Main function to anonymize patient data in a data frame or data.table. Automatically detects and anonymizes columns based on data types and naming patterns, or you can manually specify columns. Different datasets get different anonymized values for better privacy.
Usage
anonymize_dataframe(
data,
id_cols = NULL,
name_cols = NULL,
date_cols = NULL,
location_cols = NULL,
age_cols = NULL,
auto_detect = TRUE,
detect_by_type = TRUE,
date_method = "shift",
date_granularity = "month",
location_method = "generalize",
age_method = "10year",
use_uuid = TRUE,
seed = NULL,
dataset_specific = TRUE
)Arguments
- data
A data frame or data.table containing patient data
- id_cols
Character vector of column names containing patient IDs
- name_cols
Character vector of column names containing patient names
- date_cols
Character vector of column names containing dates
- location_cols
Character vector of column names containing locations
- age_cols
Character vector of column names containing ages
- auto_detect
Logical, if TRUE (default), automatically detects columns based on data types and common naming patterns
- detect_by_type
Logical, if TRUE (default), detects columns by their R data types (Date, character, etc.) in addition to name patterns
- date_method
Method for date anonymization: "shift" or "round" (default: "shift"). Use "round" to enable granularity options including "month_year" (YYYYMM format).
- date_granularity
For date rounding (when date_method = "round"): "day", "week", "month", "month_year" (returns YYYYMM format, e.g., "202005"), "quarter", or "year" (default: "month")
- location_method
Method for location anonymization: "remove" or "generalize"
- age_method
Method for age anonymization: "10year" (default) uses 10-year buckets (0-9, 10-19, 20-29, ..., 80-89, 90+) for better research utility, or "hipaa" for HIPAA-compliant buckets (0-17, 18-64, 65-89, 90+)
- use_uuid
Logical, if TRUE uses short UUIDs for IDs, names, and locations instead of sequential identifiers (default: TRUE). Dates and ages are not affected.
- seed
An optional seed for reproducible anonymization. Different datasets will still get different anonymized values even with the same seed.
- dataset_specific
Logical, if TRUE (default), generates dataset-specific seeds so different datasets get different anonymized values
Value
A data frame with anonymized patient data (preserves data.table class if input was data.table)
Examples
# Basic usage with auto-detection
patient_data <- data.frame(
patient_id = c("P001", "P002", "P003"),
name = c("John Doe", "Jane Smith", "Bob Johnson"),
dob = as.Date(c("1980-01-15", "1975-03-20", "1990-06-10")),
location = c("New York, NY", "Los Angeles, CA", "Chicago, IL"),
diagnosis = c("A", "B", "A")
)
anonymize_dataframe(patient_data, seed = 123)
#> patient_id name dob location diagnosis
#> 1 ID_ouSbzEgq Patient_6k2nsVas 1980-02-23 Location_Q98vAvyL A
#> 2 ID_E9JiNTER Patient_7g3kma75 1975-04-28 Location_sBSspjmx B
#> 3 ID_H9dZM6iQ Patient_A7K9Ea9J 1990-07-19 Location_8MA8fu9H A
# With month_year date granularity (YYYYMM format)
anonymize_dataframe(patient_data, date_method = "round", date_granularity = "month_year")
#> patient_id name dob location diagnosis
#> 1 ID_w2EsGPA6 Patient_wfvcEyLg 198001 Location_BKPB9n3n A
#> 2 ID_qLcguMtS Patient_pA3Na8gU 197503 Location_peX4F6Vi B
#> 3 ID_apswz96L Patient_6mNgRVDu 199006 Location_zx9FzuBK A
# Works with data.table
if (requireNamespace("data.table", quietly = TRUE)) {
dt <- data.table::as.data.table(patient_data)
anonymize_dataframe(dt)
}
#> patient_id name dob location diagnosis
#> <char> <char> <Date> <char> <char>
#> 1: ID_w2EsGPA6 Patient_wfvcEyLg 1980-09-10 Location_BKPB9n3n A
#> 2: ID_qLcguMtS Patient_pA3Na8gU 1975-11-14 Location_peX4F6Vi B
#> 3: ID_apswz96L Patient_6mNgRVDu 1991-02-04 Location_zx9FzuBK A
# With UUID anonymization (default)
anonymize_dataframe(patient_data, seed = 123)
#> patient_id name dob location diagnosis
#> 1 ID_ouSbzEgq Patient_6k2nsVas 1980-02-23 Location_Q98vAvyL A
#> 2 ID_E9JiNTER Patient_7g3kma75 1975-04-28 Location_sBSspjmx B
#> 3 ID_H9dZM6iQ Patient_A7K9Ea9J 1990-07-19 Location_8MA8fu9H A
# Without UUID (sequential IDs)
anonymize_dataframe(patient_data, use_uuid = FALSE, seed = 123)
#> patient_id name dob location diagnosis
#> 1 ID3663 Patient 3663 1980-02-23 Location 3663 A
#> 2 ID3664 Patient 3664 1975-04-28 Location 3664 B
#> 3 ID3665 Patient 3665 1990-07-19 Location 3665 A