WARNING: This package is under development.
- addDailyDose function works for the following patterns in the drug_strength table:
| pattern_id | amount_unit | numerator_unit | denominator_unit |
|---|---|---|---|
| 1 | international unit | NA | NA |
| 2 | microgram | NA | NA |
| 3 | milliequivalent | NA | NA |
| 4 | milligram | NA | NA |
| 5 | millimiter | NA | NA |
| 6 | NA | microgram | hour |
| 7 | NA | milligram | actuation |
| 8 | NA | microgram | hour |
| 9 | NA | milligram | hour |
Package overview
DrugUtilisation contains functions to instantiate and characterize the cohorts used in a Drug Utilisation Study in the OMOP common data model. Main functionalities are:
Create DrugUtilisation cohorts
Add indications to this cohort
Add the dosage of a certain ingredient (subseted for a list of drugs)
Calculate the daily dose
Create Concept based cohorts
Read concepts from json files
Summarise the drug use in a certain cohort
Summarise the indications in a certain cohort
Summarise the patients characteristics in a certain cohort
Summarise the patients large scale characterics in a certain cohort
Example
First, we need to create a cdm reference for the data we´ll be using. Here we´ll generate an example with simulated data, but to see how you would set this up for your database please consult the CDMConnector package connection examples.
The package also provides a functionality to generate a mockDrugUtilisation cdm reference:
library(DrugUtilisation)
cdm <- mockDrugUtilisation(numberIndividual = 100)Create a cohort of drug use
To create a cohort we will need a conceptList, this can be read from json files:
conceptList <- readConceptList(here::here("Concepts"), cdm)Or we can build our own list using other packages (e.g. CodelistGenerator)
library(CodelistGenerator)
conceptList <- getDrugIngredientCodes(cdm, "acetaminophen")
conceptList
#> $`Ingredient: acetaminophen (1125315)`
#> [1] 1125315 1125360 2905077 43135274To generate the cohort of drug use we will use generateDrugUtilisationCohortSet:
cdm <- generateDrugUtilisationCohortSet(
cdm = cdm,
name = "dus_cohort",
conceptSet = conceptList,
limit = "first",
priorObservation = 365,
gapEra = 30,
priorUseWashout = 0,
imputeDuration = "none",
durationRange = c(0, Inf)
)Cohort attributes
The generated cohort will have the GeneratedCohortSet as seen in CDMConnector
class(cdm[["dus_cohort"]])
#> [1] "GeneratedCohortSet" "tbl_duckdb_connection" "tbl_dbi"
#> [4] "tbl_sql" "tbl_lazy" "tbl"Cohort set:
library(CDMConnector)
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
cohortSet(cdm[["dus_cohort"]]) %>% glimpse()
#> Rows: 1
#> Columns: 11
#> $ cohort_definition_id <int> 1
#> $ cohort_name <chr> "Ingredient: acetaminophen (1125315)"
#> $ limit <chr> "first"
#> $ prior_observation <chr> "365"
#> $ gap_era <chr> "30"
#> $ prior_use_washout <chr> "0"
#> $ cohort_date_range_start <chr> NA
#> $ cohort_date_range_end <chr> NA
#> $ impute_duration <chr> "none"
#> $ duration_range_min <chr> "0"
#> $ duration_range_max <chr> "Inf"Cohort count:
cohortCount(cdm[["dus_cohort"]])
#> # A tibble: 1 × 3
#> cohort_definition_id number_records number_subjects
#> <int> <dbl> <dbl>
#> 1 1 35 35Cohort attrition:
cohortAttrition(cdm[["dus_cohort"]]) %>% glimpse()
#> Rows: 8
#> Columns: 7
#> $ cohort_definition_id <int> 1, 1, 1, 1, 1, 1, 1, 1
#> $ number_records <dbl> 90, 90, 70, 70, 41, 41, 41, 35
#> $ number_subjects <dbl> 62, 62, 62, 62, 35, 35, 35, 35
#> $ reason_id <dbl> 1, 2, 3, 4, 5, 6, 7, 8
#> $ reason <chr> "Qualifying initial records", "Duration imputatio…
#> $ excluded_records <dbl> 0, 0, 20, 0, 29, 0, 0, 6
#> $ excluded_subjects <dbl> 0, 0, 0, 0, 27, 0, 0, 0Indication
Indications will always be cohorts. An option that the package has is to create concept based cohorts using generateConceptCohortSet.
indications <- list(headache = 378253, influenza = 4266367)
cdm <- generateConceptCohortSet(cdm, indications, "indications_cohort")
cohortCount(cdm[["indications_cohort"]])
#> # A tibble: 2 × 3
#> cohort_definition_id number_records number_subjects
#> <int> <dbl> <dbl>
#> 1 2 46 46
#> 2 1 52 52Then we can add the indication using the function addIndication. That will add a new column for each indication gap and indication.
x <- cdm[["dus_cohort"]] %>%
addIndication(
cdm = cdm, indicationCohortName = "indications_cohort", indicationGap = c(0, 30, 365),
unknownIndicationTable = c("condition_occurrence")
)
glimpse(x)
#> Rows: ??
#> Columns: 16
#> Database: DuckDB 0.9.0 [martics@Windows 10 x64:R 4.2.3/:memory:]
#> $ cohort_definition_id <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ subject_id <dbl> 7, 48, 57, 77, 3, 58, 73, 23, 40, 45, 51,…
#> $ cohort_start_date <date> 1969-09-09, 1972-04-14, 2020-04-17, 1967…
#> $ cohort_end_date <date> 1970-02-14, 1972-12-18, 2020-11-19, 1967…
#> $ indication_gap_0_influenza <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_0_headache <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_0_none <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ indication_gap_30_headache <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_30_influenza <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_30_none <dbl> 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ indication_gap_365_headache <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,…
#> $ indication_gap_365_influenza <dbl> 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_365_none <dbl> 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1,…
#> $ indication_gap_0_unknown <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_30_unknown <dbl> 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_365_unknown <dbl> 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,…We can combine the indications in a single column using the indicationToStrata() function. This column can be used as stratification of the results if needed:
x <- x %>% indicationToStrata(keep = TRUE)
glimpse(x)
#> Rows: ??
#> Columns: 19
#> Database: DuckDB 0.9.0 [martics@Windows 10 x64:R 4.2.3/:memory:]
#> $ cohort_definition_id <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ subject_id <dbl> 7, 48, 57, 77, 3, 58, 73, 23, 40, 45, 51,…
#> $ cohort_start_date <date> 1969-09-09, 1972-04-14, 2020-04-17, 1967…
#> $ cohort_end_date <date> 1970-02-14, 1972-12-18, 2020-11-19, 1967…
#> $ indication_gap_0_influenza <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_0_headache <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_0_none <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ indication_gap_30_headache <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_30_influenza <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_30_none <dbl> 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ indication_gap_365_headache <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,…
#> $ indication_gap_365_influenza <dbl> 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_365_none <dbl> 1, 1, 0, 0, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1,…
#> $ indication_gap_0_unknown <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_30_unknown <dbl> 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_365_unknown <dbl> 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,…
#> $ indication_gap_0 <chr> "None", "None", "None", "None", "None", "…
#> $ indication_gap_30 <chr> "None", "None", "None", "Unknown", "None"…
#> $ indication_gap_365 <chr> "None", "None", "Influenza", "Unknown", "…Summarise the indication
We can summarise the indication results using the summariseIndication function:
summariseIndication(x, cdm)
#> # A tibble: 40 × 11
#> group_name group_level strata_name strata_level variable variable_level
#> <chr> <chr> <chr> <chr> <chr> <chr>
#> 1 cohort_name Ingredient: ace… Overall Overall number … <NA>
#> 2 cohort_name Ingredient: ace… Overall Overall number … <NA>
#> 3 cohort_name Ingredient: ace… Overall Overall Indicat… Influenza
#> 4 cohort_name Ingredient: ace… Overall Overall Indicat… Influenza
#> 5 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> 6 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> 7 cohort_name Ingredient: ace… Overall Overall Indicat… None
#> 8 cohort_name Ingredient: ace… Overall Overall Indicat… None
#> 9 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> 10 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> # ℹ 30 more rows
#> # ℹ 5 more variables: variable_type <chr>, estimate_type <chr>, estimate <chr>,
#> # cdm_name <chr>, result_type <chr>
summariseIndication(x, cdm) %>% glimpse()
#> Rows: 40
#> Columns: 11
#> $ group_name <chr> "cohort_name", "cohort_name", "cohort_name", "cohort_na…
#> $ group_level <chr> "Ingredient: acetaminophen (1125315)", "Ingredient: ace…
#> $ strata_name <chr> "Overall", "Overall", "Overall", "Overall", "Overall", …
#> $ strata_level <chr> "Overall", "Overall", "Overall", "Overall", "Overall", …
#> $ variable <chr> "number subjects", "number records", "Indication on ind…
#> $ variable_level <chr> NA, NA, "Influenza", "Influenza", "Headache", "Headache…
#> $ variable_type <chr> NA, NA, "binary", "binary", "binary", "binary", "binary…
#> $ estimate_type <chr> "count", "count", "count", "percentage", "count", "perc…
#> $ estimate <chr> "35", "35", "0", "0", "0", "0", "35", "100", "0", "0", …
#> $ cdm_name <chr> "DUS MOCK", "DUS MOCK", "DUS MOCK", "DUS MOCK", "DUS MO…
#> $ result_type <chr> "Summary indication", "Summary indication", "Summary in…Add strata
All summarise functions have the option to add strata. Strata will always point to preexisting columns. Here we can see an example where we create a age_group and sex columns using PatientProfiles and then we use it as strata
library(PatientProfiles)
x <- x %>%
addAge(cdm, ageGroup = list(c(0, 19), c(20, 39), c(40, 59), c(60, 79), c(80, 150))) %>%
addSex(cdm)
summariseIndication(x, cdm, strata = list("age_group", "sex", c("age_group", "sex")))
#> # A tibble: 440 × 11
#> group_name group_level strata_name strata_level variable variable_level
#> <chr> <chr> <chr> <chr> <chr> <chr>
#> 1 cohort_name Ingredient: ace… Overall Overall number … <NA>
#> 2 cohort_name Ingredient: ace… Overall Overall number … <NA>
#> 3 cohort_name Ingredient: ace… Overall Overall Indicat… Influenza
#> 4 cohort_name Ingredient: ace… Overall Overall Indicat… Influenza
#> 5 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> 6 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> 7 cohort_name Ingredient: ace… Overall Overall Indicat… None
#> 8 cohort_name Ingredient: ace… Overall Overall Indicat… None
#> 9 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> 10 cohort_name Ingredient: ace… Overall Overall Indicat… Headache
#> # ℹ 430 more rows
#> # ℹ 5 more variables: variable_type <chr>, estimate_type <chr>, estimate <chr>,
#> # cdm_name <chr>, result_type <chr>
summariseIndication(x, cdm, strata = list("age_group", "sex", c("age_group", "sex"))) %>% glimpse()
#> Rows: 440
#> Columns: 11
#> $ group_name <chr> "cohort_name", "cohort_name", "cohort_name", "cohort_na…
#> $ group_level <chr> "Ingredient: acetaminophen (1125315)", "Ingredient: ace…
#> $ strata_name <chr> "Overall", "Overall", "Overall", "Overall", "Overall", …
#> $ strata_level <chr> "Overall", "Overall", "Overall", "Overall", "Overall", …
#> $ variable <chr> "number subjects", "number records", "Indication on ind…
#> $ variable_level <chr> NA, NA, "Influenza", "Influenza", "Headache", "Headache…
#> $ variable_type <chr> NA, NA, "binary", "binary", "binary", "binary", "binary…
#> $ estimate_type <chr> "count", "count", "count", "percentage", "count", "perc…
#> $ estimate <chr> "35", "35", "0", "0", "0", "0", "35", "100", "0", "0", …
#> $ cdm_name <chr> "DUS MOCK", "DUS MOCK", "DUS MOCK", "DUS MOCK", "DUS MO…
#> $ result_type <chr> "Summary indication", "Summary indication", "Summary in…Daily dose
We can compute daily dose for a certain ingredient from a subset of drug_exposure or the whole drug exposure (can be very computationally expensive).
#cdm[["drug_exposure"]] %>%
# addDailyDose(ingredientConceptId = 1125315) %>%
# glimpse()DrugUse
You can add columns related to the drug use using addDrugUse. You always have to provide a reference ingredient.
#x <- x %>%
# addDrugUse(
# cdm = cdm,
# ingredientConceptId = 1125315,
# initialDailyDose = TRUE,
# numberExposures = TRUE,
# duration = TRUE,
# cumulativeDose = TRUE,
# numberEras = TRUE
# )Summarise the drug use
You can summarise the drug use using summariseDrugUse function
#summariseDrugUse(x, cdm)Summarise patient characteristics
You can summarise the patient characteristics with summariseCharacteristics function:
summariseCharacteristics(
x, cdm, ageGroup = list(c(0, 24), c(25, 49), c(50, 74), c(75, 150)),
tableIntersect = list(
"Visits" = list(
tableName = "visit_occurrence", value = "count", window = c(-365, 0)
)
),
cohortIntersect = list(
"Indications" = list(
targetCohortTable = "indications_cohort", value = "flag",
window = c(-365, 0)
)
)
)
#> # A tibble: 46 × 11
#> cdm_name result_type group_name group_level strata_name strata_level variable
#> <chr> <chr> <chr> <chr> <chr> <chr> <chr>
#> 1 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Number …
#> 2 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Number …
#> 3 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 4 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 5 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 6 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 7 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 8 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 9 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> 10 DUS MOCK Summary ch… cohort_na… Ingredient… Overall Overall Cohort …
#> # ℹ 36 more rows
#> # ℹ 4 more variables: variable_level <chr>, variable_type <chr>,
#> # estimate_type <chr>, estimate <chr>Summarise patients large scale characteristics
You can summarise the patient characteristics with summariseLargeScaleCharacteristics function:
summariseLargeScaleCharacteristics(
cohort = x,
window = list(c(-Inf, Inf)),
eventInWindow = "condition_occurrence",
episodeInWindow = "drug_exposure"
)
#> Warning in checkNewName(ageName, x): age already exists in x, it was renamed to
#> age_1
#> Warning in checkNewName(sexName, x): sex already exists in x, it was renamed to
#> sex_1
#> # A tibble: 26 × 14
#> result_type cdm_name group_name group_level strata_name strata_level
#> <chr> <chr> <chr> <chr> <chr> <chr>
#> 1 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 2 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 3 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 4 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 5 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 6 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 7 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 8 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 9 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> 10 Summarised Large Sc… DUS MOCK Cohort na… Ingredient… Overall Overall
#> # ℹ 16 more rows
#> # ℹ 8 more variables: table_name <chr>, type <chr>, analysis <chr>,
#> # concept <dbl>, variable <chr>, variable_level <chr>, estimate_type <chr>,
#> # estimate <chr>