Authors: Dylan E. Kirsch (1Department of Psychology, University of California, Los Angeles, 1285 Franz Hall, Box 951563, Los Angeles, CA 90095-1563, USA), Erica N. Grodin (1Department of Psychology, University of California, Los Angeles, 1285 Franz Hall, Box 951563, Los Angeles, CA 90095-1563, USA; 2Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, Los Angeles, CA, USA), Danielle Keenan-Miller (3Private Practice, Los Angeles, CA, USA), Lara A. Ray (1Department of Psychology, University of California, Los Angeles, 1285 Franz Hall, Box 951563, Los Angeles, CA 90095-1563, USA; 2Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, Los Angeles, CA, USA)
Categories: Article, alcohol use disorder, substance use, addiction, functioning, stress
Source: Journal of studies on alcohol and drugs
Authors: Dylan E. Kirsch, Erica N. Grodin, Danielle Keenan-Miller, Lara A. Ray
The alcohol field has increasingly recognized that functioning plays a key role in the clinical course of alcohol use disorder (AUD). However, there remains a need for comprehensive instruments that capture functioning in this population. This study examined the utility of the UCLA Life Stress Interview Chronic Stress Assessment (LSI)—a functional domains-based assessment widely used in mood disorder research—as a tool to measure functioning in key life domains in AUD. We: (1) compare LSI scores between individuals with and without current AUD; and (2) examine whether LSI scores are associated with recent alcohol use and AUD symptomology.
This secondary case-control study included 38 adults with AUD and 35 age- and biological sex-matched controls. The LSI was used to assess functioning in interpersonal, occupational, personal health, and family health domains over the prior 6-months. The Timeline Followback was used to assess recent alcohol use, and the Structured Clinical Interview for DSM-5 was used to assess AUD symptomology.
Individuals with AUD exhibited significantly poorer interpersonal, occupational, and personal health functioning, compared to controls. Among individuals with AUD, worse interpersonal, occupational, and personal health functioning correlated with a higher number of AUD symptoms, and both interpersonal and occupational functioning also correlated with greater recent alcohol use.
This secondary study provides preliminary support for the LSI as a tool to assess functioning in AUD, offering an unbiased, addiction-independent evaluation of functioning across key domains. Future work should replicate results in larger samples, conduct formal psychometric evaluation, and examine the effects of biological, demographic, and clinical factors.
Alcohol use disorder (AUD) is associated with impairments across multiple functional domains, including interpersonal, occupational, and health domains (Ray et al., 2019). Such functional impairments act as both a consequence of and contributor to alcohol use (Witkiewitz et al., 2019; Witkiewitz et al., 2021; Witkiewitz et al., 2018). This bidirectional relationship contributes to the chronic relapsing nature of AUD (Blaine & Sinha, 2017; Collins et al., 2007; French et al., 2011; Read et al., 2006). In recognition of this dynamic, recovery frameworks have increasingly recognized improvements in biopsychosocial functioning as a core indicator of AUD recovery (Ashford et al., 2019; SAMHSA, 2012; Witkiewitz et al., 2021; Witkiewitz et al., 2020). The National Institute on Alcohol Abuse and Alcoholism (NIAAA) proposed an operational definition of AUD recovery, which includes improvements in biopsychosocial functioning and quality-of-life, in addition to the cessation of heavy drinking and remission from DSM-5 AUD (Hagman et al., 2022). Despite this conceptual shift, validated tools to assess functioning in AUD remain limited.
Several conceptual frameworks and corresponding assessments have been developed to measure functioning in AUD (Best & Hennessy, 2022; Carlon et al., 2022; Cloud & Granfield, 2008; Kiluk et al., 2019; Kirouac & Witkiewitz, 2018). These include instruments that evaluate substance-related consequences [e.g., Addiction Severity Index (ASI) (McLellan et al., 1992), Drinkers Inventory of Consequences (DrInC) (Miller et al., 1995)], recovery capital [e.g., Assessment of Recovery Capital (Groshkova et al., 2013)], and quality-of-life [e.g., World Health Organization Quality of Life (Group, 1998)]. Each of these approaches provides valuable insight into specific aspects of functioning. Consequence measures assess the negative outcomes of alcohol use, recovery capital measures capture the social, human, and community resources that support recovery, and quality-of-life measures reflect an individual’s subjective perception of well-being and life satisfaction. However, while each framework contributes to our understanding of functioning in AUD, none offer a comprehensive, integrated assessment of overall functioning. For example, substance-related consequences may mediate or exacerbate dysfunction, but they do not fully capture impairments that persist even in the absence of acute alcohol use. Therefore, the field stands to benefit from instruments that comprehensively assess functioning.
The UCLA Life Stress Interview Chronic Stress Assessment (LSI) offers a unique opportunity to assess functioning in AUD in domains that intersect with key components of these frameworks. This semi-structured, clinician-administered interview evaluates functioning across interpersonal relationships, work, financial stability, and health domains (Hammen et al., 1987), which overlap conceptually with both recovery capital and quality-of-life constructs. Unlike traditional measures of recovery capital and quality-of-life, which largely rely on self-report, the LSI provides behaviorally anchored, contextually-grounded ratings based on objective behavioral indicators rather than personal perception. This allows for a more nuanced understanding of the tangible challenges people experience. Additionally, unlike substance-related consequences measures, the LSI assesses functioning independently of substance use, allowing for an unbiased evaluation of how impairments may precede, or exacerbate, alcohol use. Importantly, it does not require individuals to acknowledge a direct causal link between alcohol use and impairment, reducing reliance on insight (Kiluk et al., 2019). Therefore, the LSI can bridge together contemporary frameworks by capturing recovery-supporting resources (recovery capital), real-world well-being (quality-of-life), and substance-related consequences.
While extensively implemented in mood disorder research (Hammen, 1992; Hammen & Brennan, 2001; Hammen & Brennan, 2002; Hammen et al., 1989; Rao et al., 1999; Vrshek-Schallhorn et al., 2014), the LSI’s use in addiction research, particularly AUD, has been limited (Eames et al., 2014). Given its focus on real-world functioning in domains that overlap meaningfully with existing frameworks, the LSI offers a promising method for assessing functioning in AUD. The LSI’s well-established validity and reliability in mood disorders support its potential applicability to other psychiatric conditions, such as AUD, though additional validation in this population is needed. This preliminary study examined whether the LSI can detect group-level differences in functioning between individuals with and without AUD, and whether LSI scores relate to individual differences in AUD severity and alcohol use. This effort leverages 40 years of research examining functioning in mood disorders and represents an initial step towards extending the LSI to AUD populations.
This study is a secondary analysis of data collected for a parent study (R21AA030643; see Supplemental Material). Seventy-three men and women (N=38 with AUD; N=35 controls) were enrolled through the parent study between June 2023 and August 2024. The sample size for the current study was determined based on a priori power analyses conducted for the parent study, which required 64 completers. The current analyses included more than 64 participants due to participant dropout before completion of the parent study. The current sample reflects participants who completed all relevant measures for this analysis, which is a preliminary extension of the parent study. Participants were recruited through online and newspaper advertisements, campaigns in mass transit in Los Angeles, and targeted recruitment of previous participants who agreed to be contacted for future studies. Study procedures were approved by the UCLA IRB. After a full explanation of study procedures, all participants provided written informed consent.
Initial screening procedures were conducted via telephone. Eligible participants were invited for an in-person screening assessment. As part of the enrollment criteria for the parent study, participants in both groups were required to be between the ages of 18 – 65 years old. Participants in the AUD group were eligible if they (1) active (past 3-months) moderate-to-severe AUD; (2) reported heavy drinking levels, as defined by the NIAAA as ≥14 drinks/week for men or ≥7 drinks/week for women; and (3) Alcohol Use Disorder Identification Test (AUDIT) score ≥8. Eligible participants in the control group had (1) no history of moderate-or-severe AUD; (2) no AUD of any severity over the past 5 years; and (3) AUDIT score ≤7. Additional exclusion criteria for all participants (1) current treatment for alcohol use, a history of treatment in the 30-days before enrollment, or those seeking treatment; (2) past-year DSM-5 diagnosis of substance use disorder (SUD) for any psychoactive substances other than nicotine; (3) lifetime DSM-5 diagnosis of schizophrenia, bipolar disorder, or any psychotic disorder; (4) current DSM-5 major depressive disorder (MDD) with suicidal ideation; and (5) and current use of psychoactive drug, except for tetrahydrocannabinol (THC), as determined by urine toxicology screen. As the parent study had a neuroimaging component, additional exclusion criteria for all participants (1) pregnancy, nursing, or refusal to use reliable method of birth control, if female; (2) nonremovable ferromagnetic objects in the body; (3) claustrophobia; and (4) serious head injury or prolonged period of unconsciousness (>30 minutes).
Clinical and demographic characteristics were measured using a battery of standard clinician-administered and self-report assessments related to alcohol use, alcohol-related consequences, and depressive symptoms (see Supplemental Material). Participant demographic and clinical characteristics are detailed in Table 1.
The LSI was used to assess functioning over the past 6-months on nine quality of intimate relationships, close friendships, social life, relationships with family members, relationships with children, work, finances, health of self, health of family members. The psychometric properties and clinical utility of the LSI have been established across 40-years of research in mood disorders (Hammen, 1992; Hammen & Brennan, 2001; Hammen & Brennan, 2002; Hammen et al., 1992; Hammen et al., 1989; Metts et al., 2021; Rao et al., 1999; Vrshek-Schallhorn et al., 2014). This semi-structured interview generates clinician-ratings of functioning in each domain on a 5-point scale in half-point increments with behaviorally specific anchors, where 1 represents superior/exceptional functioning and 5 indicates severe difficulties. Ratings of functioning are based on factual data in each domain (e.g. frequency of conflict, number of friends) and are scored according to standardized anchor criteria, rather than the individual’s perceptions of their own functioning. Additional scoring detail is described in the Supplemental Material. Certain domains were only administered if they were relevant to an individual’s current circumstances (see Supplemental Material for detail). Supplemental Table 1 presents the probes used for each domain, the number of participants who received a rating in each domain, and average scores. Supplemental Table 2 displays the intercorrelations of the individual domain scores. The LSI was administered and rated by master’s-level clinicians trained by a graduate of the lab where the measure was developed (DKM).
We generated two composite an interpersonal functioning score and an occupational functioning score. The interpersonal functioning score was generated by averaging scores on quality of intimate relationships, close friendships, social life, relations with family members, and relations with children, if applicable, domains. The occupational functioning score was generated by averaging scores on work and financial domains. These two subscales demonstrated acceptable internal interpersonal functioning (Cronbach’s α = 0.79) and occupational functioning (Cronbach’s α = 0.77).
Analyses were conducted using IBM SPSS Statistical Software V28. We examined between-group differences in LSI functional domain scores, interpersonal functioning, occupational functioning, self-health, and health of family members. Univariate ANCOVA was used to evaluate group (AUD, control) differences in these LSI outcomes, each modeled separately. Age, sex at birth, and ethnicity (Y/N Hispanic/Latino) were covariates. Partial eta squared (ηp^2^) was used as a measure of effect size. Based on conventional guidelines, values of 0.01, 0.06, and 0.14 are interpreted as small, medium, and large effects. Given the novel combination of LSI variables to generate interpersonal and occupational summary scores, we also repeated these models for each individual LSI domain (Supplemental Table 1). We conducted two sensitivity analyses examining the role of depression. First, we repeated primary models including BDI-II scores as a covariate to account for subthreshold depression symptoms. Second, we repeated primary models after excluding participants with current MDD (N=7 with AUD). Results are presented in Supplemental Table 3. We then examined correlations between LSI outcome variables with recent alcohol use and number of AUD DSM-5 symptoms within the AUD group. Several alcohol use variables, including total drinks and drinks per drinking day (DPDD) were non-normally distributed (total skewness=2.6, kurtosis=9.1; DPDD: skewness=1.9, kurtosis=3.9); both variables were logarithmically transformed. Significant results (p<0.05) are reported below. We also explored correlations between LSI outcomes and (1) DrInC scores among individuals with AUD, and (2) recent depressive symptoms across all participants; results are presented in the Supplemental Material.
The AUD group reported greater past-month alcohol use (p’s<0.001) and alcohol-related consequences (p’s<0.001). The AUD group also had a greater percentage of individuals who smoke cigarettes (p=0.007) and individuals with positive THC toxicology screens (p=0.01). Additionally, the AUD group had a greater prevalence of current MDD (p=0.01) and greater recent depression symptoms (p<0.001). Lastly, the groups differed in ethnicity such that the AUD group had a greater prevalence of individuals who identified as Hispanic/Latino (p=0.03).
The AUD group exhibited poorer interpersonal functioning (higher LSI scores) compared with the control group (F~(1,68)=10.7, p=0.002, ηp^2^=0.14). Across both groups, females exhibited worse interpersonal functioning compared with males (F(1,68)=5.3, p=0.024, ηp^2^=0.07). Participants who identified as Hispanic/Latino exhibited better interpersonal functioning than those who did not (F(1,68)=7.6, p=0.007, ηp^2^=0.1). The AUD group also exhibited worse occupational functioning compared with the control group (F(1,68)=13.8, p=0.001, ηp^2^=0.14). Additionally, the AUD group exhibited worse self-health compared to the control group (F(1,68)=5.1, p=0.027, ηp^2^=0.07). Across both groups, there was a significant effect of age such that older age was associated with worse self-health (F(1,68)=4.9, p=0.030, ηp^2^=0.07). Groups did not differ on the health of family members domain (F(1,68)=1.2, p=0.28, ηp^2^=0.02), however, participants who identified as Hispanic/Latino reported worse health of family compared to those who did not (F(1,68)~=4.9, p=0.030, ηp^2^=0.07). See Figure 1 and Table 2. Supplemental table 1 shows average scores and group differences on individual LSI domains.
Among individuals with AUD, poorer interpersonal functioning correlated with greater alcohol use (total r=0.57, p<0.001; drinking r=0.33, p=0.04; heavy drinking r=0.61, p<0.001; DPDD: r=0.41, p=0.01) and more AUD symptoms (r=0.32, p=0.049). Similarly, worse occupational functioning correlated with greater alcohol use (total r=0.53, p<0.001; heavy drinking r=0.44, p=0.006; DPDD: r=0.54, p<0.001) and more AUD symptoms (r=0.44, p=0.005). Lastly, poorer self-health correlated with more AUD symptoms (r=0.50, p=0.001), but not with recent alcohol use. See Figure 2 and table 3.
The alcohol field has increasingly recognized that functioning plays a key role in the course of AUD; however, the field stands to benefit from instrumentation to capture functioning in this population. This secondary study aimed to provide preliminary support for the LSI as a comprehensive measure of biopsychosocial functioning in AUD. This effort leverages over 40 years of research assessing functioning in mood disorders (Hammen, 1992; Hammen & Brennan, 2001; Hammen & Brennan, 2002; Hammen et al., 1989; Rao et al., 1999; Vrshek-Schallhorn et al., 2014). We found that individuals with AUD exhibited significantly poorer interpersonal, occupational, and personal health functioning compared to age- and sex-matched controls, while the health of family domain did not differ between groups. Furthermore, the LSI effectively captured individual-level variability in functioning as it relates to AUD severity and alcohol use. Specifically, worse interpersonal, occupational, and personal health functioning correlated with greater AUD severity, and interpersonal and occupational functioning correlated with greater alcohol use. Notably, female sex was associated with poorer interpersonal functioning, while identifying as Hispanic/Latino was linked to better interpersonal functioning. These findings suggest that biological and demographic factors may influence functional outcomes in both healthy adults and those with AUD. Given the relatively small sample size, future work is needed to replicate and further interrogate these effects. In summary, these findings provide preliminary support for the LSI’s utility in detecting functional impairments in AUD; however, additional research in larger samples and formal psychometric evaluation is needed.
A notable strength of the LSI is its ability to capture real-world functioning across domains that intersect with existing frameworks such as recovery capital and quality-of-life. Specifically, the LSI captures both the presence/absence of recovery-supporting resources and real-world challenges that may shape an individual’s lived experience. Additionally, the LSI complements existing measures of substance-related consequences by offering an addiction-independent evaluation of functioning. We found that LSI domains showed stronger associations with DrInC subscales in conceptually related areas (e.g., LSI interpersonal and DrInC interpersonal) and weaker associations in unrelated areas (e.g., LSI interpersonal and DrInC physical), supporting both convergent and divergent validity (Supplemental Table 4). These findings suggest the LSI captures broader dimensions of functioning, while also reflecting impairments commonly attributed to alcohol use. Additionally, while no clinician-administered tool is entirely free of interpretation, the LSI offers a more objective assessment of functioning, relative to self-report formats, because of its structured format, emphasis on observable indicators, and anchored scale.
The AUD and control groups were matched on age and sex, but differed on ethnicity, cigarette smoking, and MDD/depression symptoms. Given the high prevalence of cigarette use and depression in AUD, excluding participants with this presentation would limit generalizability of findings and reduce their applicability to real-world populations where such presentations are common. However, given the well-established association between depression and biopsychosocial functioning, we conducted a complementary set of sensitivity analyses examining the influence of depression symptoms, ranging from subclinical (BDI-II scores) to clinically significant depression (MDD) (Supplemental Table 3). When individuals with MDD were excluded, group differences in functioning largely persisted, suggesting that the relationship between AUD and functional impairment is not solely attributable to co-occurring MDD. However, when controlling for BDI-II scores, group differences in LSI outcomes were attenuated and became non-significant. While AUD and depression are distinct conditions, emerging frameworks such as the Addictions Neuroclinical Assessment (ANA) conceptualize negative emotionality as a core domain of AUD (Kwako et al., 2017, Kwako et al., 2016, Kwako et al., 2019). Critically. The BDI-II has been consistently shown to be a key indicator of the negative emotionality domain (Votaw et al., 2020, Nieto et al., 2021, Kwako et al., 2016). Therefore, controlling for BDI-II scores may partially remove variance that is inherently linked to AUD phenomenology through negative emotionality. Indeed, negative emotionality is a key predictor of addiction recovery (Linn et al., 2023), perhaps suggesting the LSI is capable of capturing clinically relevant heterogeneity within AUD samples. Nevertheless, future work, with larger samples is needed to disentangle the additive and/or interactive effects of AUD and depression on functioning.
Our results indicate that greater functional impairment is associated with more alcohol use among individuals with AUD. While speculative, these results suggest the LSI could be informative in the context of AUD recovery research. For instance, the LSI may have the potential to serve as an outcome-tracking tool throughout recovery, or could help identify individuals at greatest risk for alcohol use based on their functional profile. More broadly, the LSI could contribute to a better understanding of treatment efficacy, with a particular emphasis on the lived experience of individuals with AUD. However, this study was cross-sectional and conducted in a sample not currently engaged in recovery, therefore, we acknowledge that further research, particularly longitudinal research in the context of recovery. Prior work in mood disorders has shown the LSI is sensitive to within-subject changes over time (Hammen et al., 2008; Kim et al., 2007; Metts et al., 2021; Mineka et al., 2020), supporting its potential utility to track functional recovery in AUD.
The relatively small sample size, particularly for within-AUD analyses, is a limitation of this study. Coupled with the number of statistical comparisons, this increases the risk of Type I and Type II error. However, the observed effect sizes were in the medium range, with ηp^2^ values ranging from 0.07 to 0.14 for group comparisons, and correlation coefficients (r) ranging from 0.33 to 0.61 for within-AUD analyses, supporting the potential clinical relevance of findings. To reduce the number of statistical comparisons, we grouped individual LSI domains into composite scores reflecting interpersonal and occupational functioning. While this is a novel analytic approach, the domains grouped within each composite were highly correlated and demonstrated acceptable internal consistency. Future work should focus on continued development of the LSI in AUD through (1) replication in larger samples; (2) formal psychometric evaluation to establish the reliability and validity of the LSI in AUD populations; and (3) efforts to broaden the LSI’s application, increase accessibility and ease of implementation.
This secondary study evaluated the LSI as a tool for assessing functional impairment across life domains in AUD. Findings suggest the LSI can detect differences in interpersonal, occupational, and personal health functioning between individuals with and without AUD, as well as individual-level variability in functioning as it relates to AUD symptom severity and alcohol use. This study provides preliminary support for the LSI as a tool to assess functioning in AUD. As an interviewer-administered instrument, the LSI provides an assessment of functioning extending beyond self-report. Additional research is needed to establish the reliability and validity of the LSI in larger samples and to tease apart the additive and/or interactive effects of biological (sex), demographic (ethnicity), and clinical (depression) factors.