max ratings weren't exceeded. Personally, when I optimize things I like to oversize and derate when possible.
Now all you have to do is decide if the load you are powering will survive any delays in protective action. There can be a big difference between 20 amps for 1/1000th of a second and 20 amps for 1/16th of a second. If shorting is something you do frequently, maybe it's time for a current limiter.
The maximum interrupting current and maximum voltage rating are two of the most important characteristics of fuses. (All tables below from Littlefuse, a major supplier of fuses and resettable polyfuses).
A cheap glass fuse might be able to interrupt some tens of amperes at 250V. It's thus not very suitable for a fault on the mains, which is why good (as in safe) multimeters intended for use on mains use massive cartridge fuses with 10,000A interrupting capacity. Even a residential electrical system can supply more than 100A under fault conditions.
Here's a cheap glass fuse:
It can interrupt only 35A at 250VAC, and drops 1.1V at rated current (quite a bit).
A (possibly) better ceramic type (still 5 x 20mm) can interrupt 1500A at 250VAC. Much safer- but look at that voltage drop- 2.8V.
Finally, I didn't see any polyfuses with both 250VAC rating and 250mA hold current, but the below chart shows one with 180mA hold and 250VAC rating- it can only interrupt 10A.
If you exceed the interrupting current or voltage on any of these devices, there is no guarantee the device will actually open- it may arc away causing severe damage or safety issues. I've made glass fuses literally explode sending shards of glass and molten metal from the element everywhere.
Bottom line- unless you really, really, really, know what you are doing and have all the information the original designer had, it's best to not muck with the protection devices.
Also, philosophically, you should not be blowing fuses often enough to care. You should find out why it's happening (if its not obvious) and figure out how not do that anymore. You may be damaging your meter and shifting the calibration, for just one consideration. Some meters have an audible alert that tells you if you've left the probe in a current socket and switched to volts, but really it's best to get into good habits that don't depend on tricks in the meter.
Best Answer
That's not really a fuse holder, although it might look a bit like one. The whole thing is the fuse itself. The 'cups' that hold the actual fuse can't usually be taken off.
The only way to replace this type of fuse is to desolder it from the board, and replace it with a new one. The new one should also have wires already connected to it, that you can bend so they will fit through the holes. Then you can solder the wires and cut off the excess length.
(Fuses usually break for a reason though, so there is a possibility that the fuse is not the only problem. Although you might be lucky.)