Methods
In health economic evaluations and economic burden of disease
assessments, health impacts may need to be converted into monetary
values. For this purpose, you can use monetize().
Several valuation metrics are available, depending on how outcomes
are quantified in natural or health units (e.g. cases reduced, deaths
prevented, reductions in mortality risk, life-years gained, QALYs
gained, DALYs averted). Common metrics include the Value of a
Statistical Life (VSL) (OECD 2025), the
Value of a Life-Year (VOLY) (Hammitt 2007)
and the Value of a Quality-Adjusted Life-Year (VAQALY) (Bobinac et al. 2010).
Discounting is the practice of converting future costs (or health
impacts as previous step to valuating them) into their present value.
The underlying rationale is that the value placed on outcomes declines
as they occur further in the future. Therefore, future costs and effects
are converted into present-value terms to make them comparable over time
(Attema et al. 2018).
Discounting is implemented by selecting a discount rate, which is
used to compute a discount factor for each time period. This factor is
then multiplied by the corresponding future cost (or effect) to express
it in present-value terms.
If you need the discounted values of a cost or health outcome, you
can call the healthiar function discount(). If you
just need the discount factor, you can alternatively call
get_discount_factor(). If you just need the inflation
factor, you can call get_inflation_factor(), or the
deflation factor by entering is_deflation = TRUE in the
same function.
Different functional forms can be used to apply discounting. The most
common is exponential discounting, also referred to as constant
discounting, since outcomes are discounted proportionally as time
increases. An alternative is hyperbolic discounting, which tends to
better capture human behavior by discounting the near present more
heavily than outcomes further in the future (Lipman and Attema 2024)
See below the equations that are used behind these functions.
In health economic evaluations and economic burden of disease
assessments, health impacts may need to be converted into monetary
values. For this purpose, you can use monetize().
Several valuation metrics are available, depending on how outcomes
are quantified in natural or health units (e.g. cases reduced, deaths
prevented, reductions in mortality risk, life-years gained, QALYs
gained, DALYs averted). Common metrics include the Value of a
Statistical Life (VSL) (OECD 2025) , the
Value of a Life-Year (VOLY) (Hammitt 2007)
and the Value of a Quality-Adjusted Life-Year (VAQALY) (Bobinac et al. 2010).
Discounting is the practice of converting future costs (or health
impacts as previous step to valuating them) into their present value.
The underlying rationale is that the value placed on outcomes declines
as they occur further in the future. Therefore, future costs and effects
are converted into present-value terms to make them comparable over time
(Attema et al. 2018).
Discounting is implemented by selecting a discount rate, which is
used to compute a discount factor for each time period. This factor is
then multiplied by the corresponding future cost (or effect) to express
it in present-value terms.
If you need the discounted values of a cost or health outcome, you
can call the healthiar function discount(). If
you just need the discount factor, you can alternatively call
get_discount_factor() (entering
is_deflation = TRUE). If you just need the inflation
factor, you can get_inflation_factor().
Different functional forms can be used to apply discounting. The most
common is exponential discounting, also referred to as constant
discounting, since outcomes are discounted proportionally as time
increases. An alternative is hyperbolic discounting, which tends to
better capture human behavior by discounting the near present more
heavily than outcomes further in the future (Lipman and Attema 2024)
See below the equations that are used behind these functions.
\[monetized\_impact = impact \times
valuation \times discount\_factor \times deflator\_factor \times
real\_growth\_factor\] The arguments
discount_factor, deflator_factor and
real_growth_factor are only used if a value is entered in
the arguments discount_rate, inflation_rate
and real_growth_rate respectively
(otherwise ignored).
Discount factor
Exponential discounting
As suggested by Frederick et al.
(2002)
\[discount\_factor = \frac{1}{(1 +
discount\_rate)^{n\_years}}\]
Hyperbolic discounting Harvey
As suggested by Harvey (1986)
\[discount\_factor = \frac{1}{(1 +
n\_years)^{discount\_rate}}\]
Hyperbolic discounting Mazur
As suggested by Mazur (1987)
\[discount\_factor = \frac{1}{1 +
(discount\_rate \times n\_years)}\]
Deflation factor
Inflation can be handled in monetize() by applying a
deflator on future values, projected in nominal terms, in order to
convert them into real values (i.e., express these values in terms of
constant prices of a single base year). Therefore, if the user of the
function provides a value for the inflation_rate argument,
a deflator factor (HM Treasury 2022; Brealey et
al. 2023; Samuelson 1937) is applied according to the following
formulas
\[deflator\_factor = \frac{1}{(1 +
inflation\_rate)^{n\_years}}\]
The formula above assumes that inflation is constant over time.
Because inflation usually varies from year to year,
inflation_rate also accepts a vector of year-specific rates
(at least as many values as years to be considered, the first value
referring to the first year after the present). In that case, the
deflator factor is calculated as the product of the year-specific
factors
\[deflator\_factor =
\frac{1}{\prod_{t=1}^{n\_years}(1 + inflation\_rate_t)}\]
Real valuation growth
For this use case, you have to enter a value in the argument
real_growth_rate in monetize(), which allows
you to project this growth by applying a valuation growth factor to
base-year unit values:
\[real\_growth\_factor =(1 +
real\_growth\_rate)^{n\_years}\]
Where \(real\_growth\_rate\)
represents the annual real growth rate in health valuation. As in
inflation_rate, a vector of year-specific rates can be
entered instead of one single value, in which case the factor is the
product of the year-specific factors. This ensures that long-term
environmental impacts are not undervalued.
The argument output_attribute accepts the output of
attribute_health(), attribute_lifetable(),
compare() or daly(). With the output of
attribute_lifetable(), the impacts are monetized year by
year and the number of years is taken from the life table. With the
output of daly(), the YLL and the YLD are monetized
separately and then summed, i.e. the YLL from the life table are
discounted year by year, while n_years is used for the
YLD.
Example 3: multiple exposure-outcome pairs
In the example below, we monetize the COPD cases attributable to
PM2.5 and the asthma cases attributable to NO2 from the chapter
Multiple exposure-outcome pairs in one single call, with a
different value and discounting per pair.
The arguments describing the value of the health impact
(valuation, discount_rate and
discount_shape) are entered as vectors parallel to the
arguments of attribute_health(), exactly as
exp_central or bhd_central. They must have one
single value per geographic unit and dimension entered in
main_results_by, here per pair.
real_growth_rate can be entered in the same way
(e.g. one rate per country, since it reflects income growth), constant
over the years. A vector refers to year-specific rates only if
output_attribute contains one single assessment.
In contrast, inflation_rate is the same for all pairs,
so that all valuations share one price base: a vector entered there
refers to year-specific rates.
Function call
results_pairs <- attribute_health(
info = data.frame(
pair = rep(c("pm2.5_copd", "no2_asthma"), each = 2)),
main_results_by = "pair",
exp_central = c(8.85, 9.20, 22.1, 24.5),
cutoff_central = c(5, 5, 10, 10),
rr_central = c(1.369, 1.369, 1.041, 1.041),
rr_increment = 10,
erf_shape = c("log_linear", "log_linear", "linear", "linear"),
bhd_central = c(30747, 31500, 12000, 12500),
geo_id_micro = rep(c("a", "b"), times = 2)
)
monetized_pairs <- monetize(
output_attribute = results_pairs,
valuation = c(50000, 50000, 3000, 3000), # E.g. EURO per case
discount_rate = c(0.03, 0.03, 0.01, 0.01),
discount_shape = rep(c("exponential", "hyperbolic_mazur_1987"), each = 2),
n_years = 5
)
Main results
One row per exposure-outcome pair and geographic unit, each monetized
with its own value and discounting.
monetized_pairs$monetization_main |>
dplyr::select(info_pair, geo_id_micro, valuation, discount_rate,
discount_shape, monetized_impact_rounded) |>
knitr::kable()
| pm2.5_copd |
a |
50000 |
0.03 |
exponential |
151041153 |
| pm2.5_copd |
b |
50000 |
0.03 |
exponential |
167901617 |
| no2_asthma |
a |
3000 |
0.01 |
hyperbolic_mazur_1987 |
1620520 |
| no2_asthma |
b |
3000 |
0.01 |
hyperbolic_mazur_1987 |
2004072 |
Social aspects
Health impact attributable to social indicator
Goal
To quantify the differences in attributable health impacts between more and less deprived areas.
Methods
Taking into account socio-economic indicators, e.g. a multiple deprivation index (Mogin et al. 2025), the differences in attributable health impacts across the study areas can be estimated (Renard et al. 2019; Otavova et al. 2022).
Social inequalities are quantified as the difference between the least deprived areas (the last n-quantile) and
the most deprived areas or
the population overall.
These differences can be
absolute or
relative.
Geographic units without a value in the social indicator (
NA), e.g. because the deprivation index is not available for small areas, cannot be assigned to a quantile. They are left out of the comparison between the most and least deprived quantiles. However, they still count in the overall values, because the overall values refer to the whole study area, i.e. the health impacts that were estimated. If you prefer to exclude them from the overall values too, remove them before calling the attribute function.Difference most deprived vs. least deprived
\[ absolute\_quantile = first - last \] Where:
\[ relative\_quantile = \frac{absolute\_quantile}{last} \]
Difference overall vs. least deprived
\[ absolute\_overall = overall - last \] Where:
If you assume that the least deprived areas are similar to counterfactual cases (no exposure to deprivation), the relative difference regarding the overall average health impact could be interpreted as some kind of relative risk attributable to social inequalities.
Example
In the example below, we compare the mortality attributable to PM2.5 exposure across deciles of a deprivation score.
Function call
First, quantify health impacts.
Second, use the function
socialize()entering the whole output ofattribute_health()in the argumentoutput_attribute.Alternatively, you can directly enter the health impact in the
socialize()argumentimpact.Main results
Multiple deprivation index
Goal
To create a multiple deprivation index (MDI) from several socio-economic indicators, e.g. to enter it in the argument
social_indicatorofsocialize().Methods
Socio-economic indicators (e.g., education level, employment status and family structure) can be condensed into a multiple deprivation index (MDI) (Mogin et al. 2025). For this purpose, the indicators can be normalized using min-max scaling.
The reliability of the MDI can be assessed using Cronbach’s alpha (Cronbach 1951).
\[ \alpha = \frac{k}{k - 1} \left( 1 - \frac{\sum_{i=1}^{k} \sigma^2_{y_i}}{\sigma^2_x} \right) \] where:
To apply this approach, you should ensure that the data set is as complete as possible. Otherwise, you can try to impute missing data using:
Time-Based Imputation: Linear regression based on historical trends if prior years’ data is complete.
Indicator-Based Imputation: Multiple linear regression if the missing indicator correlates strongly with others.
To apply this approach, you should ensure that the data set is as complete as possible. Otherwise, you can try to impute missing data using: - Time-Based Imputation: Linear regression based on historical trends if prior years’ data is complete. - Indicator-Based Imputation: Multiple linear regression if the missing indicator correlates strongly with others.
Imputation models should have an R^2 greater than or equal to 0.7. If R^2 lower than 0.7, consider alternative data sources or methods.
Example
In the example below, we create an MDI from five indicators (education, unemployment, single parents, population change and lack of heating) and split it into deciles.
Function call
verbose = FALSEsuppresses any output to the console (default:verbose = TRUE, i.e. with printing turned on).Main results
Function output includes:
mdi_main, a tibble containing the BEST-COST MDIThe function assesses the reliability of the MDI based on the Cronbach’s alpha value as follows:
0.9 and higher: Excellent reliability
between 0.8 (included) and 0.9: Good reliability
between 0.7 (included) and 0.8: Acceptable reliability
between 0.6 (included) and 0.7: Questionable reliability
lower than 0.6: Poor reliability
Detailed results
mdi_detailedDESCRIPTIVE STATISTICS
PEARSON’S CORRELATION COEFFICIENTS
CRONBACH’S α, including the reliability rating this value indicates
Code for boxplots of the single indicators
Code for histogram of the MDI’s for the geo units with a normal distribution curve
To reproduce the boxplots run