Creates A-bomb survivor incidence rates and confidence intervals.
incidAbomb(d)Tibble, typically grouped, with DG column ending 1st block and py starting the last.
A tibble data frame, summarized by groups, with cancers after py in a new cancers column, and new columns O (observed cases), incid and incid 95% CI limits LL and UL.
The columns DG and py must exist in d, in that order.
Person-year weighted means will be formed of any columns between them.
Its OK if none exist. Sums are formed on py and anything to their right. It is assumed
that cancer types begin after py, with at most upy and/or subjects intervening them.