jawalonoski
commented
3 years ago When filtering an EHR record, what data should be keep in the last N years of history is more complicated than just drawing a line in the sand (e.g., on 28 September 2020) and deleting everything before then.
It is clinically important to include chronic conditions (e.g., Diabetes diagnosed a decade previously), a list of active medications (even if originally prescribed before the filter date), a vaccination history (some vaccines require boosters every X years), and so forth --- even if someone only wants detail from the last N years.
So, the "years of history" feature can be thought of as a "clinically-aware filter" rather than merely just a "time filter." It is not going to drop important items from the record.
ProfAaronBaird
I may just be confused about the Years of History configuration setting vs. the purpose of encounter dates. But, I've set the exporter.years_of_history = 1 and also generate.only_alive_patients = true. After generating, in CSV, I looked through "encounters.csv". Many of the encounter (start and stop) dates are in many different years more than one year back (i.e., a lot of variation, sometimes back to 1965). My primary goal is to create all of the encounters for the synthetic patients for the last year. I can always just update the dates to be in the current year, but I thought I would see if this is a bug or perhaps a misunderstanding of the relationship between years of patient history and encounter data created.