Upshifting alcohol consumption

Upshift the average weekly units of alcohol consumed to adjust for the under-reporting of alcohol consumption in survey data.

alc_upshift(
  data,
  country = c("England", "Scotland")[1],
  year_select,
  pcc_data = c("HMRC", "MESAS")[1],
  proportion = 0.8
)

Arguments

data

Data.table - "Population" - the individual level data on alcohol consumption to be upshifted. The variable to be upshifted should be named "weekmean" and contain the average weekly alcohol consumption of an individual in UK standard units of ethanol.

country

Character string - either "England" or "Scotland".

year_select

Integer - Year for which upshifting will be done

pcc_data

Data.table - "PCC" - the values of per capita alcohol consumption calculated from HMRC data on duty receipts disaggregated by UK nation or MESAS monitoring report on alcohol sales 2022 (Scotland only) Stored as package data in hseclean::per_capita_alc_for_upshift.

proportion

Numeric - the proportion of this 'true' value to shift consumption data up to (default is 80 percent (Stockwell et al. 2018) ).

Value

Returns the input data with a column weekmean_adj added containing the upshifted values of weekly mean alcohol consumption.

Details

A function of the form f(PCC, Proportion, Population) where PCC=the 'true' Per Capita Consumption being aimed for in units of per capita litres of pure ethanol per year, Proportion=the proportion of this 'true' value to shift consumption data up to (default is 80 percent, taken from Stockwell et al. (2018) which is based on World Health Organisation assumptions) and Population=the population whose consumption is to be upshifted.

The fixed values used to generate the standard deviations of the gamma distributions are taken from Kehoe et al. (2012) .

Note that the result could be an upshift or a downshift depending on the reference pcc value and the distribution of consumption in the survey data.

References

Kehoe T, Gmel G, Shield KD, Gmel G, Rehm J (2012). “Determining the best population-level alcohol consumption model and its impact on estimates of alcohol-attributable harms.” Population Health Metrics, 10(1), 1--19. doi:10.1186/1478-7954-10-6 . Stockwell T, Zhao J, Sherk A, Rehm J, Shield K, Naimi T (2018). “Underestimation of alcohol consumption in cohort studies and implications for alcohol's contribution to the global burden of disease.” Addiction, 113(12), 2245--2249. doi:10.1111/add.14392 .

Examples

if (FALSE) {

# Scottish Health Survey example
# 2018

library(hseclean)
library(data.table)
library(magrittr)

# Location of Scottish data
root_dir <- "X:/HAR_PR/PR/Consumption_TA/HSE/Scottish Health Survey (SHeS)/"

data <- read_SHeS_2018(root = root_dir) %>%
  clean_age %>% clean_demographic %>%
  alc_drink_now_allages %>%
  alc_weekmean_adult %>%
  select_data(ages = 16:89, years = 2018,
    keep_vars = c("wt_int", "year", "age", "sex", "weekmean"),
    complete_vars = c("wt_int", "sex", "weekmean"))

data <- alc_upshift(data, country = "Scotland",
  pcc_data = hseclean::per_capita_alc_for_upshift,
  proportion = 0.8)

}