Flag non-opioid deaths that involved opioids

Given an MCOD dataframe, will search through contributory causes for opioid related codes, but will only flag rows for which the underlying cause is **NOT** an opioid-related death.

Usage

flag_opioid_contributed(processed_df, year = NULL)

Arguments

processed_df

processed dataframe

year

if NULL, will attempt to detect

Value

new dataframe with an `opioid_contributed` column (`NA` for ICD-9-era data; see the note above)

Details

For ICD-9-era data (pre-1999) this flag is undefined: narcan's ICD-9 opioid-death rule fires on any opioid code in any field, so an opioid recorded only in a contributory cause already makes the death an opioid death – there is no "opioid contributed but not the underlying opioid death" subset to flag. On ICD-9 input the function therefore warns and returns `NA` rather than a misleading 0/1. (For ICD-10 the underlying-cause and contributory sets are disjoint, so the flag is well defined.)

Examples

df <- data.frame(year = 2019, ucod = "I250", f_records_all = "T401")
flag_opioid_contributed(df, year = 2019)
#>   year ucod f_records_all opioid_contributed
#> 1 2019 I250          T401                  1