I would presume that mPORTDSetPinsDigitalOut is a function that takes a parameter that is type compatible with BIT_x (which I also presume are constants).
Given that BIT_0, BIT_1, and BIT_2 will have some numeric representation (possibly 0x0000001, 0x00000010, and 0x00000100) the compiler will generate (for constants) or calculate at runtime (for variables) the logical OR of the bit patterns.
If we presume the constants are as I suggested (and they may not be) then the three values are:
00000001
00000010
00000100
which OR together as
00000111
That last value is passed to the function which performs the implemented function, presumably setting the 3 low order bits of Port D as digital outputs.
In this case the actual value passed is not really interesting. We presume that BIT_0 defines the value representing bit 0, and so on for bits 1 and 2 and we know that the OR operation performs the correct "summing" operation so that the value passed (whatever it is) represents all three bits.