Validity of the brief executive-function assessment tool in an outpatient substance use disorder setting

Abstract

The Brief Executive-function Assessment Tool (BEAT) was developed and validated for use in residential substance use disorder treatment settings, where participants are mostly abstinent. It is therefore unclear whether the BEAT is valid for use in outpatient settings, where participants may be actively using substances. The effects of acute intoxication and withdrawal have the potential to alter the results of the BEAT. The current study sought to establish construct and criterion validity of the BEAT in an outpatient substance use disorder sample and to detect its sensitivity to substance use over the previous 24 hours and also over the past month. A total of 74 clients of a New South Wales-based outpatient substance use disorder service participated in the current study. Construct validity was demonstrated by significant correlations between the BEAT and three performance-based tests of executive functioning. Criterion validity was established in that the BEAT discriminated between those deemed impaired or not on a criterion composite measure of executive functioning. Test operating characteristics (88% sensitivity, 69% specificity, 44% PPV, and 95% NPV) were also established relative to this composite measure as a reference standard. The BEAT was insensitive to use/abstinence over the previous 24 hours and the past month.

Keywords:

• Alcohol and drug
• cognitive screening
• executive function
• substance use disorder
• validity

It is widely assumed that the acute intoxication and/or withdrawal effects of alcohol and other substances, such as cannabis, methamphetamine and opioids, impacts the cognition of people with Substance Use Disorders (SUD; Ekhtiari et al., 2021; Volkow& Blanco, 2023). Whilst this is a reasonable assumption given the well documented effects of substances on healthy individuals, there has been some evidence to suggest that people with SUD show no significant effects (Bruijnen et al., 2019a). In a study of a large and heterogeneous group of outpatients with SUD, years of regular use, days abstinent, severity of dependence and/or abuse, polysubstance use, depression, anxiety and stress were all unrelated to cognitive outcomes (Bruijnen, 2019a). This has implications for cognitive testing and screening in this population, especially in outpatient populations, where substance use is uncontrolled.

An increase in the effect of a drug is termed sensitization and a decrease in the effect is known as tolerance. When a given dose of a drug yields weaker effects relative to previous doses of the same size, tolerance is said to have occurred. Tolerance can be acute or chronic with acute tolerance being observed within the duration of a single dose whereas chronic tolerance is a decrease in the effect relative to previous doses (Comley & Dry, 2020). Neurocognitive deficits in SUD fall into three domains: positive valence, negative valence and cognitive control, which are variably influenced by the three stages of the drug use cycle: preoccupation and anticipation, binge and intoxication and withdrawal and negative affect (Ekhtiari et al., 2021). Thus, different aspects of cognition may be affected at different stages of the cycle.

A brief review of the potential impacts of tolerance and intoxication on cognition in commonly used substances is provided below. Following alcohol consumption, blood alcohol concentration (BAC) rapidly rises, peaks, and gradually falls, forming a curve characterized by two “limbs”: ascending and descending (Williams et al., 2021). Acute tolerance to alcohol is associated with differential neurobehavioral and subjective effects depending on the limb during which it is assessed, such that subjective intoxication tends to be higher on the ascending limb and has been associated with impairments in response time during cognitive task performance. On the descending limb, errors on cognitive tasks have been reported; as well as observations of participants underestimating their degree of intoxication and degree of impairment; poorer inhibitory control and impaired perceptions of the dangerousness of driving while intoxicated; and greater willingness to drive (Williams et al., 2021).

Comley and Dry (2020) in their investigation of acute behavioral tolerance to alcohol noted a trend wherein simpler tasks exhibited a decrease in dose effect, while measures of speed in cognitive tasks proved more sensitive to acute tolerance than measures of accuracy. The review provided evidence of acute tolerance spanning a variety of doses, implying its occurrence across a broad range. However, the authors cautioned that the tested dose range likely falls considerably below real-world situations, particularly in cases of binge drinking. They further speculated that acute tolerance might manifest at higher doses, but emphasized the limited generalizability of their findings, given the scarcity of understanding regarding effects at higher doses and in real-world situations.

In a study by Didier et al. (2023) investigating alcohol tolerance among light drinkers, heavy drinkers, and individuals with alcohol use disorders, it was discovered that the alcohol use disorder and heavy drinker groups, in comparison to the light drinker group, perceived less impairment and exhibited greater behavioral tolerance to a dose typically associated with a binge drinking episode. However, when exposed to a very high alcohol dose commensurate with high-intensity drinking, individuals with alcohol use disorder displayed significant impairment on both psychomotor and cognitive tasks. The conclusion drawn was that behavioral tolerance observed in drinkers with alcohol use disorder may be dose-dependent, indicating that tolerance resulting from regular excessive drinking can be surpassed by consuming alcohol at higher intensities.

Zhong et al. (2022) in an umbrella review of prospective meta-analyses about the associations between chronic alcohol consumption and health outcomes, showed that in terms of neurocognitive outcome, low alcohol consumption decreased the incidence of dementia with high epidemiologic evidence. However, they noted that there is extensive evidence available on excessive alcohol as a risk factor for dementia.

McCartney et al. (2023) in their review of the effects of cannabis reported there is a "window of impairment" lasting from 3 to 10 hours after use, the exact duration of which depends on the dose, route of administration, and regularity of use. Occasional cannabis users were found to experience more impairment compared to regular users, suggesting greater tolerance in the latter. While some lower-quality studies have noted negative "next day" effects on cognitive function and safety-sensitive tasks, it was noted higher-quality studies have not observed such effects. They concluded that the overall scientific evidence supporting the idea that cannabis use impairs "next day" performance is limited.

Figueiredo et al. (2020) using meta-analytic techniques to estimate the consequences of chronic cannabis use on a range of neurocognitive domains found inconclusive evidence that there are cognitive impairments associated with chronic cannabis use. They concluded that there was only a limited association between chronic cannabis use and cognitive impairments in adults.

Amphetamine-type stimulants (ATS) such as amphetamine (AMPH, “Speed”, “Pep”), methamphetamine (METH, “Crystal Meth”, “Ice”), and 3,4 methylenedioxymethamphetamine (MDMA, “Ecstasy”, “Molly”) are the third most prevalent group of illicit substances. The chronic use of all of these ATS may be associated with long-term cognitive impairments, including executive dysfunction (Opitz et al., 2023). These authors conducted a meta-analysis to investigate the relationship between the recreational use of ATS (namely AMPH, METH, and MDMA) and executive dysfunction (cognitive inhibition or interference control) in adults. They reported robust small effect sizes for ATS-related deficits in interference control, which were mainly seen in methamphetamine, as compared to MDMA users. They concluded that their meta-analyses could not provide a definitive answer to the question of whether impaired interference control is caused by ATS use or a premorbid cognitive control deficit, which might make the individual more vulnerable to ATS use.

In a study conducted by Gooden et al. (2022), the biopsychosocial and neuropsychological profiles of methamphetamine (MA)-polysubstance users reporting cognitive impairment were compared to those of an alcohol-using group. The neuropsychological profiles were largely similar between the two groups across various cognitive domains. However, individuals in the MA-polydrug group exhibited a higher risk of overall harm from substance use at a significantly younger age. The researchers concluded that factors such as the age of first use, emotional distress, and indirect substance-related harms, including overdose and bloodborne virus infection, may be pertinent to the experiences of cognitive difficulty in MA-polydrug users.

Tolomeo et al. (2021), in their review, concluded that chronic heroin use is associated with limited evidence of impairments in memory, impulsivity, compulsivity, and decision-making. They noted insufficient evidence regarding whether these impairments and disorders recover after abstinence. While acknowledging the limited research in this area, they suggested a general improvement in at least some areas of neuropsychological functioning after a minimum of 2 weeks of abstinence from heroin use. Darke et al. (2012) showed that patients on opioid maintenance had poorer cognitive performance than abstinent ex-users and non-heroin using controls. This suggests that intoxication and/or withdrawal of opioids has an impact on cognition, which has been suggested by others (e.g., Tolomeo et al., 2021).

Crowe and Stranks (2018), in their review exploring the impact of benzodiazepines use on cognitive functioning in long-term current users in the age range 38–71 years, identified negative effects across multiple cognitive domains, including working memory, processing speed, divided attention, visuoconstruction, recent memory, and expressive language. They observed deficits in individuals who had recently withdrawn and those who successfully abstained following withdrawal, with impairments noted in recent memory, processing speed, visuoconstruction, divided attention, working memory, and sustained attention. The conclusion drawn is that various neuropsychological functions are impaired due to prolonged benzodiazepine use, and these impairments are likely to persist even after withdrawal.

Liu et al. (2020) conducted a systematic review and meta-analysis on benzodiazepine use and abuse in an elderly population and reported results counter to those of Crowe and Stranks (2018). The results consistently showed impairment in elderly benzodiazepine users’ processing speed but not in memory function. They suggested that the interaction between age and benzodiazepine’s effect on memory requires more robust validation. In a benzodiazepine abusers subgroup they found significantly lower Mini-Mental Status Examination scores compared to the controls, while the benzodiazepine regular users’ scores were not significantly different from controls. They concluded that impairment in global cognition occurs after the benzodiazepine use develops into abuse. However, they recommended more experiments are needed to reach more reliable conclusions.

In the real world, people with SUD are commonly polysubstance users and the acute effects of polysubstance use are less clear (Meyerhoff, 2017), with interaction effects across substances. For example, Hayley et al. (2023) have shown that MA enhances psychomotor speed and masks the sedative and performance effects of low doses of alcohol.

When assessing the cognition of outpatients who have access to drug and alcohol services, it is particularly important to know whether the cognitive test is sensitive to the effects of acute intoxication or withdrawal, as results indicating cognitive impairment may be influenced not only by the chronic effects of substances, but also by these more acute and transient effects. In a clinical context, this is important, as establishing a correct baseline of cognitive function is important in developing appropriate support strategies. If the cognitive test is sensitive to the effects of intoxication or withdrawal, this may lead to spurious findings of neurocognitive (i.e., stable) impairment when the reduced cognitive performance is due to the relatively transient intoxication or withdrawal effects. Bruijnen et al. (2019a) found no relationship between MoCA scores and days of abstinence in outpatient polysubstance users. We are unaware of other studies that have investigated the sensitivity of cognitive tests to the effects of substance use in an outpatient SUD population with varying levels of use/abstinence.

We have shown a newly developed screening tool, the Brief Executive-function Assessment Tool (BEAT), for executive impairment in a SUD population living in residential treatment facilities has acceptable psychometric properties (Berry et al., 2022). The goal of the present study was to examine the construct and criterion validity of the BEAT in an outpatient population. We also sought to investigate the extent to which the BEAT was sensitive to use/abstinence over the previous 24 hours and also over the past month. Secondarily, we examined the test operating characteristics of the BEAT.

Method

Participants

Participants were 74 outpatients from the following six clinical programs provided by the Illawarra and Shoalhaven alcohol and other drug outpatient service located in Wollongong and Nowra, Australia: assessment (n = 11), counseling (n = 26), day program (n = 2), Magistrates Early Referral into Treatment – MERIT program (n = 5), opioid substitution program (n = 27), withdrawal management (n = 3). Sample characteristics are provided in Table 1. Recruitment took place from June to September 2019. Case managers at each site were asked to invite their clients who met the eligibility criteria to participate in the study.

Table 1. Sample characteristics.

Characteristic	Mean	SD	Min	Max	n	%
Demographics
Age	41.7	10.5	18	65
Education (years)	10.6	2.4	6	18
Gender (% male)					43	58%
TOPF	93.3	14	64	124
Australian born					62	84%
Aboriginal or Torres Strait Islander					12	16%
Substance use
Age at initial substance use	13.2	2.9	5	18
Age at first alcohol consumption	14	4.4	6	43
Years of alcohol consumption	28.1	12.7	1	62
Polysubstance use					68	92%
Total substances*	4.2	1.6	1	6
1 substance					6	8%
2 substances					6	8%
3 substances					12	16%
4 substances					15	20%
5 substances					14	19%
6 substances					21	28%
Previous treatment					41	59%
Previous treatment attempts	2.2	3	0	15
Daily tobacco use					56	77%
Primary substance
Alcohol					31	42%
Methamphetamine					14	19%
Cannabis					14	19%
Heroin/opiates					13	18%
Benzodiazepines					2	3%
Substance use in previous month					41	55%
Substance use in previous 24 hours					27	36%
Relationship status
Single (never married)					40	54%
Defacto relationship					16	22%
Married					7	9%
Separated					5	7%
Divorced					5	7%
Widowed					1	1%
Living children
0					20	27%
1					16	22%
2					21	28%
3					7	9%
4					7	9%
5					2	3%
6					1	1%
Children in care under age 5					6	8%
Children in care over age 5					10	14%
Social
Employed					21	28%
Homeless in previous month					10	14%
Risk of eviction in previous month					7	9%
Arrested in previous month					7	9%
Perpetrator of violence in previous month					6	8%
Victim of violence in previous month					13	18%
Self-rated health / quality of life
ATOP psychological health	5.4	2	0	9
ATOP physical health	5.6	2	1	10
ATOP overall quality of life	5.8	1.9	2	10

Notes: TOPF: test of premorbid functioning; ATOP: Australian treatment outcome profiles.

Eligibility criteria were that the person was a current client of the service, aged between 18 and 65 years and had full BEAT data. Exclusion criteria were acute mental health phenomena, inability to provide consent, physically unable or experiencing acute withdrawal symptoms that barred participation, and insufficient English language skills to engage in the study. No other exclusion criteria were applied, to increase the ecological validity of the results.

Materials

Test of premorbid functioning (TOPF; Pearson Assessment, 2009)

The TOPF is an estimate of overall current or premorbid intelligence comprising 70 words that have atypical grapheme to phoneme translations. The examinee must read the words out loud and is credited with one point for each word read correctly. Scores are converted to standard scores with a mean of 100 and standard deviation of 15.

Brief executive-function Assessment Tool (BEAT; Berry et al., 2022)

The BEAT is a screening tool that was designed to assess cognitive functioning in people with substance use disorder. It comprises 20 items, each of which can be scored 0 to 3, giving a total score of 60. A score of 30 or less indicates cognitive impairment. It includes executive tasks such as clock drawing, figure copying, working memory, fluency and delay discounting. All of the necessary material, including a Response Form, Record Form and Stimulus Card, to administer and score the BEAT are available in the on-line Supplementary Appendices to the original BEAT validation study (Berry et al., 2022).

Alpha span Task (AS; Craik et al., 2018)

The Alpha Span Task is a measure of working memory in which participants are read a list of words and are asked to repeat the words in alphabetical order. The outcome variable was the total alpha score.

Five-Point test (FPT; Goebel et al., 2009)

The Five-Point Test assesses nonverbal idea generation and consists of producing novel designs under time constraints. The task comprises a sheet of paper with 40 five-dot matrices. Participants are asked to produce as many different figures as possible by connecting the dots within each rectangle within three minutes. The outcome variable was total number of unique designs correctly completed.

Stroop Task (AS; Golden, 1978; Golden & Freshwater, 2002)

The Stroop Task is a measure of response inhibition. Participants must respond as quickly and accurately as possible across three conditions. The first condition presents the words “red,” “blue,” and “green”; the second presents patches of colors; the third condition (interference) presents words printed in incongruent colors and requires the participant to ignore the word and say the color. The outcome variable was total score on the interference condition.

Executive function composite binary (EFCB)

The EFCB was used as the reference standard to examine criterion-related validity and to establish the test operating characteristics. It comprised the mean z-scores across the AS, FPT, and Stroop measures. A cut score of 1.5, representing 1.5 standard deviations below the mean was used to transform EFCB into a binary variable with two levels; impaired (<=1.5 SD below the mean) and intact (>1.5 SD below the mean). This variable was used in the original validation of the BEAT in a residential treatment setting. It was selected on the basis of both conceptual and statistical grounds. That is, each of the component tests are purported and validated to be tests of executive functioning, and they each correlate significantly with the BEAT, with Pearson r coefficients all being above 0.5 (Berry et al., 2022).

Australian treatment outcome profiles (ATOP; Ryan et al., 2014)

The ATOP is a structured interview that has been validated to measure treatment outcomes in Australian drug and alcohol populations. It contains two sections. Section one details the quantity and frequency of substance use. Section two details health and wellbeing variables such as days in paid work/study, homelessness, eviction risk, violence and arrest, as well as ratings of psychological health, physical health and overall quality of life.

Procedure

Five volunteer undergraduate psychology students were trained and supervised in test administration and assisted the first author in data collection for the study. Data were collected during 90-minute in-person sessions with the researchers.

Design and data analysis

The study employed a cross sectional, within-subjects design. Construct validity was investigated by examining Pearson correlations between the BEAT and established tests of executive functioning (AS, FPT, and Stroop Task). Based on previously established correlations at or above 0.5 between the BEAT and these tests (Berry et al., 2022), a one-tailed power calculation with 80% power and alpha level of .05 indicated a sample size of 23 was required. Criterion validity was investigated by examining differences in BEAT scores across impaired and unimpaired groups on the criterion measure of EFCB. Based on the effect size of 1.37 between the impaired and intact groups on a similar criterion measure from previous research, a one-tailed power calculation with 80% power and alpha level of .05 indicated a sample size of 16 was required. Test-operating characteristics were established by observing BEAT classifications across the two levels of EFCB. Independent samples t-tests were used to investigate the sensitivity of the BEAT to self-reported abstinence over both the previous 24 hours and the past month. A one-tailed power calculation with 80% power and alpha level of .05 for an effect equivalent to detecting the difference between individuals with some and no impairment (d = .65; Berry et al., 2022) indicated a sample size of 60 was required.

Results

Construct validity

As can be seen in Table 2, the BEAT correlated significantly and moderately with all tests of executive functioning.

Table 2. Pearson correlations between the BEAT and tests of executive functioning.

Test	Pearson r	p-value
Stroop	0.491	< .001
Alpha Span	0.504	< .001
Five Point Test	0.609	< .001

Criterion validity

Independent samples t-test revealed a significant difference in BEAT scores between the EFCB impaired (n = 16) and unimpaired (n = 58) groups, t(72) = 4.322, p < .001. As can be seen in Figure 1, all participants in the EFCB impaired group scored less than the established BEAT cut-score of < = 30 (Berry et al., 2022).

Figure 1. BEAT scores across the impaired and intact Executive function composite binary (EFCB) groups.

Test operating characteristics

At an optimal cut score of < = 22, the BEAT yielded 87.5% (61.7–98.4) sensitivity, 69% (55.5–80.5) specificity, 43.8% (33.7–54.4) positive predictive value (PPV), 95.2% (84.4–98.7) negative predictive value (NPV) and explained 81.8% (71.6–92.0) area under the curve. Base rate was 21.6%. It should be noted that because base rates will change across definitions of impairment and settings, positive and negative predictive values will change accordingly.

Differences in BEAT scores for 24-hour and one-month use versus abstinence

Independent samples t-tests revealed no differences in BEAT scores between the 24-hour abstinent (n = 47) and non-abstinent (n = 27) groups, t(72) = −.484, p = .63, and the one-month abstinent (n = 33) and non-abstinent (n = 41) groups, t(72) = −.046, p = .963.

Discussion

The primary purpose of the current study was to validate the BEAT for use in an outpatient SUD population. Whilst this included investigating construct and criterion validity, another important aspect of validating a cognitive test in this setting includes demonstrating that the test is sufficiently sensitive to detecting stable cognitive impairment but not so sensitive that it also detects transient cognitive changes caused by the effects of substance intoxication or withdrawal. Construct validity was established in demonstrating significant and moderate correlations between the BEAT and established tests of executive functioning. Criterion validity was also established in demonstrating that the BEAT distinguished between those considered impaired and intact on a criterion composite measure of executive functioning. The BEAT showed comparable test operating characteristics as in the original validation study (Berry et al., 2022) and was insensitive to previous 24 hour and previous month substance use versus abstinence.

The demonstrated construct and criterion validity of the BEAT is consistent with that shown in a large residential and abstinent sample (Berry et al., 2022). In fact, the correlations between the BEAT and Stroop Task and Five Point Test, respectively, were higher in the current study. This provides confidence that the BEAT measures the same executive function constructs across both inpatient and outpatient settings.

The current study also established test operating characteristics that were comparable to those obtained in the original inpatient validation study (Berry et al., 2022). The optimal cut-score in the current study was < = 22, compared with < = 30 in the original validation study, which was conducted in a residential setting. This suggests greater morbidity in an outpatient setting. Use of the original cut score of < = 30 in an outpatient setting would result in 100% sensitivity but increase the false positive rate, hence we recommend application of the cut score of < = 22 when using the BEAT with outpatients.

Despite relatively high sensitivity at detecting stable cognitive impairment, the BEAT was insensitive to self-reported substance use/abstinence over both the previous 24 hour and one month periods. If the BEAT were sensitive to the more transient effects of intoxication and withdrawal, the results would be less stable and therefore less reliable in terms of identifying stable cognitive impairment. The PPV in the current study was low, meaning there is a high probability of false positive findings. However, NPV was high, indicating that there were few false negative findings. When developing the BEAT, it was desirable to identify as many true positives as possible even though this might result in misidentifying unimpaired people as being impaired (i.e., increase the false positive rate). This is because many people without cognitive impairment, defined as scores more than 1.5 standard deviations below the mean, may still have cognitive weaknesses and would therefore benefit from interventions to enhance their functional cognition during treatment.

Few other studies have examined test operating characteristics of cognitive screening tests in outpatient SUD treatment settings. Bruijnen et al. (2019b) found that the MoCA had low sensitivity and specificity in an outpatient setting when compared to an extensive neuropsychological assessment, urging particular caution in relation to the low sensitivity of the test. Given that the purpose of screening is to identify as many people with the condition of interest as possible, it is better to trade off specificity for sensitivity (Lange & Lippa, 2017). Thus, whilst the BEAT has low PPV (44%), which was in fact lower than that of the MoCA in the Bruijnen et al. (2019b) study (64%), the NPV of the BEAT (95%) was much higher than that of the MoCA (75%; Bruijnen et al, 2019b). This means the BEAT is good at detecting impairment when it is present. We suggest that a person who scores as impaired on the BEAT be referred for comprehensive neuropsychological assessment, which can then better identify those with true impairments, i.e., reduce the number of misclassified cognitively intact people.

It should be noted, though, that there are differences across ours and the Bruijnen et al. (2019b) study, disallowing a direct comparison of BEAT and MoCA utility in outpatient settings. We used EFCB, a composite measure of executive functioning based on three well validated tests of executive functioning, whereas Bruijnen et al. (2019b) used the criterion of less than one standard deviation below the mean on two of seven cognitive domains (executive functioning, visuospatial abilities, attention, abstract reasoning, memory, orientation and processing speed); domain scores were based on average z-scores for tests within each domain. Thus, our criterion variable was more specific in that it targeted executive function impairment, and the criterion for impairment was more stringent, requiring a mean z-score of less than 1.5 standard deviations below the mean among the constituent tests. Whilst Bruijnen et al. (2019b) purported that their reference standard represented DSM-5 neurocognitive disorder, two important elements of NCD diagnosis that were missing from their operationalization of NCD was evidence of a decline from a previous level of functioning (criterion A) and whether or not the cognitive impairment interferes with independence in everyday activities (criterion B), which distinguishes between major and mild NCD. Moreover, a DSM-5 diagnosis of substance induced NCD requires evidence that the cognitive decline correlates with onset of substance use. Furthermore, people with substance use disorder have multiple comorbidities including neurodevelopmental and mental health disorders that have the potential to impact cognitive functioning, and in the absence of a clinical interview or other methods of ascertaining the temporal patterns of cognitive impairment, it is not possible to rule those factors out as potential causes for observed performance-based cognitive impairment at a single point in time (criterion D). It is suggested that future studies of test operating characteristics of cognitive screening tests in an SUD setting apply the complete DSM-5 criteria for neurocognitive disorder as a reference standard, potentially presenting results for mild and major neurocognitive disorder separately.

There were several limitations to the current study. Substance use was based on retrospective self-report of participants and we did not collect substance dosage data objectively. Inaccurate or under-reporting of substance use may have contributed to the absence of an effect for 24 hour and one month use/abstinence, as may have limited power; as there was no precedent in the literature, an effect size of .65 was used to estimate the effect for the power calculation, however, if the effect were much smaller, our analyses would have been underpowered. We also lacked statistical power to investigate single versus polysubstance use patterns. There are inherent limitations to the use of composite measures, especially when unidimensionality and statistical properties such as test-retest reliability have not been established (McKenna & Heaney, 2020; Barclay et al., 2019). Barclay et al. (2019) proposed the following requirements for composite measures: transparency, purpose-led design, technical reproducibility, and statistical fitness. Our EFCB composite measure meets the first three, but not the last, of these requirements, as we did not examine test-retest reliability and therefore cannot provide measures of uncertainty such as confidence intervals. Another limitation of the current study is that we did not have sufficient power to investigate the cognitive effects by substance type and we were also therefore unable to apply our results to the BAC curve. Given the varying typical intoxication and withdrawal durations across the substance types, varying impacts on cognition is reasonably expected with past 24 hour use.

Conclusion

In the present article, we demonstrated construct and criterion validity of the BEAT in an outpatient setting and that it had comparable test operating characteristics as in the original inpatient validation study (Berry et al., 2022). We also showed that the BEAT was insensitive to past 24 hour and past month abstinence, meaning it can be applied in an outpatient setting without concern about false positive findings due to transient effects of intoxication or withdrawal.

Acknowledgements

We would like to acknowledge the support of Sarah Adams and the staff and clients of the ISLHD.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.