Sure, just use two duals and connect the grounds in parallel.
Now, note this breaks the assumption of compensated current -- in equilibrium, the ground current will split evenly (according to DCR), while the +/-15V currents won't. Also, since the supplies are opposite, the ground current will average towards zero (or more exactly, whatever average the loads are actually drawing), which is even more awkward for this arrangement.
You want the unbalanced current to be as near zero as possible; typical CMCs saturate at 10s of mA, so you lose the benefit rapidly as imbalance goes up.
If the supplies are equally loaded (so that ground current is nearly zero), swap windings around so +/-15V goes through one part, and GND goes through another (or just any regular FB of similar value).
As for the linked part, it can probably handle up to an amp of imbalance before saturating too much (evidently it's not very "typical"; "typical" ones also have a peak impedance in the low kohms, whereas this is more like... a very pricey ferrite bead :-) ).
If you can't keep imbalance within range, then you may be better off using inductors (proper), rated for load current, and value large enough to get the desired impedance. Namely, 1-2uH is enough here, maybe in parallel with a resistor (only one R is needed if you've got bypasses linking the supplies together on either side of the inductors). The resistor being about 200 ohms to get roughly similar performance to the part shown. Or use much larger values if you actually need more impedance. Don't forget to adequately dampen the inductors and bypasses (use a bypass C, then a bulk cap around 3*C with ESR = sqrt(L/C)).
Tim