Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 1.
Published in final edited form as: Health Place. 2025 May 23;94:103489. doi: 10.1016/j.healthplace.2025.103489

Investigating the Impacts of Alcohol Outlet Zoning Policy on Alcohol Consumption and Access to Non-Alcoholic Services: A Spatial Agent-Based Simulation

Tingting Ji 1,2,*, Ivana Stankov 3,4, Niles Sherman Egan 5, Kristen Hassmiller Lich 6, Rachel L J Thornton 7, Qi Wang 1, Takeru Igusa 1, Hsi-Hsien Wei 2, Pamela A Matson 8
PMCID: PMC12178808  NIHMSID: NIHMS2084866  PMID: 40411920

Abstract

Debates continue over the effectiveness of limiting alcohol outlet density in reducing alcohol consumption, and its broader impacts on access to non-alcoholic services in low-income urban communities remain underexplored. This study addresses this gap by investigating the impacts of alcohol outlet zoning policies on alcohol consumption and walkable access to non-alcoholic services in low-income urban communities with different baseline densities of liquor and grocery stores. We developed a spatial agent-based model of Baltimore City neighborhoods, simulating the closure of non-conforming liquor stores following the city’s zoning code rewrite. The model was calibrated using national survey data and empirical research on alcohol consumption and walkable access to alcohol, food, lottery, and ATM outlets by subgroups. We observed non-linear relationships and differences by gender and employment status in the effects of liquor store closures on heavy drinking, with policies showing limited effectiveness in neighborhoods with high baseline liquor store density. While the policies had minimal impact on access to food and ATMs due to high prevalence of grocery stores, they reduced access to lottery services. Our modeling approach serves as a valuable decision-making tool for policymakers to explore hypothetical scenarios, identify tipping points of policy impacts, and provide actionable insights into the complex interactions between zoning policies and neighborhood dynamics concerning alcohol consumption and access to essential goods and services.

Keywords: agent-based models, GIS, liquor stores, outlet density, public health, zoning policy

1. INTRODUCTION

Alcohol outlet oversaturation is a significant public health issue in the US, with an average density of about 10.5 per 100,000 people (Census Bureau, 2022) and even higher densities in urban area, where census tracts average 7.4 outlets per 1,000 residents (Berke et al., 2010). High alcohol outlet density is linked to excessive alcohol consumption and related health and social harms, including injuries, traffic crashes, suicides, homicides, liver disease, and cancer (Hippensteel et al., 2019; Milam et al., 2020; Shield et al., 2024; Yang et al., 2024). Excessive alcohol use caused over 140,000 deaths annually in the US from 2015 to 2019, making it the third-leading preventable cause of death (CDC, 2022b, 2022a).

Given the adverse impacts of dense alcohol environments, many studies advocate for reducing alcohol outlet density to improve public health. A notable legislative intervention in the US is alcohol outlet zoning policy, which uses zoning codes to regulate the density, location, and distribution of outlets (Furr-Holden et al., 2020). A typical example is Baltimore City’s 2017 Transform Baltimore zoning code rewrite, which required 76 non-conforming off-premise liquor stores - about 3.3% of total outlets - to close within two years (Stacy et al., 2020).

While alcohol outlet zoning policies show potential benefits, their effectiveness in reducing alcohol consumption remains debated (Mosher & Treffers, 2013). Few longitudinal studies have assessed the impact of these policies. Gmel et al.’s review of 160 studies found limited evidence that outlet density impacted individual consumption, while natural experiments demonstrated little to no effect of density restrictions (Gmel et al., 2016). Campbell et al. (2009) theorized that the impacts of density changes depend on baseline density, with decreases in alcohol outlets within high-density areas conferring minimal effects. However, few studies have explored how alcohol consumption responds to zoning policies across varying baseline liquor store densities.

The broader impact of zoning polices on access to non-alcoholic services (e.g., food, lotteries, and ATMs) also remains underexplored. Alcohol outlets are disproportionately concentrated in low-income, predominantly Black communities, as seen in cities like Baltimore, Chicago, and San Francisco(Gyimah-Brempong, 2005; Laveist & Wallace, 2000; Mair et al., 2020; Romley et al., 2007). This pattern stems from historical urban decline, during which supermarkets relocated to the suburbs, leaving some liquor and convenience stores as essential providers of household goods in underserved areas (Bodor et al., 2010; ChangeLabSolutions, 2019). For residents without cars, these outlets became crucial for accessing daily necessities. Consequently, zoning policies could have unintended consequences for these communities. Furr-Holden et al. (2020) found that non-conforming outlets in Baltimore sold more healthy foods than conforming ones, indicating that closures could reduce access to essential goods. Despite these concerns, there is a lack of quantitative research on how zoning policies influence non-alcoholic service access in these communities.

Our paper aimed to understand the impacts of alcohol outlet zoning policies on residents’ alcohol consumption and walkable access to non-alcoholic services. Using spatial agent-based modeling, we address two questions: 1) How do zoning policies affect alcohol consumption and access to services like food, lotteries, and ATMs? 2) How do these impacts vary across neighborhoods with different baseline liquor and grocery store densities? We hope our study can inform policies seeking to limit alcohol outlet density in disadvantaged communities to improve public health and reduce environmental inequities.

2. METHODS

Agent-based modeling (ABM) is a computational simulation method that creates virtual urban systems where decision-making agents interact with the environment and each other over time (Cerdá & Keyes, 2019; Stankov et al., 2019). In our ABM, agents representing residents conduct routine activities within a GIS-based neighborhood environment, including road networks, residential areas, liquor stores, and grocery stores. To present the key components of our model design, we used the PARTE framework (Ross A. Hammond, 2015): Properties, Actions, Rules, Time, and Environment.

2.1. Study Area and Model Environment

Baltimore City, with a 2020 population of 585,708 (61.6% African American), has history of racial segregation, particularly in east and west Baltimore (Furr-Holden et al., 2020). Fifty years ago, the city aimed to reduce alcohol outlet density from 2.56 to 1 per 1,000 residents, but by 2016, it remained at 2.01 per 1,000 (Hippensteel et al., 2019). In 2017, the Transform Baltimore zoning code mandated non-conforming off-premise outlets in residentially areas to cease alcohol sales by 2019, affecting 76 liquor stores across 48 neighborhoods (Figure 1).

Figure 1.

Figure 1.

Open in a new tab

Distribution of liquor and grocery stores in Baltimore City. The blue colors indicate the percentage of non-conforming liquor stores in each neighborhood. The size of the pie charts represents the total number of liquor and grocery stores per neighborhood, with orange indicating liquor stores and purple indicating grocery stores. The digital maps of Baltimore neighborhoods were downloaded from Maryland’s GIS Data Catalog. Liquor store locations were obtained from the Board of Liquor License Commissioners for Baltimore City. Grocery store locations were downloaded from the Maryland Food System Map. The distribution of other non-alcoholic services is not shown in this figure because no records are available.

To assess the policy’s neighborhood-level impact, we selected a predominantly Black neighborhood in Baltimore (95% African American) as a case study for our ABM simulation. To mitigate boundary effects, we included three surrounding blocks, encompassing 27 grocery stores and 21 liquor stores, 12 of which were non-conforming (57%). Since data on non-alcoholic services was unavailable, we used Google Street View and Photos to identify services offered by liquor and grocery stores (Figure 2). GIS shapefiles of residential and non-residential areas, road networks, and store locations were imported into the model, which represented the environment as a two-dimensional 90×90 square grid, with each square covering 80×80 square feet.

Figure 2.

Figure 2.

Open in a new tab

Photos of two liquor stores within the case study area. (a) D&M Liquors offers alcohol, lottery, and ATM, but no food. (b) B&J Liquors offers alcohol and grocery (food), but no ATM or lottery. Source: Google Street View.

2.2. Properties and Actions of Agents

We used the demographic data from the case study neighborhood to define the race, age, gender, and employment status of resident agents (Table 1). Agents were categorized as alcohol drinkers, alcohol buyers, food buyers, lottery players, and ATM users based on prevalence rates drawn from empirical data. While the legal alcohol purchasing age in the U.S. is 21, younger age groups also engage in grocery shopping and other purchases (Green et al., 2021; Larson et al., 2006; Setiono et al., 2021). To capture these dynamics, our model simulates the purchasing behaviors of 3,997 resident agents aged 12 and above, randomly assigned to residential areas within the neighborhood.

Table 1.

Demographic information of the case study neighborhood.

Demographic Category Number and percent among the total population (%)
Race Black 4,681 (94.9%)
White 117 (2.4%)
Hispanic, Asian, and others 133 (2.7%)
Age 0–4 456 (9.25%)
5–11 478 (9.69%)
12–14 205 (4.16%)
15–17 245 (4.97%)
18–24 547 (11.09%)
25–34 571 (11.58%)
35–44 588 (11.92%)
45–64 1272 (25.80%)
65 and over 569 (11.54%)
Gender Male 2217 (44.96%)
Female 2714 (55.04%)
Employment Employed 50.94% among the persons who aged 16–64
Unemployed - not in the labor force 36.33% among the persons aged 16–64
Unemployed - seeking jobs 12.73% among the persons aged 16–64
Open in a new tab

Source: Open Baltimore website(Open Baltimore, 2024).

Each simulation day, agents were probabilistically assigned behavioral parameters that triggered alcohol or food consumption and purchase activities. These parameters were based on distributions of residents’ demographic variables (e.g., age, gender, employment status) and environmental factors (e.g., proximity to stores). Table 2 shows the model parameters and calibration data for charactering agents’ demographics, alcohol consumption behaviors, food consumption behaviors, lottery play and ATM use behaviors, and store-visiting behaviors. Taking alcohol drinkers as an example, we calibrated prevalence and behavior parameters, such as drinking frequency and intensity, using the 2019 Behavioral Risk Factor Surveillance System (BRFSS) survey data (CDC, 2019), stratified by age, gender, and drinking habits (Table 3).

Table 2.

Model parameters for charactering agents’ properties and actions and calibration data.

Parameter Definition and value Data and empirical foundation
Demographics
Age Describe the age of a resident, which is 12 and over. Open Baltimore website (Open Baltimore, 2024). See Table 1. Values remain constant.
Age-group Young people: age ∈ [10, 24] Adults: age ∈ [25, 64) Older adults: age >= 65 WHO (WHO, 2024) and CDC (CDC, 2024). Values remain constant.
Gender Describe the gender of a resident. 0 = male, 1 = female. Open Baltimore website (Open Baltimore, 2024). See Table 1. Values remain constant.
Employment-status Describe the employment status of a resident. 1 = employed, 2 = non-employed. Open Baltimore website (Open Baltimore, 2024). Employed residents are those aged between 16 and 64 who are currently working. Non-employed residents include students, unemployed individuals, and retired residents. The unemployed include both individuals not in the labor force and those actively seeking employment (See Table 1).Values remain constant.
Home-address Determine the home location of a resident. Randomly assigned to a location in the residential area of the case study neighborhood. Values remain constant.
Alcohol consumption behaviors
Drinker-type Describe whether a resident is an alcohol drinker. 0 = non-drinker, 1 = non-alcohol drinker. The 2019 BRFSS survey data (CDC, 2019). For the prevalence of drinkers, see Table 3. Values remain constant.
Alcohol-type The types of alcoholic beverages consumed by drinkers. 1 = beer, 2 = wine. We assume that half the agents prefer wine, and the other half prefer beer. We consider alcohol type in our model since the weight of one standard drink in the US differs by different alcohol types, i.e., one standard drink is equal to 12 ounces of beer or 5 ounces of wine. Values remain constant.
Alcohol-inventory Describe the number of drinks remaining at home. Initially, we randomly assign 0 to 5 number of drinks for each drinker. Values are updated over time.
Prob-drinking The probability of drinking at home today, which is equal to drinking frequency divided by 30 days. The 2019 BRFSS survey data (CDC, 2019). For probability distribution of drinking frequency, see Table 3. Values are updated over time.
Expected-drink-intensity The expected number of drinks to be consumed today. The 2019 BRFSS survey data (CDC, 2019). For probability distribution of drinking intensity, see Table 3. Values are updated over time.
Actual-drink-intensity The actual number of drinks to be consumed today. If the drinker has no alcohol at home, then actual-drink-intensity = 0 They are influenced by drinkers’ actual drinking behaviors and alcohol purchase behaviors. Values are updated over time.
Food consumption behaviors
Food-buyer? Describe whether a resident primarily purchases food at corner stores. True or False. D’Angelo et al. (2011) conducted a survey of residents who lived in low-income African-American dominant communities and found that 18.3% of respondents’ primary food source was corner stores. Values remain constant.
Food-inventory Describe the level of food inventory at home, ranging from 0 to 1. Randomly assigned initially and changes every day. At time step t, food-inventory (t) = food-inventory (t-1) - food-consumption-rate (t-1). Values are updated over time.
Food-consumption-rate Describe the speed of food consumed every day by food buyers, ranging from 0 to 1. According to the study by D’Angelo et al. (2011), among corner-store shoppers, 75% of them shop 5–7 times per week. Here we assume the rest 25% shop 1–4 times per week. Food-consumption-rate is equal to food purchase frequency per week divided by 7 days. For example, if a resident purchases food three days per week, the food-consumption-rate = 3 / 7. Values remain constant.
Lottery play and ATM use behaviors
Lottery-player? Describe whether a resident is a lottery player. True or False. According to the 2017 Maryland Problem Gambling Prevalence Survey (J. Kathleen Tracy et al., 2017), around 68.8% of residents in Baltimore City play the lottery. Values remain constant.
Prob-lottery-play The probability of playing lottery at corner stores today, ranging from 0 to 1. According to the 2017 Maryland Problem Gambling Prevalence Survey(J. Kathleen Tracy et al., 2017), among lottery gamblers, 44.2% play 1–5 days per year (infrequent lottery gamblers), 14.7% play 6–12 days per year (yearly lottery gamblers), 9.8% play 3–5 days per month (monthly lottery gamblers), and 10.2% play more than 6 days per month (weekly lottery gamblers). Values are updated over time.
ATM-user? Describe whether a resident is an ATM user. True or False. Around 86% of US consumers use ATMs (PaymentsJournal, 2020b). Here we assume 92% of residents in our case study area use ATMs. This is because the residents living in poorer and minority communities rely more on cash economies due to their difficulty in obtaining credit cards from banks. Values remain constant.
Prob-ATM-use The probability of using an ATM service today, ranging from 0 to 1. Average US consumers use ATMs 3.8 times per month to get cash, 2.5 times to deposit cash, and 2.2 times to deposit checks (PaymentsJournal, 2020a). In sum, there are 8.5 times of ATM usage per month for US consumers. We assume the residents in our case study are using ATM services 12 times per month.
Store-visiting behaviors
Acceptable-walking-distance The acceptable walking distance from the agent’s home address to the store location. The acceptable walking distances for shopping purposes were based on the probability distributions conditional to different age and gender groups (See Table 4), which were calibrated using American Travel Survey data(Bureau of Transportation Statistics, 2021). Values are updated over time.
Max-purchase-day-per-week The maximum number of days that allow for purchase at stores per week. Employed ∈ [1, 7]; Non-employed ∈ [4, 7]. This parameter is used to limit the purchasing frequency of residents per week. This parameter differs depending on the employment status of the residents, as employed residents have fewer days per week available for shopping than non-employed residents. Values remain constant.
Comfortably-carried-load The comfortable load of groceries or alcoholic beverages that the agent can carry. This parameter is used to limit the amount of alcohol or food that residents can purchase in one trip. This parameter differs depending on the agent’s body weight, age, gender, and walking distance from store to home (Phonpichit et al., 2016). Values are updated over time.
Open in a new tab

Table 3.

Drinking frequency and intensity of resident agents.

Subgroup %Drinker Drinking habit Drinking frequency (days) Drinking intensity (drinks)
Probability distribution Median value Probability distribution Median value
Young male 49.4% Occasional (96.4%) Exp (0.136) 5 Exp(0.294)a 3
Every day (3.6%) 30 30 3
Young female 56.0% Occasional (98.7%) Exp (0.163) 4 Exp(0.397)a 2
Every day (1.3%) 30 30 2
Adult male 60.8% Occasional (90.9%) Exp (0.130) 5 Exp(0.422) 2
Every day (9.1%) 30 30 Exp(0.300) 3
Adult female 53.7% Occasional (94.7%) Exp(0.160) 4 Exp(0.554) 1
Every day (5.3%) 30 30 Exp(0.372) 2
Older male 56.4% Occasional (79.9%) Exp (0.119) 5 Exp(0.602) 1
Every day (20.1%) 30 30 Exp(0.491) 2
Older female 40.4% Occasional (86.2%) Exp (0.146) 4 Exp(0.752) 1
Every day (13.8%) 30 30 Exp(0.654) 1
Open in a new tab

Notes. Occasional drinker means the drinker do not drink every day per month; Everyday drinkers means the drinker drinks every day per month.

a

The sample size of everyday drinkers in the young male group and young female group is too small (less than 10) to be fitted using the probability distribution function; therefore, we combined the samples of occasional and everyday drinkers in these two subgroups to directly fit the probability distribution function of drinking frequency and intensity. Drinking frequency refers to the number of days to consume at least one drink of any alcoholic beverage per month, and drinking intensity refers to the average number of drinks consumed on drinking days. Notably, the age of young drinkers should be larger than 21 years old according to Baltimore City’s legal drinking age (Phonpichit et al., 2016). Exp (mean) refers to the exponential distribution with the mean parameter.

2.3. Model Time and Rules

Each simulation ran for 360 days, during which resident agents performed routine activities, including alcohol and food consumption at home and purchasing or using services in stores. A conceptual framework was developed to illustrate the effects of liquor store closures, which key mechanisms and feedback loops shown in the causal loop diagram (Figure A.1 in Appendix A).

Each day, drinkers decide whether to drink and how much, based on drinking frequency and expected intensity, and check their alcohol inventory to ensure sufficient supply. Food buyers’ inventories decrease according to food consumption rate. If alcohol or food is insufficient, agents may go out to replenish their supplies. Lottery buyers and ATM users determine whether to access these services at stores based on assigned probabilities. Purchase and store-visiting behaviors were simulated after consumption. An agent visits a store if the store is within their acceptable walking distance (based on age and gender) and their weekly shopping trips are below the employment-based limit. Otherwise, the agent defers shopping to the following week. The agents’ acceptable walking distance by subgroups were calibrated using 2021 American Travel Survey data (Bureau of Transportation Statistics, 2021) as indicated in Table 4.

Table 4.

Acceptable walking distance by subgroups.

Subgroup Acceptable walking distance (mile)
Probability distribution Median value
Young male Exp (2.118) 0.431
Young female Exp (2.382) 0.410
Adult male Exp (2.400) 0.412
Adult female Exp (2.762) 0.386
Older male Exp (2.595) 0.348
Older female Exp (3.361) 0.278
Open in a new tab

Notes. Exp (mean) refers to the exponential distribution with the mean parameter.

When an agent decides to visit stores, it select the one with the highest purchase utility. For example, if an agent needs both food and alcohol, it prefers a store offering both, even if a store with only alcohol is closer. If multiple stores offer the same items, the agent chooses the closest one. After each visit, the agent recalculates purchase utility for remaining stores until all needs are met, using shortest path algorithm to calculate distances. In stores, agents ensure their purchases do not exceed their comfortably carried load.

2.4. Model Validation

The status quo of alcohol consumption in the case study neighborhood was modeled for validation. Alcohol consumption was measured by the percentage of drinkers who are heavy drinkers, a group at high risk for health issues such as injury, violence, liver disease, and cancer. Following the 2019 BRFSS definition, heavy drinking includes more than 14 drinks per week for men and more than 7 for women. The percentages of heavy drinkers by gender were compared to the Maryland samples from the 2019 BRFSS data to ensure the model reflected real-world patterns.

As shown in Table 5, the simulated status quo and actual real-world percentages of heavy drinkers by gender were almost identical, with less than a 1% difference. The simulated patterns also matched the real patterns, i.e., the percentage of female heavy drinkers was about 2% higher than that of male heavy drinkers. It is plausible that the simulated heavy drinkers are slightly higher than the Maryland samples because Baltimore City has a higher prevalence of heavy drinkers than the Maryland state average.

Table 5.

Comparison between simulated status-quo scenario and actual real-world data.

Alcohol consumption indicators Simulated status-quo scenario Actual real-world data
% male drinkers are heavy drinkers 9.4% 8.8%
% female drinkers are heavy drinkers 11.9% 11.0%
Open in a new tab

Notes. Real-world data is based on Maryland samples of the 2019 BRFSS survey data.

2.5. Scenario Analyses

To examine the impact of the zoning policy, we compared the zoning policy scenario to the status quo scenario in the case study neighborhood, and the potential effects of the policy across areas with varying mixed densities of liquor and grocery stores were tested under four baseline density scenarios (Table 6). The number of liquor stores closed was calculated as a percentage of the total number of stores, with closure rates ranging from 10% to 90% in 20% increments. The area of the neighborhood was held constant at 0.35 sq. mi., reflecting its true geographic size and ensuring consistency across scenarios. Model outcomes included: 1) the proportion of residents with walkable access to alcohol (within a 5-minute walk), 2) the proportion of drinkers who are heavy drinkers, and 3) the proportion of residents with walkable access to non-alcoholic services.

Table 6.

Parameters of different density scenarios.

Scenario (S) Liquor Store density (#/sq. mi.) Grocery Store Density (#/sq. mi.)
Case study neighborhood 60 78
S2: Low liquor + low grocery 10 17
S2: High liquor + Low grocery 36 17
S3: Low liquor + High grocery 10 48
S4: High liquor + High grocery 36 48
Open in a new tab

Notes. Store density is defined as the number of stores per square mile. “Low” density refers to values below the first quartile threshold, and “high” density refers to values above the third quartile threshold, based on data from 48 Baltimore neighborhoods affected by the zoning code rewrite. These precise thresholds were utilized for simulation rather than exploring all values within the quartiles.

To ensure that the simulation results have statistical consistency and are not random results due to parameter uncertainties, we performed a consistency analysis(Alden et al., 2013) to determine the minimum number of runs for each parameter value, namely 100 runs for each parameter. The agent-based model was developed and implemented in NetLogo (version 6.3.0), and the analysis was performed in R (version 4.2.3).

3. RESULTS

3.1. Policy Impacts on the Case Study Neighborhood

We conducted t-tests to compare the policy scenario and the status quo (Table 7). The zoning policy significantly reduces the percentage of residents with walkable access to alcohol by over 20% (p = 0.000). However, this reduction does not lead to a meaningful decrease in alcohol consumption, as the percentage of heavy drinkers declines by less than 1%. Disaggregating the results by gender, age, and employment status, we find a significant effect among female drinkers and employed drinkers, whose prevalence of heavy drinking decreases by 0.5% (p = 0.003) and 0.4% (p = 0.009), respectively. For non-alcoholic services, the policy has no effect on food or ATM access due to the high density of grocery stores and conforming liquor stores. However, access to lottery services is significantly impacted, with walkable access dropping by about 30% (p = 0.000).

Table 7.

The impacts of alcohol outlet zoning policy on residents’ alcohol access, alcohol consumption, and access to food, lotteries, and ATMs in the case study neighborhood.

Outcome Mean (after policy) Mean (before policy) Mean difference with 95% CI P value
Lower Upper
Alcohol access 75.4% 99.5% 24.3% 24.0% 0.000
% Heavy drinkers 10.4% 10.7% −0.5% −0.1% 0.003
 Male 9.3% 9.4% −0.4% 0.0% 0.197
 Female 11.4% 11.9% −0.7% −0.1% 0.003
 Young 12.5% 13.2% −1.4% 0.0% 0.073
 Middle 10.1% 10.3% −0.4% 0.0% 0.062
 Old 10.5% 10.9% −0.8% 0.0% 0.064
 Employed 9.8% 10.2% −0.7% −0.1% 0.009
 Non-employed1 10.8% 11.0% −0.4% 0.4% 0.214
Food access 100.0% 100.0% 0.0% 0.0% na
Lottery access 68.3% 97.6% −29.4% −29.2% 0.000
ATM access 100.0% 100.0% 0.0% 0.0% na
Open in a new tab

Note:

1

Non-employed residents include students, unemployed individuals, and retired residents.

We explored the zoning policy’s impact by varying the proportion of liquor stores closed (Figure 3). Residents’ walkable access to alcohol declines noticeay after 30% of liquor stores close, but a reduction in heavy drinking is only observed after more than 70% of closures. Access to food services remains unaffected, even with 90% of liquor stores closed.

Figure 3.

Figure 3.

Open in a new tab

Impact of alcohol outlet zoning policy on (a) % residents within walkable alcohol access, (b) % drinkers who are heavy drinkers, and (c) % residents within walkable food access in the case study neighborhood. “Actual” refers to the actual proportion of liquor stores closed in this neighborhood.

3.2. Policy Impacts on Baltimore Neighborhoods

To assess the zoning policy’s impact on other Baltimore neighborhoods with non-conforming liquor stores, we analysed four mixed-density scenarios (Figure 4). The relationship between liquor store closures and residents’ walkable access to alcohol is nearly linear across all scenarios, but the magnitude of the reduction varies with baseline liquor store density. The policy’s effect on heavy drinking is not obvious until closures exceed 70%. Notably, reductions in heavy drinking are slightly more visible in low liquor density scenarios (S1 and S3) compared to high-density ones (S2 and S4), though significant variation remains. For food access, the impact is minimal: even in the most affected scenario (S2), less than 20% of residents lose walkable access to food after 90% of liquor stores close.

Figure 4.

Figure 4.

Open in a new tab

Impact of alcohol outlet zoning policy on (a) % residents within walkable alcohol access, (b) % drinkers who are heavy drinkers, and (c) % residents within walkable food access in neighborhoods with mixed densities of liquor and grocery stores (refer to Table 6 for parameter settings).

4. DISCUSSION

4.1. Policy Impacts on Alcohol Consumption

We found that closing non-conforming liquor stores through zoning has limited effectiveness in reducing heavy alcohol consumption, particularly in neighborhoods with high baseline liquor store density. This aligns with some US-based research, including a longitudinal study showing that changes in alcohol outlet density affected the frequency and quantity of alcohol consumption but not problem drinking, such as binge drinking (Auchincloss et al., 2022) and “highest daily alcohol use” (p18) (Brenner et al., 2015). In contrast, a Swedish longitudinal study (2012–2018) found that decreases in the distance from home or work to the nearest liquor store were associated with a higher risk of problematic alcohol use (Raza et al., 2023).

4.2. Non-linear Relationships and Tipping Points

Our simulation reveals a non-linear relationship between liquor store closures and heavy drinking, with a noticeable flattening in the curve as baseline liquor store density increases. Tipping points emerge when closures exceed 70% in high-density neighborhoods, leading to significant reductions in heavy drinking, while closures below 50% have minimal impact. This highlights the importance of baseline outlet density, and the extent of closures needed for meaningful effects. Our findings align with Foster et al., who noted that new alcohol licenses have greater impacts in areas without existing licenses, suggesting that reducing outlet density in such areas may be more effective (Foster et al., 2020). However, our results diverge from Ahern et al. (2013), who argued that reducing high density yields the largest public health benefit due to higher binge drinking prevalence in those areas. Ahern et al. (2013) may have overlooked that in high-density neighborhoods, even after reducing one outlet, alternative stores remain readily accessible, limiting the policy’s impact.

4.3. Heterogeneous Effects of Zoning Policy on Alcohol Consumption

While the overall reduction in heavy drinking is limited, our study highlights the heterogeneous effects of the zoning policy, with potential gender and employment status differences in the association between liquor store density and problem drinking. In our case study neighborhood, closing non-conforming liquor stores significantly reduces the proportion of female heavy drinkers but not male ones. This aligns with previous studies, such as Halonen et al. (2013), who found that reducing the distance to the nearest off-premise outlet increased the odds of heavy drinking in women but not men, and Raza et al. (2023), who observed a higher risk of problem drinking in women with reductions in distances between home or workplace and alcohol outlets.

A possible explanation is that reduced walkable alcohol access makes it more difficult for women to carry heavy alcoholic beverages as women are generally less able to carry heavy loads compared to men, thereby disincentivizing alcohol consumption. Other mechanisms may also contribute. In our model, we calibrated resident agents’ acceptable walking distances using survey data, which revealed that women reported shorter acceptable walking distances than men across all age groups. This suggests that women may be more sensitive to spatial barriers in alcohol access. It is also important to consider the influence of gender-specific thresholds in defining heavy drinking (i.e., more than 14 drinks per week for men and more than seven for women). Consequently, even modest reductions in consumption may be more likely to shift women below the heavy drinking threshold. This classification difference may partly explain the observed gender disparity, suggesting that women’s apparent responsiveness to zoning policies may reflect both behavioral sensitivity and measurement effects.

We also found that closing non-conforming liquor stores significantly reduces the proportion of employed drinkers, but not non-employed drinkers. Although empirical studies examining the association between alcohol outlet density and alcohol consumption by employment status are limited, partly because many surveys focus primarily on employed populations (e.g., Raza et al. (2023)), there are potential explanations for our findings. First, employed residents may face greater time and mobility constraints, reducing their opportunities to purchase alcohol, particularly when travel distances increase due to store closures. In our model, employed individuals are assumed to have fewer days per week available for store visits compared to their non-employed counterparts. In contrast, non-employed residents typically have more time and flexibility to conduct alcohol purchasing activities. Second, although we did not stratify drinking behavior by employment status, prior research has suggested that unemployed individuals are more likely to engage in problem drinking (Backhans et al., 2016; Henkel, 2011; Zins et al., 2011). Taking this into account, it is plausible that zoning policies may have a greater impact on employed residents, and a more limited effect on those who are non-employed, as the latter may be more willing or have more time to travel longer distances to purchase alcohol.

4.4. Policy Impact on Access to Food

While local community members expressed concerns that closing non-conforming liquor stores could limit food access (Stacy et al., 2020), our analysis shows minimal impact. Even with over 90% of liquor stores closed in neighborhoods with high liquor and low grocery store density, fewer than 20% of residents lose food access within a 5-minute walk. This is due to the high prevalence of grocery stores in these urban communities, providing alternative food options. However, our model does not account for the types of food (e.g., healthy or unhealthy) sold by at these stores. Furr-Holden et al. (2020) noted that some non-conforming liquor stores noted that some non-conforming liquor stores in resource-deprived neighborhoods sell healthy foods. To ensure access to healthier options, urban planners could adopt strategies like those in Los Angeles County, where licensed off-premise outlets must provide at least three varieties of fresh food items (Lacounty.gov, 2021).

4.5. Policy Impact on Access to Social Services

Our simulation shows that the zoning policy affects residents’ access to lottery services in the case study neighborhood. In low-income Baltimore communities, many liquor stores also serve as lottery retailers, acting as vital social hubs in neighborhoods lacking community centers or gathering spaces (Matson et al., 2022). This unique role makes liquor stores important for sociability, particularly in Black neighborhoods (Kwate, 2021). Losing these spaces may push residents into areas where they “aren’t supposed to be”, such as outside neighborhoods or vacant buildings, increasing the risk of police encounters, conflicts, and even violent crime (Matson et al., 2022). To mitigate these impacts, converting non-conforming liquor stores into social spaces or entrepreneurial venues could enhance community social capital, though attracting investment to under-resourced areas will require supportive local policies.

4.6. Public Health Implications

Our study advances understanding of alcohol outlet zoning policies in low-income urban communities. Contrary to some US studies, our findings suggest limited effectiveness of these policies, especially in neighborhoods with high baseline liquor store density, underscoring the need for a nuanced approach to zoning regulations for alcohol control. We identify nonlinear relationships between reducing liquor stores and alcohol consumption, emphasizing the importance of baseline density and closure proportions. This is crucial for policymakers tailoring zoning regulations to different communities’ unique characteristics. Additionally, we examine the impact of zoning on access to non-alcoholic services, offering valuable insights into broader implications for economically disadvantaged communities.

In addition to these substantive findings, our study contributes methodologically by introducing a spatial agent-based simulation approach to quantify zoning policy effects, overcoming challenges of long-term cohort data collection. ABM serves as a complementary tool to address empirical data gaps in policy impact evaluation. While the Transform Baltimore policy has already been implemented, comprehensive and neighborhood-scale post-policy data on individual-level alcohol use and service access remain limited. This is in part because the policy sought to address alcohol misuse and violence within communities, without a corresponding focus on understanding and tracking the impact of the policy on access to other services offered by non-conforming alcohol outlets. As such, the collection of these data has not been prioritized. More broadly, collecting such data through longitudinal surveys is time- and resource-intensive (Caruana et al., 2015). Unlike observational studies, ABM enables transparent and cost-effective exploration of counterfactuals and “what-if” scenarios (e.g., varying closure rates and baseline store densities) that would be difficult or impossible to examine in the real world. Direct comparison of these simulated scenarios provides an opportunity to include broader knowledge with those based solely on empirical data collection (Silverman et al., 2021). For example, our model revealed tipping points where liquor store closures substantially reduce heavy drinking and impact other essential goods and services, insights that may not be apparent through observational data alone.

Finally, we demonstrate that ABM is not only a predictive tool but also a means for systems thinking in public health (Stankov et al., 2019). It can be adapted to other chronic disease challenges, such as tobacco retailer density reduction policies (Luke et al., 2017). The ABM framework developed is also adaptable to other settings. Although our case study focuses on predominantly African American, low-income urban neighborhoods, the spatial structure, service configurations, and agent behaviors can be modified to reflect different geographic, socioeconomic, or demographic contexts. For example, our study, informed by community workshops (Matson et al., 2022), incorporated behaviors such as food and lottery purchasing, and ATM use, reflecting the specific service roles that liquor and grocery stores play in underserved communities. In other neighborhoods, such as rural or suburban areas, different local data and community consultations can be integrated to reflect unique patterns of service access and alcohol use. Thus, while the findings are most directly relevant to structurally similar communities, the model serves as a flexible platform for exploring zoning policy impacts across a range of built environments and population compositions.

4.7. Limitations

Our study has several limitations that highlight avenues for future research. First, parameterizing an ABM for specific community contexts is challenging due to limited availability of neighborhood-level data. In this study, we relied on state-level alcohol consumption patterns from Maryland, which may overlook nuanced relationships between local drinking behaviors and zoning policy effectiveness. Second, our focus on individual drinking behavior driven by reduced alcohol accessibility does not fully capture the influence of social networks. Literature suggests that changes in alcohol consumption among friends or relatives have potentials to amplify or counteract the effects of reduced alcohol availability (Knox et al., 2019; Rosenquist et al., 2010). Moreover, zoning policies may influence individual behaviors not only through social networks, but also via neighborhood norms or local culture. These potential pathways are highlighted in a recent narrative review by Booth et al. (2024), which suggested that increased exposure to alcohol venues may increase the risk of alcohol-related harm among young people through mechanisms such as alcohol advertising and observing others drinking beyond the effect of outlet density alone. Prior ABM studies on substance use have demonstrated the feasibility of incorporating such socio-cultural complexities when data are available. For example, Giabbanelli & Crutzen (2013) used agent-based social network models to predict binge drinking in the adult Dutch population by incorporating survey data on drinking behaviors and peer influence. Sukthankar & Beheshti (2019) introduced two ABM frameworks designed to model the influence of social norms on smoking cessation. Future research integrating neighborhood-specific data, social network dynamics, and representations of neighborhood norms into ABM frameworks are needed to more comprehensively capture the social complexity and contextual embeddedness of zoning policy impacts.

5. CONCLUSIONS

This study utilized agent-based modeling to examine the effects of zoning policies aimed at limiting liquor store density on alcohol consumption and access to essential non-alcoholic services. Our findings highlight the non-linear relationship between liquor store closures and heavy drinking, emphasizing the importance of considering baseline outlet densities and their impact on tipping points for effective policy interventions. Unlike prior studies that primarily focused on alcohol consumption, our approach extends the analysis to examine policy impacts on specific demographic groups, such as women, the elderly, and the non-employed, as well as on access to food and social services. This research underscores the need for tailored policies that account for the complexities of local community contexts, providing actionable evidence for policymakers to design interventions that balance public health benefits with equitable access to essential services.

Supplementary Material

1

  1. A spatial agent-based model is developed to analyze zoning policies in low-income urban communities of color.

  2. Zoning policies exhibit non-linear, context-specific impacts on alcohol consumption and access to food and social services.

  3. Baseline store density is crucial for identifying tipping points and designing effective interventions.

FUNDING

Funding for this work was provided by the National Institute on Drug Abuse [K01DA035387].

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  1. Ahern J, Margerison-Zilko C, Hubbard A, & Galea S (2013). Alcohol outlets and binge drinking in urban neighborhoods: The implications of nonlinearity for intervention and policy. American Journal of Public Health, 103(4). 10.2105/AJPH.2012.301203 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alden K, Read M, Timmis J, Andrews PS, Veiga-Fernandes H, & Coles M (2013). Spartan: A Comprehensive Tool for Understanding Uncertainty in Simulations of Biological Systems. PLoS Computational Biology, 9(2), e1002916. 10.1371/journal.pcbi.1002916 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Auchincloss AH, Niamatullah S, Adams M, Melly SJ, Li J, & Lazo M (2022). Alcohol outlets and alcohol consumption in changing environments: prevalence and changes over time. Substance Abuse: Treatment, Prevention, and Policy, 17(1). 10.1186/s13011-021-00430-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Backhans MC, Balliu N, Lundin A, & Hemmingsson T (2016). Unemployment Is a Risk Factor for Hospitalization Due to Alcohol Problems: A Longitudinal Study Based on the Stockholm Public Health Cohort (SPHC). Journal of Studies on Alcohol and Drugs, 77(6), 936–942. 10.15288/jsad.2016.77.936 [DOI] [PubMed] [Google Scholar]
  5. Berke EM, Tanski SE, Demidenko E, Alford-Teaster J, Shi X, & Sargent JD (2010). Alcohol Retail Density and Demographic Predictors of Health Disparities: A Geographic Analysis. American Journal of Public Health, 100(10). 10.2105/AJPH [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bodor JN, Ulmer VM, Futrell Dunaway L, Farley TA, & Rose D (2010). The Rationale behind Small Food Store Interventions in Low-Income Urban Neighborhoods: Insights from New Orleans. The Journal of Nutrition, 140(6), 1185–1188. 10.3945/jn.109.113266 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Booth L, Miller M, & Pettigrew S (2024). The potential adverse effects of minors’ exposure to alcohol-related stimuli via licenced venues: A narrative review. Drug and Alcohol Review, 43(1), 141–155. 10.1111/dar.13769 [DOI] [PubMed] [Google Scholar]
  8. Brenner AB, Borrell LN, Barrientos-Gutierrez T, & Diez Roux AV (2015). Longitudinal associations of neighborhood socioeconomic characteristics and alcohol availability on drinking: Results from the Multi-Ethnic Study of Atherosclerosis (MESA). Social Science and Medicine, 145, 17–25. 10.1016/j.socscimed.2015.09.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bureau of Transportation Statistics. (2021). American Travel Survey. https://www.bts.gov/browse-statistical-products-and-data/surveys/american-travel-survey
  10. Campbell CA, Hahn RA, Elder R, Brewer R, Chattopadhyay S, Fielding J, Naimi TS, Toomey T, Lawrence B, & Middleton JC (2009). The Effectiveness of Limiting Alcohol Outlet Density As a Means of Reducing Excessive Alcohol Consumption and Alcohol-Related Harms. American Journal of Preventive Medicine, 37(6), 556–569. 10.1016/j.amepre.2009.09.028 [DOI] [PubMed] [Google Scholar]
  11. Caruana EJ, Roman M, Hernández-Sánchez J, & Solli P (2015). Longitudinal studies. Journal of Thoracic Disease, 7(11), E537–E540. 10.3978/j.issn.2072-1439.2015.10.63 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. CDC. (2019). 2019 BRFSS Survey Data and Documentation. https://www.cdc.gov/brfss/annual_data/annual_2019.html
  13. CDC. (2022a). Alcohol Use in the United States. https://www.niaaa.nih.gov/publications/brochures-and-fact-sheets/alcohol-facts-and-statistics
  14. CDC. (2022b). Deaths from Excessive Alcohol Use in the United States. https://www.cdc.gov/alcohol/features/excessive-alcohol-deaths.html
  15. CDC. (2024). Older Adults. https://www.cdc.gov/cdi/indicator-definitions/older-adults.html?CDC_AAref_Val=https://www.cdc.gov/cdi/definitions/older-adults.html
  16. Census Bureau. (2022). County Business Patterns by Industry: 2020. https://www.census.gov/library/visualizations/interactive/county-business-patterns-by-industry-2020.html
  17. Cerdá M, & Keyes KM (2019). Systems Modeling to Advance the Promise of Data Science in Epidemiology. American Journal of Epidemiology, 188(5), 862–865. 10.1093/aje/kwy262 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. ChangeLabSolutions. (2019). Outlet Density & Community Characteristics. http://alcohol-psr.changelabsolutions.org/alcohol-psr-faqs/local-authority-to-regulate-the-density-of-alcohol-outlets-faq/outlet-density-community-characteristics/
  19. D’Angelo H, Suratkar S, Song H-J, Stauffer E, & Gittelsohn J (2011). Access to food source and food source use are associated with healthy and unhealthy food-purchasing behaviours among low-income African-American adults in Baltimore City. Public Health Nutrition, 14(9), 1632–1639. 10.1017/S1368980011000498 [DOI] [PubMed] [Google Scholar]
  20. Foster S, Hooper P, Divitini M, Knuiman M, & Trapp G (2020). Over the limit? Testing non-linear associations between alcohol outlets and young adults’ alcohol consumption. Drug and Alcohol Review, 39(6), 664–670. 10.1111/dar.13115 [DOI] [PubMed] [Google Scholar]
  21. Furr-Holden CDM, Milam AJ, Nesoff ED, Linton S, Reboussin B, Sadler RC, & Leaf PJ (2020). Using Zoning as a Public Health Tool to Reduce Alcohol Outlet Oversaturation, Promote Compliance, and Guide Future Enforcement: a Preliminary Analysis of Transform Baltimore. Journal of Urban Health, 97(4), 568–582. 10.1007/s11524-020-00453-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Giabbanelli P, & Crutzen R (2013). An Agent-Based Social Network Model of Binge Drinking Among Dutch Adults. Journal of Artificial Societies and Social Simulation, 16(2). 10.18564/jasss.2159 [DOI] [Google Scholar]
  23. Gmel G, Holmes J, & Studer J (2016). Are alcohol outlet densities strongly associated with alcohol-related outcomes? A critical review of recent evidence. Drug and Alcohol Review, 35(1), 40–54. 10.1111/dar.12304 [DOI] [PubMed] [Google Scholar]
  24. Green EM, Gaines A, Hill TF, & Dollahite JS (2021). Personal, proxy, and collective food agency among early adolescents. Appetite, 166, 105435. 10.1016/j.appet.2021.105435 [DOI] [PubMed] [Google Scholar]
  25. Gyimah-Brempong K (2005). Neighborhood income, alcohol availability, and crime rates. The Review of Black Political Economy, 33(3), 21–44. [Google Scholar]
  26. Halonen JI, Kivimäki M, Virtanen M, Pentti J, Subramanian SV, Kawachi I, & Vahtera J (2013). Proximity of off-premise alcohol outlets and heavy alcohol consumption: A cohort study. Drug and Alcohol Dependence, 132(1–2), 295–300. 10.1016/j.drugalcdep.2013.02.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Henkel D (2011). Unemployment and Substance Use: A Review of the Literature (1990–2010). Current Drug Abuse Reviewse, 4(1), 4–27. 10.2174/1874473711104010004 [DOI] [PubMed] [Google Scholar]
  28. Hippensteel CL, Sadler RC, Milam AJ, Nelson V, & Debra Furr-Holden C (2019). Using Zoning as a Public Health Tool to Reduce Oversaturation of Alcohol Outlets: an Examination of the Effects of the New “300 Foot Rule” on Packaged Goods Stores in a Mid-Atlantic City. Prevention Science, 20(6), 833–843. 10.1007/s11121-018-0947-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kathleen Tracy J, Louise Maranda, & Christina Scheele. (2017). Statewide Gambling Prevalence in Maryland: 2017. https://dlslibrary.state.md.us/publications/Exec/MDH/HG19-804(b)_2017.pdf
  30. Knox J, Schneider J, Greene E, Nicholson J, Hasin D, & Sandfort T (2019). Using social network analysis to examine alcohol use among adults: A systematic review. PLoS ONE, 14(8). 10.1371/journal.pone.0221360 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kwate NOA (2021). The Corner Liquor Store: Rethinking Toxicity in the Black Metropolis. Journal of Medical Humanities, 42(3), 307–323. 10.1007/s10912-020-09645-3 [DOI] [PubMed] [Google Scholar]
  32. Lacounty.gov. (2021). Safe Access to Alcohol and Food Establishments. https://planning.lacounty.gov/wp-content/uploads/2022/11/saafe_summary-brochure_English.pdf
  33. Larson NI, Story M, Eisenberg ME, & Neumark-Sztainer D (2006). Food Preparation and Purchasing Roles among Adolescents: Associations with Sociodemographic Characteristics and Diet Quality. Journal of the American Dietetic Association, 106(2), 211–218. 10.1016/j.jada.2005.10.029 [DOI] [PubMed] [Google Scholar]
  34. Laveist TA, & Wallace JM (2000). Health risk and inequitable distribution of liquor stores in African American neighborhood. Social Science & Medicine, 51(4), 613–617. www.elsevier.com/locate/socscimed [DOI] [PubMed] [Google Scholar]
  35. Luke DA, Hammond RA, Combs T, Sorg A, Kasman M, Mack-Crane A, Ribisl KM, & Henriksen L (2017). Tobacco Town: Computational Modeling of Policy Options to Reduce Tobacco Retailer Density. American Journal of Public Health, 107(5), 740–746. 10.2105/AJPH.2017.303685 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mair C, Sumetsky N, Gruenewald PJ, & Lee JP (2020). Microecological Relationships Between Area Income, Off-Premise Alcohol Outlet Density, Drinking Patterns, and Alcohol Use Disorders: The East Bay Neighborhoods Study. Alcoholism: Clinical and Experimental Research, 44(8), 1636–1645. 10.1111/acer.14387 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Matson PA, Stankov I, Hassmiller Lich K, Flessa S, Lowy J, & Thornton RLJ (2022). A systems framework depicting how complex neighborhood dynamics and contextual factors could impact the effectiveness of an alcohol outlet zoning policy. American Journal of Community Psychology, 70(1–2), 18–32. 10.1002/ajcp.12566 [DOI] [PubMed] [Google Scholar]
  38. Milam AJ, Barajas CB, Buchalski Z, Wang L, Sadler RC, & Furr-Holden CDM (2020). Discrepancies in Local, State, and National Alcohol Outlet Listings: Implications for Research and Interventions. Substance Use and Misuse, 55(14), 2348–2356. 10.1080/10826084.2020.1817080 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Mosher JF, & Treffers RD (2013). State Pre-Emption, Local Control, and Alcohol Retail Outlet Density Regulation. American Journal of Preventive Medicine, 44(4), 399–405. 10.1016/j.amepre.2012.11.029 [DOI] [PubMed] [Google Scholar]
  40. Open Baltimore. (2024). Neighborhood Statistical Area (NSA) Boundaries. https://data.baltimorecity.gov/datasets/8112521d3e284518b9fa497a188bfb45_0/explore
  41. PaymentsJournal. (2020a). The Demographics of ATM Use. https://www.paymentsjournal.com/the-demographics-of-atm-use/
  42. PaymentsJournal. (2020b). What Percentage of US Consumers Don’t Use ATMs… at All? https://www.paymentsjournal.com/what-percentage-of-us-consumers-dont-use-at-ms-at-all/
  43. Phonpichit C, Chansirinukor W, & Akamanon C (2016). The response of the body when carrying a handbag. Work, 53(3). [DOI] [PubMed] [Google Scholar]
  44. Raza A, Hanson LLM, Westerlund H, Mäkelä P, & Halonen JI (2023). Access to Alcohol Outlets From Home and Work in Sweden: Longitudinal Associations With Problem Drinking. Journal of Studies on Alcohol and Drugs, 84(1), 37–44. 10.15288/jsad.21-00365 [DOI] [PubMed] [Google Scholar]
  45. Romley JA, Cohen D, Ringel J, & Sturm R (2007). Alcohol and Environmental Justice: The Density of Liquor Stores and Bars in Urban Neighborhoods in the United States. Journal of Studies on Alcohol and Drugs, 68(1), 48–55. 10.15288/jsad.2007.68.48 [DOI] [PubMed] [Google Scholar]
  46. Rosenquist JN, Murabito J, Fowler JH, & Christakis NA (2010). The Spread of Alcohol Consumption Behavior in a Large Social Network. https://annals.org [DOI] [PMC free article] [PubMed]
  47. Hammond Ross A.. (2015). Considerations and best practices in agent-based modeling to inform policy (Wallace R, Geller A, & Ogawa VA, Eds.). National Academies Press (US). https://www.ncbi.nlm.nih.gov/books/NBK305917/ [PubMed] [Google Scholar]
  48. Setiono FJ, Gangrade N, & Leak TM (2021). U.S. Adolescents’ Diet Consumption Patterns Differ between Grocery and Convenience Stores: National Health and Nutrition Examination Survey 2011–2018. International Journal of Environmental Research and Public Health, 18(16), 8474. 10.3390/ijerph18168474 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Shield KD, Parry C, & Rehm J (2024). Chronic Diseases and Conditions Related to Alcohol Use 155 Chronic Diseases and Conditions Related to Alcohol Use. Alcohol Research: Current Reviews, 35(2), 155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Silverman E, Gostoli U, Picascia S, Almagor J, McCann M, Shaw R, & Angione C (2021). Situating agent-based modelling in population health research. Emerging Themes in Epidemiology, 18(1), 10. 10.1186/s12982-021-00102-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Stacy C, Meixell B, Lowy J, & Thornton RLJ (2020). Using Land Policy to Improve Population Health. Journal of Urban Health, 97(6), 887–898. 10.1007/s11524-020-00466-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Stankov I, Yang Y, Langellier BA, Purtle J, Nelson KL, & Diez Roux AV (2019). Depression and alcohol misuse among older adults: exploring mechanisms and policy impacts using agent-based modelling. Social Psychiatry and Psychiatric Epidemiology, 54(10), 1243–1253. 10.1007/s00127-019-01701-1 [DOI] [PubMed] [Google Scholar]
  53. Sukthankar G, & Beheshti R (2019). Using Agent-Based Models to Understand Health Related Social Norms. In Social-Behavioral Modeling for Complex Systems (pp. 633–654). Wiley. 10.1002/9781119485001.ch27 [DOI] [Google Scholar]
  54. WHO. (2024). Definition of Youth. https://www.un.org/esa/socdev/documents/youth/fact-sheets/youth-definition.pdf
  55. Yang Y, Hodge AM, Lynch BM, Dugué PA, Williamson EJ, Jayasekara H, Mishra G, & English DR (2024). Sustained Hypothetical Interventions on Midlife Alcohol Consumption in Relation to All-Cause and Cancer Mortality: The Australian Longitudinal Study on Women’s Health. American Journal of Epidemiology, 193(1), 75–86. 10.1093/aje/kwad164 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zins M, Guéguen A, Kivimaki M, Singh-Manoux A, Leclerc A, Vahtera J, Westerlund H, Ferrie JE, & Goldberg M (2011). Effect of Retirement on Alcohol Consumption: Longitudinal Evidence from the French Gazel Cohort Study. PLoS ONE, 6(10), e26531. 10.1371/journal.pone.0026531 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2084866-supplement-1.docx (158.7KB, docx)

RESOURCES