EpiStandard

1. Overview
2. Key Features
3. Installation
4. Workflow with IncidencePrevalence
5. Usage Example
6. Main Functions

The EpiStandard R package provides functions for direct age standardisation of epidemiological measures—such as incidence and prevalence rates—enabling valid comparisons across populations or time periods with differing age structures.

---

Overview

Crude incidence and prevalence rates can be misleading when comparing different healthcare systems, regions, or historical eras with varying age demographics (for example, comparing an older European population to a younger demographic).

Direct standardisation applies stratum-specific rates observed in a study population to a standard reference population (e.g., European Standard Population [ESP 2013], WHO World Standard, or US Standard), calculating the overall rate that would be observed if the study population had the reference age distribution.

EpiStandard is built to work seamlessly with results generated by the DARWIN EU IncidencePrevalence package.

---

Key Features

• Direct Standardisation: Computes directly standardised rates (DSR) with 95% confidence intervals using the Gamma method (Fay & Feuer, 1997).
• Built-in Reference Populations: Includes standard population weights for European ("Europe" / ESP 2013), World ("World" / WHO), and US populations.
• Dynamic Age Band Merging: mergeAgeGroups() consolidates fine 5-year reference bands into study-specific custom age brackets (e.g. 0-19, 20-64, 65+).
• Missing Age Group Handling: Robust imputation and diagnostic handling when specific age strata are unobserved.
• Standardized Reporting: Summarizes results into clean, tidy analytical tables.

---

Installation

Install EpiStandard from CRAN or GitHub:

# From CRAN
install.packages("EpiStandard")

# Development version from GitHub
# install.packages("pak")
pak::pkg_install("oxford-pharmacoepi/EpiStandard")

---

Workflow with IncidencePrevalence

graph TD
    subgraph "1. OMOP Study Analysis"
        IP["IncidencePrevalence::estimateIncidence()<br/>Stratified by Age Group and Sex"]
    end

    subgraph "2. Reference Population Alignment"
        SP["standardPopulation('Europe')<br/>European Standard Population (ESP 2013)"]
        MA["mergeAgeGroups()<br/>Align Reference Bands to Study Strata"]
    end

    subgraph "3. Direct Age Standardisation"
        DSR["directlyStandardiseRates()<br/>Calculate Standardised Rates & Gamma 95% CI"]
    end

    subgraph "4. Output & Visualisation"
        RES["Tidy Standardised Incidence / Prevalence Summary"]
    end

    IP --> DSR
    SP --> MA --> DSR
    DSR --> RES

---

Usage Example

library(EpiStandard)
library(IncidencePrevalence)
library(omopgenerics)
library(dplyr)

# 0. Load mock CDM reference with incidence data
cdm <- mockIncidencePrevalence(sampleSize = 10000, outPre = 0.25)

# 1. Estimate stratified incidence rates
cdm <- generateDenominatorCohortSet(
  cdm = cdm,
  name = "denominator",
  cohortDateRange = as.Date(c("2010-01-01", "2020-01-01")),
  ageGroup = list(c(0, 19), c(20, 64), c(65, 150)),
  sex = c("Male", "Female"),
  daysPriorObservation = 0
)

inc <- estimateIncidence(
  cdm = cdm,
  denominatorTable = "denominator",
  outcomeTable = "outcome",
  interval = "years"
)

# 2. Prepare tidy incidence data
incidenceTidy <- inc |>
  asIncidenceResult()

# 3. Align European Standard Population to study age groups
standardPop <- mergeAgeGroups(
  standardPopulation("Europe"),
  newGroups = c("0 to 19", "20 to 64", "65 to 150")
) |>
  rename(denominator_age_group = age_group)

# 4. Compute directly standardised incidence rates
standardised_inc <- directlyStandardiseRates(
  data = incidenceTidy,
  refdata = standardPop,
  event = "outcome_count",
  denominator = "person_years",
  age = "denominator_age_group",
  pop = "pop",
  strata = c("incidence_start_date", "denominator_sex", "outcome_cohort_name")
)

standardised_inc |>
  select(incidence_start_date, denominator_sex, crude_rate, standardised_rate, standardised_rate_95CI_lower, standardised_rate_95CI_upper) |>
  glimpse()

---

Main Functions

Function	Purpose
directlyStandardiseRates()	Calculates directly standardised rates (DSR) and 95% confidence intervals using Gamma distribution methods.
standardPopulation()	Retrieves standard population weights (e.g. "Europe" for ESP 2013, "World" for WHO 2000-2025).
mergeAgeGroups()	Aggregates fine-grained reference age bands into broader target age groups.