ADD5107: Week 10 Discussion 1: Categories of Addiction

Contents lists available at ScienceDirect

Journal of Substance Abuse Treatment

journal homepage: www.elsevier.com/locate/jsat

Changes in DSM criteria following a culturally-adapted computerized CBT for Spanish-speaking individuals with substance use disorders

Michelle A. Silva, Yudilyn Jaramillo, Manuel Paris Jr, Luis Añez-Nava, Tami L. Frankforter, Brian D. Kiluk⁎

Yale University School of Medicine, 40 Temple St (Suite 6C), New Haven, CT 06510, United States of America

A R T I C L E I N F O

Keywords: DSM-IV substance dependence CBT4CBT Clinically meaningful outcome Spanish-speakers

A B S T R A C T

This study sought to replicate and extend findings regarding change in the number of endorsed Diagnostic and Statistical Manual (DSM) criteria for substance use disorders as a meaningful outcome for clinical trials with Spanish-speakers. A secondary analysis was conducted of data from 83 treatment-seeking individuals with current DSM-IV substance dependence participating in a randomized controlled trial evaluating a culturally- adapted version of a computer-based cognitive behavioral therapy program (CBT4CBT) for Spanish-speakers. Participants were randomized to either weekly standard outpatient counseling (treatment as usual – TAU), or TAU plus access to CBT4CBT (TAU+CBT4CBT). The Structured Clinical Interview for DSM-IV (SCID-IV) was administered at baseline and at the end of the 8-week treatment period to measure change in diagnostic status and total criteria count. Frequency of substance use during treatment and throughout a 6-month follow-up period was measured by self-report using a calendar-based Timeline FollowBack method, with abstinence ver- ified through instant urine toxicology, and problem severity was measured with the Addiction Severity Index (ASI). Results of a generalized linear model with Poisson's distribution indicated significant reduction in the total count of DSM-IV dependence criteria during treatment (Wald X2 = 136.20; p < .001), and a significant in- teraction with treatment assignment (Wald X2 = 19.92, p < .001), indicating a greater reduction in endorsed criteria for those assigned to TAU+CBT4CBT compared to TAU only. Total criteria count and diagnostic status at end-of-treatment was significantly correlated with substance use outcomes during the follow-up period, such that fewer criteria endorsed were associated with greater rates of abstinence and lower problem severity. These findings paralleled the primary outcomes from the main trial, and replicated prior findings in English-speakers regarding the utility of DSM criteria count as a potential clinically meaningful outcome.

1. Introduction

Primary outcomes from substance use disorder (SUD) trials typically include the frequency of drug/alcohol use during and following treat- ment as evidence of a treatment's efficacy. However, measures of the frequency and/or quantity of substance use are considered merely surrogate endpoints for demonstrating the clinical benefit of a given treatment; that is, they do not indicate clinically meaningful improve- ment in isolation (Winchell, Rappaport, Roca, & Rosebraugh, 2012). There has been greater interest recently toward establishing whether reductions in substance use also translate into meaningful improvement in other areas, such as psychosocial functioning or physical con- sequences (Kiluk et al., 2016; Kiluk, Fitzmaurice, Strain, & Weiss, 2019; Knox et al., 2019; Witkiewitz et al., 2017). One potentially promising avenue for evaluating meaningful improvement from treatment is

through measuring changes in the diagnostic criteria for the disorder being treated.

Although the aim of most treatments for substance use is to modify patients' drug use behavior, the desired effect is improvement in the physical and psychosocial consequences that define the disorder (Winchell et al., 2012). The diagnostic features of substance use dis- orders include an emphasis on the functional consequences of use, such as interference with work/school, interpersonal problems, and impair- ments in physical or psychological health. Yet it is unclear whether these functional consequences improve in accordance with reductions in substance use (Kiluk et al., 2019; (McLellan et al., 1981)). Many clinical trials evaluating treatment for substance use require partici- pants to have a Diagnostic and Statistical Manual (DSM) substance use disorder diagnosis for inclusion in the trial, yet few, if any, report whether participants continue to meet diagnostic criteria at the end of

https://doi.org/10.1016/j.jsat.2019.12.006 Received 22 April 2019; Received in revised form 6 December 2019; Accepted 12 December 2019

⁎ Corresponding author. E-mail address: [email protected] (B.D. Kiluk).

Journal of Substance Abuse Treatment 110 (2020) 42–48

0740-5472/ © 2019 Elsevier Inc. All rights reserved.

T

treatment. This is surprising given that DSM diagnosis is a more direct measure of an individual's functioning with respect to substance-related problems, and changes in diagnostic severity or status (i.e., meeting a given threshold for diagnosis) for other psychiatric disorders are com- monly evaluated in treatment trials. Rather than evaluate treatment efficacy based solely on changes in substance use behavior, why not also measure changes in the severity of the clinical diagnosis?

In a recent secondary analysis of a randomized controlled trial evaluating the efficacy of a web-based Cognitive Behavioral Therapy program (CBT4CBT) for alcohol use disorders (Kiluk et al., 2016), we found a significant reduction in the number of DSM-5 criteria and se- verity category from baseline to end-of-treatment, which was associated with better outcomes during a follow-up period, thereby offering sup- port for change in DSM criteria count and severity level as a potential meaningful outcome (Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018). The utility of evaluating change in DSM criteria was further demonstrated in a recent observational study in Spain, whereby the levels of SUD were deemed sensitive to the impact of treatment (both inpatient and outpatient), as measured by the difference between a baseline assessment and 3-month follow-up (Dacosta-Sanchez, Fernandez-Calderon, Gonzalez-Ponce, Diaz-Batanero, & Lozano, 2019). However, there have been no studies evaluating change in DSM criteria among Spanish-speakers participating in a randomized controlled trial (RCT) evaluating treatment for SUD.

Given the diversity of individuals presenting for substance use treatment in community settings, it is important to replicate findings regarding potential measures of treatment benefit (i.e., endpoints for clinical trials) for SUDs in different populations to establish external validity/generalizability of findings. National-level data from the latest Treatment Episode Data Set (TEDS) indicated the racial/ethnic com- position of substance use treatment admissions differed from that of the U.S. population, with a greater percentage of treatment admissions represented by racial/ethnic minorities compared to the general po- pulation (Substance Abuse and Mental Health Services Administration, 2017). Individuals of Hispanic origin comprised roughly 15% (280,551) of all treatment admissions in 2017. There is also clear evidence of racial/ethnic disparities in terms of SUD treatment access, duration, and completion, with most studies indicating Hispanic/Latino samples are less likely to seek treatment, receive fewer services, and less likely to complete treatment than non-Hispanic Whites (Amaro, Arévalo, Gonzalez, Szapocznik, & Iguchi, 2006; Guerrero, Marsh, Khachikian, Amaro, & Vega, 2013; Mulvaney-Day, DeAngelo, Chen, Cook, & Alegría, 2012). Multiple literature reviews have indicated a need for greater research on SUD treatment outcomes among Hispanic/Latino populations (Amaro et al., 2006; Guerrero et al., 2013), which includes emphasis on treatment efficacy with respect to abstinence and other functional outcomes.

A recent RCT demonstrated the benefits of a cultural and linguistic adaptation of a web-based CBT4CBT for Spanish-speaking individuals at reducing rates of substance use when added to standard outpatient treatment (Paris et al., 2018). However, it is unknown whether the benefits in terms of reduced rates of substance use extend to other meaningful improvement. The purpose of the current study was to re- plicate and extend findings of Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018 regarding change in DSM criteria count through a sec- ondary analysis of data from this RCT (Paris et al., 2018), in which DSM-IV criteria for substance dependence was assessed at multiple time points. Specifically, we sought to examine whether the number of en- dorsed DSM criteria changed from baseline to end-of-treatment, whe- ther there was differential change according to treatment groups, and whether the DSM criteria count or diagnostic status at the end-of- treatment was associated with substance use outcomes during follow- up.

2. Material and methods

This is a secondary analysis of data from an RCT evaluating the efficacy of a culturally-adapted web-based CBT program for Spanish- speakers (CBT4CBT-Spanish) delivered as an add-on to standard out- patient addiction treatment in a heterogeneous population of treat- ment-seeking Latino adults. Full details of the trial have been published elsewhere (Paris et al., 2018), but are summarized below.

2.1. Participants

Participants were recruited from outpatient treatment centers of- fering services to Spanish-speaking individuals seeking treatment for substance use. Inclusion criteria were: (1) age 18 or older; (2) Spanish as preferred and primary language; and (3) current (past 30 days) DSM- IV criteria for abuse or dependence on alcohol or other drugs. Participants were excluded if they had an untreated bipolar or psy- chotic disorder or were otherwise not sufficiently stable for outpatient treatment.

2.2. Treatments

Participants were randomized to one of two treatment conditions delivered over the course of 8 weeks: (1) treatment as usual (TAU) at the outpatient facility, or (2) TAU plus CBT4CBT. TAU consisted of supportive counseling via weekly group or individual sessions, as well as access to other ancillary services as needed. Those assigned to CBT4CBT, in addition to TAU, were provided access to a culturally- adapted version of a web-based CBT program (CBT4CBT) for reducing substance use that has been validated in multiple trials with English speakers (Carroll et al., 2008; Carroll et al., 2014; Kiluk et al., 2018; Kiluk, Devore, et al., 2016). Briefly, the CBT4CBT-Spanish program retained the 7 modules/sessions focused on teaching the core CBT skills and strategies for recognizing, avoiding, and coping with triggers for substance use as in the English-language program, but now delivered entirely in Spanish with content adapted to focus on the integration of cultural values in storyline and character development (Paris et al., 2018). For this adaptation, a telenovela format was used as the skill- teaching platform to facilitate engagement with the program. Also La- tino values of respeto (respect), confianza (trust), machismo, caballerismo, and marianismo (gender-specific values), familismo (family orientation), fatalismo (fatalismo), sabiduría (wisdom), and personalismo (value on interpersonal relationships), were incorporated into development of the characters due to their relevance among diverse Latino subgroups and anticipated influence on behavior change (Anez, Paris Jr., Bedregal, Davidson, & Grilo, 2005; Anez, Silva, Paris, & Bedregal, 2008).

Participants were provided access to the CBT4CBT-Spanish program on a dedicated computer/laptop in a private space within the out- patient facility and asked to complete one module/session per week in addition to attending their weekly TAU session. Each CBT4CBT module took approximately 35–40 min to complete; participants completed on average 5 out of the 7 modules (Paris et al., 2018).

2.3. Assessments

Assessments were administered by a bilingual research assistant and included those with Spanish translations already available and vali- dated in similar populations. For DSM diagnosis, the Structured Clinical Interview for DSM-IV (SCID-IV; Torrens, Serrano, Astals, Perez- Dominguez, & Martin-Santos, 2004) was administered at study screening (i.e., baseline), end-of-treatment (8 weeks), and 6-month follow-up to determine the presence of current (past 30 days) substance abuse/dependence. Although DSM-5 was published (American Psychiatric Association, 2013) at the time this trial was initiated in January 2014, the accompanying SCID-5 was not yet available. While DSM-5 combined criteria for abuse and dependence into one

M.A. Silva, et al. Journal of Substance Abuse Treatment 110 (2020) 42–48

43

unidimensional construct (Hasin et al., 2013), DSM-IV stipulated that abuse should not be diagnosed when dependence was present (American Psychiatric Association, 1994). Therefore, we are not able to combine endorsed criteria for abuse and dependence in this study to determine an overall criteria count consistent with DSM-5 (i.e., parti- cipants met criteria for one or the other for a given substance). Criteria counts and diagnostic threshold for this study will be based on only those for substance dependence according to DSM-IV. The Addiction Severity Index (ASI; Butler, Redondo, Fernandez, & Villapiano, 2009) was also administered to assess problem severity across a range of do- mains including medical, legal, employment, social/family, psychiatric, as well as alcohol and drugs. The Substance Use Calendar, similar to the Timeline Follow Back (Robinson, Sobell, Sobell, & Leo, 2014) was ad- ministered at each research visit to collect day-by-day self-reports of drug and alcohol use during the entire study period (from the 28-day period prior to randomization, throughout the 56-day treatment phase as well as 6-month follow-up). We verified self-reported drug use/ab- stinence through instant urine toxicology screens (Yes/No), as well as breathalyzer screens at each visit. Self-report and urine results matched in 94% of urine specimens collected.

2.4. Data analysis

All analyses were conducted using SPSS version 24 (IBM Corp., Armonk, NY). We use non-parametric tests when variables were not normally distributed. Due to the potential for meeting diagnostic cri- teria for multiple substances, participants were asked to identify their “primary substance” at the time of diagnostic interview, which was used as the primary outcome indicator in the clinical trial, although data regarding any alcohol or drug use were collected throughout the trial. The presence of DSM-IV criteria for substance dependence for a participant's primary substance was determined by the SCID-IV and summed to produce a total criteria count for each participant (max- imum of 7 total criteria) at baseline and week 8 (end-of-treatment). Rates of participants meeting criteria for substance dependence (yes/ no), as well as total criteria counts for dependence on their primary substance, at baseline were compared across treatment groups (TAU vs. TAU+CBT4CBT) using Chi-square, and Mann-Whitney U test. Spearman's rho correlations were used to explore the association be- tween DSM-IV dependence total criteria count for primary substance at baseline and other indicators of substance use severity, such as days of primary drug use in 28 days prior to baseline, number of years of pri- mary drug use, age of first primary drug use, number of days attended treatment in 28 days prior to baseline, and ASI composite scores.

To explore change in total criteria count from baseline to end-of- treatment by treatment group, we used a generalized linear model with Poisson's distribution (to accommodate count data) with an un- structured covariance matrix. This model uses all available data; missing data from the SCID were minimal (data at end-of-treatment were available on 92% of intention to treat sample). Chi-square was used to explore whether there was differential change in diagnostic status from baseline to end-of-treatment by treatment group. For this analysis, a dichotomous variable was created indicating whether par- ticipants continued to meet diagnostic threshold for dependence on their primary substance at end-of-treatment or not (which includes those only meeting threshold for abuse). To evaluate the sensitivity to treatment effects, findings regarding change in total criteria count and diagnostic status were compared with those from the main study report regarding primary substance use outcomes (Paris et al., 2018). Lastly, we use Spearman's rho correlations (or Mann-Whitney U) to explore whether DSM-IV substance dependence criteria count (or diagnostic threshold status) at end-of-treatment was associated with substance outcomes during the 6-month follow-up period, such as the percentage of days abstinent from primary drug, the percentage of days abstinent from all drugs and alcohol, and ASI composite scores across drug and non-drug domains.

3. Results

3.1. Participants

From a total of 92 participants randomized in the clinical trial, 83 met DSM-IV criteria for substance dependence at baseline (90% of randomized sample) and were used in subsequent analyses. Rates of participants meeting criteria for substance dependence (yes) did not differ across treatment group (TAU n = 44 of 49, 90%; TAU+CBT4CBT n = 39 of 43, 91%; Χ2 (3, 92) = 0.02, p = .89), nor did total substance dependence criteria endorsed at baseline (TAU: Mdn = 5.5, TAU +CBT4CBT: Mdn = 6.0; U = 805.0, n1 = 44, n2 = 39, p = .62). Demographics and baseline clinical characteristics did not significantly differ across treatment group in this subsample (see Table 1), which is consistent with the larger sample reported in the main paper (Paris et al., 2018). This subsample were mostly male (66%), unemployed (66%), single/unmarried (75%), with less than a high school education (58%). Mean age of the sample was 43 years old (SD = 11.5).

In terms of substance use, 34% reported their primary substance was marijuana, 35% reported alcohol, and 27% reported cocaine, with the remainder reporting opioids (2%), benzodiazepines (1%), or other (1%). On average, participants reported using their primary substance on 12 days (SD = 10.0) out of the past 28 prior to baseline, with a

Table 1 Baseline characteristics by group.

Categorical variables TAU + CBT4CBT n = 39

TAU only n = 44

X2 p

n (%) n (%)

Female 11 (28.2) 17 (38.6) 1.01 0.32 Place of birth Puerto Rico 28 (71.8) 31 (70.5) 1.48 0.92 United States mainland 2 (5.1) 2 (4.5) South America 1 (2.6) 1 (2.3) Mexico 4 (10.3) 4 (9.1) Central America 2 (5.1) 5 (11.4) Other 2 (5.1) 1 (2.3)

Completed high school, yes 15 (38.5) 20 (45.5) 0.42 0.52 Never married/living alone 28 (71.8) 34 (77.3) 0.33 0.57 Unemployed 25 (64.1) 30 (68.2) 0.15 0.70 Referred by criminal justice system 5 (12.8) 4 (9.1) 0.30 0.59 Primary substance used, self-

reported Alcohol 13 (33.3) 16 (36.4) 2.16 0.83 Cocaine 11 (28.2) 11 (25.0) Marijuana 13 (33.3) 15 (34.1) Opiates 1 (2.6) 1 (2.3) Benzodiazepenes 0 (0) 1 (2.3) Other 1 (2.6) 0 (0)

Continuous variables TAU + CBT4CBT n = 39

TAU only n = 44

F p

mean (SD) mean (SD)

Age, years 42.0 (11.8) 43.4 (11.2) 0.30 0.59 Years living in United States

mainland 15.8 (10.0) 19.0 (13.3) 1.49 0.23

Days of primary substance use, past 28

10.0 (9.3) 13.7 (10.3) 2.85 0.10

Age first used primary substance 19.1 (8.9) 19.2 (9.7) 0.00 0.98 Years of primary substance use 20.4 (12.1) 22.9 (11.6) 0.96 0.33 ASI Medical composite 0.44 (0.36) 0.59 (0.36) 3.39 0.06 ASI Employment composite 0.88 (0.18) 0.86 (0.19) 0.19 0.67 ASI Alcohol composite 0.24 (0.23) 0.21 (0.26) 0.19 0.66 ASI Drug composite 0.14 (0.10) 0.13 (0.11) 0.44 0.51 ASI Legal composite 0.12 (0.21) 0.09 (0.17) 0.44 0.51 ASI Family composite 0.17 (0.24) 0.23 (0.26) 1.13 0.29 ASI Psych composite 0.51 (0.26) 0.53 (0.22) 0.29 0.59 Days in tx 28 days prior 5.2 (10.3) 8.2 (11.9) 1.53 0.22

M.A. Silva, et al. Journal of Substance Abuse Treatment 110 (2020) 42–48

44

history of approximately 22 years (SD = 11.9) of primary substance use, and attended 7 days (SD = 11.3) of treatment in the past month. Sixty percent reported using more than one substance.

3.2. DSM criteria endorsed and association with substance use at baseline

Rates of participants endorsing each of the seven criteria for DSM-IV substance dependence at baseline did not differ across treatment group. Internal consistency for the 7 dependence criteria in the full sample was good (Cronbach's alpha = 0.84). The most commonly endorsed criteria in the sample (N = 83) were related to impaired control, including “using for a longer period of time than intended, or using larger amounts than intended” (n = 78; 94%), “persistent desire or unsuccessful efforts to cut down or control substance use” (n = 77; 93%), and “great deal of time spent in activities necessary to obtain, use, or recover from effects” (n = 65; 78%). The next most endorsed criteria were related to pharmacological indicators of dependence, including “tolerance” (n = 66; 80%), and “withdrawal” (n = 61; 74%). The remaining criteria, in order of en- dorsement were “continued substance use despite knowledge of persistent or recurrent physical or psychological problems caused or exacerbated by substance” (n = 55; 66%), and “important social, occupational, recrea- tional activities given up or reduced because of substance use” (n = 49; 59%).

Results indicated DSM-IV dependence criteria count at baseline was significantly associated with the days of self-reported primary drug use in the 28 days prior to baseline (r = −0.27, p = .02), which suggests a greater number of DSM criteria endorsed was associated with fewer days of self-reported primary drug use at baseline. Correlations did not reveal significant associations between DSM criteria count at baseline and the number of years of primary drug use (r = −0.03), or the age of first primary drug use (r = 0.17). However, DSM criteria count was significantly correlated with the number of days attended treatment in past month (r = 0.28, p = .01), suggesting a greater number of en- dorsed criteria was associated with a greater number of treatment days in the past month.

In terms of the relationship between DSM-IV dependence criteria counts and ASI composite scores at baseline, results indicated a sig- nificant correlation with the ASI alcohol composite (r = 0.37, p < .001), suggesting a greater number of DSM criteria endorsed was associated with greater alcohol problem severity. However, there were no other significant correlations between DSM criteria counts and ASI composites scores for other domains at baseline (r's ranged from 0.03 to 0.16).

3.3. Change over time

Of the 83 participants meeting DSM-IV criteria for dependence at baseline, 76 completed the SCID interview at end-of-treatment. Results of generalized linear models with Poisson distribution indicated a sig- nificant change in DSM-IV dependence criteria count over time: Wald X2 (df = 1) = 136.20; p < .001, and a significant interaction of time by treatment group: Wald X2 (df = 1) = 19.92, p < .001. This in- dicated a significant reduction in total criteria count overall, as well as a greater reduction for those assigned to TAU+CBT4CBT compared to TAU (See Fig. 1), which is consistent with the main study findings re- garding change in frequency of primary substance used. See Appendix Table 1 for statistical output, including exponentiated beta weights from the generalized linear model.

In terms of the change in proportion of those continuing to meet diagnostic threshold for DSM-IV substance dependence, results of chi- square tests did not reveal a significant difference by treatment condi- tion at week 8: X2 (3, 76) = 1.91, p = .17. Among those meeting de- pendence criteria at baseline, 71% assigned to TAU+CBT4CBT no longer met criteria at week 8, compared to 56% of those assigned to TAU. Chi-square tests evaluating differences in rates of endorsement for each DSM-IV dependence criteria across treatment group at week 8

indicated three of the seven criteria were endorsed more frequently for those assigned to TAU compared to TAU+CBT4CBT: “using for a longer period of time than intended, or using larger amounts than intended” (56% vs. 26%), X2 (3, 76) = 7.15, p = .007; “withdrawal” (44% vs. 17%), X2

(3, 76) = 6.26, p = .01; and “tolerance” (44% vs. 20%), X2 (3, 76) = 4.89, p= .03. Rates of endorsement for the other criteria were in a similar pattern indicating more frequent endorsement for those as- signed to TAU compared to TAU+CBT4CBT, but were not significantly different at the p < .05 level: “persistent desire or unsuccessful efforts to cut down or control substance use” (68% vs. 51%), X2 (3, 76) = 2.25, p = .13; “important social, occupational, recreational activities given up or reduced because of substance use” (17% vs. 6%), X2 (3, 76) = 2.33, p = .13; “continued substance use despite knowledge of persistent or re- current physical or psychological problems caused or exacerbated by sub- stance” (34% vs. 20%), X2 (3, 76) = 1.89, p = .17; and “great deal of time spent in activities necessary to obtain, use, or recover from effects” (37% vs. 29%), X2 (3, 76) = 0.55, p = .46.

3.4. Association with outcomes during follow-up

Results of correlations between DSM-IV dependence criteria count at week 8 and substance use outcomes (n = 65) and ASI composite scores (n = 68) during a 6-month follow-up period indicated several significant relationships. Criteria count was significantly correlated with the percentage of days abstinent from primary drug during follow- up (r = −0.35, p = .003) and correlated at trend-level with the per- centage of days abstinent from all alcohol and drugs during follow-up (r = −0.24, p = .06); both indicated a greater number of dependence criteria endorsed at week 8 was associated with less abstinence during follow-up. In terms of correlations with ASI composite scores at 6- month follow-up, criteria count was significantly correlated with ASI legal composite (r = 0.28, p = .02), and the ASI drug composite score at trend-level significance (r = 0.22, p = .07), indicating a greater number of criteria endorsed at week 8 was associated with greater legal and drug problem severity 6-months later.

Results evaluating the difference in follow-up substance use out- comes and ASI composite scores according to whether individuals met diagnostic threshold for DSM-IV dependence at week 8 are displayed in Table 2. Compared to those meeting diagnostic threshold for DSM-IV dependence at week 8, those not meeting diagnostic threshold reported a greater percentage of days abstinent from their primary drug (Mdn: 95% vs. 72%), as well as from all alcohol and drugs during the follow- up period (Mdn: 86% vs. 66%). Also, those not meeting diagnostic threshold reported less legal problem severity at 6-month follow-up (at trend-level significance), as measured by the ASI legal composite score, than those meeting diagnostic threshold (Legal Composite Mdn:<0.01 vs. 0.01, respectively).

Fig. 1. Change in DSM-IV criteria count.

M.A. Silva, et al. Journal of Substance Abuse Treatment 110 (2020) 42–48

45

4. Discussion

This study explored whether the number of endorsed DSM-IV sub- stance dependence criteria changed from baseline to end-of-treatment, whether there was differential change according to treatment group, and whether criteria count at end-of-treatment was associated with follow-up outcomes in a Spanish-speaking sample enrolled in a rando- mized controlled trial. A significant reduction in DSM-IV substance dependence criteria count from baseline to end-of-treatment found here in the sample overall, as well as greater reduction for those assigned to TAU+CBT4CBT compared to TAU, consistent with findings from the main trial indicating an overall reduction in frequency of substance use and greater reduction for those assigned to TAU+CBT4CBT (Paris et al., 2018). Furthermore, both a continuous indicator of total criteria count, as well as a binary indicator of meeting diagnostic threshold, at end-of-treatment was associated with rates of substance use during a subsequent 6-month follow-up period, which provides support as meaningful outcome indicators. Overall, results replicated and ex- tended prior findings in English-speakers regarding the utility of DSM diagnosis criteria count as a potential clinically meaningful outcome (Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018).

There has been significant interest in the evaluation and establish- ment of an alternative to sustained abstinence as a clinically meaningful outcome for use in RCTs of treatments for substance use disorders (Hasin et al., 2017; Kiluk, Carroll, et al., 2016; Volkow et al., 2018), and change in DSM-5 criteria counts or severity level may be a promising indicator (Dacosta-Sanchez et al., 2019; Kiluk et al., 2019; Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018; Nielsen, 2019). Evalu- ating change in DSM-5 criteria to detect change in the severity of the disorder may provide a more direct measure of clinical benefit than change in drug use frequency. It could also be used to inform clinical actions using a measurement-based care approach. In this approach, clinicians monitor the signs and symptoms of the disorder to provide feedback about progress in treatment and direct clinical decisions (Marsden et al., 2019). Monitoring acute changes in DSM-5 criteria for

substance use disorders in combination with changes in frequency of substance use may provide a more comprehensive indicator to guide clinical decision-making and detect signs of early remission or wor- sening severity.

DSM has consistently been used across large U.S. epidemiological studies (e.g., National Co-morbidity-Survey – Replication; the National Latino and Asian American Study; the Hispanic Americans Baseline Alcohol Survey; and the National Epidemiological Survey on Alcohol and Related Conditions) and smaller-scale studies to establish substance use severity among Latino subgroups (Caetano & Schafer, 1996; Villalobos & Bridges, 2018), yet there are no known reports evaluating change over time within an RCT in this population. The findings pre- sented here not only replicated those supporting the potential utility of examining change in criteria count for alcohol use disorder (Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018), but extended the sup- port to include change in criteria counts for various substances of abuse, as well as within an entirely Spanish-speaking Latino sample. Further- more, these results were consistent with findings regarding differential change in primary substance use by treatment condition from the main trial (Paris et al., 2018), which was not apparent in the English- speaking study (Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018). Although these findings were based on DSM-IV substance dependence criteria counts, DSM-5 includes all seven of the dependence criteria, as well as a measure of severity based on criteria count (Hasin et al., 2013), suggesting these results are applicable to the current diagnostic system. Replication of findings in different samples and settings is im- portant for establishing the validity of this outcome.

Despite the notable findings regarding the significant reduction in dependence criteria count over time and relationship with follow-up outcomes, the findings regarding the limited relationship with in- dicators of substance use severity at baseline were puzzling. The number of endorsed DSM-IV dependence criteria at baseline was not correlated with the ASI drug composite, nor years of primary drug use or age of first use, which are classic indicators of drug problem severity one would expect to be related to dependence criteria. Also, the sig- nificant negative correlation between DSM-IV dependence criteria and days of primary drug use at baseline was surprising, as the relationship was in the opposite direction than anticipated (i.e., greater number of endorsed criteria associated with fewer days of drug use). The reason for this finding is unknown.

One explanation may be that some participants with more frequent substance use failed to recognize or acknowledge the negative impact of their substance use and thus denied most of the dependence symptom criteria (i.e., those with high rates of substance use endorsed fewer criteria). An alternative explanation may be related to the study in- clusion criterion regarding current (past 30 days) substance abuse or dependence, yet no criterion regarding days of substance use. This is relevant because some individuals may have significantly reduced their substance use from prior months to a current pattern of intermittent use, yet still endorsed multiple dependence criteria at the time of study enrollment (i.e., those with relatively low rates of substance use en- dorsed a high number of criteria). Participants reported an average of 12 days of primary drug use in the 28 days prior to baseline and en- dorsed an average of 5 out of 7 DSM-IV substance dependence criteria, so this was not necessarily a sample with low level substance use se- verity. However, the limited variability in substance dependence cri- teria count at baseline compared to level of variability in days of drug use may also be a contributing factor. Lastly, there may be variation in the underlying latent severity associated with a substance use disorder (dependence in this case) or in the severity of specific symptoms used to assess a given DSM criterion that may not be captured by simple criteria counts (Boness, Lane, & Sher, 2019). In principle, someone may have a low total criteria count, but the relative severity of the endorsed symptoms (and associated drug use), may be high (Lane & Sher, 2015; Lane, Steinley, & Sher, 2016). Measurement of severity within each DSM criterion, rather than a sum of criteria counts, may provide a more

Table 2 Follow-up outcomes according to DSM-IV dependence threshold at week 8.

Outcome during 6-month follow-up

Met DSM-IV substance dependence threshold at week 8

U

No (n = 43) Yes (n = 22)

Median Median

% days abstinent from primary drug

94.6 72.3 258.0⁎⁎

% days abstinent from all alcohol and drugs

86.3 66.1 323.0⁎

Met DSM-IV substance dependence threshold at week 8

U

No (n = 44) Yes (n = 24)

Median Median

ASI medical composite 0.51 0.48 503.5 ASI employment

composite 0.95 0.94 509.5

ASI alcohol composite 0.02 0.03 507.0 ASI drug composite 0.07 0.14 404.5 ASI legal composite < 0.01 0.01 441.0±

ASI family/social composite

0.02 0.01 444.0

ASI psychiatric composite 0.43 0.48 426.5

± p < .10. ⁎ p < .05. ⁎⁎ p < .01.

M.A. Silva, et al. Journal of Substance Abuse Treatment 110 (2020) 42–48

46

fine-grained assessment of substance use disorder severity. Interestingly, there were relatively few significant correlations be-

tween DSM-IV dependence criteria counts and ASI composite scores. At baseline, no significant correlations were present for dependence cri- teria counts and ASI composite scores for the domains of medical, employment, legal, family/social, psychiatric, or drug problem severity. This may be reflective of the notion that substance users may have problems in various life domains that are not directly related to their substance use (Borders et al., 2009; Kiluk et al., 2019; (McLellan et al., 1981); McLellan & Bartlett, 2007). Of note, the items on the ASI that contribute to the composite scores for the non-drug and alcohol do- mains, do not distinguish between problems related to substance use and those that are not. Therefore, while the ASI is a valuable instrument for evaluating problem severity across a number of life domains that may be relevant for treatment planning (McLellan, Cacciola, Alterman, Rikoon, & Carise, 2006), it is not a proxy for substance use related consequences. The lack of relationship between DSM-IV dependence criteria count and ASI composite scores at baseline should not be as- sumed to reflect low construct validity. When examining end-of-treat- ment criteria count and ASI composite scores at the 6-month follow-up period, there was a significant correlation with the ASI legal composite, and trend-level significance with the drug composite, which suggested lower criteria count at end-of-treatment was associated with less legal and drug-related problems 6-months later.

This study is not without limitations. First and foremost, the criteria count was based on DSM-IV substance dependence only, which includes a maximum of 7 total criteria, rather than DSM-5 substance use dis- order, which includes 11 criteria. Although data taken from a nationally representative sample of U.S. adults demonstrated that there is agree- ment between DSM-IV and DSM-5 substance use diagnoses, the best match for DSM-IV Alcohol, Cocaine, and Opioid Dependence occurred when four or more DSM-5 criteria were endorsed at the moderate level, and six or more DSM-5 criteria at the severe level for match with DSM- IV Cannabis Dependence (Compton, Dawson, Goldstein, & Grant, 2013). Additionally, the sample size here is relatively small, with 83 treatment seeking individuals, and there was some loss of data when examining substance use outcomes during the 6-month follow-up period (78% of the 83 participants had substance use data collected

during follow-up). Despite this, a greater reduction in endorsed criteria for those assigned to TAU+CBT4CBT compared to TAU only was de- tected, which paralleled primary outcomes in the main trial (Paris et al., 2018). Lastly, this trial included ‘current’ DSM-IV diagnosis based on symptom criteria endorsed in the past 30 days versus the DSM-5 defi- nition for ‘current’ disorder, which is based on the presence of criteria endorsed over the past 90 days. This may have served to artificially inflate the amount of change in symptoms.

Despite the limitations, this is the first study to report on the change in the total count of endorsed DSM criteria for substance use disorders, and whether individuals continue to meet diagnostic threshold, within a randomized controlled trial for Spanish-speaking populations. The results replicated and extended prior findings in English-speakers (Kiluk, Frankforter, Cusumano, Nich, & Carroll, 2018), and suggest that a total count of endorsed DSM criteria may be a useful indicator of treatment outcome in both English and Spanish-speaking populations. This is an important step toward the acceptance of non-abstinence- based outcome measures in the context of treatment efficacy trials in multiple populations.

Sources of funding

This work was supported by the National Institute on Drug Abuse [grants R01 DA030369, P50 DA09241] and the National Institute on Alcohol Abuse and Alcoholism [grant R01 AA024122]. The funding sources had no involvement in the study design, collection, analysis, or interpretation of data, in the writing of the report, or in the decision to submit for publication.

CRediT authorship contribution statement

Michelle A. Silva:Conceptualization, Writing - original draft, Writing - review & editing.Yudilyn Jaramillo:Project administra- tion, Writing - review & editing.Manuel Paris:Funding acquisition, Writing - review & editing.Luis Añez-Nava:Writing - review & editing.Tami L. Frankforter:Data curation, Formal analysis, Writing - review & editing.Brian D. Kiluk:Conceptualization, Writing - ori- ginal draft, Writing - review & editing.

Appendix A

Appendix Table 1 Results of generalized linear model.

Parameter Beta weight Std. Error 95% Wald confidence interval Wald Chi-Square df p Exp (B) 95% Wald Confidence Interval for Exp (B)

Lower Upper Lower Upper

Intercept 1.671 0.0654 1.543 1.799 653.487 1 0.000 5.318 4.679 6.045 Tx Group 0.053 0.0947 −0.133 0.238 0.308 1 0.579 1.054 0.875 1.269 Tx Week −0.810 0.1208 −1.047 −0.573 44.991 1 0.000 0.445 0.351 0.564 Tx Group ∗ Tx Week −1.003 0.2248 −1.444 −0.563 19.923 1 0.000 0.367 0.236 0.570

Group 2 (TAU only) is the reference group.

References

Amaro, H., Arévalo, S., Gonzalez, G., Szapocznik, J., & Iguchi, M. Y. (2006). Needs and scientific opportunities for research on substance abuse treatment among Hispanic adults. Drug and Alcohol Dependence, 84, S64–S75. https://doi.org/10.1016/j. drugalcdep.2006.05.008.

American Psychiatric Association (1994). Diagnostic and statistical manual of mental dis- orders (4th ed.). Washington DC: APA Press.

American Psychiatric Association (2013). Diagnostic and statistical manual of mental dis- orders (5th ed.). Arlington, VA: American Psychiatric Association.

Anez, L. M., Paris, M., Jr., Bedregal, L. E., Davidson, L., & Grilo, C. M. (2005). Application of cultural constructs in the care of first generation Latino clients in a community mental health setting. Journal of Psychiatric Practice, 11(4), 221–230.

Anez, L. M., Silva, M. A., Paris, M., & Bedregal, L. (2008). Engaging Latinos through the

integration of cultural values and motivational interviewing principles. Professional Psychology: Research and Practice, 39(2), 153–159.

Boness, C. L., Lane, S. P., & Sher, K. J. (2019). Not all alcohol use disorder criteria are equally severe: Toward severity grading of individual criteria in college drinkers. Psychology of Addictive Behaviors, 33(1), 35–49. https://doi.org/10.1037/ adb0000443.

Borders, T. F., Booth, B. M., Falck, R. S., Leukefeld, C., Wang, J., & Carlson, R. G. (2009). Longitudinal changes in drug use severity and physical health-related quality of life among untreated stimulant users. Addictive Behaviors, 34(11), 959–964.

Butler, S. F., Redondo, J. P., Fernandez, K. C., & Villapiano, A. (2009). Validation of the Spanish Addiction Severity Index Multimedia Version (S-ASI-MV). Drug and Alcohol Dependence, 99(1–3), 18–27. https://doi.org/10.1016/j.drugalcdep.2008.06.012.

Caetano, R., & Schafer, J. (1996). DSM-IV alcohol dependence in a treatment sample of white, black, and Mexican-American men. Alcoholism, Clinical and Experimental Research, 20(2), 384–390.

M.A. Silva, et al. Journal of Substance Abuse Treatment 110 (2020) 42–48

47

Carroll, K. M., Ball, S. A., Martino, S., Nich, C., Babuscio, T., Gordon, M. A., ... Rounsaville, B. J. (2008). Computer-assisted cognitive-behavioral therapy for ad- diction. A randomized clinical trial of “CBT4CBT”. American Journal of Psychiatry, 165(7), 881–888.

Carroll, K. M., Kiluk, B. D., Nich, C., Gordon, M. A., Portnoy, G., Marino, D., & Ball, S. A. (2014). Computer-assisted delivery of cognitive-behavioral therapy: Efficacy and durability of CBT4CBT among cocaine-dependent individuals maintained on me- thadone. American Journal of Psychiatry, 171, 436–444.

Compton, W. M., Dawson, D. A., Goldstein, R. B., & Grant, B. F. (2013). Crosswalk be- tween DSM-IV dependence and DSM-5 substance use disorders for opioids, cannabis, cocaine and alcohol. Drug and Alcohol Dependence, 132(1–2), 387–390. https://doi. org/10.1016/j.drugalcdep.2013.02.036.

Dacosta-Sanchez, D., Fernandez-Calderon, F., Gonzalez-Ponce, B., Diaz-Batanero, C., & Lozano, O. M. (2019). Severity of substance use disorder: Utility as an outcome in clinical settings. Alcoholism, Clinical and Experimental Research. https://doi.org/10. 1111/acer.14020.

Guerrero, E. G., Marsh, J. C., Khachikian, T., Amaro, H., & Vega, W. A. (2013). Disparities in Latino substance use, service use, and treatment: Implications for culturally and evidence-based interventions under health care reform. Drug and Alcohol Dependence, 133(3), 805–813. https://doi.org/10.1016/j.drugalcdep.2013.07.027.

Hasin, D. S., O’Brien, C. P., Auriacombe, M., Borges, G., Bucholz, K., Budney, A., ... Grant, B. F. (2013). DSM-5 criteria for substance use disorders: Recommendations and ra- tionale. The American Journal of Psychiatry, 170(8), 834–851. https://doi.org/10. 1176/appi.ajp.2013.12060782.

Hasin, D. S., Wall, M., Witkiewitz, K., Kranzler, H. R., Falk, D., Litten, R., ... Anton, R. (2017). Change in non-abstinent WHO drinking risk levels and alcohol dependence: A 3 year follow-up study in the US general population. Lancet Psychiatry, 4(6), 469–476. https://doi.org/10.1016/s2215-0366(17)30130-x.

Kiluk, B. D., Carroll, K. M., Duhig, A., Falk, D. E., Kampman, K., Lai, S., ... Strain, E. C. (2016). Measures of outcome for stimulant trials: ACTTION recommendations and research agenda. Drug and Alcohol Dependence, 158, 1–7. https://doi.org/10.1016/j. drugalcdep.2015.11.004.

Kiluk, B. D., Devore, K. A., Buck, M. B., Nich, C., Frankforter, T. L., LaPaglia, D. M., ... Carroll, K. M. (2016). Randomized trial of computerized cognitive behavioral therapy for alcohol use disorders: Efficacy as a virtual stand-alone and treatment add-on compared with standard outpatient treatment. Alcoholism, Clinical and Experimental Research, 40(9), 1991–2000. https://doi.org/10.1111/acer.13162.

Kiluk, B. D., Fitzmaurice, G. M., Strain, E. C., & Weiss, R. D. (2019). What defines a clinically meaningful outcome in the treatment of substance use disorders: Reductions in direct consequences of drug use or improvement in overall func- tioning? Addiction, 114(1), 9–15. https://doi.org/10.1111/add.14289.

Kiluk, B. D., Frankforter, T. L., Cusumano, M., Nich, C., & Carroll, K. M. (2018). Change in DSM-5 alcohol use disorder criteria count and severity level as a treatment outcome indicator: Results from a randomized trial. Alcoholism, Clinical and Experimental Research. https://doi.org/10.1111/acer.13807.

Kiluk, B. D., Nich, C., Buck, M. B., Devore, K. A., Frankforter, T. L., LaPaglia, D. M., ... Carroll, K. M. (2018). Randomized clinical trial of computerized and clinician-de- livered CBT in comparison with standard outpatient treatment for substance use disorders: Primary within-treatment and follow-up outcomes. The American Journal of Psychiatry, 175(9), 853–863. https://doi.org/10.1176/appi.ajp.2018.17090978.

Knox, J., Scodes, J., Wall, M., Witkiewitz, K., Kranzler, H. R., Falk, D., ... Hasin, D. S. (2019). Reduction in non-abstinent WHO drinking risk levels and depression/anxiety disorders: 3-year follow-up results in the US general population. Drug and Alcohol Dependence, 197, 228–235. https://doi.org/10.1016/j.drugalcdep.2019.01.009.

Lane, S. P., & Sher, K. J. (2015). Limits of current approaches to diagnosis severity based

on criterion counts: An example with DSM-5 alcohol use disorder. Clinical Psychological Science: A Journal of the Association for Psychological Science, 3(6), 819–835. https://doi.org/10.1177/2167702614553026.

Lane, S. P., Steinley, D., & Sher, K. J. (2016). Meta-analysis of DSM alcohol use disorder criteria severities: Structural consistency is only “skin deep”. Psychological Medicine, 46(8), 1769–1784. https://doi.org/10.1017/S0033291716000404.

Marsden, J., Tai, B., Ali, R., Hu, L., Rush, A. J., & Volkow, N. (2019). Measurement-based care using DSM-5 for opioid use disorder: Can we make opioid medication treatment more effective? Addiction, 114(8), 1346–1353. https://doi.org/10.1111/add.14546.

McLellan, A. T., & Bartlett, J. (2007). Outcome performance, and quality–What’s the difference. Journal of Substance Abuse Treatment, 32, 331–340.

McLellan, A. T., Cacciola, J. C., Alterman, A. I., Rikoon, S. H., & Carise, C. (2006). The addiction severity index at 25: Origins, contributions and transitions. The American Journal on Addictions, 15(2), 113–124. https://doi.org/10.1080/ 10550490500528316.

McLellan, A. T., Luborsky, L., Woody, G. E., O’Brien, C. P., & Kron, R. (1981). Are the addiction-related problems of substance abusers really related? Journal of Nervous and Mental Disease, 169, 232–239.

Mulvaney-Day, N., DeAngelo, D., Chen, C.-N., Cook, B. L., & Alegría, M. (2012). Unmet need for treatment for substance use disorders across race and ethnicity. Drug and Alcohol Dependence, 125 Suppl 1(Suppl. 1), S44–S50. https://doi.org/10.1016/j. drugalcdep.2012.05.005.

Nielsen, A. S. (2019). Don’t ask apple trees to give pears. Addiction, 114(1), 16–17. https://doi.org/10.1111/add.14395.

Paris, M., Silva, M., Anez-Nava, L., Jaramillo, Y., Kiluk, B. D., Gordon, M. A., ... Carroll, K. M. (2018). Culturally adapted, web-based cognitive behavioral therapy for Spanish- speaking individuals with substance use disorders: A randomized clinical trial. American Journal of Public Health, 108(11), 1535–1542. https://doi.org/10.2105/ ajph.2018.304571.

Robinson, S. M., Sobell, L. C., Sobell, M. B., & Leo, G. I. (2014). Reliability of the timeline Followback for cocaine, cannabis, and cigarette use. Psychology of Addictive Behaviors, 28(1), 154–162. https://doi.org/10.1037/a0030992.

Substance Abuse and Mental Health Services Administration (2017). Treatment episode data set (TEDS): 2017. Admissions to and discharges from publicly-funded substance use treatment. Rockville, MD: Substance Abuse and Mental Health Services Administration.

Torrens, M., Serrano, D., Astals, M., Perez-Dominguez, G., & Martin-Santos, R. (2004). Diagnosing comorbid psychiatric disorders in substance abusers: Validity of the Spanish versions of the psychiatric research interview for substance and mental disorders and the structured clinical interview for DSM-IV. The American Journal of Psychiatry, 161(7), 1231–1237. https://doi.org/10.1176/appi.ajp.161.7.1231.

Villalobos, B. T., & Bridges, A. J. (2018). Prevalence of substance use disorders among Latinos in the United States: An empirical review update. Journal of Latina/o psy- chology, 6(3), 204–219. https://doi.org/10.1037/lat0000097.

Volkow, N. D., Woodcock, J., Compton, W. M., Throckmorton, D. C., Skolnick, P., Hertz, S., & Wargo, E. M. (2018). Medication development in opioid addiction: Meaningful clinical end points. Science Translational Medicine, 10(434), https://doi.org/10.1126/ scitranslmed.aan2595.

Winchell, C., Rappaport, B. A., Roca, R., & Rosebraugh, C. J. (2012). Reanalysis of me- thamphetamine dependence treatment trial. CNS Neuroscience & Therapeutics, 18(5), 367–368.

Witkiewitz, K., Roos, C. R., Pearson, M. R., Hallgren, K. A., Maisto, S. A., Kirouac, M., ... Heather, N. (2017). How much is too much? Patterns of drinking during alcohol treatment and associations with post-treatment outcomes across three alcohol clinical trials. Journal of Studies on Alcohol and Drugs, 78(1), 59–69.

M.A. Silva, et al. Journal of Substance Abuse Treatment 110 (2020) 42–48

48