Flag drug deaths according to ISW7 rules

Given an MCOD dataframe, will apply ISW7 rules to flag drug deaths for both ICD9 and ICD10 codes.For ICD9, it is true if any poison code is in any record field. For ICD10, it is true if there is a specific UCOD code **and** at least one specified T-code.

Usage

flag_drug_deaths(processed_df, year = NULL, keep_cols = FALSE)

Arguments

processed_df

processed dataframe

year

if NULL, will attempt to detect

keep_cols

keep intermediate columns

Value

new dataframe with a drug_death column

Note

The ICD-10 rule requires BOTH a drug-poisoning underlying cause (X40-X44, X60-X64, X85, Y10-Y14) AND at least one drug T-code (T36-T50) on the record. This is marginally stricter than the CDC WONDER / NCHS "drug overdose" count and most other studies, which key on the underlying cause alone. narcan therefore excludes the small number of drug-poisoning-UCOD deaths that carry no drug T-code at all – on the order of 0.1 example, 31 of 27,424 such deaths, 0.11 This is deliberate: every death narcan flags as a drug death carries an identifiable drug T-code.

Examples

df <- data.frame(
    year = 2019,
    ucod = c("X42", "I250"),
    f_records_all = c("T400 T401", "I250")
)
flag_drug_deaths(df, year = 2019)
#>   year ucod f_records_all drug_death
#> 1 2019  X42     T400 T401          1
#> 2 2019 I250          I250          0