DrugUtilisation

An R package to conduct Drug Utilisation Studies in the OMOP CDM

Introduction

Drug Utilisation Studies (DUS) were defined by the World Health Organization as studying the marketing, distribution, prescription, and use of medicinal products in a society, with special emphasis on the resulting medical and socioeconomic consequences (WHO, 2003).

This package aims to standardise and provide the tools to conduct Drug Utilisation studies as of the Darwin-EU Catalogue of Standard Analytics.

The package

install.packages("DrugUtilisation")

• v1.1.0 on cran.

• package website

• Report an issue/Feature request

Create a reference to a cdm object

For this example we are going to use the GiBleed synthetic test database:

library(omock)
library(duckdb)
library(CDMConnector)
library(dplyr)
library(CodelistGenerator)
library(DrugUtilisation)
library(CohortConstructor)
library(PatientProfiles)
library(ggplot2)
library(clock)
library(omopgenerics)
library(CohortSurvival)

cdm <- mockCdmFromDataset(datasetName = "GiBleed", source = "duckdb")
cdm$drug_exposure <- cdm$drug_exposure |>
  mutate(quantity = sql("floor(random() * 6) + 1")) |>
  mutate(quantity = case_when(
    quantity == 1 ~ 0,
    quantity == 2 ~ 1,
    quantity == 3 ~ 5,
    quantity == 4 ~ 10,
    quantity == 5 ~ 30,
    quantity == 6 ~ 100
  )) |>
  compute(name = "drug_exposure")

Generate a drug cohort

For this example we are going to use all records of acteaminophen:

codelist <- getDrugIngredientCodes(cdm = cdm, name = "acetaminophen", nameStyle = "{concept_name}")

cdm$dus_cohort <- conceptCohort(
  cdm = cdm,
  name = "dus_cohort",
  conceptSet = codelist
)

The results is the cdm object with the new cohort instantiated.

cdm

── # OMOP CDM reference (duckdb) of GiBleed ────────────────────────────────────────────────────────────────────────────

• omop tables: care_site, cdm_source, concept, concept_ancestor, concept_class, concept_relationship, concept_synonym,
condition_era, condition_occurrence, cost, death, device_exposure, domain, dose_era, drug_era, drug_exposure,
drug_strength, fact_relationship, location, measurement, metadata, note, note_nlp, observation, observation_period,
payer_plan_period, person, procedure_occurrence, provider, relationship, source_to_concept_map, specimen, visit_detail,
visit_occurrence, vocabulary

• cohort tables: dus_cohort

• achilles tables: -

• other tables: -

Cohort objects

The cohort objects have some attributes:

class(cdm$dus_cohort)

[1] "cohort_table"          "cdm_table"             "GeneratedCohortSet"    "tbl_duckdb_connection"
[5] "tbl_dbi"               "tbl_sql"               "tbl_lazy"              "tbl"                  

names(attributes(cdm$dus_cohort))

[1] "names"            "class"            "tbl_source"       "tbl_name"         "cohort_set"       "cohort_attrition"
[7] "cohort_codelist"  "cdm_reference"   

• Settings
• Counts
• Attrition
• Codelist

Cohort attributes

Settings:

settings(cdm$dus_cohort)

# A tibble: 1 × 4
  cohort_definition_id cohort_name   cdm_version vocabulary_version
                 <int> <chr>         <chr>       <chr>             
1                    1 acetaminophen 5.3         v5.0 18-JAN-19    

Cohort counts:

cohortCount(cdm$dus_cohort)

# A tibble: 1 × 3
  cohort_definition_id number_records number_subjects
                 <int>          <int>           <int>
1                    1          13908            2679

Cohort attributes

Cohort attrition:

attrition(cdm$dus_cohort)

# A tibble: 4 × 7
  cohort_definition_id number_records number_subjects reason_id reason                excluded_records excluded_subjects
                 <int>          <int>           <int>     <int> <chr>                            <int>             <int>
1                    1          14205            2679         1 Initial qualifying e…                0                 0
2                    1          14205            2679         2 Record in observation                0                 0
3                    1          14205            2679         3 Not missing record d…                0                 0
4                    1          13908            2679         4 Merge overlapping re…              297                 0

Cohort codelist:

cohortCodelist(cdm$dus_cohort, cohortId = 1)

- acetaminophen (7 codes)

gapEra parameter

gapEra parameter can be used to concatenate different records.

Number of days between two continuous exposures to be considered in the same era. Records that have fewer days between them than this gap will be concatenated into the same cohort record.

gapEra parameter

Generate a drug cohort

cohortCount(cdm$dus_cohort)

# A tibble: 1 × 3
  cohort_definition_id number_records number_subjects
                 <int>          <int>           <int>
1                    1          13908            2679

cdm$dus_cohort <- cdm$dus_cohort |>
  collapseCohorts(gap = 30)

cohortCount(cdm$dus_cohort)

# A tibble: 1 × 3
  cohort_definition_id number_records number_subjects
                 <int>          <int>           <int>
1                    1          13860            2679

attrition(cdm$dus_cohort)

# A tibble: 5 × 7
  cohort_definition_id number_records number_subjects reason_id reason                excluded_records excluded_subjects
                 <int>          <int>           <int>     <int> <chr>                            <int>             <int>
1                    1          14205            2679         1 Initial qualifying e…                0                 0
2                    1          14205            2679         2 Record in observation                0                 0
3                    1          14205            2679         3 Not missing record d…                0                 0
4                    1          13908            2679         4 Merge overlapping re…              297                 0
5                    1          13860            2679         5 Collapse cohort with…               48                 0

Inclusion criteria

Four functions to add inclusion criteria to our cohort:

• Washout requirePriorDrugWashout() (requireConceptIntersect)
• First observation requireIsFirstDrugEntry() (requireIsFirstEntry)
• Prior observation requireObservationBeforeDrug() (requirePriorObservation)
• Date range requireDrugInDateRange() (requireInDateRange)

Important considerations

Not all inclusion criteria are commutable operations (A + B != B + A):

Not commutable:

• prior washout + prior observation != prior observation + prior washout
• prior washout + first != first + prior washout
• prior washout + date range != date range + prior washout
• first + prior observation != prior observation + first
• first + date range != date range + first

Commutable:

• date range + prior observation == prior observation + date range

Recommended order

Having all this into account the recommended order to apply criteria would be:

1. Require a prior drug washout or require first drug entry (particular case).

2. Require a prior observation before the drug episode.

3. Require the drugs to be in a certain date range.

Vignette to create cohorts

https://darwin-eu-dev.github.io/DrugUtilisation/articles/create_cohorts.html

Drug usage

result <- cdm$dus_cohort |>
  summariseDrugUtilisation(
    ingredientConceptId = 1125315L, 
    indexDate = "cohort_start_date",
    censorDate = "cohort_end_date", 
    restrictIncident = TRUE, 
    gapEra = 30, 
    numberExposures = TRUE, 
    numberEras = TRUE,
    daysExposed = TRUE,
    daysPrescribed = TRUE,
    timeToExposure = FALSE, 
    initialQuantity = TRUE, 
    cumulativeQuantity = TRUE, 
    initialDailyDose = TRUE, 
    cumulativeDose = TRUE,
    estimates = c("q25", "median", "q75")
  )

Drug usage

result |>
  suppress(minCellCount = 5) |>
  tableDrugUtilisation()

Concept set	Ingredient	Variable name	Estimate name	CDM name
				GiBleed
acetaminophen
overall	overall	number records	N	13,860
		number subjects	N	2,679
ingredient_1125315_descendants	overall	number exposures	Median (Q25 - Q75)	1 (1 - 1)
		cumulative quantity	Median (Q25 - Q75)	10.00 (1.00 - 30.00)
		initial quantity	Median (Q25 - Q75)	10.00 (1.00 - 30.00)
		initial exposure duration	Median (Q25 - Q75)	15 (8 - 15)
		number eras	Median (Q25 - Q75)	1 (1 - 1)
		days exposed	Median (Q25 - Q75)	15 (8 - 15)
		days prescribed	Median (Q25 - Q75)	15 (8 - 15)
	Acetaminophen	cumulative dose milligram	Median (Q25 - Q75)	1,625.00 (217.00 - 9,750.00)
		initial daily dose milligram	Median (Q25 - Q75)	108.33 (13.56 - 443.18)

Indications

summariseIndication():

• To summarise mutually exclusive indications.

• Define a window respect to the ‘cohort_start_date’.

• Indications must be instantiated beforehand as cohorts.

• Unknown indication (check a table to see if there is a record).

Indications

Let’s instantiate the cohorts of interest:

cdm$indications <- conceptCohort(
  cdm = cdm,
  conceptSet = list(
    arteriosclerosis = 317576, 
    myocardial_infarction = 4329847, 
    pain = getCandidateCodes(cdm = cdm, keywords = "pain")$concept_id
  ),
  name = "indications"
)

Indciations

result <- cdm$dus_cohort |>
  summariseIndication(
    indicationCohortName = "indications", 
    indicationWindow = list(c(0, 0), c(-30, 7)), 
    unknownIndicationTable = "condition_occurrence"
  )
result |>
  glimpse()

Rows: 40
Columns: 13
$ result_id        <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ cdm_name         <chr> "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBl…
$ group_name       <chr> "cohort_name", "cohort_name", "cohort_name", "cohort_name", "cohort_name", "cohort_name", "co…
$ group_level      <chr> "acetaminophen", "acetaminophen", "acetaminophen", "acetaminophen", "acetaminophen", "acetami…
$ strata_name      <chr> "overall", "overall", "overall", "overall", "overall", "overall", "overall", "overall", "over…
$ strata_level     <chr> "overall", "overall", "overall", "overall", "overall", "overall", "overall", "overall", "over…
$ variable_name    <chr> "Indication on index date", "Indication on index date", "Indication on index date", "Indicati…
$ variable_level   <chr> "arteriosclerosis", "arteriosclerosis", "myocardial_infarction", "myocardial_infarction", "pa…
$ estimate_name    <chr> "count", "percentage", "count", "percentage", "count", "percentage", "count", "percentage", "…
$ estimate_type    <chr> "integer", "percentage", "integer", "percentage", "integer", "percentage", "integer", "percen…
$ estimate_value   <chr> "0", "0", "64", "0.461760461760462", "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", "13633…
$ additional_name  <chr> "window_name", "window_name", "window_name", "window_name", "window_name", "window_name", "wi…
$ additional_level <chr> "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 …

Indications

result |>
  suppress() |>
  tableIndication()

		CDM name
		GiBleed
Indication	Estimate name	Cohort name
		acetaminophen
Indication on index date
arteriosclerosis	N (%)	0 (0.00 %)
myocardial_infarction	N (%)	64 (0.46 %)
pain	N (%)	0 (0.00 %)
arteriosclerosis and myocardial_infarction	N (%)	0 (0.00 %)
arteriosclerosis and pain	N (%)	0 (0.00 %)
myocardial_infarction and pain	N (%)	0 (0.00 %)
arteriosclerosis and myocardial_infarction and pain	N (%)	0 (0.00 %)
unknown	N (%)	13,633 (98.36 %)
none	N (%)	163 (1.18 %)
not in observation	N (%)	0 (0.00 %)
Indication from 30 days before to 7 days after the index date
arteriosclerosis	N (%)	5 (0.04 %)
myocardial_infarction	N (%)	64 (0.46 %)
pain	N (%)	0 (0.00 %)
arteriosclerosis and myocardial_infarction	N (%)	0 (0.00 %)
arteriosclerosis and pain	N (%)	0 (0.00 %)
myocardial_infarction and pain	N (%)	0 (0.00 %)
arteriosclerosis and myocardial_infarction and pain	N (%)	0 (0.00 %)
unknown	N (%)	13,784 (99.45 %)
none	N (%)	7 (0.05 %)
not in observation	N (%)	0 (0.00 %)

Indications

plotIndication(result)

Indications

Output of ploting functions are ggplot2 objects.

plotIndication(result) +
  theme(legend.position = "none")

Indications

Currently considering to add support for plotly.

library(plotly)
plotIndication(result) |>
  ggplotly()

Indications (new plot)

Let’s say we conducted the analysis stratified by sex:

result <- cdm$dus_cohort |>
  addSex() |>
  summariseIndication(
    strata = "sex",
    indicationCohortName = "indications", 
    indicationWindow = list(c(0, 0), c(-30, 7)), 
    unknownIndicationTable = "condition_occurrence"
  )
result |>
  glimpse()

Rows: 120
Columns: 13
$ result_id        <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ cdm_name         <chr> "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBl…
$ group_name       <chr> "cohort_name", "cohort_name", "cohort_name", "cohort_name", "cohort_name", "cohort_name", "co…
$ group_level      <chr> "acetaminophen", "acetaminophen", "acetaminophen", "acetaminophen", "acetaminophen", "acetami…
$ strata_name      <chr> "overall", "overall", "overall", "overall", "overall", "overall", "overall", "overall", "over…
$ strata_level     <chr> "overall", "overall", "overall", "overall", "overall", "overall", "overall", "overall", "over…
$ variable_name    <chr> "Indication on index date", "Indication on index date", "Indication on index date", "Indicati…
$ variable_level   <chr> "arteriosclerosis", "arteriosclerosis", "myocardial_infarction", "myocardial_infarction", "pa…
$ estimate_name    <chr> "count", "percentage", "count", "percentage", "count", "percentage", "count", "percentage", "…
$ estimate_type    <chr> "integer", "percentage", "integer", "percentage", "integer", "percentage", "integer", "percen…
$ estimate_value   <chr> "0", "0", "64", "0.461760461760462", "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", "13633…
$ additional_name  <chr> "window_name", "window_name", "window_name", "window_name", "window_name", "window_name", "wi…
$ additional_level <chr> "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 to 0", "0 …

Indications (new plot)

plotIndication(result = result, x = "sex")

Indications (new plot)

You can do the same plot by year:

result <- cdm$dus_cohort |>
  requireDrugInDateRange(
    name = "new_cohort",
    dateRange = as.Date(c("1990-01-01", "1999-12-31"))
  ) |>
  mutate(year = get_year(cohort_start_date)) |>
  summariseIndication(
    strata = "year",
    indicationCohortName = "indications", 
    indicationWindow = c(-Inf, 0), 
    unknownIndicationTable = "condition_occurrence"
  )
result |>
  filterStrata(year != "overall") |>
  plotIndication(x = "year")

Indications (new plot)

Treatment persistence

To analyse treatment persistence and adherence there are two standard pipelines:

• Survival analysis (implemented in the CohortSurvival package)

• Proportion of patients covered

Survival Analysis

survivalResults <- estimateSingleEventSurvival(
  cdm = cdm, 
  targetCohortTable = "dus_cohort", 
  outcomeCohortTable = "dus_cohort", 
  outcomeDateVariable = "cohort_end_date", 
  followUpDays = 90
)

Survival analysis

plotSurvival(result = survivalResults)

Proportion of patients covered

Proportion of treated individuals between index date and followUpDays. Percentages will be reported daily.

result <- cdm$dus_cohort |>
  summariseProportionOfPatientsCovered(followUpDays = 90)

Proportion of patients covered

tableProportionOfPatientsCovered(result)

Time	Estimate name	Cohort name
		acetaminophen
GiBleed
0	PPC (95%CI)	100.00% [99.86% - 100.00%]
1	PPC (95%CI)	98.69% [98.19% - 99.06%]
2	PPC (95%CI)	98.69% [98.19% - 99.06%]
3	PPC (95%CI)	98.69% [98.19% - 99.06%]
4	PPC (95%CI)	98.69% [98.19% - 99.06%]
5	PPC (95%CI)	98.69% [98.19% - 99.06%]
6	PPC (95%CI)	98.69% [98.19% - 99.06%]
7	PPC (95%CI)	98.66% [98.15% - 99.03%]
8	PPC (95%CI)	74.73% [73.05% - 76.34%]
9	PPC (95%CI)	74.73% [73.05% - 76.34%]
10	PPC (95%CI)	74.73% [73.05% - 76.34%]
11	PPC (95%CI)	74.73% [73.05% - 76.34%]
12	PPC (95%CI)	74.73% [73.05% - 76.34%]
13	PPC (95%CI)	74.73% [73.05% - 76.34%]
14	PPC (95%CI)	74.62% [72.93% - 76.23%]
15	PPC (95%CI)	19.48% [18.03% - 21.03%]
16	PPC (95%CI)	19.37% [17.92% - 20.91%]
17	PPC (95%CI)	19.26% [17.81% - 20.80%]
18	PPC (95%CI)	19.26% [17.81% - 20.80%]
19	PPC (95%CI)	19.22% [17.78% - 20.76%]
20	PPC (95%CI)	19.22% [17.78% - 20.76%]
21	PPC (95%CI)	19.19% [17.74% - 20.72%]
22	PPC (95%CI)	14.37% [13.09% - 15.75%]
23	PPC (95%CI)	14.33% [13.06% - 15.71%]
24	PPC (95%CI)	14.33% [13.06% - 15.71%]
25	PPC (95%CI)	14.30% [13.02% - 15.67%]
26	PPC (95%CI)	14.30% [13.02% - 15.67%]
27	PPC (95%CI)	14.30% [13.02% - 15.67%]
28	PPC (95%CI)	14.30% [13.02% - 15.67%]
29	PPC (95%CI)	11.35% [10.20% - 12.60%]
30	PPC (95%CI)	11.27% [10.13% - 12.53%]
31	PPC (95%CI)	8.40% [7.41% - 9.51%]
32	PPC (95%CI)	8.36% [7.37% - 9.47%]
33	PPC (95%CI)	8.36% [7.37% - 9.47%]
34	PPC (95%CI)	8.36% [7.37% - 9.47%]
35	PPC (95%CI)	8.32% [7.34% - 9.43%]
36	PPC (95%CI)	5.97% [5.14% - 6.93%]
37	PPC (95%CI)	5.97% [5.14% - 6.93%]
38	PPC (95%CI)	5.71% [4.89% - 6.65%]
39	PPC (95%CI)	5.64% [4.82% - 6.57%]
40	PPC (95%CI)	5.56% [4.76% - 6.49%]
41	PPC (95%CI)	5.56% [4.76% - 6.49%]
42	PPC (95%CI)	5.56% [4.76% - 6.49%]
43	PPC (95%CI)	5.56% [4.76% - 6.49%]
44	PPC (95%CI)	5.52% [4.72% - 6.45%]
45	PPC (95%CI)	5.45% [4.65% - 6.37%]
46	PPC (95%CI)	5.45% [4.65% - 6.37%]
47	PPC (95%CI)	5.45% [4.65% - 6.37%]
48	PPC (95%CI)	5.56% [4.76% - 6.49%]
49	PPC (95%CI)	5.56% [4.76% - 6.49%]
50	PPC (95%CI)	5.56% [4.76% - 6.49%]
51	PPC (95%CI)	5.52% [4.72% - 6.45%]
52	PPC (95%CI)	5.52% [4.72% - 6.45%]
53	PPC (95%CI)	5.52% [4.72% - 6.45%]
54	PPC (95%CI)	5.52% [4.72% - 6.45%]
55	PPC (95%CI)	5.52% [4.72% - 6.45%]
56	PPC (95%CI)	5.49% [4.69% - 6.41%]
57	PPC (95%CI)	5.49% [4.69% - 6.41%]
58	PPC (95%CI)	5.75% [4.93% - 6.69%]
59	PPC (95%CI)	5.75% [4.93% - 6.69%]
60	PPC (95%CI)	5.82% [5.00% - 6.77%]
61	PPC (95%CI)	3.28% [2.67% - 4.03%]
62	PPC (95%CI)	3.40% [2.77% - 4.15%]
63	PPC (95%CI)	3.32% [2.71% - 4.07%]
64	PPC (95%CI)	3.32% [2.71% - 4.07%]
65	PPC (95%CI)	3.32% [2.71% - 4.07%]
66	PPC (95%CI)	3.40% [2.77% - 4.15%]
67	PPC (95%CI)	3.40% [2.77% - 4.15%]
68	PPC (95%CI)	3.17% [2.57% - 3.91%]
69	PPC (95%CI)	3.17% [2.57% - 3.91%]
70	PPC (95%CI)	3.10% [2.51% - 3.82%]
71	PPC (95%CI)	3.10% [2.51% - 3.82%]
72	PPC (95%CI)	3.06% [2.47% - 3.78%]
73	PPC (95%CI)	2.87% [2.31% - 3.58%]
74	PPC (95%CI)	2.84% [2.27% - 3.54%]
75	PPC (95%CI)	2.80% [2.24% - 3.49%]
76	PPC (95%CI)	2.76% [2.21% - 3.45%]
77	PPC (95%CI)	2.76% [2.21% - 3.45%]
78	PPC (95%CI)	2.84% [2.27% - 3.54%]
79	PPC (95%CI)	2.87% [2.31% - 3.58%]
80	PPC (95%CI)	2.87% [2.31% - 3.58%]
81	PPC (95%CI)	2.80% [2.24% - 3.49%]
82	PPC (95%CI)	2.80% [2.24% - 3.49%]
83	PPC (95%CI)	2.80% [2.24% - 3.49%]
84	PPC (95%CI)	2.80% [2.24% - 3.49%]
85	PPC (95%CI)	2.80% [2.24% - 3.49%]
86	PPC (95%CI)	2.80% [2.24% - 3.49%]
87	PPC (95%CI)	2.80% [2.24% - 3.49%]
88	PPC (95%CI)	2.84% [2.27% - 3.54%]
89	PPC (95%CI)	2.91% [2.34% - 3.62%]
90	PPC (95%CI)	2.91% [2.34% - 3.62%]

Proportion of patients covered

plotProportionOfPatientsCovered(result)

gapEra parameter can have a big impact in survival analysis, whereas it does not have such big impact on proportion of patients covered.

Drug restart and drug switching

We have the ability to study drug restart or drug switching after a discontinuation:

• Switching cohorts must be defined in advance.

• We have to define the windows of interest.

• Index date will be the date of discontinuation.

• Reported percentages in each window will be:

• Restart: individuals that restarted and not switched.

• Switch: individuals that switched and not restarted.

• Restart and switch: individuals that restarted and switched.

• Untreated: individuals that did not restart or switch.

Drug restart and drug switching

Let’s define the cohorts of interest for switching:

codelist <- getDrugIngredientCodes(
  cdm = cdm, 
  name = c("clopidogrel", "simvastatin", "warfarin"), 
  nameStyle = "{concept_name}"
)
cdm$switch <- conceptCohort(
  cdm = cdm, 
  name = "switch",
  conceptSet = codelist
)

Drug restart and drug switching

We have to be careful with the inclusion criteria of the input cohort, as restart is derived from it.

result <- cdm$dus_cohort |>
  summariseDrugRestart(
    switchCohortTable = "switch", 
    followUpDays = c(90, 180, 270, 360), 
    restrictToFirstDiscontinuation = TRUE
  )
result |>
  glimpse()

Rows: 32
Columns: 13
$ result_id        <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
$ cdm_name         <chr> "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBleed", "GiBl…
$ group_name       <chr> "cohort_name", "cohort_name", "cohort_name", "cohort_name", "cohort_name", "cohort_name", "co…
$ group_level      <chr> "acetaminophen", "acetaminophen", "acetaminophen", "acetaminophen", "acetaminophen", "acetami…
$ strata_name      <chr> "overall", "overall", "overall", "overall", "overall", "overall", "overall", "overall", "over…
$ strata_level     <chr> "overall", "overall", "overall", "overall", "overall", "overall", "overall", "overall", "over…
$ variable_name    <chr> "Drug restart in 90 days", "Drug restart in 90 days", "Drug restart in 90 days", "Drug restar…
$ variable_level   <chr> "restart", "restart", "switch", "switch", "restart and switch", "restart and switch", "untrea…
$ estimate_name    <chr> "count", "percentage", "count", "percentage", "count", "percentage", "count", "percentage", "…
$ estimate_type    <chr> "integer", "percentage", "integer", "percentage", "integer", "percentage", "integer", "percen…
$ estimate_value   <chr> "43", "1.60507652108996", "0", "0", "0", "0", "2636", "98.39492347891", "97", "3.620754012691…
$ additional_name  <chr> "follow_up_days", "follow_up_days", "follow_up_days", "follow_up_days", "follow_up_days", "fo…
$ additional_level <chr> "90 days", "90 days", "90 days", "90 days", "90 days", "90 days", "90 days", "90 days", "180 …

Drug restart and drug switching

tableDrugRestart(result)

		CDM name
		GiBleed
Treatment	Estimate name	Cohort name
		acetaminophen
Drug restart in 90 days
restart	N (%)	43 (1.61 %)
switch	N (%)	0 (0.00 %)
restart and switch	N (%)	0 (0.00 %)
untreated	N (%)	2,636 (98.39 %)
Drug restart in 180 days
restart	N (%)	97 (3.62 %)
switch	N (%)	0 (0.00 %)
restart and switch	N (%)	0 (0.00 %)
untreated	N (%)	2,582 (96.38 %)
Drug restart in 270 days
restart	N (%)	175 (6.53 %)
switch	N (%)	0 (0.00 %)
restart and switch	N (%)	0 (0.00 %)
untreated	N (%)	2,504 (93.47 %)
Drug restart in 360 days
restart	N (%)	244 (9.11 %)
switch	N (%)	1 (0.04 %)
restart and switch	N (%)	0 (0.00 %)
untreated	N (%)	2,434 (90.85 %)

Drug restart and drug switching

plotDrugRestart(result)

Summarise treatments

summariseTreatment() is a general function to analyse presence of treatments (cohorts) after an index date. This can be used with many different purposes:

• Summarise treatments after a certain condition

• Analyse comedications

• Analyse treatments after discontinuation

• …

Let’s create some treatments cohorts

codelist <- getDrugIngredientCodes(
  cdm = cdm, 
  name = c("clopidogrel", "simvastatin", "warfarin", "nitroglycerin"),
  nameStyle = "{concept_name}"
)
cdm$treatments <- conceptCohort(
  cdm = cdm, 
  name = "treatments",
  conceptSet = codelist
)
cdm$osteoarthritis <- conceptCohort(
  cdm = cdm,
  conceptSet = list(osteoarthritis = 80180L),
  name = "osteoarthritis"
)

Summarise treatments after condition

result <- cdm$osteoarthritis |>
  summariseTreatment(
    window = list(c(1, 90), c(91, 180), c(181, 270), c(271, 360)), 
    treatmentCohortName = "treatments", 
    indexDate = "cohort_start_date", 
    censorDate = NULL
  )

Summarise treatments after condition

tableTreatment(result, style = "darwin")

		CDM name
		GiBleed
Treatment	Estimate name	Cohort name
		osteoarthritis
Medication from 1 day after to 90 days after the index date
clopidogrel	N (%)	9 (0.34 %)
nitroglycerin	N (%)	20 (0.76 %)
simvastatin	N (%)	1 (0.04 %)
warfarin	N (%)	0 (0.00 %)
untreated	N (%)	2,608 (99.16 %)
not in observation	N (%)	0 (0.00 %)
Medication from 91 days after to 180 days after the index date
clopidogrel	N (%)	10 (0.38 %)
nitroglycerin	N (%)	20 (0.76 %)
simvastatin	N (%)	0 (0.00 %)
warfarin	N (%)	0 (0.00 %)
untreated	N (%)	2,583 (98.21 %)
not in observation	N (%)	25 (0.95 %)
Medication from 181 days after to 270 days after the index date
clopidogrel	N (%)	9 (0.34 %)
nitroglycerin	N (%)	20 (0.76 %)
simvastatin	N (%)	1 (0.04 %)
warfarin	N (%)	0 (0.00 %)
untreated	N (%)	2,569 (97.68 %)
not in observation	N (%)	40 (1.52 %)
Medication from 271 days after to 360 days after the index date
clopidogrel	N (%)	11 (0.42 %)
nitroglycerin	N (%)	21 (0.80 %)
simvastatin	N (%)	1 (0.04 %)
warfarin	N (%)	0 (0.00 %)
untreated	N (%)	2,547 (96.84 %)
not in observation	N (%)	60 (2.28 %)

Summarise treatments after condition

plotTreatment(result)

Summarise treatments after discontinuation

result <- cdm$dus_cohort |>
  summariseTreatment(
    window = list(c(1, 90), c(91, 180), c(181, 270), c(271, 360)), 
    treatmentCohortName = "treatments", 
    indexDate = "cohort_end_date", 
    censorDate = NULL
  )

Summarise treatments after discontinuation

plotTreatment(result)

Summarise comedications

comedications <- cdm$dus_cohort |>
  summariseTreatment(
    window = c(0, Inf), 
    treatmentCohortName = "treatments", 
    indexDate = "cohort_start_date", 
    censorDate = "cohort_end_date"
  )

Summarise comedications

plotTreatment(comedications, style = "darwin")

DrugUtilisation

👉 Packages website
👉 CRAN link
👉 Manual

📧 marti.catalasabate@ndorms.ox.ac.uk