Can I incorporate data from a previous trial in a new trial without any adjustment for bias (individual subject data)?

If not, which is the best approach?