Targeting alcohol use in high-risk population groups: a US microsimulation study of beverage-specific pricing policies

Abstract

Background: Raising retail prices on alcoholic beverages preferred by high-risk groups (males, those of low socioeconomic status, and those with heavy alcohol use) might selectively reduce their alcohol consumption. However, the differential impact of beverage-specific price increases on US population groups has yet to be studied. This study aimed to simulate the effect of beverage-specific price increases on alcohol use within subgroups of the adult US population defined by sex, educational attainment, and alcohol use category.

Methods: An individual-level microsimulation of the US population (aged 18-79 years) was used to simulate alcohol consumption from 2000 to 2019 based on individual characteristics (ie, sex, age, race, ethnicity, and educational attainment as a proxy for socioeconomic status categorised as high school degree or less, some college, and college degree or more) and previous alcohol use. The microsimulation model was generated via integration of diverse data sources including decennial US Census data, annual data from the American Community Survey, annual data from the National Vital Statistics System, annual data from the Behavioral Risk Factor Surveillance System, and biennial, longitudinal data from the Panel Study of Income Dynamics. Policy parameters were informed by the existing literature. Four national policy scenarios were compared with a reference scenario without price change in 2019: a uniform price increase of 10% (scenario 1), a uniform price increase of 30% (scenario 2), a beverage-specific price increase of 30% for beer and spirits and 10% for wine (scenario 3), and a beverage-specific price increase of 50% for beer and spirits and 10% for wine (scenario 4). Individual-level effects on alcohol consumption were simulated using beverage-specific own-price elasticities. Sensitivity analysis assessed assumption-based correlation coefficient between alcohol consumption and the individual-level percent reduction in alcohol consumed; and the application of the beverage-non-specific own-price participation elasticity.

Findings: Scenario 4 had the strongest effect on alcohol use overall and most effectively reduced consumption in high-risk groups: males and females with high alcohol use (more than 60 g of pure alcohol per day for males and 40 g of pure alcohol per day for females) and low educational attainment (high school degree or less) reduced their alcohol use by -17·30% (-17·62 g per day, credible interval [CI] -21·77 to -13·20) and -17·49% (-12·25 g per day, CI -14·72 to -9·58), respectively. In comparison, smaller relative changes were observed among groups at less risk of harm.

Interpretation: Disproportionate increases in retail prices for the cheapest beverages, beer and spirits, might lead to a greater decline in consumption among high-risk groups. Pricing policies could thus be used as a powerful public health tool to mitigate the unequal alcohol-attributable burden of disease.

Funding: National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health.

Figure 1:. Simulated average alcohol consumption under different pricing policy scenarios within population subgroups (including abstinent individuals)

Reference: no price change. Scenario 1: 10% increase for beer, wine, and spirits. Scenario 2: 30% increase for beer, wine, and spirits. Scenario 3: 30% increase for beer and spirits and 10% increase for wine. Scenario 4: 50% increase for beer and spirits and 10% increase for wine. Price changes are being introduced in 2019. Error bars represent credible intervals based on the minimum and maximum values across 60 unique parameter combinations.

Figure 2:. Simulated prevalence of alcohol abstinence under different pricing policy scenarios within population subgroups

Reference: no price change. Scenario 1: 10% increase for beer, wine, and spirits. Scenario 2: 30% increase for beer, wine, and spirits. Scenario 3: 30% increase for beer and spirits and 10% increase for wine. Scenario 4: 50% increase for beer and spirits and 10% increase for wine. Reference scenario without changes is shown in grey. Alcohol beverage prices are increased in 2019.

Figure 3:. Change in simulated average alcohol consumption in grams per day by alcohol use category in reference to 2018 under different pricing policy scenarios withing population subgroups

Reference: no price change. Scenario 1: 10% increase for beer, wine, and spirits. Scenario 2: 30% increase for beer, wine, and spirits. Scenario 3: 30% increase for beer and spirits and 10% increase for wine. Scenario 4: 50% increase for beer and spirits and 10% increase for wine. Alcohol use category 1: up to 20 g per day for females and up to 40 g per day for males. Alcohol use category 2: more than 20 g per day up to 40 g per day for females and more than 40 g per day up to 60 g per day for males. Alcohol use category 3: more than 40 g per day for females and more than 60 g per day for males. Alcohol beverage prices are increased in 2019. Error bars represent credible intervals based on the minimum and maximum values across 60 unique parameter combinations.

Figure 4:. Distribution of the differences in change in simulated average alcohol consumption in grams per day by alcohol use category between reference (no price change) and scenario 4: 50% increase for beer and spirits and 10% increase for wine within population subgroups

Alcohol use category 2: more than 20 g per day up to 40 g per day for females and more than 40 g per day up to 60 g per day for males. Alcohol use category 3: more than 40 g per day for females and more than 60 g per day for males. Negative values indicate a larger reduction for the high school degree or less or the some college group following price changes compared with the college degree or more group. The whiskers of the box plot indicate the 2·5th and 97·5th percentile across 60 unique parameter combinations.