You can use a macro like the following to perform the merge, but only if
your mail merge main document does not have any <<Next record>> fields and
similar stuff such as <<Next record if>>

Peter Jamieson

Sub ProduceOneDocPerSourceRec()
'

' NB, needs bettor error management and doubtless other things a VBA expert
' will point out.

Dim intSourceRecord
Dim objMerge As Word.MailMerge
Dim strEman As String
Dim strOutputDocumentName As String
Dim TerminateMerge As Boolean

' Need to set up this object as the ActiveDocument changes when the
' merge is performed. Besides, it's clearer.

Set objMerge = ActiveDocument.MailMerge
With objMerge

' If no data source has been defined, do it here using OpenDataSource.
' But if it is already defined in the document, you should not need to
define it here.

'  .OpenDataSource _
'    Name:="whatever"

 intSourceRecord = 1
 TerminateMerge = False

 Do Until TerminateMerge
   .DataSource.ActiveRecord = intSourceRecord

   ' if we have gone past the end (and possibly, if there are no records)
   ' then the Activerecord will not be what we have just tried to set it to

   If .DataSource.ActiveRecord <> intSourceRecord Then
     TerminateMerge = True
   ' the record exists
   Else

    ' while we are looking at the correct activerecord,
     ' create the document path name
     ' e.g. - you will need to change this -
    strEman = StrReverse(objMerge.Name)
 StrReverse(Mid(strEman, InStr(1, strEman & ".", ".") + 1))

     strOutputDocumentName = _
       "c:\mydoc\" & _
       StrReverse(Mid(strEman, InStr(1, strEman & ".", ".") + 1)) & _
       .DataSource.Datafields("CaseID").Value & ".doc"

     .DataSource.FirstRecord = intSourceRecord
     .DataSource.LastRecord = intSourceRecord
     .Destination = wdSendToNewDocument
     .Execute

     ' The Activedocument is always the output document
     ' Add any parameters you need to these calls
     ActiveDocument.SaveAs strOutputDocumentName
     ActiveDocument.Close
     intSourceRecord = intSourceRecord + 1
   End If
 Loop
End With
End Sub

Hmmm. If it's a single record merge then the original split merge add-in
solution should work. This names the file from a field (CaseID) in the
document?