Frankly it sounds right. If the sensor is sending 4mA, I would expect 0.004*220=0.88V. So you're getting about 0.994V, or 4.5mA. Pretty reasonable.
Converting this to an A/D reading:
0.994 / 5 * 1023 --> 203
Again, pretty reasonable compared to what you're getting.
As for the variation, there could be noise in the current signal. Do you expect/know that it's relatively noise-free? Did you put the 10k/0.1uF components in?
--
The Quick Shield: breakout all 28 pins to quick-connect terminals