Supplementary Online Content - JAMA€¦ · It incorporated elements of MI,3 motivational enhancement therapy,4 and brief ... session (that could be conducted by phone if the participant
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
eMethods. Intervention details Both the Adaptation of motivational interviewing (MOTIV) manual and the brief negotiated interview (BNI) manual are available from the authors on request. Brief negotiated interview (BNI)
Training. Training (120 hours initially then 1-2 booster sessions a year) was didactic and experiential (with direct observation) with competency demonstrated by video recordings scored for content using a standardized form.1 Interventionists had two 90-minute group and one or two 1-hour individual meetings a month with a Master’s trained, licensed independent clinical social worker supervisor to discuss job performance and cases, and for direct observation.
Description. The BNI began with becoming acquainted (e.g. tell me a bit about yourself). To enhance motivation for change the counselor next reviewed “pros and cons” of drug use and used reflective listening to develop discrepancy between participant values and actions. Feedback of screening results was followed by discussion of any perceived risks and discussion of readiness and confidence for change (“readiness ruler”). Finally, goals (e.g. cut down, quit, seek further help) and next steps were negotiated, and a change plan generated and discussed, which could include a recommendation to discuss with their physician. Adaptation of motivational interviewing (MOTIV)
Training. Didactic and experiential training (40 hours) was followed by practice towards, and confirmation of proficiency using audio recordings coded for motivational interviewing content. Counselors, who were clinical psychology doctoral students, had weekly 1-hour meetings to discuss issues surrounding the implementation of the intervention and to review with their supervisor (licensed doctoral level faculty clinical psychologist) audio recordings facilitated by a Motivational Interviewing(MI) assessment instrument.2 Description. The MOTIV was less structured than the BNI, with the interchange guided by patient responses. It incorporated elements of MI,3 motivational enhancement therapy,4 and brief intervention for drug use.5 The main goals of the intervention were to enhance motivation and self-efficacy to change drug use and, if relevant, help the participant choose drug-related behavior change goals and increase commitment to them. Although elements varied by patient, objectives generally included developing rapport, increasing an understanding of the link between drug use and participant health concerns, helping the patient become aware of how drug use related to values and life goals, identifying strengths that could support change, supporting and affirming participants’ beliefs and abilities related to behavior change, and providing treatment choices and options to facilitate change.
The MOTIV was designed to incorporate relevant patient health information and structured assessment information into motivational interviewing. The MOTIV utilized the key principles of MI such as expressing empathy, rolling with resistance, developing discrepancy, and supporting self-efficacy.3 MOTIV was collaborative—the patients’ views are respected, valued and integrated into efforts to promote change. The counselor focused on providing the patient a sense of autonomy by emphasizing their control, choice and decision-making options. The interview began with an open-ended question to elicit the patient’s perspective on his/her drug use. Strategies in the interview included reflective listening, open-ended questions, and decisional balance to develop the discrepancy between important goals/values and drug use. Through these and other strategies, the interventionist sought to elicit and elaborate on change talk from patients (i.e., patient statements reflecting personal reasons and advantages of changing drug use). The counselor was also able to use assessment information provided by the patient to facilitate these objectives including items from the ASSIST (The Alcohol, Smoking and Substance Involvement Screening Test), readiness and confidence for change, and drug use consequences experienced by the participant (SIP-D - The Short Inventory of Problems—Modified for Drug Use). Where relevant, the counselor also provided personalized information and feedback about medical conditions that may be influenced by drug use based on electronic medical record review of emergency visits, hospitalizations, problem lists, medications and visit notes. If appropriate, change plans were developed, including the change goal (cut down, quit, treatment referral), reasons for change, the specific plan, potential barriers and identification of those who could help with change plan efforts. In order to facilitate referral, those who identified outpatient counseling as a preferred option were provided with a specific contact person (e.g., outpatient counselor) who was called at the end of the interview. Electronic medical record documentation was supplemented with a “flag” to the primary care physician that also included patient reasons for change, barriers, the
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
change plan and any scheduled follow-up. All patients were offered a follow-up 20-30 minute counseling session (that could be conducted by phone if the participant preferred) that was to be completed over the subsequent 2 weeks. Intervention fidelity
There were 14 BNI and 4 MOTIV interventionists who performed 175 BNI and 177 MOTIV interventions. MOTIV counselors performed 8, 42, 42 and 85 interventions each. BNI counselors performed between 2 and 65 interventions each (7 interventionists performed between 2 and 4 interventions; the remaining counselors performed 8-65 sessions). All audible intervention recordings from both conditions were coded by at least one of three trained coders (master’s and doctoral level psychology students) using both an established measure of motivational interviewing fidelity and a study specific measure--the Motivational Interviewing Treatment Integrity (MITI 3.1.1) code system and a 9-item coding system developed for this study (ASPIRE codes, available from author TP on request), respectively. For each intervention, the complete recording was coded rather than specific segments. Twenty percent of the recordings were coded by at least 2 coders. To prevent drift, coders met weekly with a psychologist to review key issues in coding and review reliability recordings.
The MITI contains a set of 5 global codes rated on a 1 (low) to 5 (high) scale and a number of therapist behavior categories that are represented by counts (e.g., number of open questions, reflections). Summary scores are also calculated to determine MI-consistent behavior (e.g., open-ended questions/total questions).6 Motivational interviewing proficiency is typically determined by thresholds associated with global coder ratings and summary scores from behavioral counts. The ASPIRE codes included a list of common and unique intervention elements that were rated on a scale of 1 (not implemented/implemented incorrectly) to 5 (implemented in an ideal/exceptional manner). The rating scale designated a score of 3 as performance to standard, 2 as successful implementation of an element but performance below standard, and 4 as performance above standard.
Inter-rater reliability for the key MITI codes were in the good to excellent range according to Cicchetti’s (1994) criteria of intraclass correlation coefficients (ICCs)(i.e., below 0.40 as poor, 0.40-0.59 as fair, 0.60-0.74 as good and above 0.75 as excellent).7 For the ASPIRE codes, all ICCs were > 0.60 (see eTable 4 for complete set of ratings).
To evaluate treatment fidelity, we used relevant codes from the MITI and ASPIRE coding systems. The interventions were adaptations of motivational interviewing in that they were structured, single session brief interventions delivered in a primary care context; the content of the protocol differed between BNI and MOTIV, as described above. However, both were constructed to be delivered using an MI style. Thus, we expected that the MI Spirit variable (i.e., mean of evocation, autonomy, collaboration) and the percent MI adherent variable (pMIA) would be appropriate indices of treatment adherence along with the intervention specific ASPIRE code items developed for this study. (See eTable 5 for the complete list of ratings).
Examination of global codes from the MITI suggested that MOTIV interventions were delivered consistent with an MI style as indicated by MI Spirit composite, mean (M) standard deviation (SD)=4.56 (0.46) and percent MI adherent scores (pMIA) of 0.96 (0.13). Using commonly used MITI thresholds for proficiency, 98% of the MOTIV interventions showed adequate MI Spirit and 90% were proficient on pMIA. As anticipated, the BNI condition showed lower MI Spirit scores M (SD) = 3.58 (0.71) and had a lower rate of MI adherence in the interviews as indicated by pMIA scores, M (SD) = 0.70 (0.26). Indeed only, 64% attained proficiency for MI Spirit and 28% for pMIA.
Other elements of treatment fidelity were calculated based on the study specific ASPIRE codes. eTable 5 shows mean (SD) ratings for each code by intervention condition and percent of interventions that met thresholds for implementation (score of 2 or greater) and implementation to standard (score of 3 or greater) by intervention condition. For the MOTIV group, mean ratings for the relevant ASPIRE codes ranged from 2.88 (1.39) to 3.97 (1.08) [median = 3.56]. As anticipated, the BNI showed lower rates of treatment fidelity as mean scores ranged from 2.21 (1.19) to 3.01 (1.08) [median = 2.51]. Intervention Discriminability
There were a number of structural features (e.g., time, provider training, supervision) that discriminated between the two interventions delivered. Although the interventions shared an emphasis on motivational interviewing style, the MOTIV had a stronger emphasis on motivational interviewing and had fewer structured elements. As shown in eTable 5, motivational interviewing style (as indicated by
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
MI-Spirit and pMIA) was significantly higher for MOTIV interventions compared with BNI. On the study specific coding instrument, items discriminated between the interventions as anticipated. The BNI interventions showed higher scores on the use of the readiness ruler while the MOTIV scored higher on use of medical record assessment, affirmation, and developing discrepancy. Although the change plan was specifically part of the BNI intervention, it was an optional component of the MOTIV and consequently did not differ between groups.
All Participants
The websites provided to participants were www.alcoholscreening.org and www.drugscreening.org.
References
1. Boston University Brief Negotiated Interview-Active Referral to Treatment Institute. Tools for the brief negotiated interview (BNI). http://www.bu.edu/bniart/sbirt-in-health-care/sbirt-educational-materials/sbirt-brief-intervention/. Accessed February 18, 2014.
2. Martino S, Ball S, Gallon S, et al. Motivational interviewing supervisory tool for enhancing proficiency. Salem, OR: Northwest Frontier Addiction Technology Transfer Center, Oregon Health and Science University; 2006.
3. Miller WR, Rollnick S. Motivational interviewing: Preparing people for change. New York, NY: Guilford press; 2002.
4. Miller WR. Motivational enhancement therapy with drug abusers. Albuquerque, NM: Department of Psychology and Center on Alcoholism, Substance Abuse, and Addictions, The University of New Mexico; 1995.
5. Henry-Edwards S, Humeniuk R, Ali R, Poznyak V. The alcohol, smoking and substance involvement screening test (ASSIST): guidelines for use in primary care (draft version 1.1 for field testing). Updated 2003. http://www.who.int/substance_abuse/activities/en/Draft_The_ASSIST_Guidelines.pdf. Accessed February 18, 2014.
6. Moyers TB, Martin T, Manuel JK, Hendrickson SM, Miller WR. Assessing competence in the use of motivational interviewing. J Subst Abuse Treat. 2005;28(1):19-26.
7. Ciccetti DV. Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment. 1994;6(4):284-290.
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
eTable 1. Effects of brief intervention on drug use, drug use consequences, sex and drug risk behaviors, healthcare utilization, mutual help group attendance and hair drug test results at 6-month follow-up
BNIa MOTIVb
Control
BNI vs. Control MOTIV vs. Control
Predicted Valuesc
Negative Binomial Regression Analyses
n
Means
IRR (95% CI)
p-valued
IRR (95% CI)
p-valued
Days >1 time use main druge in last 30 days
516 8.6 8.5 7.2 1.20
(0.86,1.66) 0.31
1.19 (0.86,1.65)
0.31
Days use main drug in last 90 days
516 33.5 38.4 36.0 0.93
(0.72, 1.20) 0.61
1.07 (0.83, 1.37)
0.61
Days > 1 time use of main drug in last 90 days
516 26.5 26.9 23.0 1.15
(0.80, 1.65) 0.45
1.17 (0.81, 1.68)
0.45
Drug use consequences (SIP-Df)
513 5.8 6.8 6.1 0.95
(0.71,1.26) 0.71
1.11 (0.83,1.47)
0.71
Number of unsafe sexual encounters in past 3 months
434 10.4 10.3 10.2 1.02
(0.63,1.65) 0.97
1.01 (0.62,1.63)
0.97
Median Regression Analyses
n
Median
β (95% CI)
p-value
β (95% CI)
p-value
Main drug ASSISTg score
516 16.4 17.5 16.0 0.41
(-1.73,2.54) 0.71
1.50 (-
0.81,3.81) 0.41
Global ASSIST score
516 21.7 23.4 22.7 -1.00
(-3.62,1.62) 0.50
0.73 (-
1.41,2.87) 0.50
90th Quantile Regression Analysesh,i
n
90th percentile
β (95% CI)
p-value
β (95% CI)
p-value
Marijuana ASSIST score
516 21.4 23.5 22.2 -0.79
(-4.18, 2.61)0.65
1.29 (-2.59, 5.16)
0.65
Cocaine ASSIST score
516 11.8 11.8 14.7 -2.89
(-6.47, 0.70)0.12
-2.89 (-6.51, 0.74)
0.12
Opioid ASSIST score
516 10.2 10.2 10.2 0 (0, 0) 1.00 0 (0, 0) 1.00
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
a BNI-Brief negotiated interview. b MOTIV-Adaptation of motivational interviewing. c Model-based expected values, unless otherwise indicated, calculated for population with observed means and proportions of the overall sample. d All p-values were adjusted for multiple comparisons using the Hochberg procedure. e Main drug - Drug of most concern, as determined by the participant. f SIP-D - Short Inventory of Problems—Alcohol and Drugs modified for Drug Use (Ranges from 0-45, higher score means worse consequences). gASSIST - The Alcohol, Smoking and Substance Involvement Screening Test (substance-specific scores range from 0-39, global scores range from 0-273; higher scores means riskier use/greater severity). h 90th quantile was modeled due to a highly skewed distribution with excess zero’s. i Quantile regression was not possible for the other substance specific ASSIST scores (sedative, amphetamines, hallucinogens, and inhalants) due to large number of excess zero’s, all p-values were >0.3 in unadjusted analyses. j These are observed proportions not adjusted for covariates. k We were not able to do adjusted analyses in these models due to low numbers of events. l All of the utilization questions (including mutual help group) asked about visits since study entry, which was assessed at the 6-month interview.
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
eTable 2 – Secondary outcomes, stratified by main drug and severity, at 6-month follow-up: effects of brief intervention in primary care patients with drug use identified by screening
BNIa MOTIVb Control BNI vs. Control MOTIV vs. Control
a BNI-Brief negotiated interview. b MOTIV-Adaptation of motivational interviewing. c p-values adjusted for multiple comparisons using the Hochberg procedure. d Main drug - Drug of most concern, as evaluated by subject. e All models adjusted for main drug (unless stratified by main drug), CIDI-SF drug dependence, and outpatient addiction or mental health treatment during the past 3 months, and the baseline value of the outcome.f ASSIST - The Alcohol, Smoking and Substance Involvement Screening Test. g SIP-D - Short Inventory of Problems—Alcohol and Drugs modified for Drug Use. h Quantile regression was not possible for the substance specific ASSIST scores for opioid, sedative, amphetamines, hallucinogens, and inhalants for which all median values were 0 and all p-values were >0.3 in unadjusted analyses. i These are observed proportions not adjusted for covariates. j Stratification for logistic regression analyses by main drug and by ASSIST were not done. k Unadjusted – We were not able to do adjusted analyses in these models due to low numbers. *Risky alcohol=>3 (>4 for men) standard drinks in any one day
Stratified by severity (ASSIST)
ASSIST < 27
Unadjusted 424 20
(14.0) 21
(14.5) 14
(10.3) 1.42
(0.68, 2.93) 0.35
1.48 (0.72, 3.03)
0.35
Adjustede 423 1.48
(0.59, 3.83) 0.64
1.34 (0.53, 3.50)
0.64
ASSIST >= 27
Unadjusted 93 14
(53.9) 10
(35.7) 14
(35.9) 2.08
(0.76, 5.73) 0.31
0.99 (0.36, 2.73)
0.99
Adjustede 93 2.02
(0.52, 8.51) 0.79
0.74 (0.18, 2.96)
0.87
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
eTable 3 – Secondary outcomes stratified by main drug and severity at 6-week follow-up: effects of brief intervention in primary care patients with drug use identified by screening
BNIa MOTIVb Control BNI vs. Control MOTIV vs. Control
Negative Binomial Regression Analyses
N Means IRR
(95% CI) p-
valc IRR
(95% CI) p-
valc Days use main drugd in last 30 days
Unadjusted 525 12.17 12.48 12.15 1.00
(0.76,1.32) 0.99 1.03(0.78,1.35) 0.99
Adjustede 524 7.81 9.42 8.92 0.88
(0.71,1.08) 0.44 1.06(0.85,1.31) 0.62
Stratified by main drug Opioids
Unadjusted 90 7.00 7.89 6.61 1.06
(0.42,2.64) 0.90 1.19(0.47,3.05) 0.90
Adjustede 90 3.67 6.68 6.05 0.61
(0.26,1.43) 0.51 1.11(0.48,2.55) 0.81
Cocaine
Unadjusted 97 5.71 4.73 5.33 1.07
(0.54,2.12) 0.85 0.89(0.45,1.74) 0.85
Adjustede 97 3.62 3.79 4.47 0.81
(0.46,1.44) 0.57 0.85(0.48,1.49) 0.57
Marijuana
Unadjusted 329 15.73 16.53 15.80 1.00
(0.76,1.31) 0.97 1.05(0.80,1.37) 0.97
Adjustede 328 12.13 14.12 12.57 0.97
(0.78,1.19) 0.74 1.12(0.91,1.39) 0.57
Stratified by severity (ASSIST)f
ASSIST < 27
Unadjusted 428 12.22 12.69 12.73 0.96
(0.71,1.29) 0.98 1.00(0.74,1.34) 0.98
Adjustede 427 8.14 9.36 9.55 0.85
(0.68,1.07) 0.34 0.98(0.78,1.23) 0.86
ASSIST >= 27
Unadjusted 97 11.97 11.39 10.15 1.18
(0.58,2.42) 0.75 1.12(0.54,2.32) 0.75
Adjustede 97 5.47 9.37 5.58 0.98
(0.57,1.68) 0.94 1.68(0.96,2.94) 0.14
Days use of main drug >1 time in last 30 days
Unadjusted 525 8.35 8.34 8.80 0.95
(0.63,1.42) 0.80 0.95(0.64,1.42) 0.80
Adjustede 524 5.01 5.15 5.42 0.92
(0.66,1.29) 0.77 0.95(0.68,1.34) 0.77
Stratified by main drug Opioids
Unadjusted 90 3.71 3.39 2.35 1.58
(0.41,6.11) 0.61 1.44(0.36,5.80) 0.61
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
a BNI-Brief negotiated interview. b MOTIV-Adaptation of motivational interviewing. c all p-values adjusted for multiple comparisons using the Hochberg procedure. d Main drug - Drug of most concern, as determined by the subject e All models adjusted for drug of most concern (DOMC)(unless stratified by DOMC), CIDI-SF drug dependence, and outpatient addiction or mental health treatment during the past 3 months, and the baseline value of the outcome.f ASSIST - The Alcohol, Smoking and Substance Involvement Screening Test. g SIP-D - Short Inventory of Problems—modified for Drug Use. h Stratification for logistic regression analyses by DOMC and by ASSIST were only done for Inpatient or outpatient treatment for addiction or mental health in past 6 weeks and mutual help group in the past 6 weeks.
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
a BNI-Brief negotiated interview. b MOTIV-Adaptation of motivational interviewing c p-values adjusted for multiple comparisons using the Hochberg procedure. d Adjusted on baseline covariates: Drug of most concern (DOMC), Number days heavy use of DOMC in past 30 days, CIDI-SF drug dependence, and outpatient treatment during past 3 months for addiction or mental health. e Unable to complete and insufficient samples assumed to be positive. f Adjusted regression analysis was not done as there were too few samples. *% with positive value of difference between baseline and follow-up quantitative drug test
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020
aIndicates the percent of interventions by group that were conducted at level of beginning proficiency or better (MI Spirit variable > 3.5; pMIA variable > 90%) bIndicates the percent of intervention by group that were conducted at level competency (MI Spirit variable > 4.0; pMIA variable =100%) cDifference between groups on the variable pMIA was assessed using a t-test. All other group comparisons were conducted using Wilcoxon non-parametric tests. dASPIRE intervention component implemented eASPIRE intervention component implemented to standard fDenotes intervention component for specific intervention gDenotes optional intervention component for specific intervention
Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 11/26/2020