ICD code scoring systems

A common use of ICD codes is calculation as a Charlson score, which gives a measure of how well a patient is, albeit based on the limited available in admission and discharge diagnoses. The Charlson scoring system attributes scores based on presence of diseases falling into any of the Charlson comorbidities. Quan updated the scores given to each comorbidity to better reflect morbidity and mortality in a more recent population. Van Walraven provides a similar scoring methodology for the Elixhauser comorbidities (as used by the US AHRQ).

More complicated scoring systems may use lab values, patient demographic information, and so on. Any contributions to this package for calculations of scoring systems based on comorbidities and other data would be welcome.

Vermont example data, Charlson scores

The Vermont data are actually discharge, not admission diagnoses, but can be used to demonstrate generating Charlson scores.

head(icd.data::vermont_dx[1:10])
#>   visit_id   age_group    sex death DRG   DX1   DX2   DX3   DX4   DX5
#> 1        7       40-44   male  TRUE 640 27801 03842 51881 41519 99591
#> 2       10 75 and over female FALSE 470 71526 25000 42830  4280  4019
#> 3       13 75 and over female FALSE 470 71535 59651 78052 27800 V8537
#> 4       16       55-59 female FALSE 470 71535 49390 53081 27800  V140
#> 5       37       70-74   male FALSE 462 71536  4241  2859  2720  4414
#> 6       41       70-74   male FALSE 462 71536 V1259 V1582  V160  V171
v <- wide_to_long(icd.data::vermont_dx)
head(v)
#>   visit_id age_group  sex death DRG icd_code
#> 1        7     40-44 male  TRUE 640    27801
#> 2        7     40-44 male  TRUE 640    03842
#> 3        7     40-44 male  TRUE 640    51881
#> 4        7     40-44 male  TRUE 640    41519
#> 5        7     40-44 male  TRUE 640    99591
#> 6        7     40-44 male  TRUE 640    42842
charlson(v) %>% summary
#>    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
#>   0.000   0.000   1.000   1.573   2.000   9.000
head(charlson(v))
#>  7 10 13 16 37 41 
#>  5  3  0  1  1  0
head(names(charlson(v)))
#> [1] "7"  "10" "13" "16" "37" "41"

This default result is a numeric vector with the names (top numbers) as the patient identifiers. Those who like working with ‘tidy’ data frames can use:

head(charlson(v, return_df = TRUE))
#>   visit_id Charlson
#> 1        7        5
#> 2       10        3
#> 3       13        0
#> 4       16        1
#> 5       37        1
#> 6       41        0

Behind the scenes, icd calculates the Charlson comorbidities for those ICD codes, applies the Charlson scoring system, and returns the Charlson score for each patient.

Vermont example data, Van Walraven scores

The same principle can be used to calculate the Van Walraven score, which is the Charlson score counterpart for Elixhauser comorbidities.

`Vermont Van Walraven Scores` <- van_walraven(v)
hist(`Vermont Van Walraven Scores`)

Working with mixed ICD-9 and ICD-10 codes

All the functions in icd work with one code type. They are tolerant of having different sub-types of ICD-9 or ICD-10 codes together for comorbidity calculations, but patients with mixed data can be combined. E.g. patient A has two ICD-9 and two ICD-10 codes:

icd9 <- data.frame(pts = c("A", "A"), c("041.04", "244.9"))
icd10 <- data.frame(pts = c("A", "A"), c("C82.28", "M08.979"))
both <- comorbid_elix(icd9) | comorbid_elix(icd10)
van_walraven_from_comorbid(both)
#> A 
#> 9

More commonly, some patients before a certain date will have ICD-9 codes, and others will have ICD-10 codes:

icd9 <- data.frame(pts = c("A", "A"), c("041.04", "244.9"))
icd10 <- data.frame(pts = c("B", "B"), c("C82.28", "M08.979"))
both <- rbind(comorbid_elix(icd9), comorbid_elix(icd10))
van_walraven_from_comorbid(both)
#> A B 
#> 0 9