Alcohol Consumption

Alcohol consumption is a significant risk factor for various non-communicable diseases (NCDs), injuries, and other health problems. This document explains how alcohol consumption and related interventions are modeled, focusing on the mechanisms connecting interventions to changes in health outcomes.

Used in

Alcohol consumption is incorporated as a risk factor in the following disease models:

• Diabetes
• CVD

Mechanism

Alcohol consumption influences health outcomes in several ways:

• Incidence: Alcohol increases the incidence (new cases) of several NCDs, including liver cirrhosis, certain cancers (e.g., liver, breast, colorectal), and cardiovascular diseases.
• Mortality: Alcohol contributes to mortality (death rates) directly (e.g., alcohol poisoning, liver failure) and indirectly through increased risk of NCDs and injuries.
• Disability: The model considers impact to Years Lived with Disability (YLD)

The model represents these relationships using:

1. Prevalence of Alcohol Consumption: The model tracks the amount of alcohol consumed by different age and sex groups.

2. Relative Risks (RRs) and Population Attributable Fractions (PAFs): Similar to tobacco and physical inactivity, the model uses RRs and PAFs derived from epidemiological studies.

   • Relative Risk (RR): Quantifies the increased risk of a specific disease (or death) for individuals who consume alcohol compared to those who don't, or who consume less. The RR often varies with the level of consumption.
   • Population Attributable Fraction (PAF): Represents the proportion of disease incidence (or mortality) in the entire population attributable to alcohol consumption.

How the Model Calculates Changes (Intervention Impact):

The interventions primarily aim to reduce the prevalence and level of alcohol consumption. These reductions, in turn, lead to changes in PAFs, incidence, and mortality.

1. Baseline Consumption: The model starts with a baseline level of alcohol consumption for each age and sex group.

2. Intervention Impact: Each intervention is assigned an impact, which reflects its effect on alcohol consumption. This impact can be expressed in different ways:

   • Relative Reduction: Interventions (e.g., taxes, advertising bans) reduce prevalence of harmful consumption by a certain percentage, which is scaled by the difference in coverage from baseline.
   • Drive Laws. For NC_RFAlcoholDriveLaws, the impact is specifically linked to mortality and Years Lived with Disability(YLD). This captures the effect of reducing alcohol-related injuries (and deaths) from road traffic accidents.
   • Additional YLD and Deaths: Effects that alcohol have that aren't modelled through diabetes or CVD are added to the total healthy years lived in the CVD model through adjusting the total years lived with disability attributable to non-modelled effects by the change in prevalence.

3. Coverage: The model incorporates the coverage of each intervention - the percentage of the target population that is actually reached.

4. Calculating Overall Consumption Change: The model combines the impacts and coverage of the various interventions to determine the overall change in alcohol consumption patterns in the population.

5. Calculating the PAF Adjustment: After calculating the reduction in alcohol consumption, the model recalculates the Population Attributable Fractions (PAFs) using the new, lower consumption levels. The PAF represents how much of the disease burden can be attributed to alcohol use. When consumption decreases due to interventions, the PAF also decreases proportionally, which means less disease can be attributed to alcohol exposure in the population.

6. New Incidence, Mortality, and Disability: The adjusted PAFs are then used (in other parts of the model) to calculate new incidence, mortality, and disability rates for alcohol-related conditions.

Additional Non-Modelled Alcohol Effects:

Beyond the direct disease-specific effects modelled through diabetes and CVD, alcohol interventions also affect health outcomes through conditions not explicitly modelled in the system. The model accounts for these additional effects by:

1. Baseline Health Impact: The model calculates the number of healthy years lived per population and determines what proportion of Years Lived with Disability (YLDs) can be attributed to alcohol consumption based on country-specific attributable fractions.

2. Prevalence-Based Adjustment: As alcohol interventions reduce consumption prevalence, the model applies this reduction proportionally to the non-modelled health effects. The formula used is: Additional_Health_Benefit = Baseline_YLDs × Alcohol_Attributable_Fraction × Prevalence_Change_Proportion, where the prevalence change proportion is calculated as (baseline_prevalence - current_prevalence) / baseline_prevalence.

3. Comprehensive Impact: This ensures that the model captures the full spectrum of alcohol's health effects, not just those flowing through the explicitly modelled disease pathways of diabetes and cardiovascular disease.

Specific Intervention: Drunk Driving Laws (NC_RFAlcoholDriveLaws)

This intervention has a unique mechanism. It doesn't directly reduce overall alcohol consumption. Instead, it reduces the harm associated with a given level of consumption by reducing the risk of alcohol-related traffic injuries and deaths.

The model calculates this impact as follows:

1. Baseline Risk: The model has a baseline risk of alcohol-related mortality and YLD (years lived with disability) specifically associated with road traffic accidents and injuries.

2. Intervention Impact: NC_RFAlcoholDriveLaws is assigned a percentage reduction in road traffic injury risk. This reduction represents the effectiveness of drunk driving laws and enforcement in preventing alcohol-related traffic accidents. The impact is calculated as the percentage reduction in both mortality (NC_RiskMortality) and YLD (NC_RiskYLD) attributable to road traffic injuries.

3. Coverage: The coverage of the intervention reflects the extent to which the laws are enforced and the population is affected.

4. PAF Correction to Mortality and YLD:

   • PAF_Mortality_Correction = 1 - (impact_mortality * coverage_change)
   • PAF_YLD_Correction = 1 - (impact_YLD * coverage_change)
   • The coverage_change is the difference between the coverage at the current time step and the coverage at the baseline year.

5. Adjusted Mortality and YLD: The baseline alcohol-related mortality and YLD are then multiplied by these correction factors. This results in lower mortality and YLD, even if overall alcohol consumption remains the same.

In Summary

The model simulates the impact of alcohol control interventions by:

• Reducing the prevalence and/or level of alcohol consumption (for most interventions).
• Specifically reducing alcohol-related mortality and disability from road traffic accidents (for drunk driving laws).
• These changes lead to reductions in the Population Attributable Fractions (PAFs) for alcohol-related conditions.
• The reduced PAFs, in turn, lead to lower incidence, mortality, and disability rates.

Interventions

Intervention Table

Null Scenario

A1

Intervention: Increase excise taxes on alcoholic beverages

A2

Intervention: Enact and enforce bans or comprehensive restrictions on exposure to alcohol advertising (across multiple media)

A3

Intervention: Enact and enforce restrictions on the physical availability of retailed alcohol (via reduced hours of sale)

A4

Intervention: Enact and enforce drink-driving laws and blood alcohol concentration limits via sobriety checkpoints

A5

Intervention: Provide brief psychosocial intervention for persons with hazardous and harmful alcohol use

Assumptions

Relative Risk

Impact on Prevalence

*Note: Drink driving laws primarily impact mortality and disability rather than prevalence. *Note: Alcohol impacts also vary by region, South Asia shown here.