Skip to contents

Identify CCSR categories for ICD-10-CA diagnosis codes

icd_to_ccsr.Rd

CCSR Clinical Classifications Software Refined provides a grouping of individual ICD-10 diagnosis codes into broader, clinically meaningful disease categories. The original CCSR grouping was developed based on US ICD-10-CM codes. GEMINI developed an algorithm mapping Canadian diagnosis codes (ICD-10-CA) to CCSR categories. The full mapping procedure is described here.

This function returns the GEMINI-derived CCSR mapping for each ICD-10-CA diagnosis in the diagnosis_table input.

By default, the function will only return the CCSR category(s) for the most responsible discharge diagnosis (MRDx). This typically refers to type-M diagnoses, unless a patient has a type-6 diagnosis (see CIHI definition of MRDx page 13). Users can choose to obtain CCSR categories for all diagnosis types by setting type_mrdx to FALSE.

Usage

icd_to_ccsr(
  dbcon,
  dxtable,
  type_mrdx = TRUE,
  unique_mrdx = FALSE,
  replace_invalidpdx = TRUE
)

Arguments

dbcon

(DBIConnection)
A database connection to any GEMINI database.

dxtable

(data.frame | data.table) Table containing ICD-10-CA diagnosis codes of interest. Typically, this refers to the ipdiagnosis table, which contains the CIHI in-patient diagnoses for each encounter (see GEMINI database schema).

If a different type of diagnosis table is provided as input (e.g., erdiagnosis), please make sure the table contains a column named diagnosis_code (character) where each row refers to a single, alphanumeric diagnosis code consisting of 3-7 characters. If type_mrdx is set to TRUE, the following additional columns are required:

Note, each encounter may have multiple rows, referring to diagnosis codes of different types. However, typically, each encounter should only have a single MRDx. In case type_mrdx is set to TRUE, the function will return a warning if any encounters have multiple MRDx codes. By default, those encounters will be returned with multiple rows (one per MRDx). However, users may specify unique_mrdx = TRUE (see below), which will cause the function to quit in case multiple MRDx codes were found.

type_mrdx

(logical) Flag indicating whether to filter by most responsible discharge diagnosis. If TRUE (default), only MRDx codes (type 6 if present, otherwise, type M) will be returned. This option is recommended and should typically return a single diagnosis code per encounter. If multiple MRDx codes are found, a warning message will be shown (and if unique_mrdx = TRUE, the function will stop). Additionally, the function will show a warning if there are any encounters with missing MRDx code. If type_mrdx = FALSE, the function Will return all diagnosis codes (of any diagnosis type) that are present in dxtable. In that case, any customized filtering for specific diagnosis types should be applied by the user outside of this function. When using this option, please refer to the function's vignette to find out more about potential issues when interpreting default CCSR categories for non-MRDx diagnosis types.

unique_mrdx

(logical) Flag indicating whether the requirement for a unique MRDx code per encounter is strict. This is only relevant when type_mrdx = TRUE. If FALSE (default), multiple MRDx codes/encounter will only result in a warning to inform the user. The function will finish running and will return multiple rows per encounter. If TRUE, the function will quit if any genc_ids with multiple MRDx codes are found and the user is advised to investigate the cause of multiple MRDx codes.

replace_invalidpdx

(logical) Flag indicating whether to replace invalid default CCSR categories (default = TRUE). If set to FALSE, the function will return the CCSR default categories as derived based on the original US CCSR tool, which may include some ICD-10-CA codes where ccsr_default = 'XXX000' ("invalid PDX"; see below for details). If set to TRUE (default), XXX000 values will be replaced with one of the other (valid) CCSR categories 1-6 that have been mapped to a given code. Specifically, if a given ICD-10-CA code has been mapped to a single CCSR category (ccsr_1), XXX000 will be replaced with that category. Otherwise, if a code has been mapped to multiple CCSR 1-6 categories, the function will check which one of those is the most frequent CCSR default category among ICD-10-CA codes that start with the same first 3 characters (same ICD-10 chapter). If other codes from the same ICD-10 chapter don't share the same CCSR default category (or default is XXX000 for all of them), ccsr_1 is used as default.

Value

data.table This function returns a table containing the ICD-10-CA diagnosis codes of interest, together with their corresponding (GEMINI-derived) CCSR category(s). For each row in the output table, the following variables are returned:

  • diagnosis_code: ICD-10-CA code

  • diagnosis_code_desc: description of the ICD-10-CA code

  • ccsr_default: default CCSR category, which is the main disease group that was assigned to each diagnosis code and is typically the main category of interest to be used in further analyses.

  • ccsr_default_desc: description of the CCSR default category

  • ccsr_1 through ccsr_6: All 1-6 CCSR categories that were mapped to a given code. Although ccsr_default is typically the main category of interest, some ICD-10-CA codes are mapped to more than 1 (up to 6) CCSR categories. In some circumstances, it may be of interest to analyse all 1-6 CCSR categories, instead of only analyzing the default CCSR. If required, the descriptions for CCSR 1-6 categories can be obtained from the lookup_ccsr table.

Note

The GEMINI-derived CCSR mappings were validated by clinical experts; however, users are advised to spot-check the mappings to ensure accuracy. Please notify the GEMINI team if you encounter any issues. Some mappings were derived based on an older version of CCSR (v2020.3) and may be subject to change with future updates.

For some diagnosis codes, ccsr_default will be NA (ccsr_default_desc = 'Unmapped'), which indicates that the diagnosis code has not been mapped to any CCSR category yet. This is less likely to happen if type_mrdx is set to TRUE (default) since the vast majority of MRDx codes have been mapped by GEMINI. Note: Type-4 morphology codes will always be returned as 'Unmapped' since they are numeric codes based on ICD-0 (oncology) diagnoses, which are not meant to be mapped to CCSR categories.

Some diagnosis codes will be returned with ccsr_default = 'XXX000', which indicates that the diagnosis code is not a valid in-patient principal diagnosis (PDX) according to US diagnostic coding following Medicare Code Edits guidelines (see here. Due to mapping ambiguity and differences in coding between the US and Canada, you may find that some ICD-10-CA MRDx codes are returned as 'XXX000'(invalid PDX). Additionally, for most clinical research projects, the distinction between valid vs. invalid principle diagnosis coding may not be relevant. Similarly, if users are interested in CCSR groups for all diagnosis types (type_mrdx = FALSE), the definition of CCSR default categories as valid vs. invalid PDX categories might not be meaningful. Therefore, by default, replace_invalidpdx is set to TRUE. However, please carefully read the documentation to understand how exactly invalid PDX categories are replaced within this function.

See also

vignette("icd_to_ccsr", package = "Rgemini") The vignette provides further context and some example code illustrating how to use CCSR categories in analyses.

Examples

if (FALSE) { # \dontrun{
drv <- dbDriver("PostgreSQL")
dbcon <- DBI::dbConnect(drv,
                        dbname = "db",
                        host = "domain_name.ca",
                        port = 1234,
                        user = getPass("Enter user:"),
                        password = getPass("password"))

dxtable <- dbGetQuery(dbcon, "select * from ipdiagnosis") %>% data.table()
icd_to_ccsr(dbcon, dxtable)
} # }