This is a companion dataset for pima_diabetes
. The
pima_diabetes
dataset is real; this dataset is synthetic. You can see
how it was generated here:
https://docs.healthcare.ai/articles/site_only/best_levels.html#appendix-data-generation.
Briefly, each patient in pima_diabetes
is assigned 0-4 medications
from the following six: insulin and metformin are more common among
diabetics, prednisone and metoprolol are less common among diabetics, and
nexium and tiotropium are equally likely among diabetic and non-diabetic
patients. Each patient-medication has a years_taken
value associated
with it, which is a random number drawn from an exponential distribution.
pima_meds
A tibble data frame with 1,604 rows and 3 variables:
Unique identifier, used to join pima_diabetes
One of the six medications described above
Numeric value indicating the duration the medication has been used