Such discrepancies are quite common upon first determinations. For those of you who have been watching earthquake for a while, you will recognize that this is an everyday occurrence.
Even after several months, epicentral determinations by such groups as ISC and NEIS can differ by up to 50 to 100 km.
Reasons for this include an inaccurate determination of how fast waves get to particular seismographs (velocity model), poor seismograph coverage - perhaps all the seismographs are located in the same region and a bias is developed; a depth determination which is poorly developed leading to an error in horizontal determination, etc.
It has been noted on this board many time before that the first epicentral determinations should never be considered accurate. All determinations are no closer than 10-30 km and often farther, especially in areas of poor coverage. The 3-decimal point locations given by\
NEIS are an artifact of the computer program used and have nothing to do with the accuracy of the determination.