Abstract
Brief Interventions (Bis) for problematic drug use in general medical settings, including in Emergency Departments (EDs), have shown disappointing results compared to those that target problematic alcohol use. Telephone booster calls may augment the impact of a BI delivered in the ED. The current study uses data from the National Drug Abuse Treatment Clinical Trials Network (CTN) Protocol 0047, “Screening, Motivational Assessment, Referral, and Treatment in Emergency Departments (SMART-ED)”, a multisite randomized clinical trial conducted in six EDs in the U.S. We examine dose effects of telephone boosters (0, 1, or 2 calls completed) with non–treatment seeking patients who we randomized to the BI-Booster condition and who endorsed problematic drug use during their ED visit (N=427). We assessed primary outcomes at 3-, 6-, and 12-month follow-ups, which included past month use of the primary drug of choice, use of any drug, and heavy drinking. There were no significant differences among those completing 0, 1, or 2 booster calls on any of the three main outcomes at 3-, 6-, and 12-months post-BI in the ED. Patients who were older were significantly more likely to complete booster calls. Taken together, these findings raise questions about the clinical utility of booster phone calls following screening and Bis targeting heterogeneous drug use in the ED.
Keywords: Brief interventions, Booster calls, Problematic drug use
1. Introduction
Brief interventions (BIs) for alcohol or drug use, delivered in emergency department (ED) or primary care settings to individuals who are not actively seeking help for drinking or drug use, have been well-studied but have shown mixed results. In a recent review of 34 ED setting trials of BIs versus a control condition for alcohol misuse, all studies reported significant decreases in alcohol use three months after ED staff delivered BI (Landy, Davey, Quintero, Pecora, & McShane, 2016). However, the relative benefit of BI was unclear: in some studies, reduced drinking was seen only in the BI condition, whereas in others both the BI and control conditions showed comparable reductions in alcohol use. Most studies did not show any between-group differences in drinking at 6- or 12-month follow-up, although some studies showed individuals in the BI condition were less likely to have a subsequent alcohol-related injury resulting from driving under the influence of alcohol.
A review of BIs for problematic drug use in general medical settings ranging from primary care to EDs revealed even more consistent but disappointing results (Saitz, 2014). Of the nine studies that we examined, only two reported decreases in drug use but had small effects or no biological outcomes to corroborate self-report. The review concluded that little evidence exists in support of BI for drug use, likely due to its greater complexity compared to alcohol misuse.
In response to the equivocal results of BIs, one recommendation for increasing BI effectiveness has been to use in-person or telephone booster sessions following the initial BI (Academic ED SBIRT Research Collaborative, 2007; Bernstein & Bernstein, 2008; Longabaugh et al., 2001). Boosters may augment the impact of a BI that is typically delivered in a general medical setting unrelated to the individual’s substance use, such as the ED or primary care.
The BI may prompt an individual to consider potential change in drinking or substance use and their associated risk behaviors, but might be insufficient to motivate the individual to create a change plan (Donovan et al., 2015). Thus, the booster functions to remind and reinforce the BI. Because the booster can occur after and outside the ED, the individual may also be less distracted and better able to focus on exploring and problem solving the barriers to change. A number of studies have now employed boosters as part of their ED-delivered Bis, including via face-to-face (Blow et al., 2017; Longabaugh et al., 2001; Mello et al., 2005), and telephone (Bernstein et al., 2009, 2010; Bogenschutz et al., 2014; Choo et al., 2016; D’Onofrio et al., 2012; Field et al., 2014; Monti et al., 2007; Soderstrom et al., 2007; Sommers et al., 2006, 2013).
A significant problem, however, has been that the effect of boosters on drinking, drug use, or risk behavior, independent of the BI, is often not tested. Only four studies have examined the impact of boosters above and beyond the Bis that they are designed to follow; three did not find that boosters were associated with significant benefit (Blow et al., 2017; Bogenschutz et al., 2014; D’Onofrio et al., 2012) and one did (Longabaugh et al., 2001). Although these studies were all conducted in EDs, they are difficult to compare due to variations in booster delivery method (face-to-face, phone); booster session length (unreported, 10-22 minutes); target population (injured hazardous drinkers, drug-using adults); and target substance use (alcohol, drugs). The nature of boosters as a BI extender remains unclear.
Another area of uncertainty is whether some individuals might engage with and benefit from boosters more than others, and relatedly, whether the number, or dose, of booster sessions relates to substance use or risk behavior outcomes. To our knowledge, no studies have examined these variables to further inform the utility of Bis. Yet these are critical questions given that booster implementation is no small feat. One study has described in detail the staff and time resources involved in conducting the booster arm of a randomized clinical trial evaluating screening, BI, and referral to treatment for problematic drug use in six EDs in the U.S. (Donovan et al., 2015).
This trial was designed to provide two telephone booster follow-ups post-BI. The authors reported that booster interventionists on average made 4.9 calls to complete the first booster and 3.7 calls to complete the second. Interventionists made an average of 11.4 calls before giving up on the first booster, and an additional 5.8 calls before giving up on the second booster. Interventionists also made an average of 4.1 additional calls to locators to help connect with the participant. The booster call interventionists did not limit their availability to business hours, but instead made efforts to accommodate participants’ schedules in the evenings and on weekends whenever possible.
Despite this labor-intensive effort, of the 425 participants randomized to the booster condition, only 56.2% completed a first booster and 37.7% completed the second. Calls were spread out over a range of 0 to 67 days for the first booster and 6 to 40 days for the second. Other reported challenges included participant variables, such as inconsistent and unreliable phone access, disinterest due to lack of monetary compensation, unstable housing, unavailability due to additional medical treatment or incarceration, and confusion around who the booster interventionist was. These results illustrate the workload involved, and persistence needed in conducting a booster intervention following a BI delivered in the ED for drug use.
This investment of resources in conducting boosters makes it imperative to understand whether the dose of booster contacts is related to substance use outcomes, and whether certain individuals are likely to engage with boosters more than others. The current study uses data from the National Drug Abuse Treatment Clinical Trials Network (CTN) Protocol 0047, “Screening, Motivational Assessment, Referral, and Treatment in Emergency Departments (SMART-ED)” (Bogenschutz et al., 2014), a multisite randomized clinical trial conducted in six EDs in the U.S. Our study is the first to examine dose effects of telephone boosters in a large-scale trial of Screening, Brief Intervention, and Referral to Treatment (SBIRT) in EDs with non–treatment seeking patients who endorse problematic drug use during their ED visit.
2. Method
2.1. Sample
This study describes the methodology for conducting brief, Motivational Interviewing (MI) interventions via booster telephone follow-up calls (0, 1, or 2 calls completed) to participants in one arm of a multisite randomized clinical trial on screening and BI with drug users in six EDs across the U.S. (Bogenschutz, et al., 2011; Donovan et al., 2012). The trial’s primary objective was to compare substance use and related outcomes among substance abusing ED patients randomized to either 1) minimal screening only (MSO); 2) screening, assessment, and referral to treatment if indicated (SAR); or 3) screening, assessment, and referral plus a BI with two telephone follow-up booster calls (BI-B). There were no statistically significant differences among groups (MSO, SAR, BI-B) in primary drug of choice, any drug use, and heavy drinking at 3-, 6-, and 12-months follow-up (Bogenschutz et al., 2014). We conducted the trial within the CTN between 2010 and 2012.
The BI-B group was the only group in the SMART-ED study that received telephone booster calls. As such, the current report focuses only on the BI-B arm of the study, evaluating the association between booster phone call dose (2 calls), substance use outcomes, and predictors of call completion for 427 adults who had received a BI in the ED. Within the BI-B group, we considered three intervention classification categories concerning the number of booster calls completed – no calls, one call, or two calls.
2.2. Booster counselors
One male and two female counselors conducted booster calls. One had a master’s degree in social work and had also worked in an ED; the other two had master’s degrees in counseling. All previously had been certified as MI practitioners and had experience conducting brief MI interventions in both clinical and research settings, with approximately 5–10 years of BI experience. All booster calls were made from the study’s centralized Booster Call Center located at the University of Washington in Seattle.
2.3. Study participant recruitment
We recruited male and female adult patients from six geographically diverse EDs across the U.S. (one each from the Southwest and Midwest, and two each from the southeast and northeast). Study research staff screened potential participants for study participation upon admission to the ED for medical treatment. Study inclusion criteria were: 1) registration as a patient in the ED during study screening hours; 2) positive screen (≥3) for problematic use of a non-alcohol, non-nicotine drug based on the Drug Abuse Screening Test (Skinner, 1982); 3) at least one day of problematic drug use (excluding alcohol and nicotine) in the past 30 days; 4) age 18+; 5) adequate English proficiency and ability to provide informed consent; and 6) access to a phone (for booster telephone sessions).
Reasons for exclusions included: 1) inability to participate due to emergency treatment; 2) significant impairment of cognition or judgment rendering the person incapable of informed consent (e.g., delirium, intoxication, traumatic brain injury); 3) status as a prisoner or in police custody at the time of treatment; 4) current engagement in addiction treatment; 5) residing more than 50 miles from the location of follow-up visits; 6) inability to provide at least two reliable locators as contacts; and 7) prior participation in the current study.
2.4. Main outcomes
The main outcomes were days of use of the patient-defined a) primary problem drug, b) any drug, and c) heavy drinking, assessed by the Time-Line Follow-back interview (TLFB; Sobell & Sobell, 1992) for the 30-day period prior to the 3-, 6-, and 12-month follow-up, as well as a similar baseline assessment. Each outcome in the analysis was a difference score calculated as outcome at each of the three follow-up time points minus baseline. Thus, negative scores indicate a decrease in substance use. We selected difference scores over other statistical approaches (e.g., mixed effects models), because distributions of the baseline and the three follow-up time points violated normality due to extreme ceiling and floor effects on all outcomes. That is, these values generally had a large proportion of zero days and a large proportion at the maximum score (30 days), and difference scores were more normally distributed.
2.5. Predictors
We examined eleven potential predictors of completing or not completing a booster call. Categorical variables included gender, ethnicity, race, marital status, usual employment in both the past 3 years and the past 30 days, relation of substance use to ED visit, and drug of choice. Cell sizes were too small to include sedatives/sleeping pills or hallucinogens as part of drug of choice categories in the analyses (see Table 1). For the variable of usual employment pattern in the past 3 years and past 30 days, the category of homemaker/unemployed consisted mostly of unemployed cases (94.32% in past 3 years and 96.97% in past 30 days) and we interpreted it as such. Continuous variables included age in years, level of education, and annual household income.
Table 1.
Baseline Characteristics Overall and by Number of Booster Calls
| Number of Booster Calls | |||||
|---|---|---|---|---|---|
| Characteristic | None | One | Two | Total Sample | |
| n = 184 | n = 77 | n = 166 | N = 427 | ||
| Gender | |||||
| Male | 129 (70.11%) | 48 (62.34%) | 124 (74.70%) | 301 (70.49%) | |
| Female | 55 (29.89%) | 29 (37.66%) | 42 (25.30%) | 126 (29.51%) | |
| Age [Mean (Std.)] | 34.90 (11.69) | 37.34 (11.87) | 37.88 (11.87) | 36.50 (11.85) | |
| Ethnicity | |||||
| Hispanic or Latino | 50 (27.17%) | 15 (19.48%) | 35 (21.08%) | 100 (23.42%) | |
| Not Hispanic or Latino | 133 (72.28%) | 61 (79.22%) | 130 (78.31%) | 324 (75.88%) | |
| Chose not to answer | 1 (0.54%) | 1 (1.30%) | 1 (0.60%) | 3 (0.70%) | |
| Race | |||||
| American Indian or Alaska Native/ Asian/Native Hawaiian or Pacific Islander/Multiracial/Other | 26 (14.13%) | 14 (18.18%) | 18 (10.84%) | 58 (13.58%) | |
| Black or African American | 60 (32.61%) | 25 (32.47%) | 59 (35.54%) | 144 (33.72%) | |
| White | 90 (48.91%) | 34 (44.16%) | 83 (50.00%) | 207 (48.48) | |
| Chose not to answer/Unknown | 8 (4.35%) | 4 (5.19%) | 6 (3.61%) | 18 (4.22%) | |
| Education Completed | |||||
| 1-11 Years | 63 (34.24%) | 24 (31.17%) | 46 (27.71%) | 133 (31.15%) | |
| GED/12 Years | 51 (27.72%) | 28 (36.36%) | 57 (34.34%) | 136 (31.85%) | |
| Some College | 52 (28.26%) | 17 (22.08%) | 41 (24.70%) | 110 (25.76%) | |
| College Degree | 16 (8.70%) | 6 (7.79%) | 14 (8.43%) | 36 (8.43%) | |
| Some Graduate | 2 (1.09%) | 1 (1.30%) | 4 (2.41%) | 7 (1.64%) | |
| Graduate Degree/Post-Graduate Degree | 0 | 1 (1.30%) | 4 (2.41%) | 5 (1.17%) | |
| Marital Status | |||||
| Married/Cohabiting, not married | 34 (18.48%) | 22 (28.57%) | 25 (15.06%) | 81 (18.97%) | |
| Widowed/Separated/Divorced | 37 (20.11%) | 16 (20.78%) | 43 (25.90%) | 96 (22.48%) | |
| Never Married | 113 (61.41%) | 39 (50.65%) | 98 (59.04%) | 250 (58.55%) | |
| Usual Employment Pattern in Past 3 years | |||||
| Full time/Part time regular | 82 (44.57%) | 33 (42.86%) | 72 (43.37%) | 187 (43.79%) | |
| Part time irregular/Student/Controlled environment/Retired or on disability | 62 (33.70%) | 29 (37.66%) | 62 (37.35%) | 153 (35.83%) | |
| Homemaker/Unemployed | 40 (21.74%) | 15 (19.48%) | 32 (19.28%) | 87 (20.37%) | |
| Usual Employment Pattern in Past 30 days | |||||
| Full time/Part time regular | 45 (24.46%) | 22 (28.57%) | 58 (34.94%) | 125 (29.27%) | |
| Part time irregular/Student/Controlled environment/Retired or on disability | 55 (29.89%) | 25 (32.47%) | 58 (34.94%) | 138 (32.32%) | |
| Homemaker/Unemployed | 84 (45.65%) | 30 (38.96%) | 50 (30.12%) | 164 (38.41%) | |
| Annual Household Income | |||||
| $0 - $15,000 | 119 (64.67%) | 46 (59.74%) | 104 (62.65%) | 269 (63.00%) | |
| $15,001 - $30,000 | 26 (14.13%) | 10 (12.99%) | 25 (15.06%) | 61 (14.29%) | |
| $30,001 - $50,000 | 16 (8.70%) | 3 (3.90%) | 12 (7.23%) | 31 (7.26%) | |
| $50,001 - $75,000 | 1 (0.54%) | 4 (5.19%) | 4 (2.41%) | 9 (2.11%) | |
| $75,001 - $100,000 | 3 (1.63%) | 0 (0.00%) | 3 (1.81%) | 6 (1.41%) | |
| $100,001+ | 0 (0.00%) | 2 (2.60%) | 3 (1.81%) | 5 (1.17%) | |
| Declined to answer | 19 (10.33%) | 12 (15.58%) | 15 (9.04%) | 46 (10.77%) | |
| Does participant think emergency room visit related to any substances used? | |||||
| Visit not at all related to substance use | 116 (63.04%) | 55 (71.43%) | 121 (72.89%) | 292 (68.38%) | |
| Substance use played a minor role | 20 (10.87%) | 10 (12.99%) | 15 (9.04%) | 45 (10.54%) | |
| Substance use played a major role/Visit happened because of substance use | 48 (26.09%) | 12 (15.58%) | 30 (18.07%) | 90 (21.08%) | |
| Drug of Choice | |||||
| Cannabis | 76 (41.30%) | 30 (38.96%) | 80 (48.48%) | 186 (43.66%) | |
| Cocaine | 49 (26.63%) | 18 (23.38%) | 46 (27.88%) | 113 (26.53%) | |
| “Street” opioids | 39 (21.20%) | 15 (19.48%) | 21 (12.73%) | 75 (17.61%) | |
| Prescription opioids | 10 (5.43%) | 7 (9.09%) | 6 (3.64%) | 23 (5.40%) | |
| Methamphetamine | 3 (1.63%) | 5 (6.49%) | 8 (4.85%) | 16 (3.76%) | |
| Sedatives or sleeping pills | 4 (2.17%) | 1 (1.30%) | 4 (2.42%) | 9 (2.11%) | |
| Hallucinogens | 3 (1.63%) | 1 (1.30%) | 0 (0.00%) | 4 (0.94%) | |
| Covariates | |||||
| DAST-10 score | |||||
| Moderate level of problems | 86 (46.74%) | 36 (46.75%) | 92 (55.42%) | 214 (50.12%) | |
| Substantial/Severe level of problems | 98 (53.26%) | 41 (53.25%) | 74 (44.58%) | 213 (49.88%) | |
| AUDIT-C score (Mean (Std.) | 5.88 (3.73) | 5.05 (3.56) | 5.39 (3.76) | 5.54 (3.72) | |
2.6. Covariates
We included five covariates in general linear models (GLMs) assessing potential differences in substance use outcomes between three groups (no calls, 1 call, or 2 calls completed):—DAST-10 (Skinner, 1982) scores, AUDIT-C (Bradley et al., 2007) scores, participant age, usual employment pattern in the past 30 days, and participants’ belief that the ED visit was related to their substance use. Because one of the eligibility criteria for inclusion in the study required a DAST-10 score of ≥ 3, scores were skewed and had very small sample sizes in some of the cells. Subsequently, we collapsed scores into a categorical variable where 1=moderate level (scores of 3–5) and 2=substantial/severe level (scores of 6–10) of problems related to drug use (Skinner, 1982) (see Table 1).
2.7. Booster calls
Detailed booster call methodology is published elsewhere (Donovan et al., 2015). We made available to booster callers substance use–related information collected as part of the initial screening in the ED, and ED counselors’ summaries of issues raised and change plans developed as part of the BI in the ED. This information helped to orient the booster callers to the unique issues of each participant who they were contacting, facilitated the introduction of the caller to the participant, and helped to frame and guide the initial booster call. To minimize potential surprise or confusion, we told participants at the end of the BI in the ED to expect a phone call from study staff within the next three days.
Trained interventionists made the telephone booster calls from a centralized, study-wide intervention booster call center. The calls were intended to “boost” the effects of the initial BI in the ED. We chose the proposed number of booster sessions (2 calls), in combination with the BI in the ED, to replicate the structure of the standard Motivational Enhancement Therapy (MET) (Miller, Zweben, Diclemente, & Rychtarik, 1992), with the goal of maximizing the magnitude of the therapeutic effect while keeping the intervention short enough to be practical. Each call was targeted to last approximately 20 minutes.
We patterned the content of the booster calls after sessions in MET (Longabaugh et al., 2001; Miller et al., 1992). The goal of the first booster call (B1) was to re-engage the participant, reinforce the change plan that originated in the ED, explore potential barriers to change, and support continuing efforts to change. The goal of the second booster call (B2) was to check-in and address barriers to treatment engagement. To provide continuity, we assigned the same booster counselor to make both calls to a given participant.
The target window for B1 was within three days of discharge from the ED, and for B2 within seven days of discharge; however, booster counselors had a 30-day window to complete both calls. Booster counselors were instructed to make any reasonable attempt to locate participants and invite them to complete the B1 and B2 sessions. Each counselor had a cell phone to make and receive calls at any time from any location. The cell phone was programmed with a toll-free phone number that participants could use to return calls to the counselor.
2.8. Analytic approach
We used a general linear model (GLM) to assess potential differences in outcomes among three groups (no calls, 1 call, or 2 calls completed). We conducted the same statistical analysis on each of three specific outcomes: Participant reported days of use of primary drug of choice, any drug, and heavy drinking; and with difference scores at 3-, 6-, and 12-month follow-ups. We used a logistic regression model to evaluate the eleven predictors of completing (1 or 2 calls) or not completing a booster call.
3. Results
3.1. Baseline characteristics
Baseline characteristics for the total BI-B sample and by number of booster call groups are provided in Table 1. Participants were mostly male (70.49%), not Hispanic or Latino (75.88%), never married (58.55%), with annual household incomes at or below $15,000 (63.00%), and with a mean age of 36.50 years (range of 18 to 72 years). The majority of participants also reported that they did not believe their ED visit was related to their substance use (68.38%). About half of participants identified as white (48.48%) and one-third as black/African American (33.72%). Almost one-third had not completed high school (31.15%) and roughly one-third had obtained a high school degree (31.85%). Regarding usual employment patterns, 43.79% held full time or part-time regular jobs in the past 3 years; in the past 30 days, 29.27% held full time or part-time regular jobs. Cannabis was the main drug of choice (43.66%), followed by cocaine (26.53%), and street opioids (17.61%). Roughly half of participants had moderate levels of drug use problems (50.12%) and half had substantial or severe problems (49.88%). Mean AUDIT-C scores were 5.54 (range of 0–12), which suggest alcohol misuse for both men and women.
3.2. Descriptive statistics of substance use outcomes
Table 2 shows descriptive statistics for the number of days in the past 30 days that participants reported use of drug of choice, any drug, and heavy drinking. We provide means and standard deviations overall as well as by number of booster calls at baseline and at 3-, 6-, and 12-months follow-up.
Table 2.
Descriptive Statistics of Substance Use Outcomes Overall and by Number of Booster Calls
| Number of Booster Calls | |||||
|---|---|---|---|---|---|
| Characteristic | Nonea | Oneb | Twoc | Total Sampled |
|
| Days Drug of Choice Use (M, SD) | |||||
| Baseline | 15.50 (11.37) | 15.25 (10.51) | 13.72 (11.34) | 14.76 (11.21) | |
| 3 Months Follow-Up | 9.59 (12.04) | 10.04 (11.80) | 8.89 (11.27) | 9.37 (11.65) | |
| 6 Months Follow-Up | 8.12 (11.25) | 9.27 (11.77) | 7.70 (10.81) | 8.15 (11.15) | |
| 12 Months Follow-Up | 8.11 (11.27) | 8.05 (10.61) | 9.22 (11.35) | 8.57 (11.17) | |
| Days Any Drug Use (M, SD) | |||||
| Baseline | 17.04 (11.11) | 16.90 (10.38) | 15.35 (11.20) | 16.35 (11.02) | |
| 3 Months Follow-Up | 12.05 (12.15) | 12.51 (11.96) | 11.28 (11.99) | 11.80 (12.03) | |
| 6 Months Follow-Up | 10.47 (11.70) | 11.00 (12.14) | 10.90 (12.34) | 10.75 (12.03) | |
| 12 Months Follow-Up | 10.61 (12.03) | 9.33 (11.09) | 11.36 (11.96) | 10.69 (11.82) | |
| Days Heavy Drinking (M, SD) | |||||
| Baseline | 4.41 (7.95) | 3.56 (7.51) | 5.42 (9.18) | 4.65 (8.39) | |
| 3 Months Follow-Up | 2.65 (6.29) | 1.91 (5.75) | 3.59 (7.44) | 2.92 (6.73) | |
| 6 Months Follow-Up | 3.01 (6.57) | 3.18 (7.26) | 3.60 (7.65) | 3.29 (7.16) | |
| 12 Months Follow-Up | 2.92 (6.24) | 1.79 (5.21) | 4.31 (8.72) | 3.30 (7.30) | |
Note.
N=180 at baseline, n=146 at 3-months, n=141 at 6-months, n=131 at 12-months follow-up.
N=77 at baseline, n=68 at 3-months, n=66 at 6-months, n=63 at 12-months follow-up.
N=165 at baseline, n=160 at 3-months, n=154 at 6-months, n=144 at 12-months follow-up.
N=422 at baseline, n=374 at 3-months, n=361 at 6-months, n=338 at 12-months follow-up.
3.3. GLMs assessing differences in substance use outcomes between 0 Calls, 1 Call, or 2 booster calls
The results of the GLMs are presented in Table 3. Of the original 427 participants in the BI-B intervention, due to the calculation of difference scores, we observed reduced sample sizes. That is, N = 373 (87.36%) at 3-month follow-up, N = 360 (84.31%) at 6-month follow-up, and N = 337 (78.92%) at 12-month follow-up. There were no statistically significant relationships (p > .05) regarding number of booster calls and change in number of days using drug of choice, any drug, or heavy drinking in the past 30 days between baseline and any of the follow-up assessments.
Table 3.
Generalized Linear Model Results: Differences in Substance Use Outcomes between 0 Calls, 1 Call, or 2 Booster Calls
| Drug of Choice | Any Drug | Heavy Drinking | ||||||
|---|---|---|---|---|---|---|---|---|
| F-value | Pr. > F | F-value | Pr. > F | F-value | Pr. > F | |||
| Baseline – 3 Month Follow-Up (N = 373) | ||||||||
| Number of booster calls (0, 1, or 2) | 0.09 | 0.914 | 0.02 | 0.980 | 0.74 | 0.477 | ||
| DAST-10 | 1.75 | 0.187 | 0.25 | 0.615 | 0.54 | 0.463 | ||
| AUDIT-C | 0.44 | 0.509 | 0.85 | 0.357 | 32.52 | <.0001 | ||
| Participant Age | 0.14 | 0.705 | 0.09 | 0.762 | 1.01 | 0.316 | ||
| Usual employment pattern past 30 days | 1.16 | 0.315 | 1.68 | 0.189 | 0.36 | 0.697 | ||
| ED visit related to substance use | 0.98 | 0.377 | 1.59 | 0.206 | 1.06 | 0.349 | ||
| Baseline – 6 Month Follow-Up (N = 360) | ||||||||
| Number of booster calls (0, 1, or 2) | 0.11 | 0.900 | 0.34 | 0.709 | 0.70 | 0.4991 | ||
| DAST-10 | 9.29 | 0.003 | 8.11 | 0.005 | 2.29 | 0.131 | ||
| AUDIT-C | 2.04 | 0.154 | 2.91 | 0.089 | 19.90 | <.0001 | ||
| Participant Age | 0.71 | 0.401 | 0.16 | 0.693 | 1.06 | 0.305 | ||
| Usual employment pattern past 30 days | 2.67 | 0.070 | 1.95 | 0.144 | 0.83 | 0.437 | ||
| ED visit related to substance use | 1.39 | 0.250 | 0.93 | 0.397 | 2.18 | 0.115 | ||
| Baseline – 12 Month Follow-Up (N = 337) | ||||||||
| Number of booster calls (0, 1, or 2) | 0.96 | 0.382 | 1.29 | 0.278 | 0.86 | 0.423 | ||
| DAST-10 | 3.75 | 0.054 | 1.57 | 0.211 | 0.04 | 0.837 | ||
| AUDIT-C | 0.04 | 0.840 | 0.67 | 0.413 | 13.59 | 0.000 | ||
| Age | 0.00 | 0.972 | 0.14 | 0.707 | 1.12 | 0.300 | ||
| Usual employment pattern past 30 days | 0.49 | 0.615 | 0.64 | 0.528 | 1.49 | 0.227 | ||
| ED visit related to substance use | 2.49 | 0.085 | 3.64 | 0.027 | 2.00 | 0.137 | ||
Note. Bold indicates significant relationship (p < .05)
Regarding the covariates, there were significant negative relationships between DAST scores and change scores in both days of drug of choice use and days of any drug use at 6-months follow-up. There were also significant negative relationships between AUDIT-C scores and change scores in heavy drinking at all follow-ups. Negative relationships indicate that greater baseline substance use scores are related to a greater decrease in substance use at follow-up (i.e., lower difference scores). Additionally, there was a negative relationship between participants’ belief that their ED visit was related to substance use and days of any drug use at 12-month follow-up. Specifically, participants who believed that substance use played a major role in their ED visit had a greater decrease in number of days of any drug use at 12-month follow-up compared to those who did not think that substance use played any role in their ED visit.
To explore these relationships further, we examined whether there were interaction effects between number of booster calls and DAST scores for drug of choice and any drug use at 6-month follow-up, between number of booster calls and AUDIT-C scores for heavy drinking at 3-, 6-, and 12-month follow-up, and between number of booster calls and beliefs that ED visits are related to substance use for any drug use at 12 month follow-up (results not shown). There were only two significant interaction effects (p < .05)—drug of choice use and any drug use at 6-month follow-up. Post-hoc analyses indicated that among participants with substantial/severe baseline drug problems who had 1 booster call, there was a significantly greater decrease in drug of choice use and any drug use compared to those who had 2 booster calls; we found no significant differences compared to those who received no booster calls. Also, we found no significant differences in reports among participants with moderate baseline drug problems in relation to number of booster calls.
3.4. Logistic regression of predictors of completing or not completing a booster call
The results of the logistic regression model are presented in Table 4. Only one of the predictor variables—age—reached statistical significance (p < .05). That is, the odds of completing at least 1 booster call increased with age. Although not statistically significant (p >.05), an examination of the Confidence Intervals (CIs) suggests that participants who believed that their substance use played a major role in their ED visit were less likely to receive a booster call than those who thought that their substance use was not at all related to their ED visit. Additionally, participants who reported that they were homemakers/unemployed compared to those who were employed full or part time in the past 30 days were less likely to receive a booster call.
Table 4.
Logistic Regression Results: Predictors of Odds of Completing a Booster Call (N = 355)
| Predictor Variables | Parameter Estimate (Standard Error) |
p- value |
Odds Ratio (95% CI) |
|---|---|---|---|
| Gender | |||
| Males (Females) | −0.00 (0.12) | 0.989 | 1.00 (0.62, 1.62) |
| Age in years | 0.03 (0.01) | 0.013 | 1.03 (1.01, 1.05) |
| Ethnicity | |||
| Hispanic or Latino (Not Hispanic or Latino) | −0.16 (0.14) | 0.241 | 0.72 (0.42, 1.25) |
| Race | |||
| Minority Race (White) | −0.04 (0.12) | 0.748 | 0.92 (0.57, 1.50) |
| Level of Education | 0.02 (0.11) | 0.874 | 1.02 (0.82, 1.27) |
| Marital Status | |||
| Widowed/Separated/Divorced (Married/Cohabiting) | −0.03 (0.20) | 0.870 | 1.02 (0.49, 2.12) |
| Never Married (Married/Cohabiting) | 0.09 (0.16) | 0.596 | 1.15 (0.63, 2.12) |
| Usual Employment Pattern in past 3 years | |||
| Part time irregular/Student/Controlled environment/Retired/Disabled (Full time/Part time regular) | 0.16 (0.20) | 0.431 | 1.45 (0.78, 2.68) |
| Homemaker/Unemployed (Full time/Part time regular) | 0.05 (0.22) | 0.813 | 1.30 (0.67, 2.54) |
| Usual Employment Pattern in past 30 days | |||
| Part time irregular/Student/Controlled environment/Retired/Disabled (Full time/Part time regular) | −0.10 (0.21) | 0.630 | 0.58 (0.28, 1.22) |
| Homemaker/Unemployed (Full time/Part time regular) | −0.33 (0.18) | 0.068 | 0.46 (0.24, 0.88) |
| Annual Household Income | 0.07 (0.14) | 0.628 | 1.07 (0.82, 1.40) |
| Does participant think ED visit related to substances used? | |||
| Substance use played a minor role (Visit not at all related to substance use) | 0.11 (0.25) | 0.669 | 0.90 (0.43, 1.89) |
| Substance use played a major role/Visit happened because of substance use (Visit not at all related to substance use) | −0.31 (0.21) | 0.139 | 0.59 (0.32, 1.09) |
| Drug of Choice | |||
| Cannabis (Street or Prescription Opioids) | −0.01 (0.18) | 0.972 | 1.21 (0.64, 2.28) |
| Cocaine/Methamphetamine (Street or Prescription Opioids) | 0.20 (0.18) | 0.263 | 1.48 (0.79, 2.79) |
Note. Reference group is listed in parentheses. Bold indicates significant relationship (p < .05)
4. Discussion
This secondary analysis of data from the SMART-ED study (Bogenschutz et al., 2014) examined 1) whether the dose of booster telephone calls is related to substance use outcomes in a multi-site sample of ED patients in six U.S. cities, and 2) whether certain individuals are likely to engage with boosters more than others based on characteristics identified at baseline. Results showed no relationships between booster call dose and any of the three primary outcomes. That is, completing 0, 1, or 2 booster calls was unrelated to participants’ drug of choice use, any drug use, or of heavy drinking at 3-, 6-, and 12-months post-BI in the ED. Additionally, there were two interaction effects among number of booster calls, level of drug use at baseline (DAST), and drug of choice and any drug use at 6-month follow-up. Results also showed that patients who were older were significantly more likely to complete booster calls.
These results build on the findings of Bogenschutz et al. (2014) and Blow et al. (2017), the only two studies to date that examined the relative benefit of a booster contact added to a BI in the ED to reduce drug use. Neither study found that adding a booster was associated with improved performance of the BI. Other studies targeting drug use with a BI plus booster have reported similar nonsignificant results (Bernstein et al., 2005, 2009; Choo et al., 2016), though these studies did not test independently the role of the booster. The current study is unique in examining the dose of booster phone calls, yet the lack of significant differences in drug use outcomes at follow-up associated with engaging in 0, 1, or 2 calls is further evidence for the uncertain benefit of booster contacts post-BI in the ED.
More specifically, although studies have suggested use of booster contacts to augment the impact of the BI, especially pertaining to alcohol use (Landy et al., 2016), the current results do not provide clear support for including boosters in SBIRT design for interventions targeting drug use in the ED. It may be that booster calls have different associations with outcomes depending on the drug of choice and level of substance use problem.
In fact, post-hoc analyses in the current study showed interaction effects among number of booster calls, level of drug use at baseline (DAST), and both drug of choice use and any drug use. Specifically, participants with substantial/severe drug use problems and 1 booster call reported a greater drop in number of days of drug use between baseline and 6-month follow-up than other participants, with the exception of those with substantial/severe drug use problems and no booster calls. On one hand, these results could suggest that a single booster call is advantageous for those with substantial/severe drug use problems. On the other hand, findings might be explained by the concept of “regression to the mean”, which suggests that extreme values at baseline tend to revert back to more normal levels at follow-up. More research is needed on Bis and optimal number of boosters in relation to drug of choice and severity of use.
The current results also showed that only age was related to completing or not completing booster calls. This is surprising, given past findings linking lower socioeconomic status and being black or Hispanic to poorer addiction treatment completion (Saloner & Lê-Cook, 2013), and well-established literature documenting links among health outcomes and race, ethnicity, socioeconomic and biological factors (Buka, 2002). Donovan and colleagues (2015), for example, illustrated how disparities in socioeconomic status and health can translate to concrete barriers to booster call completion: instability due to lack of consistent housing, reliable cell phone access, availability and freedom to take a call due to being in additional treatment or incarceration, and confusion around the purpose of the call and identity of the caller. In contrast, socioeconomic factors in the current study were not predictive of call completion.
4.1. Limitations
We should note several important limitations in this study. First, as noted, this secondary analysis examines data from the SMART-ED study, which contained a large proportion of people endorsing cannabis as their primary drug (Bogenschutz et al., 2014). The extent to which this accounted for the lack of any main intervention effects is unknown. Second, it is important to note that we collected this sample between 2010 and 2012, at an earlier time point in the opioid epidemic. Rates of opioid use may have been lower and the relationship among Bis, telephone booster dose, and people who use opioids may look different today. There is evidence that ED Bis targeting opioid misuse and dependence may be more effective than those such as in the SMART-ED trial that targeted more heterogeneous drug use (Hawk & D’Onofrio, 2018).
Third, although this study utilized secondary data from a randomized clinical trial, participants in the BI-B arm were not randomized to receive 0, 1, or 2 booster calls. Thus, findings from this study can only be interpreted in terms of association rather than causation. It is possible that the number of booster calls explains the lack of significant changes in substance use outcomes. However, it is equally likely that substance use outcomes may explain whether a participant decides to accept one or more booster calls. Fourth, other confounders of the association between telephone booster calls and drug use, which might merit further investigation (e.g., disparities in socioeconomic status, unstable living situation, cell phone access). Finally, the use of telephone booster calls rather than face-to-face interactions and follow-ups may yield different results. Therefore, it would be interesting for future research to ascertain whether and how substance use and heavy drinking might change based on different booster methods (e.g., combining calls with face-to-face interactions).
4.2. Conclusions
The parent SMART-ED study (Bogenschutz et al., 2014) had found no benefit of BI or BI plus telephone booster calls over and above minimal screening. The current secondary analysis of data from the SMART-ED study further indicates that the number of booster calls that participants completed in the BI-B condition, ranging from 0 to 2, was not associated with different drug use or heavy alcohol use outcomes. These findings raise questions about the clinical utility of booster phone calls following screening and BIs targeting heterogeneous drug use in the ED.
Highlights.
Phone booster dose effects after ED brief interventions for drug use were examined
Individual patient characteristics predicting booster completion were also examined
Main outcomes were past month drug of choice use, any drug use, and heavy drinking
Number of booster calls had no effect on main outcomes at 3-, 6-, and 12-months post-ED
Acknowledgments
The present paper is based on a component of a multisite clinical trial, Screening, Motivational Assessment, Referral, and Treatment in Emergency Departments (SMART-ED; NIDA CTN Protocol 0047), conducted within the National Institute on Drug Abuse National Drug Abuse Treatment Clinical Trials Network (CTN). Its preparation was support by grant # 5UG1DA013714, Clinical Trials Network: Pacific Northwest Node, Donovan & Hatch-Maillette, MPI. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NIDA.
Role of Funding: This work was supported by the National Institute on Drug Abuse (5UG1DA013714, Clinical Trials Network: Pacific Northwest Node, Donovan & Hatch-Maillette, MPI). The funding source had no role in the design, implementation, analysis, or description of this project.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declarations of interest: none.
References
- Academic ED SBIRT Research Collaborative. (2007). The impact of screening, brief intervention, and referral for treatment on emergency department patients’ alcohol use. Annals of Emergency Medicine, 50, 699–710. doi: 10.1016/j.annemergmed.2007.06.486 [DOI] [PubMed] [Google Scholar]
- Bernstein J, Bernstein E, Tassiopoulos K, Heeren T, Levenson S, & Hingson R (2005). Brief motivational intervention at a clinic visit reduces cocaine and heroin use. Drug and Alcohol Dependence, 77, 49–59. doi: 10.1016/j.drugalcdep.2004.07.006 [DOI] [PubMed] [Google Scholar]
- Bernstein E, & Bernstein J (2008). Effectiveness of alcohol screening and brief motivational intervention in the emergency department setting. Annals of Emergency Medicine, 51, 751–754. 10.1016/j.annemergmed.2008.01.325. [DOI] [PubMed] [Google Scholar]
- Bernstein E, Edwards E, Dorfman D, Heeren T, Bliss C, & Bernstein J (2009). Screening and brief intervention to reduce marijuana use among youth and young adults in a pediatric emergency department. Academy of Emergency Medicine, 16, 1174–1185. doi: 10.1111/j.1553-2712.2009.00490.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bernstein J, Heeren T, Edward E, Dorfman D, Bliss C, Winter M, & Bernstein E (2010). A brief motivational interview in a pediatric emergency department, plus 10-day telephone follow-up, increases attempts to quit drinking among youth and young adults who screen positive for problematic drinking. Academy of Emergency Medicine, 17, 890–902. doi: 10.1111/j.1553-2712.2010.00818.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Blow FC, Walton MA, Bohnert ASB, Ignacio RV, Chermack S, Cunningham RM, … Ilgen M (2017). A randomized controlled trial of brief interventions to reduce drug use among adults in a low-income urban emergency department: The HealthiER You study. Addiction, 112, 1395–1404. doi: 10.1111/add.13773. [DOI] [PubMed] [Google Scholar]
- Bogenschutz MP, Donovan DM, Adinoff B, Crandall C, Forcehimes AA, Lindblad R, … Walker R (2011). Design of NIDA CTN Protocol 0047: Screening, motivational assessment, referral, and treatment in emergency departments (SMART-ED). American Journal of Drug and Alcohol Abuse, 37(5), 417–425. doi: 10.3109/00952990.2011.596971. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bogenschutz MP, Donovan DM, Mandlers RN, Perl HI, Forcehimes AA, Crandall C, … Douaihy AB (2014). Brief intervention for patients with problematic drug use presenting in emergency departments: A randomized clinical trial. Journal of the American Medical Association, 174, 1736–1745. doi: 10.1001/jamainternmed.2014.4052. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bradley KA, DeBenedetti AF, Volk RJ, Williams EC, Frank D, & Kivlahan DR (2007). AUDIT-C as a brief screen for alcohol misuse in primary care. Alcoholism: Clinical and Experimental Research, 31(7), 1208–1217. doi: 10.1111/j.1530-0277.2007.00403.x [DOI] [PubMed] [Google Scholar]
- Buka SL (2002). Disparities in health status and substance use: Ethnicity and socioeconomic factors. Public Health Reports, 117(Suppl 1), S118–S125. [PMC free article] [PubMed] [Google Scholar]
- Choo EK, Tape C, Glerum KM, Mello MJ, Zlotnick C, & Guthrie KM (2016). “That’s where the arguments come in”: A qualitative analysis of booster sessions following a brief intervention for drug use and intimate partner violence in the emergency department. Substance Abuse: Research and Treatment, 10, 77–87. doi: 10.4137/SART.S33388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Donovan DM, Bogenschutz MP, Perl H, Forcehimes A, Adinoff B, Mandler R, … Walker R (2012). Study design to examine the potential role of assessment reactivity in the Screening, Motivational Assessment, Referral, and Treatment in Emergency Departments (SMART-ED) protocol. Addiction Science and Clinical Practice, 7 (1), 16 10.1186/1940-0640-7-16 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Donovan DM, Hatch-Maillette MA, Phares MM, McGarry E, Peavy M, & Taborsky J (2015). Lessons learned for follow-up booster counseling calls with substance abusing emergency department patients. Journal of Substance Abuse Treatment, 50, 67–75. 10.1016/j.sat.2014.10.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- D’Onofrio G, Fiellin DA, Pantalon MV, Chawarski MC, Owens PH, Degutis LC, … O’Connor PG (2012). A brief intervention reduces hazardous and harmful drinking in emergency department patients. Annals of Emergency Medicine, 60, 181–192. doi: 10.1016/j.annemergmed.2012.02.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Field C, Walters S, Marti CN, Jun J, Foreman M, & Brown C (2014). A multisite randomized controlled trial of brief intervention to reduce drinking in the trauma care setting: How brief is brief? Annals of Surgery, 259, 873–880. doi: 10.1097/SLA.0000000000000339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hawk K, & D'Onofrio G (2018). Emergency department screening and interventions for substance use disorders. Addiction Science & Clinical Practice, 13(1), 18. doi: 10.1186/s13722-018-0117-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Landy MSH, Davey CJ, Quintero D, Pecora A, & McShane KE (2016). A systematic review on the effectiveness of brief interventions for alcohol misuse among adults in emergency departments. Journal of Substance Abuse Treatment, 61, 1–12. http://dx.doi.Org/10.1016/j.sat.2015.08.004. [DOI] [PubMed] [Google Scholar]
- Longabaugh R, Woodlard RE, Nirenberg TD, Minugh AP, Becker B, Clifford PR, … Gogineni A (2001). Evaluating the effects of a brief motivational intervention for injured drinkers in the emergency department. Journal of Studies on Alcohol, 62, 806–816. [DOI] [PubMed] [Google Scholar]
- Mello MJ, Nirenberg TD, Longabaugh R, Woolard R, Minugh A, Becker B, … Stein L (2005). Emergency department brief motivational interventions for alcohol with motor vehicle crash patients. Annals of Emergency Medicine, 45, 620–625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miller WR, Zweben A, Diclemente CC, & Rychtarik RG (1992). Motivational enhancement therapy manual: A clinical research guide for therapists treating individuals with alcohol abuse and dependence . Vol 2 National Institute on Alcohol Abuse and Alcoholism: Rockville, MD. [Google Scholar]
- Monti PM, Barnett NP, Colby SM, Gwaltney CJ, Spirito A, Rohsenow DJ, & Woolard R (2007). Motivational interviewing vs. feedback only in emergency care for young adult problem drinking. Addiction, 102, 1234–1243. doi: 10.1111/j.1360-0443.2007.01878.x [DOI] [PubMed] [Google Scholar]
- Saitz R (2014). Screening and brief intervention for unhealthy drug use: Little or no efficacy. Frontiers in Psychiatry, 5, Article 12. 10.3389/fpsyt.2014.00121 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Saloner B, & Lê Cook B (2013). Blacks and Hispanics are less likely than whites to complete addiction treatment, largely due to socioeconomic factors. Health Affairs, 32(1), 135–45. doi: 10.1377/hlthaff.2011.0983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Skinner HA (1982). The drug abuse screening test. Addictive Behavior, 7(4), 363–371. [DOI] [PubMed] [Google Scholar]
- Sobell LC, & Sobell MB (1992). Timeline follow-back: A technique for assessing self-reported alcohol consumption In: Litten RA, & Allen JP (Eds). Measuring alcohol consumption: Psychosocial and biological methods. Totowa, NJ: Humana Press. [Google Scholar]
- Soderstrom CA, DiClemente CC, Dischinger PC, Hebel JR, McDuff DR, Auman KM, & Kufera JA (2007). A controlled trial of brief intervention versus brief advice for at-risk drinking trauma center patients. Journal of Trauma Injury, Infection, and Critical Care, 62,1102–1112. doi: 10.1097/TA.0b013e31804bdb26 [DOI] [PubMed] [Google Scholar]
- Sommers MS, Dyehouse JM, Howe SR, Fleming M, Fargo JD, & Schafer JC (2006). Effectiveness of brief interventions after alcohol-related vehicular injury: A randomized controlled trial. Journal of Trauma: Injury, Infection, and Critical Care, 61, 523–531. doi: 10.1097/01.ta.0000221756.67126.91. [DOI] [PubMed] [Google Scholar]
- Sommers MS, Lyons MS, Fargo JD, Sommers BD, McDonald CC, Shope JT, & Fleming MF (2013). Emergency department-based brief intervention to reduce risky driving and hazardous/harmful drinking in young adults: A randomized controlled trial. Alcoholism, Clinical ad Experimental Research, 37, 1753–1762. doi: 10.1111/acer.12142 [DOI] [PMC free article] [PubMed] [Google Scholar]