Title: Distributional Cost-Effectiveness Analysis for Health Technology Assessment
Version: 0.1.0
Description: Implements distributional cost-effectiveness analysis (DCEA) as described in Cookson et al. (2020, ISBN:9780198838197) and the methods endorsed by NICE (2025) for health technology evaluation. Provides functions for both aggregate and full-form DCEA, inequality measurement (Atkinson index, Gini coefficient, slope index of inequality, relative index of inequality), social welfare function evaluation, equity-efficiency impact plane visualisation, and sensitivity analysis over inequality aversion parameters. Includes baseline health distributions for England (by IMD quintile), Canada (income quintile), and global WHO regions. Suitable for academic research, health technology assessment submissions, and public health policy analysis.
License: MIT + file LICENSE
Encoding: UTF-8
LazyData: true
RoxygenNote: 7.3.3
Language: en-GB
URL: https://heorlytics.github.io/dceasimR/
BugReports: https://github.com/heorlytics/dceasimR/issues
Imports: ggplot2 (≥ 3.4.0), dplyr (≥ 1.1.0), rlang (≥ 1.1.0), tibble (≥ 3.2.0), scales (≥ 1.3.0), stats
Suggests: knitr, rmarkdown, testthat (≥ 3.0.0), openxlsx, flextable, covr, vdiffr
VignetteBuilder: knitr
Config/testthat/edition: 3
Depends: R (≥ 4.1.0)
NeedsCompilation: no
Packaged: 2026-04-16 16:40:07 UTC; ShubhramPandey
Author: Shubhram Pandey ORCID iD [aut, cre]
Maintainer: Shubhram Pandey <shubhram.pandey@heorlytics.com>
Repository: CRAN
Date/Publication: 2026-04-21 19:02:26 UTC

dceasimR: Distributional Cost-Effectiveness Analysis for Health Technology Assessment

Description

Implements distributional cost-effectiveness analysis (DCEA) as described in Cookson et al. (2020) and methods endorsed by NICE (2025). Provides functions for aggregate and full-form DCEA, inequality measurement, social welfare function evaluation, equity-efficiency impact plane visualisation, and sensitivity analysis.

Main functions

Key references

Cookson R, Griffin S, Norheim OF, Culyer AJ (2020). Distributional Cost-Effectiveness Analysis. Oxford University Press. Oxford University Press (ISBN:9780198838197).

Love-Koh J, Asaria M, Cookson R, Griffin S (2019). The Social Distribution of Health: Estimating Quality-Adjusted Life Expectancy in England. Value in Health 22(5): 518-526. doi:10.1016/j.jval.2018.10.007

Asaria M, Griffin S, Cookson R (2016). Distributional Cost-Effectiveness Analysis: A Tutorial. Medical Decision Making 36(1): 8-19. doi:10.1177/0272989X15583266

Author(s)

Maintainer: Shubhram Pandey shubhram.pandey@heorlytics.com (ORCID)

See Also

Useful links:

Build inequality staircase data from component inputs

Description

Constructs the five-step inequality staircase data frame from individual component vectors. The staircase traces how the distribution of health gains is shaped at each stage: prevalence, eligibility, uptake, clinical effect, and net opportunity cost.

Usage

build_staircase_data(
  group,
  group_labels,
  prevalence,
  eligibility,
  uptake,
  clinical_effect,
  opportunity_cost
)

Arguments

group

Integer vector of group identifiers (1 = most deprived).

group_labels

Character vector of group labels.

prevalence

Numeric vector: disease prevalence by group (proportion).

eligibility

Numeric vector: proportion eligible for the intervention by group.

uptake

Numeric vector: uptake rate by group (0-1).

clinical_effect

Numeric vector: incremental QALY gain by group.

opportunity_cost

Numeric vector: QALYs displaced per group via budget impact.

Value

A tibble in long format suitable for plot_inequality_staircase.

Examples

build_staircase_data(
  group         = 1:5,
  group_labels  = paste("IMD Q", 1:5),
  prevalence    = c(0.08, 0.07, 0.06, 0.05, 0.04),
  eligibility   = c(0.70, 0.72, 0.74, 0.76, 0.78),
  uptake        = c(0.60, 0.64, 0.68, 0.72, 0.76),
  clinical_effect = c(0.30, 0.38, 0.45, 0.52, 0.58),
  opportunity_cost = c(0.05, 0.05, 0.05, 0.05, 0.05)
)

Calculate all inequality indices in one call

Description

Convenience wrapper that computes SII, RII, concentration index, Atkinson index (for multiple \varepsilon values), and Gini coefficient and returns them as a tidy data frame.

Usage

calc_all_inequality_indices(
  data,
  health_var,
  group_var,
  weight_var,
  epsilon_values = c(0.5, 1, 2)
)

Arguments

data

A data frame with health and group columns.

health_var

Name of the health variable column (character).

group_var

Name of the socioeconomic group column (ordered integer, 1 = most deprived).

weight_var

Name of the population share column (sums to 1).

epsilon_values

Numeric vector of \varepsilon values for the Atkinson index (default: c(0.5, 1, 2)).

Value

A tibble with columns index, value, and description.

Examples

df <- tibble::tibble(
  group      = 1:5,
  mean_hale  = c(60, 63, 66, 69, 72),
  pop_share  = rep(0.2, 5)
)
calc_all_inequality_indices(df, "mean_hale", "group", "pop_share")

Calculate Atkinson Index of Health Inequality

Description

The Atkinson index measures the extent of inequality in the health distribution, explicitly incorporating a parameter \varepsilon representing inequality aversion. Higher \varepsilon values give more weight to health differences at the bottom of the distribution.

Usage

calc_atkinson_index(health_dist, pop_weights, epsilon = 1)

Arguments

health_dist

Numeric vector of health values across population groups.

pop_weights

Numeric vector of population weights (need not sum to 1; will be normalised internally).

epsilon

Inequality aversion parameter (default = 1). Must be non-negative. When \varepsilon = 0, returns 0 (no aversion).

Value

Atkinson index value in [0, 1]. A value of 0 indicates perfect equality; a value approaching 1 indicates maximum inequality.

References

Atkinson AB (1970). On the Measurement of Inequality. Journal of Economic Theory 2(3): 244-263. doi:10.1016/0022-0531(70)90039-6

Examples

# Perfect equality
calc_atkinson_index(rep(70, 5), rep(0.2, 5), epsilon = 1)

# Gradient across groups
calc_atkinson_index(c(60, 63, 66, 69, 72), rep(0.2, 5), epsilon = 1)

Calculate Concentration Index

Description

Measures the degree to which a health variable is concentrated among socioeconomically advantaged or disadvantaged groups. A negative value indicates the health variable (e.g., illness) is concentrated among the deprived; a positive value indicates concentration among the advantaged.

Usage

calc_concentration_index(
  data,
  health_var,
  group_var,
  weight_var,
  rank_var = NULL,
  type = c("standard", "erreygers", "wagstaff")
)

Arguments

data

A data frame with health and group columns.

health_var

Name of the health variable column (character).

group_var

Name of the socioeconomic group column (ordered integer, 1 = most deprived).

weight_var

Name of the population share column (sums to 1).

rank_var

Name of the socioeconomic rank variable (ridit scores, 0 = lowest, 1 = highest). If NULL, computed from group_var and weight_var using ridit scoring.

type

Concentration index variant: "standard" (Kakwani), "erreygers" (bounded), or "wagstaff" (normalised).

Value

A named list with ci (concentration index), se, and type.

References

Erreygers G (2009). Correcting the Concentration Index. Journal of Health Economics 28(2): 504-515. doi:10.1016/j.jhealeco.2008.02.003

Examples

df <- tibble::tibble(
  group      = 1:5,
  mean_hale  = c(60, 63, 66, 69, 72),
  pop_share  = rep(0.2, 5)
)
calc_concentration_index(df, "mean_hale", "group", "pop_share")

Calculate Equally Distributed Equivalent Health (EDE)

Description

Uses the Atkinson social welfare function to calculate EDE health — the level of health that, if equally distributed, would generate the same social welfare as the actual distribution given inequality aversion parameter \eta.

Usage

calc_ede(health_dist, pop_weights, eta = 1)

Arguments

health_dist

Numeric vector of health values by group (must be strictly positive).

pop_weights

Numeric vector of population weights (will be normalised to sum to 1).

eta

Inequality aversion parameter (numeric scalar, default = 1).

  • \eta = 0: returns arithmetic mean (no inequality aversion).

  • \eta = 1: returns geometric mean (moderate aversion).

  • \eta > 1: increasing inequality aversion.

  • NICE relevant range: 0 to 10.

Details

\text{EDE}(\eta) = \left(\sum_i w_i h_i^{1-\eta} \big/ \sum_i w_i\right)^{1/(1-\eta)} \quad \text{for } \eta \neq 1

\text{EDE}(1) = \exp\!\left(\sum_i w_i \ln(h_i)\right) \quad \text{(geometric mean, } \eta = 1\text{)}

Value

EDE health value (numeric scalar). Returns NA with a warning if any health values are non-positive.

References

Atkinson AB (1970). On the Measurement of Inequality. Journal of Economic Theory 2(3): 244-263. doi:10.1016/0022-0531(70)90039-6

Examples

health  <- c(60, 63, 66, 69, 72)
weights <- rep(0.2, 5)

# eta = 0: arithmetic mean
calc_ede(health, weights, eta = 0)

# eta = 1: geometric mean
calc_ede(health, weights, eta = 1)

# eta = 5: high inequality aversion
calc_ede(health, weights, eta = 5)

Calculate EDE over a range of eta values

Description

Evaluates calc_ede across a vector of \eta values and returns a tidy tibble. This is the basis for EDE profile plots as described in the York DCEA handbook.

Usage

calc_ede_profile(health_dist, pop_weights, eta_range = seq(0, 10, 0.1))

Arguments

health_dist

Numeric vector of health values by group.

pop_weights

Numeric vector of population weights.

eta_range

Numeric vector of \eta values to evaluate (default: seq(0, 10, 0.1)).

Value

A tibble with columns eta and ede.

Examples

health  <- c(60, 63, 66, 69, 72)
weights <- rep(0.2, 5)
calc_ede_profile(health, weights)

Equity weights over a range of eta values

Description

Convenience function that returns equity weights for each group across a range of \eta values. Useful for understanding how the choice of inequality aversion changes the implied weights.

Usage

calc_equity_weight_profile(
  baseline_health,
  pop_weights,
  eta_range,
  group_labels = NULL
)

Arguments

baseline_health

Numeric vector of baseline health by group.

pop_weights

Numeric vector of population weights.

eta_range

Numeric vector of \eta values.

group_labels

Optional character vector of group labels.

Value

A tidy tibble with columns eta, group, group_label (if provided), and equity_weight.

Examples

baseline <- c(60, 63, 66, 69, 72)
weights  <- rep(0.2, 5)
calc_equity_weight_profile(baseline, weights, eta_range = 0:5)

Calculate equity-weighted Net Health Benefit (NHB)

Description

Applies equity weights to per-group NHB values to obtain the population-level equity-weighted NHB. This is the key summary statistic from the social welfare perspective.

Usage

calc_equity_weighted_nhb(nhb_by_group, equity_weights, pop_weights)

Arguments

nhb_by_group

Numeric vector of net health benefit per group.

equity_weights

Numeric vector of equity weights from calc_equity_weights.

pop_weights

Numeric vector of population weights.

Value

Scalar equity-weighted NHB (numeric).

Examples

baseline <- c(60, 63, 66, 69, 72)
weights  <- rep(0.2, 5)
ew       <- calc_equity_weights(baseline, weights, eta = 1)
nhb      <- c(100, 150, 200, 250, 300)
calc_equity_weighted_nhb(nhb, ew, weights)

Calculate Equity Weights

Description

Derives equity weights from the Atkinson social welfare function. Equity weights represent the relative social value of a one-unit health gain in each socioeconomic group given inequality aversion \eta.

Usage

calc_equity_weights(baseline_health, pop_weights, eta = 1, normalise = TRUE)

Arguments

baseline_health

Numeric vector of baseline health (HALE) by group (ordered from most to least deprived).

pop_weights

Numeric vector of population weights.

eta

Inequality aversion parameter (default = 1).

normalise

Logical. If TRUE (default), weights are normalised so their population-weighted mean equals 1. If FALSE, returns raw marginal welfare derivatives.

Details

For the Atkinson SWF, the equity weight for group i is proportional to h_i^{-\eta}: groups with lower baseline health receive higher weights when \eta > 0.

Value

Named numeric vector of equity weights, one per group.

References

Cookson R, Griffin S, Norheim OF, Culyer AJ (2020). Distributional Cost-Effectiveness Analysis. Oxford University Press. Oxford University Press (ISBN:9780198838197).

Robson M, Asaria M, Cookson R, Tsuchiya A, Ali S (2017). Eliciting the Level of Health Inequality Aversion in England. Health Economics 26(10): 1328-1334. doi:10.1002/hec.3386

Examples

baseline <- c(60, 63, 66, 69, 72)
weights  <- rep(0.2, 5)
calc_equity_weights(baseline, weights, eta = 1)

Calculate Gini Coefficient for Health Distribution

Description

Computes the Gini coefficient as a measure of health inequality across socioeconomic groups. The Gini ranges from 0 (perfect equality) to 1 (maximum inequality).

Usage

calc_gini(health_dist, pop_weights = NULL)

Arguments

health_dist

Numeric vector of health values (ordered from lowest to highest group).

pop_weights

Optional numeric vector of population weights. If NULL, equal weights are assumed.

Value

Gini coefficient (numeric scalar in [0, 1]).

Examples

calc_gini(c(60, 63, 66, 69, 72), pop_weights = rep(0.2, 5))

Calculate Relative Index of Inequality (RII)

Description

The RII is the SII expressed relative to the mean health level. An RII of 0.20 means the most deprived group has health 20 across the full socioeconomic range.

Usage

calc_rii(data, health_var, group_var, weight_var)

Arguments

data

A data frame with health and group columns.

health_var

Name of the health variable column (character).

group_var

Name of the socioeconomic group column (ordered integer, 1 = most deprived).

weight_var

Name of the population share column (sums to 1).

Value

A named list with elements rii, sii, se_rii, p_value, and model.

Examples

df <- tibble::tibble(
  group      = 1:5,
  mean_hale  = c(60, 63, 66, 69, 72),
  pop_share  = rep(0.2, 5)
)
calc_rii(df, "mean_hale", "group", "pop_share")

Calculate Slope Index of Inequality (SII)

Description

Fits a weighted regression of health on ridit scores to estimate the absolute health difference between the most and least deprived groups. The SII is the regression coefficient on the ridit score, interpretable as the total health gap across the full socioeconomic range.

Usage

calc_sii(data, health_var, group_var, weight_var)

Arguments

data

A data frame with health and group columns.

health_var

Name of the health variable column (character).

group_var

Name of the socioeconomic group column (ordered integer, 1 = most deprived).

weight_var

Name of the population share column (sums to 1).

Value

A named list with elements:

sii

Slope Index of Inequality (numeric)

rii

Relative Index of Inequality (numeric)

se_sii

Standard error of SII

p_value

p-value for SII

model

The underlying lm object

References

Mackenbach JP, Kunst AE (1997) Measuring the magnitude of socioeconomic inequalities in health: an overview of available measures illustrated with two examples from Europe. Social Science and Medicine 44(6): 757-771. doi:10.1016/S0277-9536(96)00073-1

Examples

df <- tibble::tibble(
  group      = 1:5,
  mean_hale  = c(60, 63, 66, 69, 72),
  pop_share  = rep(0.2, 5)
)
calc_sii(df, "mean_hale", "group", "pop_share")

Calculate Social Welfare Function value

Description

Wraps calc_ede to compute social welfare before and after an intervention and decomposes the welfare change into efficiency and equity components.

Usage

calc_social_welfare(baseline_health, post_health, pop_weights, eta = 1)

Arguments

baseline_health

Numeric vector of pre-intervention health by group.

post_health

Numeric vector of post-intervention health by group.

pop_weights

Numeric vector of population weights.

eta

Inequality aversion parameter (default = 1).

Value

A named list with elements:

ede_baseline

EDE health before intervention.

ede_post

EDE health after intervention.

delta_ede

Change in EDE (welfare gain).

efficiency_component

Change in mean health.

equity_component

Change in EDE minus change in mean.

Examples

pre  <- c(60, 63, 66, 69, 72)
post <- c(61, 64, 66.5, 69.2, 72.1)
w    <- rep(0.2, 5)
calc_social_welfare(pre, post, w, eta = 1)

Canada income quintile HALE data

Description

Health-Adjusted Life Expectancy for Canada, stratified by household income quintile. For use in Canadian DCEA analyses (CADTH workflow).

Usage

canada_income_hale

Format

A tibble with 5 rows and 9 variables analogous to england_imd_hale but with income-based stratification.

Source

Statistics Canada, Health-Adjusted Life Expectancy by income quintile.

Compute Generalised Lorenz curve data

Description

The Generalised Lorenz Curve (GLC) scales the Lorenz curve by mean health, making it sensitive to both inequality and average health level.

Usage

compute_generalised_lorenz_data(
  health_dist,
  pop_weights,
  label = "Distribution"
)

Arguments

health_dist

Numeric vector of health values by group.

pop_weights

Numeric vector of population weights.

label

Optional character label for this curve.

Value

A tibble with columns cum_pop, cum_health_generalised, and label.

Examples

compute_generalised_lorenz_data(c(60, 63, 66, 69, 72), rep(0.2, 5))

Compute Lorenz curve data

Description

Returns the coordinates of the Lorenz curve for a health distribution. Groups are ordered from lowest to highest health (most to least deprived).

Usage

compute_lorenz_data(health_dist, pop_weights, label = "Distribution")

Arguments

health_dist

Numeric vector of health values by group.

pop_weights

Numeric vector of population weights.

label

Optional character label for this curve.

Value

A tibble with columns cum_pop (cumulative population share), cum_health (cumulative health share), and label.

Examples

compute_lorenz_data(c(60, 63, 66, 69, 72), rep(0.2, 5))

Compute ridit scores (cumulative midpoint ranks) for inequality measures

Description

Ridit scores are used as the socioeconomic rank variable in concentration index calculations. For group i, the ridit is the cumulative population share up to the midpoint of group i.

Usage

compute_ridit_scores(pop_shares)

Arguments

pop_shares

Numeric vector of population proportions (ordered from most to least deprived, sums to 1).

Value

Numeric vector of ridit scores in [0, 1].

Examples

compute_ridit_scores(rep(0.2, 5))

England IMD quintile Health-Adjusted Life Expectancy data

Description

Baseline HALE (Health-Adjusted Life Expectancy) at birth for England, stratified by Index of Multiple Deprivation (IMD) quintile. Quintile 1 is the most deprived; quintile 5 is the least deprived.

Usage

england_imd_hale

Format

A tibble with 5 rows and 14 variables:

imd_quintile

Integer (1-5). 1 = most deprived.

group

Integer (1-5). Standard group identifier (same as imd_quintile).

quintile_label

Character. Human-readable quintile label.

group_label

Character. Standard group label (same as quintile_label).

mean_hale

Numeric. HALE at birth (years), both sexes (standard name).

mean_hale_all

Numeric. HALE at birth (years), both sexes.

mean_hale_male

Numeric. HALE at birth (years), males.

mean_hale_female

Numeric. HALE at birth (years), females.

se_hale

Numeric. Standard error of mean_hale (standard name).

se_hale_all

Numeric. Standard error of mean_hale_all.

pop_share

Numeric. Proportion of population in quintile (sums to 1).

cumulative_rank

Numeric. Ridit score for concentration index.

year

Integer. Reference data year.

source

Character. Data source.

Source

Office for Health Inequalities and Disparities (OHID) / Public Health England Health Profiles Plus. Proxy values based on published PHE data and interpolation from peer-reviewed literature.

References

Love-Koh J et al. (2019). Value in Health 22(5): 518-526. doi:10.1016/j.jval.2018.10.007

England IMD quintile EQ-5D utility norms

Description

Age- and IMD-stratified EQ-5D-3L utility norms for England. Useful for assigning baseline quality of life weights in full-form DCEA.

Usage

england_imd_qol

Format

A tibble with 40 rows (5 IMD quintiles x 8 age bands) and 6 variables:

imd_quintile

Integer (1-5).

age_band

Character. Age band label.

mean_eq5d_utility

Numeric. Mean EQ-5D-3L utility score.

se_eq5d

Numeric. Standard error.

mean_qale_remaining

Numeric. Quality-Adjusted Life Expectancy remaining (years).

source

Character. Data source citation.

Source

Adapted from Ara R & Brazier JE (2010) with IMD gradient adjustments from Petrou et al. (Population Health Metrics).

Example CEA model output

Description

A hypothetical cost-effectiveness analysis output for a lung cancer (NSCLC) treatment versus standard of care. Used in package examples and vignettes to demonstrate DCEA functions without requiring real data.

Usage

example_cea_output

Format

A list with two elements:

deterministic

A tibble with columns: strategy, total_qaly, total_cost, inc_qaly, inc_cost, icer, nhb_at_20k, nhb_at_30k.

psa

A data frame of 1000 PSA iterations with columns inc_qaly and inc_cost.

Source

Hypothetical data generated for illustration purposes only.

Export DCEA results to Excel (NICE submission format)

Description

Writes a multi-sheet Excel workbook with NICE-formatted DCEA output, including the summary table, per-group results, inequality indices, and social welfare profile.

Usage

export_dcea_excel(dcea_result, filepath, include_plots = FALSE)

Arguments

dcea_result

Object of class "aggregate_dcea" or "full_dcea".

filepath

Output .xlsx file path (character).

include_plots

Logical. Embed plots as images in the Excel workbook (default: FALSE).

Value

Invisibly returns filepath.

Examples


result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
export_dcea_excel(result, file.path(tempdir(), "my_dcea_results.xlsx"))

Generate a full DCEA report

Description

Renders a complete DCEA report as an HTML, Word, or PDF document using R Markdown.

Usage

generate_dcea_report(
  dcea_result,
  format = "html",
  filepath = NULL,
  template = "nice_submission"
)

Arguments

dcea_result

Object of class "aggregate_dcea" or "full_dcea".

format

Output format: "html" (default), "word", "pdf".

filepath

Output file path. If NULL, a temporary file is used and the report is opened in a browser.

template

Report template: "nice_submission" (default), "academic", "cadth".

Value

Invisibly returns the output file path.

Examples


result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
generate_dcea_report(result, format = "html")

Generate NICE-formatted DCEA submission table

Description

Creates a summary table formatted according to NICE (2025) Methods Support Document guidance for DCEA as supplementary evidence in technology appraisals.

Usage

generate_nice_table(dcea_result, format = "tibble", include_psa = FALSE)

Arguments

dcea_result

Object of class "aggregate_dcea" or "full_dcea".

format

Output format: "tibble" (default), "flextable", "html", or "xlsx".

include_psa

Logical. Include probabilistic uncertainty columns if PSA data are available (default: FALSE).

Value

A formatted table object of the requested type.

Examples

result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
generate_nice_table(result, format = "tibble")

Get baseline health distribution for a country

Description

Returns pre-loaded HALE (Health-Adjusted Life Expectancy) data stratified by equity subgroup for use in DCEA. Data are sourced from ONS/OHID (England), Statistics Canada, and the WHO Global Health Observatory.

Usage

get_baseline_health(
  country = "england",
  equity_var = "imd_quintile",
  age_group = "all",
  sex = "all",
  year = NULL
)

Arguments

country

Character. One of "england", "canada", "who_regions", "scotland", "wales".

equity_var

Character. Stratification variable. Options depend on country:

  • England: "imd_quintile" (default), "imd_decile"

  • Canada: "income_quintile"

  • WHO: "who_region"

age_group

Character. Age filter (default "all").

sex

Character. Sex filter: "all" (default), "male", "female".

year

Integer. Data year. Uses most recent available if NULL.

Value

A tibble with columns: group, group_label, mean_hale, se_hale, pop_share, cumulative_rank, source, year.

Examples

england_baseline <- get_baseline_health("england", "imd_quintile")
england_baseline

Merge CEA model output with baseline health distribution

Description

Joins a CEA result data frame (one row per equity subgroup) with a baseline health distribution returned by get_baseline_health. This is the key data-preparation step before running DCEA.

Usage

merge_cea_with_baseline(cea_output, baseline, by = "group")

Arguments

cea_output

Data frame of CEA results with at least one column matching the baseline by variable.

baseline

Tibble returned by get_baseline_health.

by

Column name to join on (default: "group").

Value

A merged tibble suitable for run_full_dcea.

Examples

baseline <- get_baseline_health("england", "imd_quintile")
cea_out  <- tibble::tibble(
  group    = 1:5,
  inc_qaly = c(0.3, 0.4, 0.5, 0.55, 0.6),
  inc_cost = rep(10000, 5)
)
merge_cea_with_baseline(cea_out, baseline, by = "group")

Normalise population weights to sum to 1

Description

Normalise population weights to sum to 1

Usage

normalise_weights(weights)

Arguments

weights

Numeric vector of weights.

Value

Normalised numeric vector.

Examples

normalise_weights(c(1, 2, 3, 4, 5))

NSCLC DCEA worked example

Description

A full DCEA worked example based on published literature for a non-small-cell lung cancer treatment. Includes subgroup-level CEA results by IMD quintile for use in full-form DCEA demonstrations.

Usage

nsclc_dcea_example

Format

A list with elements:

subgroup_cea

Tibble of per-IMD-quintile CEA results.

baseline

Baseline health distribution for NSCLC population.

staircase

Staircase data for the inequality staircase plot.

Source

Adapted from published NSCLC DCEA literature for illustration.

Plot method for aggregate_dcea

Description

Dispatches to plot_equity_impact_plane by default.

Usage

## S3 method for class 'aggregate_dcea'
plot(x, type = "impact_plane", ...)

Arguments

x

An object of class "aggregate_dcea".

type

Plot type: "impact_plane" (default), "lorenz", or "ede_profile".

...

Additional arguments passed to the underlying plot function.

Value

A ggplot2 object.

Plot method for full_dcea

Description

Plot method for full_dcea

Usage

## S3 method for class 'full_dcea'
plot(x, type = "impact_plane", ...)

Arguments

x

An object of class "full_dcea".

type

Plot type (default "impact_plane").

...

Additional arguments passed to the underlying plot function.

Value

A ggplot2 object.

Plot Tornado Diagram for DCEA Sensitivity Analysis

Description

Creates a tornado diagram showing the influence of each parameter on the net health benefit. Parameters are sorted by range (most influential at the top).

Usage

plot_dcea_tornado(sensitivity_result)

Arguments

sensitivity_result

Output from run_dcea_sensitivity.

Value

A ggplot2 tornado diagram.

Examples

result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
sa <- run_dcea_sensitivity(result)
plot_dcea_tornado(sa)

Plot EDE Profile

Description

Shows how equity-weighted NHB (or EDE health) changes as inequality aversion (\eta) increases. The profile reveals the critical \eta at which an intervention becomes welfare-improving after accounting for equity concerns.

Usage

plot_ede_profile(
  dcea_result,
  eta_range = seq(0, 15, 0.1),
  comparators = NULL,
  show_benchmark_eta = TRUE
)

Arguments

dcea_result

DCEA result object.

eta_range

Numeric vector of \eta values (default: seq(0, 15, 0.1)).

comparators

Optional list of additional DCEA result objects to overlay on the same plot.

show_benchmark_eta

Logical. Mark commonly used \eta values (0 = efficiency only; 1 = Robson et al. UK estimate; 10 = strong aversion; default: TRUE).

Value

A ggplot2 object.

Examples

result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
plot_ede_profile(result)

Plot Equity-Efficiency Impact Plane

Description

Creates the equity-efficiency impact plane as described in Cookson et al. (2017) Value in Health. The x-axis shows health inequality impact (change in a chosen inequality index), the y-axis shows net health benefit (efficiency). Four quadrants represent: Win-Win (NE), equity gain and efficiency loss (NW), equity loss and efficiency gain (SE), Lose-Lose (SW).

Usage

plot_equity_impact_plane(
  dcea_result,
  comparators = NULL,
  x_axis = "sii_change",
  y_axis = "nhb",
  show_psa_cloud = TRUE,
  show_quadrant_labels = TRUE,
  show_threshold_lines = TRUE,
  point_labels = NULL,
  colour_palette = NULL,
  theme_style = "publication"
)

Arguments

dcea_result

Object of class "aggregate_dcea" or "full_dcea".

comparators

Optional list of additional DCEA result objects to overlay on the same plane (for multi-comparator plots).

x_axis

Inequality metric for x-axis. One of "sii_change" (default), "atkinson_change", "gini_change".

y_axis

Health outcome for y-axis. One of "nhb" (default), "net_monetary_benefit".

show_psa_cloud

Logical. Show probabilistic scatter cloud if PSA data are available (default: TRUE).

show_quadrant_labels

Logical (default: TRUE).

show_threshold_lines

Logical. Show NHB = 0 and inequality = 0 reference lines (default: TRUE).

point_labels

Optional character vector of labels for points.

colour_palette

Optional named character vector of hex colours.

theme_style

Visual theme: "publication" (default) or "ggplot_default".

Value

A ggplot2 object.

References

Cookson R, Asaria M, Ali S, Shaw R, Doran T, Goldblatt P (2017). Health equity monitoring for healthcare quality assurance. Social Science & Medicine 198: 148-156.

Examples

result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
plot_equity_impact_plane(result)

Plot Inequality Staircase

Description

Visualises the causal pathway from intervention access to health inequality impact across the five staircase steps: (1) disease prevalence by group, (2) eligibility, (3) uptake/access, (4) clinical effect, (5) opportunity cost distribution.

Usage

plot_inequality_staircase(staircase_data, equity_var = "imd_quintile")

Arguments

staircase_data

Data frame with columns: step (integer 1-5), step_label (character), group (integer), group_label (character), value (numeric).

equity_var

Equity stratification variable name (for axis label).

Value

A ggplot2 faceted plot.

Examples

staircase_df <- data.frame(
  step       = rep(1:5, each = 5),
  step_label = rep(c("Prevalence", "Eligibility", "Uptake",
                     "Clinical effect", "Opportunity cost"), each = 5),
  group      = rep(1:5, times = 5),
  group_label = rep(paste("Q", 1:5), times = 5),
  value      = c(0.30, 0.28, 0.25, 0.22, 0.18,
                 0.90, 0.88, 0.85, 0.82, 0.80,
                 0.70, 0.65, 0.60, 0.55, 0.50,
                 0.45, 0.44, 0.43, 0.42, 0.40,
                 0.20, 0.18, 0.17, 0.15, 0.12)
)
plot_inequality_staircase(staircase_df)

Plot Lorenz Curve for Health Distribution

Description

Plots the Lorenz curve showing health concentration across the socioeconomic distribution. The 45-degree line represents perfect equality.

Usage

plot_lorenz_curve(dcea_result, show_pre_post = TRUE, show_generalised = FALSE)

Arguments

dcea_result

DCEA result object, or a data frame with health and weight columns.

show_pre_post

Logical. Overlay pre- and post-intervention curves (default: TRUE).

show_generalised

Logical. Overlay the Generalised Lorenz Curve (default: FALSE).

Value

A ggplot2 object.

Examples

result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
plot_lorenz_curve(result)

Prepare subgroup CEA data for DCEA

Description

Validates and standardises a data frame of subgroup-specific CEA results for use in run_full_dcea.

Usage

prepare_subgroup_cea(
  data,
  group_var,
  inc_qaly_var,
  inc_cost_var,
  pop_share_var
)

Arguments

data

Data frame with per-subgroup CEA results.

group_var

Name of the group identifier column.

inc_qaly_var

Name of the incremental QALY column.

inc_cost_var

Name of the incremental cost column.

pop_share_var

Name of the population share column.

Value

A validated and normalised tibble.

Examples

df <- tibble::tibble(
  group     = 1:5,
  inc_qaly  = c(0.3, 0.4, 0.5, 0.55, 0.6),
  inc_cost  = rep(10000, 5),
  pop_share = rep(0.2, 5)
)
prepare_subgroup_cea(df, "group", "inc_qaly", "inc_cost", "pop_share")

Print method for aggregate_dcea

Description

Print method for aggregate_dcea

Usage

## S3 method for class 'aggregate_dcea'
print(x, ...)

Arguments

x

An object of class "aggregate_dcea".

...

Further arguments (ignored).

Value

Invisibly returns x.

Print method for full_dcea

Description

Print method for full_dcea

Usage

## S3 method for class 'full_dcea'
print(x, ...)

Arguments

x

An object of class "full_dcea".

...

Further arguments (ignored).

Value

Invisibly returns x.

Run Aggregate DCEA

Description

Implements the aggregate DCEA approach of Love-Koh et al. (2019) Value in Health. Uses disease-level healthcare utilisation patterns to distribute average health benefits from a standard CEA across socioeconomic groups. This is the method supported by NICE (2025) as a supplementary analysis for technology appraisals.

Usage

run_aggregate_dcea(
  icer,
  inc_qaly,
  inc_cost,
  population_size,
  wtp = 20000,
  disease_icd = NULL,
  subgroup_distribution = NULL,
  baseline_health = NULL,
  equity_var = "imd_quintile",
  wtp_for_equity = NULL,
  opportunity_cost_threshold = 13000,
  psa_results = NULL
)

Arguments

icer

Incremental cost-effectiveness ratio (GBP per QALY).

inc_qaly

Incremental QALYs per patient (from base-case CEA).

inc_cost

Incremental cost per patient (from base-case CEA).

population_size

Total eligible population size (integer).

wtp

Willingness-to-pay threshold in GBP/QALY (default: 20000).

disease_icd

ICD-10 code or description for HES utilisation lookup. Used to distribute benefits across IMD groups if subgroup_distribution is NULL. Example: "C34" for lung cancer.

subgroup_distribution

Optional named numeric vector (length = number of equity subgroups) giving the proportion of patients in each group. Names should match group labels in the baseline dataset. Must sum to 1. If NULL, derived from disease_icd via internal lookup.

baseline_health

Optional tibble from get_baseline_health. If NULL, uses England IMD quintile data.

equity_var

Equity stratification variable (default: "imd_quintile").

wtp_for_equity

Optional second WTP threshold for equity-weighted analysis.

opportunity_cost_threshold

Cost per QALY of care displaced by the intervention's budget impact (default: 13000, i.e., NICE's k threshold).

psa_results

Optional data frame of PSA iteration results (one row per iteration, columns inc_qaly and inc_cost).

Value

An object of class "aggregate_dcea", a named list with:

summary

Key scalar DCEA outputs.

by_group

Per-group tibble: health gain, opportunity cost, NHB.

inequality_impact

Pre/post inequality indices.

social_welfare

Social welfare results over eta.

equity_plane_data

Data frame for plot_equity_impact_plane.

metadata

Inputs and assumptions.

References

Love-Koh J, Asaria M, Cookson R, Griffin S (2019). The Social Distribution of Health: Estimating Quality-Adjusted Life Expectancy in England. Value in Health 22(5): 518-526. doi:10.1016/j.jval.2018.10.007

See Also

plot_equity_impact_plane, run_full_dcea

Examples

result <- run_aggregate_dcea(
  icer            = 25000,
  inc_qaly        = 0.5,
  inc_cost        = 12500,
  population_size = 10000,
  wtp             = 20000
)
summary(result)

Run DCEA Sensitivity Analysis

Description

Performs systematic one-way and multi-way sensitivity analysis across key DCEA parameters: inequality aversion (\eta), WTP threshold, opportunity cost threshold, subgroup distribution assumptions, and equity measure choice.

Usage

run_dcea_sensitivity(
  dcea_result,
  params_to_vary = "all",
  eta_range = 0:10,
  wtp_range = NULL,
  occ_range = NULL
)

Arguments

dcea_result

Object of class "aggregate_dcea" or "full_dcea".

params_to_vary

Character vector of parameter names to vary. Options: "eta", "wtp", "occ_threshold", "subgroup_distribution", "equity_measure", "all" (default: "all").

eta_range

Numeric vector of \eta values for eta sensitivity (default: 0:10).

wtp_range

Numeric vector of WTP values to test (default: varies ±50% around base case).

occ_range

Numeric vector of opportunity cost threshold values (default: c(8000, 10000, 13000, 15000, 20000)).

Value

An object of class "dcea_sensitivity" with elements:

eta_profile

Tibble: NHB and SII change across eta range.

one_way

Tibble: results of one-way sensitivity for all parameters.

tornado_data

Data frame ready for tornado plot.

parameters

List of parameter ranges used.

Examples

result <- run_aggregate_dcea(
  icer = 25000, inc_qaly = 0.5, inc_cost = 12500,
  population_size = 10000, wtp = 20000
)
sa <- run_dcea_sensitivity(result, params_to_vary = "eta")
plot_dcea_tornado(sa)

Run Full-Form DCEA

Description

Implements full-form DCEA where subgroup-specific model parameters are available (e.g., differential uptake, differential quality-of-life gains, differential survival by socioeconomic group). Full-form DCEA is appropriate when:

Usage

run_full_dcea(
  subgroup_cea_results,
  baseline_health,
  wtp = 20000,
  opportunity_cost_threshold = 13000,
  uptake_by_group = NULL,
  adherence_by_group = NULL,
  comorbidity_adjustment = FALSE,
  psa_iterations = NULL
)

Arguments

subgroup_cea_results

Data frame with one row per equity subgroup. Required columns: group (integer), group_label (character), inc_qaly (numeric), inc_cost (numeric), pop_share (numeric, sums to 1).

baseline_health

Tibble from get_baseline_health.

wtp

Willingness-to-pay threshold in GBP/QALY (default: 20000).

opportunity_cost_threshold

Cost per QALY of displaced care (default: 13000).

uptake_by_group

Optional named numeric vector of uptake rates (0-1) per group. If NULL, assumes equal uptake across all groups.

adherence_by_group

Optional named numeric vector of adherence rates (0-1) per group. Applied as a multiplier on inc_qaly.

comorbidity_adjustment

Logical. If TRUE, applies an SES-based comorbidity adjustment to QoL gains (experimental; default: FALSE).

psa_iterations

Optional integer. Number of PSA iterations. If provided, returns probabilistic NHB distribution by group.

Value

An object of class "full_dcea" with elements analogous to run_aggregate_dcea plus subgroup-level detail.

Examples

baseline <- get_baseline_health("england", "imd_quintile")
subgroup_data <- tibble::tibble(
  group       = 1:5,
  group_label = paste("IMD Q", 1:5),
  inc_qaly    = c(0.30, 0.38, 0.45, 0.52, 0.58),
  inc_cost    = c(12000, 11500, 11000, 10500, 10000),
  pop_share   = rep(0.2, 5)
)
result <- run_full_dcea(subgroup_data, baseline)
summary(result)

Summary method for aggregate_dcea

Description

Summary method for aggregate_dcea

Usage

## S3 method for class 'aggregate_dcea'
summary(object, ...)

Arguments

object

An object of class "aggregate_dcea".

...

Further arguments (ignored).

Value

A tibble of key DCEA outputs.

Summary method for full_dcea

Description

Summary method for full_dcea

Usage

## S3 method for class 'full_dcea'
summary(object, ...)

Arguments

object

An object of class "full_dcea".

...

Further arguments (ignored).

Value

Invisibly returns object.

Validate and format DCEA input data

Description

Checks that a data frame has required columns, correct types, and that population weights sum to 1. Returns the data frame invisibly if valid, or throws an informative error.

Usage

validate_dcea_data(data, required_cols, weight_var, tol = 1e-06)

Arguments

data

A data frame to validate.

required_cols

Character vector of required column names.

weight_var

Name of the population weight column.

tol

Tolerance for checking that weights sum to 1 (default: 1e-6).

Value

The input data frame, invisibly.

Examples

df <- tibble::tibble(
  group      = 1:5,
  mean_hale  = c(60, 63, 66, 69, 72),
  pop_share  = rep(0.2, 5)
)
validate_dcea_data(df, c("group", "mean_hale", "pop_share"), "pop_share")

WHO regional HALE data

Description

Health-Adjusted Life Expectancy at birth for the six WHO regions. Useful for international equity analyses and global burden of disease perspectives.

Usage

who_regions_hale

Format

A tibble with 6 rows and 8 variables:

who_region

Character. WHO region code.

region_label

Character. Full region name.

mean_hale

Numeric. HALE at birth (years).

se_hale

Numeric. Standard error.

pop_share

Numeric. Proportion of world population.

cumulative_rank

Numeric. Ridit score.

year

Integer. Reference year.

source

Character. WHO GHO data citation.

Source

WHO Global Health Observatory. https://www.who.int/data/gho