Before answering to you, I want to state, that I have changed the circuit a little bit. EMI filtering at '+V' input used too high capacitances, which would degrade common mode rejection ratio. Capacitances should be decreased to 4.7nF, as shown here:




Your argument is correct, but as you can see from the schematic above, the circuit is far away from a true instrumentation amplifier, means low pass filter at inverting input of INA128 is not identical to low pass filter at noninverting input of INA128. So, the discussion in the application note, you gave us, cannot be totally applied here. What application note is meaning is:

If you want to low-pass-filter the signal in order to accomplish bandwidth limiting, then a differential filter should be used, not a common mode one. But the low pass filter at the inputs of INA128, as shown in the schematic, is not intended to do this. This sort of signal filtering is provided by the low pass filter behind the INA128. What's connected to the inputs of INA128 is a common mode filter, which is urgently needed to suppress EMI. EMI filtering can only be done by the help of a common mode filter, not by a differential one.
Have a look at the different corner frequencies involved: Common mode filtering at inputs of INA128 shows a corner frequency of about 1.6kHz, while signal filtering behind the INA128 shows a corner frequency of about 11Hz.

Kai