Alcohol use disorder (AUD) can lead to cognitive impairments in 50–80% of affected individuals (Bates et al., 2002; Hayes et al., 2016). These deficits may persist even after periods of acute intoxication and withdrawal (Vik et al., 2004) and up to 6 months into abstinence (Fernandez-Serrano et al., 2011). As a result, cognitive issues can significantly affect people in recovery from alcohol use and are a common cause of treatment drop-out (Brorson et al., 2013). Individuals with comorbid substance use and cognitive deficits also experience poorer treatment retention rates (Bates et al., 2006; Copersino et al., 2012; Shulman et al., 2018) and lower self-efficacy (Bates et al., 2006). Disruptions to cognitive processes can significantly impact daily functioning and interfere with the effectiveness of typical treatment modalities for substance dependence, such as learning new coping skills (Kiluk et al., 2011). Given these challenges, it is crucial to develop interventions targeted at remediating underlying cognitive functions. Doing so may help decrease substance use and improve re-entry into the community (Monds et al., 2021; Sullivan et al., 2021). One component identified to be particularly sensitive to alcohol use is executive function (EF; Harper, 1998, 2007; Perry, 2016; Sullivan et al., 2000). EF includes key components of working memory, inhibition and flexibility, which facilitate the execution of more complex processes such as planning and reasoning (Barkley, 2001; Miyake et al., 2000; Suchy, 2009). EF is thought to be associated with the development and maintenance of new behaviours (Perry, 2016; Vik et al., 2004). Deficits in EF can hinder an individual’s ability to actively participate in and gain benefits from treatment for AUD. Such deficits have also been linked to poorer treatment outcomes (Domínguez-Salas et al., 2016).

One potential approach for cognitive remediation in an AUD population is goal management training (GMT), a structured, therapist-led manualised intervention targeting EF (Levine et al., 2000; Robertson et al., 2005; Stamenova & Levine, 2019). This is typically delivered in a group format over 9 weekly 2-h sessions that includes psychoeducation, group brain-storming activities and hands-on activities focused on building cognitive awareness, strategies and confidence (Levine et al., 2000). GMT has demonstrated positive preliminary findings when delivered in substance use disorder (SUD) residential and outpatient programs (Alfonso et al., 2011; Valls-Serrano et al., 2016). In residential settings, GMT was associated with reductions in impulsivity and time taken to plan tasks and an increase in working memory (Valls-Serrano et al., 2016). In outpatient settings, GMT was associated with improvements in inhibition and working memory (Alfonso et al., 2011). However, there remain gaps in the literature regarding the efficacy of GMT with respect to program adherence and client retention in an alcohol outpatient setting. For instance, in Alfonso et al. (2011), people with comorbid DSM-IV Axis 1 disorders were excluded. Given psychiatric comorbidity is common in alcohol treatment populations (García-Carretero et al., 2017), it was of interest to pilot this program with more naturalistic inclusion criteria. In addition, cognitive remediation programs have been more extensively studied in inpatient programs (Marceau et al., 2017) where factors such as current substance use and travel to the treatment site are not barriers to engagement.

The primary aim of this study was to recruit participants to undertake GMT in a specialist Alcohol and Other Drug (AoD) outpatient setting and determine (1) the outcomes on cognition, using a neuropsychological battery and waitlist control design, and (2) how feasible it would be in terms of service engagement. Feasibility was operationalised as ability to meet the target recruitment rate (detailed below), with at least 80% attendance of the GMT sessions. We focused specifically on people with AUD in the first instance, as a relatively high proportion of this cohort is reported to have less polysubstance use than other drug groups (Kedia et al., 2007). The latter eligibility criterion was proposed to minimise the variation between clients, which would aid in drawing conclusions about the effectiveness of GMT in enhancing executive function. This aligns with the original goal of recruiting sufficient participants to conduct statistical analyses on neuropsychological baseline and follow-up measures. The initial aim was to analyse baseline and follow-up neuropsychological data for up to 20 participants in a waitlist control design group GMT intervention. The amended aim, due to low recruitment, was to detail a case series of five clients recruited to a once-per-week group GMT program for 9 weeks. This case series aims to examine the cognitive profiles of individuals in this group, as well as their level of engagement in the intervention and any factors that may facilitate or hinder their adherence to the treatment.

Methods

Participants

All components of the study were ethically approved by New South Wales (NSW) Health Research Ethics Committee. Participants were recruited between February and May 2018 from an AoD outpatient setting in Sydney, Australia. Participants were required to be between 18 years of age and the maximum age of 65 years to reduce confounds of natural cognitive decline associated with normal ageing (Harada, Natelson Love & Triebel, 2013). The initial referral eligibility criteria required clinical indication of the risk of cognitive deficits. This was determined by scoring below the cutoff point of < 26 on the Montreal Cognitive Assessment (MoCA), the screening tool used to assess cognition at the service, or relevant collateral clinical information identified by senior clinicians or addiction medicine specialists. This included factors such as history of emergency department presentation related to alcohol use, risk of Wernicke’s or relevant history of clinical presentation at the service, e.g. severe disorganisation and memory concerns. To determine the eligibility for the study, participants underwent a brief telephone screening conducted by a researcher or clinician who was part of the research team. Participants were required to meet the International Classification of Diseases, Tenth Revision (ICD10; WHO, 2010) criteria for SUD—alcohol; have been abstinent from alcohol for 2 or more weeks; have availability to attend the intervention weekly at the centre; and have consent to completing a breath test analysis for alcohol and a saliva swab for the presence of drugs prior to completing baseline and follow-up neuropsychological testing. They also needed to have adequate English language skills to give valid consent. Exclusion criteria included recent initiation (within preceding month) of psychotropic medication (e.g. antidepressants, antipsychotics, benzodiazepines); any other substance dependence (other than nicotine or caffeine); unable to participate in study procedures due to impending incarceration or travel during the trial; lack of stable housing and/or contact phone number; and participants who were participating in any other clinical study were also excluded.

In total, 34 individuals were referred to participate in the study. Of the 34, nine were lost to follow-up, three reported a change of interest in participating, two were not interested due to length of program, two had competing treatment priorities, three were unable to attend due to employment/training (two were recruited to control), nine were excluded due to current substance use, and one was excluded due to being > 65 years of age. In total, 11 individuals were invited to take part in the study, three were lost to follow-up and one was no longer interested due to length. The final group that was included in the study consisted of five individuals who received the intervention and two individuals who were part of the waitlist control group.

Cognitive Measures

The measures are validated neuropsychological tests (Strauss et al., 2006) chosen to capture key domains of cognitive functioning known to be impacted in AUD, while maximising brevity. The method from the Ridley et al. (2018) paper was used to establish impairment from performance on the neuropsychological test battery. This method was chosen to ensure consistency, as it aligns with the criteria that are typically used for research within the service where the current study was conducted.

The Test of Premorbid Functioning (TOPF)

The participant read a list of words aloud that are phonetically atypical and cannot be correctly pronounced by sounding them out (Wechsler, 2011). This is commonly used as a measure of premorbid intelligence.

Impairment threshold: N/A, used to establish premorbid intelligence.

The Montreal Cognitive Assessment (MoCA)

Form A was used to screen for the study and Form B for the baseline cognitive assessment (Nasreddine et al., 2005). A score of < 26 (included education adjustment) was classified as impaired on initial screening.

Executive Functioning Measures

WAIS-IV: Digits Backwards

This subtest of the WAIS-IV (Wechsler, 2008) assesses working memory. Participants are required to repeat a sequence of numbers verbally in reverse order to their presentation.

D-KEFS: Color-Word Interference (Inhibition Trial)

The Color-Word Interference task was used to measure inhibitory control with conditions 2 and 3 administered (Delis et al., 2001).

D-KEFS: Design Fluency

This was used to capture non-verbal fluency where participants are required to generate as many unique designs as possible in 60 s (Delis et al., 2001).

D-KEFS: Verbal Fluency

This was a letter fluency task, where participants are presented with target letters F, A and S and are required to recount as many words as they can think of that begin with the letter within 60 s (Delis et al., 2001).

D-KEFS: Trail-Making (Switching Trial)

This was used to assess mental flexibility; participants are required to switch between connecting ascending numbers and ordering letters of the alphabet (Delis et al., 2001).

Impairment threshold for all executive functioning measures: < 1.5 standard deviations below age-adjusted score; impairment required on two tasks to be classified as impaired in executive functioning domain.

General Cognitive Functioning Measures

The Repeatable Battery for the Assessment of Neuropsychological Status (RBANS)

This is administered as a screening measure of cognitive functioning in adults with suspected impairment. The measure contains tasks evaluating immediate memory, visuospatial/constructional, language, attention and delayed memory domains (Randolph et al., 1998). Form B was used.

Impairment threshold for general cognitive measure: < 1.5 standard deviations below age-adjusted score for each summary domain (immediate memory, visuospatial/constructional, language, attention, delayed memory).

Intervention—Goal Management Training (GMT) (Robertson, 1996; Levine et al., 2000, 2007)

GMT is a manualised intervention developed for individuals with deficits in EF. The intervention consists of 9 modules and can be delivered in either individual or group format. The intervention was delivered at the service in a group format. Participants engaged in weekly 2-h sessions, which were led by two facilitators who were psychologists. The facilitators led participants through the components of the intervention including psychoeducation, group brain-storming activities, individual pen and paper activities and hands-on activities.

Procedure

Recruitment was via self-referrals from poster advertising around the service and through the service’s existing internal referral process. Clinical staff at the service discussed the research project with clients and provided the research team with contact details of consenting potential participants. Potential participants were contacted to complete the brief telephone screen to determine eligibility as described above.

All participants screened as eligible were scheduled to attend a face-to-face appointment to complete the baseline testing prior to intervention. Prior to all testing, following written consent, participants underwent breath test analysis for alcohol intoxication and saliva analysis for the presence of illicit substances. Any positive returns required that the participant attends another day and be able to return a negative screen to complete testing. A comprehensive neuropsychological test battery was used to establish profiles of objective cognitive deficits. Interviews took approximately 2 h and breaks were provided as necessary to mitigate fatigue. The neuropsychological test battery order was selected to provide an appropriate time delay between similar tasks to avoid contamination, e.g. visuospatial tasks being confused on MoCA and RBANS. Testing was completed by the clinical neuropsychologist and appropriately trained researchers. Participants were reimbursed with a $40 groceries gift card for their time.

Results

Given low recruitment and high attrition rates, statistical analyses were not able to be performed within (i.e. before and after) or between (i.e. intervention vs control) groups for the neuropsychological results. Instead, we adopted a case series methodology. A summary is provided in Table 1 for the five participants who commenced the GMT intervention. There was a high drop-out rate, despite participants providing positive anecdotal feedback about the content and skills learnt. Of the five participants starting the intervention, only one completed all nine sessions. However, four (80%) completed at least the first four sessions. Table 2 provides further details on performance on cognitive measures.

Table 1 Case series summary of the five participants who enrolled in the GMT treatment program
Full size table
Table 2 Participant scores on neuropsychological assessment measures
Full size table

Participant 1: Dropped Out After Session 6

Participant 1 exhibited the most significant level of cognitive impairment: all cognitive domains except for language. She required a higher level of support for instructions during the program and reported difficulty concentrating and understanding tasks. Despite this, she attended the second highest number of sessions, completing sessions 1–6 out of 9.

Participant 2: Completed All 9 Sessions

Despite impairment shown via the MoCA, Participant 2 did not exhibit any impairment of any of the cognitive domains tested. However, he provided subjective reports of difficulty with memory, organising and completing tasks and citing negative affective states (frustration, anxiety etc.) as interfering. He reported benefiting from the mindfulness components of the program, as it allowed him to reduce his arousal level, helping him to utilise skills learnt for remembering and executing tasks. As he completed three sessions in a one-on-one format after all other participants had dropped out, further time was allocated to individually tailoring strategies. Participant 2 provided positive feedback regarding the program and subjective experience of increased functioning.

Participant 3: Dropped Out After Session 5

Participant 3 exhibited impairment on the MoCA and one task within the EF battery, therefore not meeting criteria for impairment for any of the domains. Recording equipment failures resulted in missing data for the MoCA, TOPF and D-KEFS verbal fluency. A pro-rated MoCA score was able to be computed but the other two domains were unable to be reported. Participant 3 completed sessions 1–5, discontinuing due to returning to employment. After session 3, he reported anecdotally applying the skills introduced in the program to reduce the number of subjective concerns he was having with cognition, e.g. remembering tasks and reading effectively.

Participant 4: Dropped Out After Session 1

Participant 4 exhibited impairment on the MoCA and in the visuospatial/constructional domain. She attended for session 1 of the program and was not interested in completing the rest.

Participant 5: Dropped Out After Session 4

Participant 5 exhibited the second highest level of impairment of the participants. He exhibited impairment on the MoCA and in the EF domain, scoring below the cutoff for two of the EF measures. He also exhibited impairment in the immediate memory and visuospatial/constructional cognition domains. He completed sessions 1–4 and, after sustaining a back injury, was unable to complete the program. He provided positive feedback about the skills taught in the program, especially about being able to learn what individually tailored strategies assisted him to improve his remembering and executing tasks.

Discussion

The aims of the study were to (1) determine the outcomes of GMT in an AoD outpatient setting using a neuropsychological battery and waitlist control design and (2) determine how feasible this intervention is in terms of service engagement. The first aim was unable to be investigated due to poor recruitment and retention across the course of the intervention. This result leads into the second aim, in that the aim of 80% attendance was achieved by only one of the five clients who started the program. This suggests that the traditional manualised 9-week GMT intervention is not suitable for roll-out in an AoD outpatient program.

Several variables may have reduced the eligibility of clients to participate in this program. The exclusion of polysubstance dependence reduced the potential participant pool considerably (9 out of 34, or 26% of potential participants). Although the exclusion criterion was implemented to reduce variability among participants and enable the investigation of cognitive changes throughout the intervention, it should be noted that the study’s sample may not be representative of the typical client population in AoD services who often present with polysubstance issues (Gooden et al., 2021; Lintzeris et al., 2016; Bonfiglio et al., 2022). Further studies should pilot interventions including polysubstance use history.

Furthermore, the content of GMT as an intervention is more specific in nature than, for example, relapse prevention skills groups, which tend to be relevant to larger numbers of clients. The specificity of the content may suggest that it applies to a narrower range of clients, resulting in a smaller proportion of clients being eligible to receive the intervention at any given time. As a result, it may only be feasible to enrol five to seven participants in the intervention simultaneously. Significantly, participants had a range of deficits in cognitive functioning across domains. One participant exhibited no impairment on the neuropsychological assessment, through to another participant exhibiting impairment in all but one cognitive domain. Although this study was provided to target EF, not all participants exhibited a deficit in EF. Only two of the five participants (Participants 1 and 5) met standard criteria for EF impairment—which is an impairment score on more than one task across the EF domain. This finding suggests that a program specifically focusing on EF may also not be the most appropriate for this cohort. On the other hand, it can be argued that measures of EF such as those employed in the current study may not capture the emotional decision-making deficits that are typical in people with AUD; thus, it may be worth considering other EF measures for future studies. Of note, higher impairment was not necessarily associated with shorter length of engagement in the program. Participant 1, exhibiting the highest level of deficit across all domains, completed the second highest number of sessions of all participants. The rate of participant drop-out from modules four to six indicates that the intervention might not be acceptable as a nine-module group format in outpatient drug and alcohol populations. If a similar pattern is observed in subsequent rounds, it may be appropriate to consider a more tailored or abbreviated version of the intervention that is specific to the deficits identified.

Participant feedback also indicated that the group-based format may be unsuitable for such a broad range in cognitive abilities (i.e. participants with severe impairment may require more intensive support). Feedback from Participants 1 and 2 suggests that a one-to-one format might be more acceptable. Participant 1 exhibited higher impairment and reported difficulty in following content as it was paced for the group. Participant 2 provided anecdotal feedback about the positive delivery in 1:1 format for the final three sessions. This finding indicates that customising the pace of the sessions to match the client’s needs may enhance adherence. Additionally, it may be beneficial to emphasise and reinforce the strategies that are most effective for the client’s cognition. Findings from the current research also provide a useful comparison between residential and outpatient settings. Previous research demonstrating effective use of group cognitive remediation approaches has predominantly occurred in inpatient settings (Marceau et al., 2017; Valls-Serrano et al., 2016). In these settings, clients stay at the location where the intervention is provided, which reduces the effort required to attend sessions and lowers the likelihood of forgetting to attend. Moreover, in a residential program, clients are expected to participate in various treatment formats throughout the day. The results of the present study suggest that the increased effort involved in remembering and attending a centre-based program may contribute to the reduced feasibility of a group format in an outpatient setting.

Conclusion

Based on the recruitment and retention rates, it appears that a group version of GMT in its current form is not feasible for this outpatient AUD population. Despite an attempt to use more naturalistic inclusion criteria than in previous studies, the results suggest that the eligibility criteria need to be even less stringent in subsequent iterations of the program to increase the number of participants. Participants generally provided positive feedback regarding the content within GMT, suggesting that the intervention itself may be acceptable for outpatient settings. However, retention rates and participant feedback from subsequent rounds suggest that a shortened or 1:1 format for GMT delivery has greater applicability for individuals with people with AUD in an outpatient clinic setting. Research with greater numbers is required to be able to statistically analyse neuropsychological data collected from participants.