Each genus/species needs to be rated as to pollution tolerance. Then you can use presence, absence, relative abundance of the species present. So you need a numbering scheme for tolerance to pollution to intolerant. Then a relative abundance numeral rating from 0 (not present) to high numbers. Your ratings can go to 5 or 10, depending on categories you set. So if intolerant is 1, tolerant is 5. If not found, it gets 0, if infrequent 1, ... , abundant, 5. Then take the tolerance times abundance value for each species present. Add the total of all, divide by (number of species evaluated times 5). Lower numbers good water quality, higher numbers bad. You develop categories.
For each sampling location, sample riffles, runs minimum, all facets as you see fit, recognizing that pools may have more sediment accumulations and more tolerant species.
Rate macroinvertegrates 1-5, tolerant ranging to intolerant to pollution. Each species may have a rating associated with it already by other scientists or those that have developed indices of biotic integrity.
For each species multiply abundance by tolerance. Locations without that species will be zero. Add up all values for location, divide by (number of species present multiplied by 5). The lower the number the better the rating, higher numbers will progress from fair to poor.
Of course, the hard part is rating the benthic invertebrates... to do that you need invertebrate samples from a wide variety of local sites, coupled with water chemistry and sediment chemistry data.
I suggest using of BMWP score.It is easy to apply, you can use Asterics software for calculation (free to download - just type in Google Asterics assessment software and you will find the site for download software and manual). The software willl calculate a lot of different metrics.
Good thing with BMWP is that you can easily work on adaptation of original system to your particular study area.