generates random but plausible healthcare utilization data using a template stored in config.json.
simply, type sim in a folder where config.json exists
The "__doc" key can be used to document the configuration file.
version: must be 1.0. n: the number of patient records to generate. Must be >0.
"population": {
"migrant_prob": 0.15,
"cancel_prob": 0.15,
"database_start_date": "1971-01-01",
"earliest_birth_date": "1920-01-01"
},
hospitalization: sets parameters for all hospitalizations regardless of disease stay_length: provides the mean and SD of the distribution of hospital length of stay in days
diseases: array of disease descriptor
chronic and recurrence are not implemented
hospital_rate: provides the mean and SD of the distribution of number of hospitalizations per year.
clinic_rate: provides the mean and SD of the distribution of number of hospitalizations per year.
rx_rate: provides the mean and SD of the distribution of the number of prescription filled per year.
dins: an array of 1 or more drugs filled. din=as per the DPD; prob= probability of getting this DIN.
locator: used to generate a random geolocation code name: is the name of the field in the generated dataset, eg. postal_code. csv_filename: the relative/absolute path of a csv file that must"