Rollup phecode counts

Takes phecodes that have been coded without rollup - E.g. existance of code 008.10 does not imply the existance of 008.00 - and produces a rolledup version of the dataset. Works with plain patient-phecode pairs and also with patient-code-count triplets.

rollup_phecode_counts(
  patient_code_counts,
  patient_col = patient,
  phecode_col = phecode,
  count_col = counts
)

Arguments

patient_code_counts	Data containing patient ids, phecodes, and counts of those phecodes.
patient_col	Unquoted name of the column containing patient ids in passed patient_code_counts's dataframe.
phecode_col	Unquoted name of the column containing phecodes in passed dataframe.
count_col	Unquoted name of the column containing counts of phecodes in passed dataframe.

Value

Dataframe in same format as passed patient_code_counts with code counts rolled up. This will most likely be longer than the original dataset because of the addition of potentially previously ommitted codes.

Examples

library(dplyr)

patient_data <- tribble(
  ~patient,    ~code, ~counts,
  1, "250.23",      7,
  1, "250.25",      4,
  1, "696.42",      1,
  1, "555.21",      4,
  2, "401.22",      6,
  2, "204.21",      5,
  2, "735.22",      4,
  2, "751.11",      2,
)

# Rolls up counts for codes if they are present
patient_data %>%
  rollup_phecode_counts(patient_col = patient,
                  phecode_col = code,
                  count_col = counts)
#> # A tibble: 22 x 3
#>    code   patient counts
#>    <chr>    <dbl>  <dbl>
#>  1 250.00       1     11
#>  2 250.20       1     11
#>  3 250.23       1      7
#>  4 250.25       1      4
#>  5 555.00       1      4
#>  6 555.20       1      4
#>  7 555.21       1      4
#>  8 696.00       1      1
#>  9 696.40       1      1
#> 10 696.42       1      1
#> # … with 12 more rows