The switcher's voltage response to current transients is a legitimate issue. Most switcher chips have plots for this purpose in the datasheet, although they can sometimes be difficult to extrapolate to any particular real world situation. Start there and see what it's supposed to be able to do.
This is also something where good testing is in order. Transient response is one of the parameters a switching power supply designer can trade off against others, particularly ripple.
Some chips have external compenstation components. This would let you chose your own tradeoff somewhat, although datasheets don't usually give you much guidance here. If your tests show this is a issue and you really want to use a particular chip for this for some other reason, you need to have a talk with a applications engineer about your situation.
Another possibility is to screw the off the shelf chip with its unknown tradeoffs and roll your own buck switcher. I've done this a bunch of times, not so much to get particular performance characteristics, but to save money. A PIC 10F204 is lots cheaper than most specialized switcher chips. If you implement a pulse on demand system the transient response will be very good with the steady state ripple will be higher, although usually well within the good enough range.
As Francois pointed out, power supply ripple is something to consider whenever there is analog electronics envolved. Most devices will have decent power supply noise immunity up to some frequency, so the strategy is not to worry about the peak to peak ripple so much but rather its frequency content. A ferrite bead or "chip inductor" followed by 20uF ceramic cap to ground feeding the Bluetooth module should help a lot with this. In fact, this is pretty much my standard power feed to any analog circuit. This is in addition to any local bypass caps.
