Thanks so much for your thoughtful response!    I have a large longitudinal dataset that I have been cleaning and all the codes are ICD 10. Basically it will be used by many researchers to answer a variety of questions. I am interested in presenting groups such as all individuals with some alcohol related dx (yes/no), all with cocaine (yes/no), opioid (yes/no) etc and also group back to main ICD 10 breakdowns simply to make it easier to subset the dataset later on and give some summary statistics of the frequency of most common dx codes.    For example, I would like to reduce the typing to write in every ICD 10 code related to alcohol so I wanted to be able to create a new indicator yes/no variable for alcohol that would be 1 if at least F10 was found in one of their dx codes without typing in all the possible F10 codes that exist.    I appreciate your input!  ... View more