Characterise a cohort using Drug Utilisation

Martí Català, Xihang Chen, Theresa Burkard

2024-07-19

Introduction

In DrugUtilisation package, there are also functions on daily dose of drugs. The best way to learn these functions is to consider a few examples.

Create mock data first

library(DrugUtilisation)
library(CodelistGenerator)
library(CDMConnector)
library(dplyr)
library(PatientProfiles)

cdm <- mockDrugUtilisation(numberIndividual  = 200)

Add daily dose information via addDailyDose() function

The following example calculates the daily dose of acetaminophen in the drug_exposure table.

cdm$drug_exposure |>
  filter(drug_concept_id == 2905077) |>
  addDailyDose(ingredientConceptId = 1125315) |>
  glimpse()
#> Rows: ??
#> Columns: 9
#> Database: DuckDB v1.0.0 [root@Darwin 23.4.0:R 4.4.1/:memory:]
#> $ drug_exposure_id         <int> 2, 6, 10, 16, 25, 26, 42, 46, 62, 64, 76, 90,…
#> $ person_id                <int> 1, 2, 3, 4, 7, 7, 17, 18, 22, 22, 25, 33, 36,…
#> $ drug_concept_id          <dbl> 2905077, 2905077, 2905077, 2905077, 2905077, …
#> $ drug_exposure_start_date <date> 2021-10-26, 2013-01-20, 2013-01-26, 2013-04-…
#> $ drug_exposure_end_date   <date> 2022-01-22, 2013-06-16, 2013-05-09, 2015-04-…
#> $ drug_type_concept_id     <dbl> 38000177, 38000177, 38000177, 38000177, 38000…
#> $ quantity                 <dbl> 90, 50, 15, 100, 100, 60, 15, 40, 50, 20, 30,…
#> $ daily_dose               <dbl> 9.707865e+03, 3.243243e+03, 1.384615e+03, 1.3…
#> $ unit                     <chr> "milligram", "milligram", "milligram", "milli…

The drug_exposure table was first according to drug_concept_id equaling \(2905077\) or not. This is to subset the drug_exposure table to the use of acetaminophen in the marketed product form. Later the ID \(1125315\) is supplied in addDailyDose() function, which corresponds to the ingredient form of acetaminophen.

We see that there are two extra columns at the end detailing the daily dose of acetaminophen of each drug exposure of interest - daily_dose and unit. Take first entry for example, we see that in that drug exposure of acetaminophen, the daily dose on average was \(48,000\) milligram. The way daily dose is being calculated is to take into consideration both the strength of the drug and the duration of the episode.

Add route information via addRoute() function

Next, the user could also obtain the information on the route of each drug used in the drug_exposure table.

cdm$drug_exposure |>
  addRoute() |>
  glimpse()
#> Rows: ??
#> Columns: 8
#> Database: DuckDB v1.0.0 [root@Darwin 23.4.0:R 4.4.1/:memory:]
#> $ drug_exposure_id         <int> 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 13, 15, 16, 1…
#> $ person_id                <int> 1, 1, 1, 1, 2, 2, 3, 3, 3, 4, 4, 4, 4, 5, 6, …
#> $ drug_concept_id          <dbl> 1503328, 2905077, 1539463, 1516980, 1516978, …
#> $ drug_exposure_start_date <date> 2022-01-11, 2021-10-26, 2021-08-10, 2021-12-…
#> $ drug_exposure_end_date   <date> 2022-01-26, 2022-01-22, 2021-11-22, 2022-01-…
#> $ drug_type_concept_id     <dbl> 38000177, 38000177, 38000177, 38000177, 38000…
#> $ quantity                 <dbl> 80, 90, 100, 35, 1, 50, 1, 100, 15, 45, 1, 80…
#> $ route                    <chr> "topical", "topical", "topical", "topical", "…

Similarly, this adds one extra column named route. For example, the drug used in the first entry of the drug_exposure table was an oral drug.

Finding out the pattern information using patternTable() function

The user could also find the patterns used in the drug_strength table. The output will also include a column of potentially valid and invalid combinations. The idea of a pattern to provide a platform to associate each drug in the drug_strength table with its constituent ingredients.

patternTable(cdm) |>
  glimpse()
#> Rows: 5
#> Columns: 12
#> $ pattern_id                  <dbl> 9, 18, 24, 40, NA
#> $ formula_name                <chr> "fixed amount formulation", "concentration…
#> $ validity                    <chr> "pattern with formula", "pattern with form…
#> $ number_concepts             <dbl> 7, 1, 1, 1, 4
#> $ number_ingredients          <dbl> 4, 1, 1, 1, 4
#> $ number_records              <dbl> 350, 67, 63, 74, 61
#> $ amount_numeric              <dbl> 1, 0, 0, 0, NA
#> $ amount_unit_concept_id      <dbl> 8576, NA, NA, NA, NA
#> $ numerator_numeric           <dbl> 0, 1, 1, 1, NA
#> $ numerator_unit_concept_id   <dbl> NA, 8576, 8510, 8510, NA
#> $ denominator_numeric         <dbl> 0, 1, 1, 0, NA
#> $ denominator_unit_concept_id <dbl> NA, 8587, 8587, 8587, NA

The output has three important columns, namely number_concepts, number_ingredients and number_records, which corresponds to count of distinct concepts in the patterns, count of distinct ingredients involved and overall count of records in the patterns respectively.

We also see that there is a column named pattern_id together with information such as number_concepts and number_ingredients. The idea is to use this output in conjunction with the data named patternWithFormula via pattern_id. Please see the data patternWithFormula for different patterns, their associated formula and combinations of amount_unit, numerator_unit and denominator_unit.

patternsWithFormula |>
  glimpse()
#> Rows: 41
#> Columns: 9
#> $ pattern_id       <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16…
#> $ amount           <chr> NA, NA, NA, NA, NA, "number", "number", "number", "nu…
#> $ amount_unit      <chr> NA, NA, NA, NA, NA, "international unit", "microgram"…
#> $ numerator        <chr> "number", "number", "number", "number", "number", NA,…
#> $ numerator_unit   <chr> "microgram", "milligram", "unit", "microgram", "milli…
#> $ denominator      <chr> "number", "number", "number", NA, NA, NA, NA, NA, NA,…
#> $ denominator_unit <chr> "hour", "hour", "hour", "hour", "hour", NA, NA, NA, N…
#> $ formula_name     <chr> "time based with denominator", "time based with denom…
#> $ formula          <chr> "if (denominator>24) {numerator * 24 / denominator} e…

Finding out the dose coverage using summariseDoseCoverage() function

This package also provides a function to check the coverage of daily dose computation for chosen concept sets and ingredients. Again let’s take acetaminophen as an example.

summariseDoseCoverage(cdm = cdm, ingredientConceptId = 1125315) |>
  glimpse()
#> Rows: 56
#> Columns: 13
#> $ result_id        <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
#> $ cdm_name         <chr> "DUS MOCK", "DUS MOCK", "DUS MOCK", "DUS MOCK", "DUS …
#> $ group_name       <chr> "ingredient_name", "ingredient_name", "ingredient_nam…
#> $ group_level      <chr> "acetaminophen", "acetaminophen", "acetaminophen", "a…
#> $ strata_name      <chr> "overall", "overall", "overall", "overall", "overall"…
#> $ strata_level     <chr> "overall", "overall", "overall", "overall", "overall"…
#> $ variable_name    <chr> "number records", "Missing dose", "Missing dose", "da…
#> $ variable_level   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, N…
#> $ estimate_name    <chr> "count", "count_missing", "percentage_missing", "mean…
#> $ estimate_type    <chr> "integer", "integer", "percentage", "numeric", "numer…
#> $ estimate_value   <chr> "174", "0", "0", "4986.89847658171", "22201.148944306…
#> $ additional_name  <chr> "overall", "overall", "overall", "overall", "overall"…
#> $ additional_level <chr> "overall", "overall", "overall", "overall", "overall"…

The output will summarise the usage of acetaminophen in the database. For example, overall there were \(13\) records of acetaminophen use and none missing in terms of the duration and strength. By default the output will also include the mean, median, lower and upper quartiles and standard deviation of the daily dose of acetaminophen (calculated using addDailyDose()). The results will also be stratified by unit, route (which we saw in addRoute() function) and pattern (which we saw in patternTable() function).

One may wish to display the output of summariseDoseCoverage() in a gt table form like so:

summariseDoseCoverage(cdm = cdm, ingredientConceptId = 1125315) |>
  tableDoseCoverage()
Database name Ingredient name Unit Route Pattern id Number records Missing dose Daily dose
N N (%) Mean (SD) Median (Q25 - Q75)
DUS MOCK Acetaminophen Overall Overall Overall 174 0 (0.00 %) 4,986.90 (22,201.15) 131.07 (18.45 - 936.23)
Milligram Overall Overall 174 0 (0.00 %) 4,986.90 (22,201.15) 131.07 (18.45 - 936.23)
Oral Overall 49 0 (0.00 %) 328.91 (956.18) 26.85 (7.81 - 188.98)
Topical Overall 125 0 (0.00 %) 6,812.83 (25,987.96) 261.78 (39.68 - 1,384.62)
Oral 9 49 0 (0.00 %) 328.91 (956.18) 26.85 (7.81 - 188.98)
Topical 18 67 0 (0.00 %) 11,277.48 (34,285.86) 970.92 (317.48 - 4,882.05)
9 58 0 (0.00 %) 1,655.39 (7,590.12) 35.55 (12.06 - 167.58)

The user also has the freedom to customize the gt table output. For example the following will suppress the cdmName:

summariseDoseCoverage(cdm = cdm, ingredientConceptId = 1125315) |>
  tableDoseCoverage(cdmName = F)
Ingredient name Unit Route Pattern id Number records Missing dose Daily dose
N N (%) Mean (SD) Median (Q25 - Q75)
Acetaminophen Overall Overall Overall 174 0 (0.00 %) 4,986.90 (22,201.15) 131.07 (18.45 - 936.23)
Milligram Overall Overall 174 0 (0.00 %) 4,986.90 (22,201.15) 131.07 (18.45 - 936.23)
Oral Overall 49 0 (0.00 %) 328.91 (956.18) 26.85 (7.81 - 188.98)
Topical Overall 125 0 (0.00 %) 6,812.83 (25,987.96) 261.78 (39.68 - 1,384.62)
Oral 9 49 0 (0.00 %) 328.91 (956.18) 26.85 (7.81 - 188.98)
Topical 18 67 0 (0.00 %) 11,277.48 (34,285.86) 970.92 (317.48 - 4,882.05)
9 58 0 (0.00 %) 1,655.39 (7,590.12) 35.55 (12.06 - 167.58)

Additionally, if the user wants to specify a title for the gt output, they could try the following:

summariseDoseCoverage(cdm = cdm, ingredientConceptId = 1125315) |>
  tableDoseCoverage(.options = list(title = "Title of summariseDoseCoverage"))
Title of summariseDoseCoverage
Database name Ingredient name Unit Route Pattern id Number records Missing dose Daily dose
N N (%) Mean (SD) Median (Q25 - Q75)
DUS MOCK Acetaminophen Overall Overall Overall 174 0 (0.00 %) 4,986.90 (22,201.15) 131.07 (18.45 - 936.23)
Milligram Overall Overall 174 0 (0.00 %) 4,986.90 (22,201.15) 131.07 (18.45 - 936.23)
Oral Overall 49 0 (0.00 %) 328.91 (956.18) 26.85 (7.81 - 188.98)
Topical Overall 125 0 (0.00 %) 6,812.83 (25,987.96) 261.78 (39.68 - 1,384.62)
Oral 9 49 0 (0.00 %) 328.91 (956.18) 26.85 (7.81 - 188.98)
Topical 18 67 0 (0.00 %) 11,277.48 (34,285.86) 970.92 (317.48 - 4,882.05)
9 58 0 (0.00 %) 1,655.39 (7,590.12) 35.55 (12.06 - 167.58)