We have now picked up 5 new diabetics out of 19 respondents, so virtually 25% positive identifications of diabetes with the same number of borderline results.  This is higher than the 14% we got in the first cohort (though probably random variation)