The newer style HL7 processing in 2017 R1 may cause older results to re-process in such as way that it creates a duplicate. This is because the newer version detects a slight difference in one of the fields, and so creates a new record.
To fix, run the following utility:
Note, as you re-review older HL7 messages, you may find that more duplicates are created. You can of course run this utility again. Alternatively you can load all results by using a wide date range - this might take a while - but then when it's done and they are showing, you can then quit and run the merge util again.