[NTG-context] formatting index entries: alphabetical order in Unicode

I seem to have a problem with Unicode in the indexing function for my book: a word like āyitu (beginning with an 'a' with a macron over it) is considered a word that starts with a 'y'; the word īga (beginning with an i with a macron) is placed as if it is 'ga'. Is there any way to tell the indexing function how to deal with such special characters? Robert