Using ecological momentary data to inform a web-based intervention for romantic partners concerned about their loved ones’ drinking: Study protocol

Abstract

Concerned partners (CPs) play a crucial role in encouraging their loved ones to moderate or seek help for their drinking but are not always equipped with the strategies to approach these conversations effectively. This research comprises three phases: First, we will conduct a dyadic ecological momentary assessment (EMA) study to establish common communication patterns associated with drinking and non-drinking behaviors. Findings from this study will be used to develop a web-based intervention (WBI) prototype, which will be used in a qualitative study with CPs assessing the WBI's feasibility and acceptability. We will evaluate the efficacy of the WBI compared to usual care on CP well-being (e.g., depression, anxiety, social support), the drinking partner's drinking, and relationship functioning via a pilot RCT. All procedures occur remotely. In Phase 1, 50 CPs and their partners (who misuse alcohol via the AUDIT-C) will be recruited from social media and participate in a dyadic EMA study, completing a baseline and follow-up survey and three daily EMA reports for 21 days. In Phase 2, we will develop the WBI and conduct qualitative interviews with 15 CPs. In Phase 3, we will recruit 80 dyads for a pilot RCT and a follow-up one-month post-intervention. Equipping CPs to effectively communicate with their loved ones around drinking may be an important catalyst for decreasing their partner's harmful drinking. The potential reach of this intervention is large such that it can be easily implemented over the web to those who may need help but would otherwise not seek care.

Alcohol misuse, defined as drinking above recommended limits or in a manner that could cause harm to the person who drinks or to those around them [1], is a pervasive health issue associated with psychological, physical, and behavioral problems for individuals and their families. Alcohol misuse imposes a significant burden to the drinking partner's (DP's) loved ones. Romantic partners of individuals who misuse alcohol (i.e., concerned partners; CPs) report lower quality of life and well-being [2,3] and higher rates of caregiver burden [4] and depression, anxiety, and somatic complaints [5,6]. These couples also report lower rates of marital satisfaction [3,7,8] and more romantic conflict [7,8], intimate partner violence [9,10], and separation and divorce [[11], [12], [13], [14]]. Given the dual impact of these harms, supporting both the DP and their CP is critical to preventing more serious alcohol problems from developing.

1. CP-DP communication and alcohol use

Given the significant amount of time CPs and DPs spend with one another, CPs are in an ideal position to exert considerable influence over the DP's drinking decisions [[15], [16], [17]]. CPs are often cited as the primary reason the DP decided to seek professional help [18,19] and can serve as a resource for the DP's treatment and recovery [19]. However, the CP's influence on the DP's drinking is not always positive and although well-intentioned, can sometimes precipitate DP drinking. When CPs use punishing behaviors like yelling, pouring out drinks, or threats to communicate the drinking problem to others [[20], [21], [22]], the CPs' behavior can paradoxically increase both relationship tension [21] and DP drinking [23]. Despite evidence of CPs' strong influence, there is a dearth of empirical evidence for specific ways they influence the DP's drinking on an event-level basis. Dyadic studies have been used to investigate the impact of partner and relationship factors on the DP's later drinking behaviors, but the follow-up timepoints often occur months or even years later, limiting our ability to understand the proximal impact of these dynamics on outcomes.

Ecological momentary assessment (EMA) is a set of longitudinal methods that researchers can use to understand complex behavioral processes in “high resolution” by collecting data as it is happening or “in the moment.” EMA is a reliable approach to identify drinking-related behaviors (e.g., [[24], [25]]), including documenting daily drinking influences among couples (e.g., [[26], [27]]), but currently none of these studies have examined how something their partner said or did immediately preceded their desire to drink or abstain from alcohol. Knowing more about these event-level sequences may help inform the development of a CP-focused intervention, as CPs are often very interested in helping their loved one but are not always equipped with the strategies to approach these conversations effectively [22,[28], [29], [30], [31]].

2. CP-focused interventions

CPs can be influential in the DP's drinking decisions and CPs are often interested in learning about web-based programs to help them or their DP [32]; however, CPs are also an understudied population in need of their own resources and support. Further, given the distress experienced, CPs may be more receptive to interventions than DPs, acting as a lever in promoting help-seeking. Community Reinforcement and Family Therapy (CRAFT; [33]) is an intervention grounded in behavioral theory and evidence suggesting that substance use behavior is shaped by its rewarding and punishing consequences and designed for CPs. CRAFT teaches CPs to provide immediate and tangible social reinforcement that can effectively compete with substance-related reinforcement to promote abstinence and alternative non-drinking behaviors. CPs learn how to communicate more effectively in a way that rewards positive (abstinent) behavior, avoids rewarding negative (drinking) behavior, and supports and encourages help-seeking. DPs whose CP received CRAFT report treatment entry rates from 40 to 86 % [34]. Of the three major CP-focused treatments (i.e., CRAFT; a 12-step intervention based on Al-Anon; Johnson intervention), CRAFT has the highest DP treatment entry rates [[33], [34], [35], [36]].

Partners Connect. Web-based interventions are emerging as a practical mechanism for reaching individuals in the privacy of their own homes and at a self-directed pace [37]. Our team developed and piloted a 4-session CP-focused intervention for active-duty military spouses called Partners Connect that was based on CRAFT principles. Partners Connect CPs reported significant reductions in anxiety and improvements in social support compared to waitlist control CPs [38]. Overall, main effects on CP-perceived DP drinking were not different by group; CPs in both conditions reported that their DP reduced their drinking over time. However, only in the Partners Connect condition, CPs who decreased their punishing responses to their DP's drinking reported their DPs drank less over time [39]. Thus, the scientific premise for a CP-focused web-based intervention (WBI) that targets communication change in the CP is strong.

3. Present research

The current study uses EMA from DP-CP dyads to better understand the communication patterns that often precede DP drinking and non-drinking behavior. We will use these insights to adapt our Partners Connect CP-focused WBI [40] for a civilian sample that is more focused on communication strategies. The WBI will include psychoeducation on interaction patterns commonly associated with DP drinking to support changes in CP communication strategies. Through a series of three studies, we will: identify how specific CP communication patterns influence their partner's alcohol-related craving, motives, and use using dyadic ecological momentary assessment techniques, apply knowledge about these patterns to iteratively develop and pilot test a CP-focused web-based intervention that provides psychoeducation about communication patterns that may influence their partner's drinking, and conduct a pilot RCT to evaluate feasibility, acceptability, and preliminary intervention effects.

4. Method

4.1. Study 1: dyadic EMA

In Study 1, we will identify specific CP communication behaviors that elicit or inhibit DP drinking using a dyadic EMA design. See Fig. 1 for study flow. We predict that CP punishing drinking (e.g., yelling; Hypothesis 1a) and CP rewarding drinking (e.g., bringing alcohol home; Hypothesis 1b) will correspond with increases in DP drinking, whereas CP rewarding sobriety (e.g., planning non-drinking activities) will be associated with decreases in DP drinking (Hypothesis 1c).

Fig. 1.

Study 1 (EMA) flow.

Study 1 Participants. Participants for Study 1 will include 50 couples where one partner (i.e., the CP) is concerned about their partner's drinking and the other partner (i.e., the drinking partner; DP) reports alcohol misuse. CP inclusion criteria are that the CP: be at least 18 years old, be in a romantic relationship with their partner, be living with their partner, have a computer, tablet, or mobile phone with Internet access, have no plans to separate from their partner in the next 60 days, and indicate no concerns they would be physically hurt by their partner on the Intimate Partner Screen [41] that has been validated in primary care outpatient and emergency care settings. We aim to include couples who are in established, committed romantic partnerships and as such, only include couples where both members have no plans to separate in the next 60 days, consistent with similar studies of the proposed intervention [38,40]. CPs will be excluded if they report drinking scores at or above the threshold for DP participation (defined below) to allow clear distinguishability of the role of CP and DP. DP inclusion criteria are that the DP: misuses alcohol as defined by their scores (4+ for women/non-gender conforming, 5+ for men) on the consumption portion of the Alcohol Use Disorder Identification Test (AUDIT-C; 41), has a computer, tablet, or mobile phone with Internet access, and indicates no concerns they would be physically hurt by their partner. The AUDIT-C is a common tool that has been validated as a screen for identifying alcohol misuse [[42], [43], [44]]. Recent research in the general population suggests that AUDIT-C thresholds of 4+ for women/non-gender-conforming and 5+ for males are optimal in sensitivity and specificity for identifying at-risk drinking and mild alcohol use disorder (AUD). Exclusion criteria for DPs include current alcohol treatment and current concern about their partner's drinking. Inclusion criteria for the CP and DP must both be met for couple eligibility; any couple where one person meets criteria for both a CP and DP will be excluded. We will not restrict CPs or DPs based on their sex.

Study 1 Procedures.Recruitment will occur via social media as done in our previous work recruiting CPs and couples with longitudinal designs [30,40]. Specifically, we will utilize our proven Facebook recruitment strategies [30,[45], [46], [47]] and extend to multiple other social media sites (e.g., X, Instagram, Snapchat). All advertisements will target CPs and include a link to the study website, which will also include a link to the screening questionnaire. If eligible, CPs will share an email with their DP, who will complete the screening questionnaire. Once both members screen in, phone calls with both members will be performed to describe study procedures and authenticate them to mitigate participant misrepresentation. Study staff will send emails to the dyad members individually based on their role. These emails will contain links to the online informed consent document, which they can click through to continue to the baseline assessment. We will ask each person to complete surveys independently within a week and then start the EMA for both DP and CP at the same time so that we may examine data for each partner from the same recall period. We have conducted these methods successfully through prompt outreach and monitoring such that over 95 % of couples have started the EMA at the same time across two large dyadic EMA studies (e.g., [48]). In the EMA phase, participants will be sent a push notification through Expiwell (www.expiwell.com) app downloaded on their smartphone that contains a link to complete a short assessment independently three times per day, at randomly selected times in the morning (10a-2p), afternoon (2p-6p), and evening (6p-10p) (see [49]). Research staff will conduct ongoing monitoring of survey completion. If a couple completes the initial survey but one or both partners fail to satisfactorily complete a week of the EMA survey (defined as completing 18 or more EMA surveys in the 7 days), research staff will reach out to the partner with missing data via email, push notification, text message, or phone call to problem-solve and ensure they are not encountering unexpected difficulties. If the partner does not complete their assessments, the couple will not be able to continue in the study and we will send both members of the couple an email that they are no longer eligible for the studyParticipants will receive an email invitation to complete a follow-up assessment immediately after the final EMA, approximately one month after baseline. DPs and CPs will each be compensated $35 in Amazon gift cards for each of the baseline and follow-up assessments. Consistent with industry standards for EMA completion [27,49], participants will also be compensated up to $123 for their EMA entries ($1 for each assessment, $10 for each completed week, and a $30 bonus for three completed weeks).

Study 1 Measures. Shortened descriptions of primary outcomes are presented in Table 1. We will also assess demographic characteristics (e.g., children in the home, race/ethnicity, income) and CP and DP substance use. At baseline, the timeframe for all measures will be past-month. EMA measures include current mood, alcohol craving, alcohol-related communication with their partner and the DP's drinking behavior. Alcohol-related communication will be measured by asking DPs about their behaviors about the frequency of their partner's behaviors toward them regarding their drinking (e.g., punishing or withdrawing when they drink, supporting their sobriety or drinking behaviors) using an adaptation of several measures, including the Significant-other Behavior Questionnaire [20], Partner Management Strategies Questionnaire [50], and others [e.g., 51]. Similarly, CPs will report on their own communications in response to their partner's drinking. Perceived partner responsiveness communications will be assessed similarly [52,53]. Drinking behaviors will include alcohol craving, motives, consumption, and related consequences via validated measures used in EMA studies [24,27,[54], [55], [56], [57], [58], [59]].

Table 1.

Study variables by target and timepoint for studies 1 and 3.

Construct	Baseline/Follow-up Variable (Studies 1 & 3)	EMA Variable (Study 1)
Alcohol	Demographics Quantity/Frequency/Peak Alcohol Use	Daily Drinking Questionnaire Social Drinking Context
	AUDIT
	AUDIT-C (CP report of DP drinking)
	Alcohol Motives	Alcohol Motives
	Drinking Consequences	Drinking Consequences
	Drinking-refusal Self-Efficacyb
		Alcohol Craving
Well-being	Mental Health Symptomsc	Mental Health Symptomsc
	Anger and Anger Expression	Anger
	Loneliness	Loneliness
		Perceived Stress
	Hopefulnessa
	Couples' Resentment
	Current or Past Treatment for Mental Health or Substances
		Momentary Moods
Relationship Functioning	General Social Supporta
	Family Environment
	Communication Patterns Importance of Improving Communicationa
	Conflicts/Disagreements	Conflicts/Disagreements
	Relationship Satisfaction	Relationship Satisfaction
	Attachment Style
	Perceived Partner Responsiveness	Perceived Partner Responsiveness
	Partner Social Support
		Events checklist
		Daily spouse events
Drinking Influence on Relationship	aPartner Drinking Attitudes	aPartner Drinking Attitudes
	Concern about Partner Drinkinga
Communication about Drinking	Drinking Communications	Drinking Communications

Note. All measures are assessed by both the CP and DP unless otherwise noted in parentheses.

^aCP only.

^bStudy 3 only.

^cAnxiety and Depression.

Study 1 Data Analysis. Power was calculated using a Monte Carlo simulation approach [60]. For the DP's EMA data, we conducted simulations using a two-level nested structure, with 63 reports (21 days x 3 reports) at Level 1 nested within 50 DPs at Level 2. We generated and analyzed 1000 data sets for each effect of interest with α = .05 and (conservatively) approximately 15 % of data missing at random. Given these parameters, we would have power greater than .80 to detect Level 1 main effects (e.g., DP's reports of CP punishing communication on DP drinking) as small as d = .0006.

We plan to analyze the EMA data using multilevel models (MLM) [61], which are also called random effects regression models [62]. The primary outcomes (i.e., drinking variables) in Study 1 will be reported only by the DP. Thus, even though some analyses will account for each partner's predictor variables and CP drinking, the analyses will focus primarily on the DP's perception of their CP's behavior and their own drinking behavior. Accordingly, these data will be analyzed using a two-level model, in which the DP's reports of drinking behaviors are modeled as repeated measures dependent variables at Level 1, the DP's reports of CP behaviors modeled as time-varying covariates (i.e., predictors) at Level 1, and characteristics that vary between people (e.g., gender, relationship quality) are modeled as time-invariant or between-person predictors at Level 2. Hypotheses will be tested by evaluating the associations between perceived CP behaviors (e.g., yelling) and DP drinking outcomes (e.g., craving; likelihood of consuming alcohol). We will conduct Level 1 analyses examining whether negative interactions reported by DPs are associated with increased craving or use (modeled using logistic regression). We will code time into intervals (e.g., morning, afternoon, evening) using the timestamp for each report. We will order reports by time of day and create lagged predictors [27,63]. This will allow us to examine the effect of CP interactions on drinking after a delay of hours or even days. We will replicate these findings using CP report at Level 1.

4.2. Study 2: intervention development and qualitative interviews

Study 2 will develop and pilot the feasibility and acceptability of a WBI providing communication-based personalized feedback (see Fig. 2). We will use the Behavioral Intervention Technology Model [64] framework to iteratively develop a four-session WBI that includes CRAFT principles and education about communication patterns learned in Study 1. During the WBI, CPs will receive personalized feedback based on their baseline data on how their communication may influence their DP's drinking. We will evaluate the WBI's feasibility and acceptability with 15 CPs to iteratively improve the WBI.

Fig. 2.

Study 2 (WBI feasibility & acceptability) flow.

Study 2 Participants. Recruitment of CPs for Study 2 (N = 15) will occur similarly to Aim 1. CPs who participated in Study 1 will be eligible to participate in Study 2 if desired. Inclusion and exclusion criteria for the CPs are identical to Study 1.

Study 2 Procedure. We will conduct usability testing procedures assessing ease of use, acceptability (whether the session is satisfactory is operationalized as 70 + average score on the Intervention Usability Scale), appropriateness (whether content is relevant to their issues), and feasibility (whether content can be successfully completed) from our previous formative work [65,66]. We will conduct an iterative beta testing phase with 15 CPs following their completion of WBI but will conduct additional rounds if qualitative input from more CPs is needed. We will conduct a 30-45-min phone interview with each CP to capture feedback on the usability, acceptability, and appeal of the recruitment strategies, and WBI content. CPs will also complete quantitative measures regarding the feasibility, acceptability, and usability of the intervention. We will iteratively revise the WBI following the first 10 interviews, pause and beta-test, and finalize the WBI after conducting an additional five interviews.

The qualitative interview will comprise of three main sections: (1) usability [67] of the web-based platform (e.g.., look and feel, if design elements are well-integrated, ease of use, recommended changes); (2) acceptability, appeal, and feasibility of the WBI information; and (3) impressions regarding the recruitment advertisements. Interviews will be recorded, transcribed, and coded to identify common themes using content analysis. CPs will be compensated $100 in an Amazon gift card for their time.

Study 2 Data Analysis Plan. Qualitative interviews will be transcribed and analyzed by two coders to evaluate the feasibility and acceptability of both the recruitment process and the WBI. Sample size will ultimately be determined by thematic saturation, which is defined as the point in which data collection becomes redundant [68]. Feasibility is defined as how accessible and convenient the processes are, while acceptability is defined as how helpful, appealing, and engaging the processes are [66]. Transcripts will be managed, coded, and analyzed using thematic content analysis [69]. Coding will be iterative and will include comments that may be positive, negative, or neutral. Coders will read each transcript; recurring ideas and important quotes will be identified and organized into groups, allowing a list of themes to emerge. Coders will complete a second reading to compare coded transcripts with the initial codebook to confirm, modify, or reject themes. Additional readings of transcripts will be completed until all categories are saturated and no additional themes emerge. Classic content analysis will be used to categorize quotes within each concept into themes (e.g., personalized feedback about CP-DP communication is helpful) [69]. We will follow procedures to maintain rigor of qualitative analyses, including an audit trail and data triangulation and following the Standards for Reporting Qualitative Research for qualitative data collection and analysis [70].

4.3. Study 3

In Study 3, we conduct a pilot randomized controlled trial (RCT) comparing WBI to psychoeducation control. Study flow is illustrated in Fig. 3. Primary outcomes include CP well-being (e.g., depression, anxiety, social support) and CP-DP relationship functioning (e.g., relationship quality). Secondary outcomes include DP drinking (e.g., consumption, related problems). Although the WBI will target CPs, we will collect data from both partners (N = 80 couples) at baseline and 1-month follow-up to evaluate effects. We predict that compared to control, couples in the WBI condition will evince improved CP well-being (Hypothesis 2a), DP drinking (Hypothesis 2b), and CP-DP relationship (Hypothesis 2c) outcomes.

Fig. 3.

Study 3 (randomized controlled trial) flow.

Study 3 Participants. We will recruit 115 CP-DP couples to retain at least 80 couples at follow-up. Participant inclusion criteria and recruitment strategies for Study 3 are identical to Study 1 with the exception of the AUDIT-C timeframe being past 90 days in Study 3. CPs who participated in Study 1 will be eligible to participate in Study 3 if desired, but not those who participated in Aim 2 due to CPs’ prior exposure to the WBI. Given the sample size and pilot nature of this project, we will not stratify by race/ethnicity, gender, or parent status. Instead, we will target social media ads in diverse areas to increase representation and be inclusive of dyads within and without children.

Study 3 Procedures.Screening. Eligible couples will undergo screening and automatic authentication processes similar to Study 1. Baseline assessment. CPs and their DPs will be sent an email with a link to complete the baseline assessment. CPs will be queried on their engagement with specific behaviors identified and supported by the Aim 1 analyses, with their responses informing the basis of WBI personalized feedback.

Randomization and RCT. Upon baseline survey completion, CPs will be randomly assigned by computer to WBI or psychoeducation control using permuted block randomization with random size blocks to ensure the number of people allocated to each group is about equal throughout recruitment [71]. WBI CPs will complete the sessions in a self-paced format, though we will recommend they complete the sessions approximately five days apart. Psychoeducation control CPs will be routed to visit self-help resources on NIAAA's website [72] immediately following randomization and asked to spend the same amount of time (20–30 min) reviewing site content. Participants will be asked to view content on the website on four different occasions to time-match the length of WBI. CPs will receive an orientation guide that includes the website URL and guidance for how they can navigate this website. This website is considered an active control condition because it contains health assessments, videos, and text-based education on topics that specifically address alcohol and drug misuse. Afterward, users will be asked to go back to the project's website (which is the participant's starting point) and complete a brief post-test measuring comprehension.

Monitoring usage and session-level data. We will collect session-level usability data from CPs about session content and satisfaction as well as data regarding their recent communication behaviors toward their DP. Each week, WBI participants will receive reminder emails to view their next available session on the project website. Control participants will also respond to questions about how long they spent on the website and questions assessing their knowledge of website content (post-test).

Follow-up assessment. CPs and DPs will receive an email invitation to complete a follow-up assessment one month after the final WBI session. CPs and DPs will each be compensated $75 in Amazon gift cards for each of their baseline and follow-up assessments.

Study 3 Measures. Measures used will be identical to Study 1 (as noted in Table 1). Primary outcomes include CP mental health and functioning (e.g., depression, anxiety, anger, social support), DP drinking (e.g., drinking quantity, frequency, motives, craving, consequences), and CP-DP relationship functioning (e.g., satisfaction, conflict, communication, perceived partner responsiveness). CP communication about drinking and perceived partner responsiveness are considered secondary outcomes. We will also ask all participants to complete five to six items after each WBI session to assess their usage and satisfaction. For example, we will ask how often they visited the WBI website and how much time they spent on the site since their last survey. Participants will rate the quality and helpfulness of the session as well as whether they intend to use the information covered in the session.

Study 3 Data Analysis Plan. Power for main effects was estimated using the G∗Power3 software program and for dyadic effects using the APIMPowerR program [73]. Based on the pilot study, we expect a 1-month dyad completion rate of at least 80 %. We used small-to-medium (d = .35) effect size estimates for changes in CP behaviors and mental health symptoms and small (d = .25) effect size estimates for changes in DP drinking to determine the minimum sample size required to achieve 80 % power at alpha = 0.05, a minimum correlation between measures of .50, and two assessment points, resulting in a final sample size of 160 (80 dyads) at 1-month post-intervention. To retain this number, 115 dyads will be recruited (∼80 % retention). With our planned dyad sample, we can detect within-dyad actor and partner effects (i.e., formally semi-partial correlations at the dyad level of analysis) as small as rcp = |.20| with power .80 and rcp = |.18| with power .71.

Analyses will use the standard intent-to-treat (ITT) approach to examine the effect of offering the intervention to all WBI CPs. Our ITT approach will analyze CPs as belonging to the group they were randomized to, regardless of their compliance, because excluding CPs that do not complete the WBI sessions could bias our results in favor of WBI and increase the Type I error rate [74]. Intervention effects will be tested by evaluating outcomes at 1-month follow-up as a function of condition, controlling for baseline to functionally model change in outcomes. Actor-Partner Interdependence Models (APIMs) [75] will be used to examine actor and partner effects, where role (CP or DP) will be used as the distinguishing characteristic within dyads. The primary analysis strategy for evaluating Aim 3 hypotheses will consist of multilevel regression analyses [76]. We will use a three-level model, in which time points and partners will be treated as cross-classified levels, nested within couples [62,77]. We will person-mean center continuous time-varying predictors, grand-mean center continuous time-invariant predictors, and include person means for categorical time-varying predictors to distinguish within-from between-couple variability. We will use a normal distribution/unit link for continuous outcomes, a binomial distribution/logit link for dichotomous outcomes, and a (zero-inflated) Poisson or negative binomial distribution/log link for count outcomes.

5. Discussion

While one of the advantages of this work is the potential reach of the web-based intervention, there are limitations. First, we are recruiting our sample through social media and thus our results will not be generalizable to couples who do not have access to social media. Second, the sample size in each of the studies does not allow for powered testing of moderation or mediation effects. Future research may be necessary to better understand how to optimize this intervention based on relationship characteristics or individual determinants. Finally, the study design precludes supporting CPs whose partner refuses to participate, as well as couples with partners meeting both CP and DP inclusion criteria.

The current study extends existing research showing how influential CPs can be on DP behaviors by integrating more contextual information from EMA data to inform a CP-focused intervention. In doing so, we hope to preliminary establish the utility of this web-based format in hopes of potentially reaching large numbers of CPs who may not otherwise seek support.

CRediT authorship contribution statement

Lindsey M. Rodriguez: Writing – review & editing, Writing – original draft, Visualization, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Cynthia D. Mohr: Writing – review & editing, Writing – original draft, Project administration, Methodology, Funding acquisition. Katherine Nameth: Writing – review & editing, Writing – original draft, Project administration, Methodology. Eric Pedersen: Project administration. Karen Chan Osilla: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Funding acquisition, Conceptualization.

Funding

This research was supported by National Institute on Alcohol Abuse and Alcoholism Grant R34AA030182.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

No data was used for the research described in the article.