Psycho-education for substance use and antisocial personality disorder

12 Pages·2015·0.61 MB·English

by  Birgitte Thylstrup

Thylstrupetal.BMCPsychiatry (2015) 15:283
DOI10.1186/s12888-015-0661-0
RESEARCH ARTICLE Open Access
Psycho-education for substance use and
antisocial personality disorder: a
randomized trial
Birgitte Thylstrup, Sidsel Schrøder and Morten Hesse*
Abstract
Background: Antisocial personality disorderoften co-exists withdrugand alcohol use disorders.
Methods: This trial examined the effectiveness ofofferingpsycho-education for antisocial personality disorder in
community substance use disorder treatmentcenters inDenmark. A total of176 patients were randomly allocated
totreatmentasusual(TAU,n=80)orTAUplusapsycho-educativeprogram,ImpulsiveLifestyleCounselling(ILC,n=96)
delivered by site clinicians (n=39). Using follow-up interviews 3 and 9 months after randomization, we examined
changesindrugandalcoholuse(AddictionSeverityIndexCompositeScores),percentdaysabstinent(PDA)withinlast
month,andaggressionasmeasuredwiththeBuss-PerryAggressionQuestionnaire-ShortFormandtheSelf-Reportof
AggressionandSocialBehaviorMeasure.
Results:Overallengagementinpsychologicalinterventionswasmodest:71(76%)ofparticipantsrandomizedto
psycho-educationattendedatleastonecounsellingsession,and21(23%)attendedallsixsessions.TheMedian
numberofsessionswas2.Allpatientsreduceddrugandalcoholproblemsat9monthswithsmallwithin-groupeffect
sizes.Intention-to-treatanalysesindicatedsignificantdifferencesbetweenILCandTAUinmeandrugscompositescore
(p=.018)andinPDA(p=.041)at3months.Aggressiondeclinedinbothgroups,butnodifferencesbetweenILCand
TAUwereobservedintermsofalcoholproblemsoraggressionatanyfollow-up.
Conclusions:Moderateshort-termimprovementsinsubstanceusewereassociatedwithrandomizationtoImpulsive
LifestyleCounselling.Thefindingssupporttheusefulnessofprovidingpsycho-educationtooutpatientswithantisocial
personalitydisorder.
Trialregistration:ISRCTNregistry,ISRCTN67266318,17/7/2012
Background which originates in childhood [20], and accordingly, the
Antisocial personality disorder (ASPD) is a serious dis- presence of conduct disorder is a prerequisite to the
turbance that imposes a major burden on individuals adult diagnosis [2].
and society [27] and for which there is no effective
treatment [25]. Affected individuals exhibit persistent
antisocial behavior and pervasive antisocial character TreatmentsforASPD
traits, such as irritability, manipulativeness, and lack of Early intervention for children with antisocial behavior
remorse [2]. The disorder affects between 1.0 and 3.6 % may prevent the development of ASPD and improve
of the general population (e.g. [11, 13]), and a substan- academic performance [46]. As for treatment for adults,
tially greater percentage of patients with substance use the treatment options that have been tested have been
disorders [13]. In terms of developmental psychopath- designed to treat comorbid substance use disorders
ology, there is evidence that ASPD is a lifespan disorder (SUDs) or have targeted a specific behavior such as
social problem-solving skills. Overall, evidence for the
efficacy of psychosocial interventions is weak [25]. In
*Correspondence:[email protected]
terms of mental health service use, individuals with
CentreforAlcoholandDrugResearch,AarhusUniversity,Artillerivej90,2nd,
2300CopenhagenS,Denmark ASPD have a high risk of needing emergency psychiatric
©2015Thylstrupetal.OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.0
InternationalLicense(http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinkto
theCreativeCommonslicense,andindicateifchangesweremade.TheCreativeCommonsPublicDomainDedicationwaiver
(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated.
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page2of12
services and inpatient hospitalization [15]. Although psychoeducation, patients reported that therapeutic alli-
they are not likely to seek treatment for their behavioral anceimproved[6].
problems, many patients with ASPD do seek treatment Given that ASPD is highly prevalent among people
for drug and alcohol problems [24, 50]. Among individ- with SUD, one context for reaching patients with ASPD
uals with SUD, ASPD is a common comorbidity across isat substanceusetreatment services.Somestudieshave
classes of substances [26] and is associated with poor indicated that treatments that integrate PD as 2an
prognosis [14, 33], even years or decades after a diagno- important component in SUD treatment may reduce
sisismade([21–23]. substance abuse among patients with comorbid SUD
The only published controlled trial of which we are and PD when compared to substance abuse treatment
aware which has tested an intervention directed at any alone, although the integration appears to have little
type of outpatients with ASPD, found a non-significant effect on symptoms and functioning [3–5, 40]. Providing
trend favoring cognitive behavior therapy over treatment treatment to people with comorbid ASPD and SUD may
as usual in terms of alcohol problems, but not in other be a challenge; in general, there is a risk of low attend-
outcomes, such as self-reported aggression and social ance to psychotherapeutic treatments and counselling at
functioning [16]. However, a potentially important ob- community-based treatment of substance use disorders
servation from this controlled trial was that a substantial [17, 43], and this risk is exacerbated in patients with
proportion of the participants were willing to engage in ASPD (e.g. [48]). In conclusion, there is a need for
treatment and that it is possible to offer treatment to further development of clinical strategies that can
individualswith ASPD. address the types of problems associated with ASPD in
settings where ASPD is common, such as substance
Psycho-education abuse treatment centers.
Onewaytoaddressapersonalitydisorder(PD)isthrough
psychoeducation. If done sensitively, providing psychoe-
Aims
ducation may help raise the patient’s awareness of his
Thepurposeofthisstudywastotesttheefficacyofabrief
behavioral difficulties and how they impact himself and
psycho-educational intervention, the Impulsive Lifestyle
others[6].Thisawareness mayinturnhelp thepatient in
Counselling program (ILC), as a supplement to standard
making informed decisions about seeking and receiving
substanceabusetreatmentinaregularoutpatientcontext.
helpforproblems.
The trial was pragmatic in the sense that the aim was to
Paradoxically, many clinicians report not providing
test the usefulness of implementing a brief intervention
psychoeducation to patients with PDs, although they
withlimiteddemandsoncostsandclinicianqualifications
perceive it to be an important aspect of treatment [44].
in a standard setting, community substance abuse treat-
Whether or not this applies to other treatment services
ment [57], in a way that is similar to a number of recent
and to ASPD is not known, but we are aware of no pub-
studies[17,54].
lished manuals or studies on how to provide psychoedu-
cation forpatientswith ASPD.
At present, there is no evidence on psychoeducation Methods
for ASPD, and little on psycho-education with patients Studydesign
with PD in general. A study of patients with borderline The study was a Phase I pragmatic randomized con-
personality disorder (BPD) found that psychoeducation trolled trial with single blind assessments, and was car-
had an effect on impulsivity and chaotic interpersonal ried out at community-based substance abuse treatment
relationships,butnot onglobal functioning[56]. clinics in 13 municipalities in Denmark between January
While BPD and ASPD are two distinct disorders, they 2012 and July 2014. Inclusion criteria were: between 18
share central features such as impulsivity and high levels and 65 years old; met criteria for ASPD using the Mini
of anger; therefore it is possible that these findings could International Neuropsychiatric Interview [47], able to
apply to the psychoeducation of patients with ASPD as providewritteninformedconsent,andseekingtreatment
well. If patients with ASPD gain a better understanding or already in treatment for a substance use disorder.
of their own personalities, they may identify dysfunc- Exclusion criteria were: plans that would interfere with
tional beliefs and behaviors that emerge in various situa- participation in the psychoeducation in the next three
tions, such as the belief that it is necessary to dominate months, such as plans to move away from the uptake
and control others [8, 38], and make informed decisions area or waiting to serve a prison sentence, plans to enter
about how to change their behavior. Another study of residential rehabilitation or hospitalization, and waiting
psychoeducation with patients with PDs was conducted to serve a prison sentence. Additionally, patients were
in two settings - community and forensic. The study did excluded if they were participating in group counselling
not have a control group, but after the course of or therapy with another patient participating in the trial,
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page3of12
were known to suffer from an acute psychosis or severe The trial coordinator informed the referring clinician
braindamage,ordidnot speakDanish. of the result of randomization immediately after being
notified that the patient had been assessed and was
found to be eligible for study participation. After this,
Ethics
the clinician informed the patient of the result. In the
The present project was reviewed by the regional ethics
cases in which patients were randomized to the ILC
committeeoftheCapitalRegionofDenmarkanddeemed
treatment, the clinician then contacted one of the ILC
exempt from a formal evaluation (J#H-3-2012-FSP45).
counsellors at the uptake unit with the participants’
Thisstudywasdoneinaccordancewiththedeclarationof
details so that the sessions could be initiated as quickly
Helsinki 2004, which states that it is the duty of the
aspossible.
researcher to protect the life, health, dignity, integrity,
Because the randomization had to take place immedi-
right to self-determination, privacy and confidentiality of
ately after the assessment interview, the trial coordinator
personal information of research subjects (WMA, 2013).
was unable to check whether the baseline assessment
All patients signed separate consent forms to participate
was complete before randomizing, and patients with
in the study and to be followed up. The Danish Data
incomplete data at baseline had to be excluded after
Protection Agency evaluated data security for the project
randomization.
and approved the procedures for data handling and stor-
age. The trial was registered in the ISRCTN register
Treatmentconditions
(#ISRCTN67266318).
TAU
All participants received whichever form of treatment
Recruitmentandrandomization they would have received at the participating treatment
Study participantswereidentified byclinicians atthe par- service if the trial had not taken place. Treatment al-
ticipating sites from new and existing patients receiving ways included: access to opioid substitution treatment
outpatient treatment for a drug or alcohol problem. After (either methadone, buprenorphine or a combination of
agreeing to be contacted, relevant participants were methadone and injectable diacetylmorphine) for pa-
invited by a trained clinician at the site to take part in an tients who needed it, and psychosocial support in the
interview to assess the diagnosis of ASPD and the other form of casework, counselling, or referral to residential
inclusion criteria. Those participants who met the inclu- rehabilitation. At some clinics, a liaison psychiatrist
sion criteria were told that their responses indicated would see the patients on-site, and at other clinics pa-
ASPD, and the counsellor would then review their tients would be referred to an off-site psychiatrist for
responsestotheMINImoduleandaskiftheyfeltthatthe diagnosis and treatment of other psychiatric conditions,
behavior described in the response constituted a problem such as attention-deficit/hyperactivity disorder, anxiety
tothem,andiftheywerewillingtospeakaboutitwithto or depression.
acounsellor.Thosewhoagreedtospeakwithacounsellor
andprovidedwritteninformedconsenttotakepartinthe ILC
study, subsequently completed the baseline assessment In addition to all of the services available to patients who
and were randomly allocated to either one of two active received TAU, patients randomized to ILC were offered
treatmentgroups:treatmentasusual(TAU)ortheImpul- up to six ILC sessions by a specially trained counsellor.
sive Lifestyle Counselling (ILC). Patients who consented The ILC program is a highly structured, manual guided
to participate in the study were also asked to provide psychoeducational intervention for people with ASPD
information for follow-up, including telephone numbers, [49]. Each session covers a specific topic and includes
home address, the addresses and telephone numbers of questions that the patient must be asked. The form and
family members or others who could help locate the contentofthesessionswereadapted from themanualfor
patient for the follow-up. Patients were also asked to theLifestyleIssuesprogram[53].InlinewiththeLifestyle
specify which of a number of alternative contact Issuesprogram,thekeyistosupportthepatientinaware-
sources other than the treatment clinic (prison services, nessraising,inrecognizingtheopportunitytochangelife-
social services, hospitals, homeless services) that they styleandintakingresponsibilityforaddressingbehavioral
would consent to being used to locate them . problems. Similar to the approach by Banerjee and col-
Randomizationwasstratifiedbyclinic.Therandomization leagues,thepsychoeducationalinterventionisintendedto
schedules were generated by the trial coordinator and kept function as a an educative and collaborative exercise that
secure and confidential at the study coordinating center in canimprovefurthertreatmentengagement[6].
Copenhagen. The randomization schedule was constructed Each session covers a specific topic and includes ques-
using the method of randomized permuted blocks of ran- tions that the patient must be asked, and pre-printed
domlyvaryingsizewitharatioof1:1(4or6perblock). handouts and worksheets are given to the patient. The
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page4of12
initial session focuses on the purpose of the ILC program interviewed, they would be asked if they would agree to
andonidentifyingthoughtsandbehaviorrelatedtoASPD. be contacted at a later point, and if they refused, they
The second session is based on an adapted version of the were notcontacted again.
Antecedents-Beliefs-Consequences model from Rational-
Emotive Behavior Therapy (Ellis & Dryden, 1997), linking Measures
the patients’ impulsive behaviors to the immediate conse- The Mini International Neuropsychiatric Interview
quences. Session 3 deals with impulsive and destructive [MINI] ASPD module was used to assess ASPD [31, 47].
behavior and how it may be related to specific value The MINI is a fully structured, brief and valid diagnostic
systems and beliefs associated with ASPD. Session 4 pre- interview that was designed to assess DSM-IVand ICD-
sents the concept of values and discusses which values 10 diagnoses [31, 47], which can be conducted by a lay
may support or prevent the patient in change of lifestyle, person and is well accepted by patients [41]. The ASPD
and session 5 focus on the patient’s social networks and module consists of six questions concerning conduct
how certain people or groups may support or challenge disorder and six questions about adult antisocial behav-
lifestyle changes. The last session is a booster session in ior. Previous research indicates that the MINI module
whichthepatientisinvitedtotalkaboutthetopicsthathe for ASPD identifies prison inmates with more serious
or she finds most relevant for future efforts to change mental health problems, more substance abuse prob-
behavior. lems, a more serious and chronic history of offending
Like the Lifestyles Issues program, the ILC program behavior compared with other inmates [9, 32, 35], and is
is designed so that no prior training or special facilities associated with illicit drug use in the general population
of any sort are necessary. However, prior to delivering [37]. For the present study, we used the official Danish
the intervention, all of the counsellors participated in translation of the MINI 5.0.0 by P. Besh, G. Bech-
two-day workshops to practice the strategies described Andersen, and T. Schütze. Also, after each adult anti-
inthemanualanddiscussissuesrelatedtotreatingpeople social item on the MINI schedule, staff members asked
with ASPD in general. All counsellors were required to about whether the behavior had occurred in the past
keep written records and make audio-recordings of the year, in order to confirm that the behavior was ongoing.
sessions. The sample internal consistency of the lifetime adult
Counsellors in the ILC group did not receive any spe- antisocial behavior items was Cronbach’s α=0.74 at
cial supervision beyond the supervision that was already baselineandα=0.65 forthe conductdisorder criteria.
available to staff in their respective clinics, but they did Additional demographic data were collected on a separ-
have the opportunity to call the study organizers with atesheet,includingeducation,employment history,history
specific questionsconcerning theintervention. of homelessness, residential treatment for substance use
disorder,incarceration,andpsychiatrichospitalizations.
Follow-upprocedures Current substance use severity was measured using
For the two follow-up waves, patients were initially con- the alcohol and drug use composite score from the
tacted through the phone number they had provided. If Addiction SeverityIndex(ASI)whichhavedemonstrated
it was not possible to establish contact with the person, high concordance with DSM-IVsubstance use disorders
the next attempt was to contact the patient through the [42], and days abstinent in the previous 30-day period.
clinic at which they had been screened for the study. If a All substance use data were collected at baseline and at
patient still could not be reached, we asked his or her eachfollow-upwave.
case manager at the clinic if there was a time when the Internalconsistencyforthedrugscompositescoreinthis
patient was expected to be at the clinic (e.g., times when sample was α=0.60 at baseline, α=0.60 at the 3-month
the patient would pick up medications). Patients who follow-up, and α=0.64 at the 9-month follow-up. Sample
could still not be reached were contacted through the internal consistency for the alcohol composite score items
telephone numbers and addresses they had provided, wasα=0.89atbaseline,α=0.92atthe3monthfollow-up,
and finally through other available sources that the pa- andα=0.77atthe9-monthfollow-up.
tient had given consent to at the study intake. In a few General aggression was measured using the 12-item
cases, the patients were finally located through the Cen- version of the Buss-Perry Aggression Questionnaire
tral Personal Register. Once a patient had been located, (BPAQ, [18]), a commonly used measure of general
and ifit was possible to speak with the patient directly,a aggression in both general population and forensic sam-
place and time for an interview was scheduled. If the ples with good psychometric properties. Sample items
patient did not show up for a face-to-face interview, a include “Given enough provocation, I may hit another
new time would be scheduled, and only after several person.” And “I often find myself disagreeing with
failed attempts was a telephone interview suggested. If people.”Theitemsarescoredonafive-pointLikert scale
the patients stated that they were not willing to be ranging from 1 (“extremely uncharacteristic of me”) to 5
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page5of12
(“extremely characteristic of me).” Sample internal effects regression analysis was used to assess the statis-
consistency for the BPAQ was α=0.82 at baseline, α= ticalsignificance ofwithin-group changes.
0.81 at 3-month follow-up and α=0.80 at 9-month Random-effects regression was used to assess the
follow-up. effects on substance use and aggression at the 3 and 9-
Interpersonal aggression was measured using the 14- month follow-up points after randomization. The out-
item version of the Self-Report of Aggression and Social come analyses were by intent-to-treat, i.e. analyzed by
Behavior Measure [36], a measure of interpersonal randomization arm irrespective of attendance or treat-
aggressive acts and dispositions. Sample items include ment compliance. Random effects were estimated for
“My friends know that I will think less of them if they both patient and site, and covariates were gender, age,
do not do what I want them to do.” And “When I am and receiving substitution treatment at baseline. The
mad at a person, I try to make sure she/he is excluded predictors were randomization status and assessment
from group activities (such as going to the movies or to wave, and the interaction of randomization status and
a bar).” Items are rated on a five-point Likert scale from assessment wave. All analyses were controlled for opioid
0 (“Never”) to 4 (“Very often”). The internal consistency substitution treatment because such medication may
for the SRASBM in this sample was α=0.78 at baseline, substantially influence both illicit drug use and use of
α=0.81 at the 3-month follow-up, and α=0.82 at the 9- treatment services, adding significant variance to the
month follow-up. dependent variables in ways that could potentially mask
effectsoftreatment.
Blinding Patients were included in the outcome analyses if they
Research technicians not affiliated with the clinics car- had complete data at baseline and at least one follow-up.
ried out all assessments at the 3 and 9-month follow-up For days abstinent, we also report the proportion at each
interviewsandwere blind totreatment groupallocation. follow-up wave who reported no days of substance use
(i.e.,currentabstinence)andtheproportionwhoreported
ILCadherencerating usealldays(i.e.,30daysofuseinthepast30days).
Two independent raters evaluated a sample of the audio-
taped ILC sessions for manual adherence. Adherence was Results
rated on a Likert scale from 1 to 5, in which 1 indicates The flow of patients through the trial is illustrated in
low adherence (that the session is largely independent of Fig.1.
the manual), and 5 indicates high adherence (that the A total of 142 patients had complete data, including at
counsellorfollowsthemanualclosely). least one follow-up wave, and could be included in the
analyses, of which 64 were assigned to TAU and 78 to
Calculationofsamplesize ILC. The sample was 87 % male, the mean age was
Assuming equal numbers of participants in intervention 32.21 years of age ([SD]=8.90), and 36.5 % received opi-
andcontrolarms,acorrelationof0.60betweenthesame oidsubstitution treatmentatthe timeofrandomization.
measure collected at two different points in time, and The most commonly used drugs in the past 30 days
two follow-up waves, a total sample size of 146 was cal- were cannabis (69.9 %), alcohol (67.1 %), benzodiaze-
culated for an effect size of 0.40, an alpha level of 0.05, pines (41.8%)andopioids (41.3%).
andapowerof0.80[19].Toadjust forpotential attrition
at follow-up, we aimed to include 200 patients in the Baselineequivalence
study. We compared the ILC and TAU groups at baseline, re-
gardless of whether they had been re-interviewed or not.
Dataanalysis The descriptive statistics are shown in Additional file 1:
The two groups were compared in terms of baseline TableS1.
characteristics using χ2 tests for dichotomous variables, Therewerenosignificantdifferencesintermsofdrugsor
andt-testsforcontinuous variables. alcohol composite scores, PDA, or aggression scores at
We report means and standard deviations for baseline.However,patientsintheILCgroupreportedmore
dependent variables at baseline, the 3 and 9-month daysofamphetamineusethanTAUpatients(p<0.05),and
follow-ups, and standard mean differences [SMD] were more often abstinent in the past month at baseline
between baseline and each follow-up wave within both (χ2(1)=3.52,p=0.029).
groups. The SMD was calculated as the difference
between baseline and follow-up mean, divided by the Exposuretotheintervention
baseline standard deviation for the group. It is common Of the 96 patients originally randomized to the ILC
to describe an SMD of 0.2 as small, 0.5 as medium, and group, 71 (76 %) attended at least one session, and 22
0.8 as a large effect size, following Cohen [10]. Fixed- (23 %) attended all six sessions. The median number of
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page6of12
Fig.1Flowdiagram
sessions attended was 2, and the mean was 2.8 (SD=2.4). Attritionanalyses
Forpatientswhoattendedthefirstsession,themedianILC We compared patients with at least one follow-up inter-
time before the first session was 26 days (inter-quartile view with patients who had never been re-interviewed in
range: 13 to 59 days). A total of 39 different counsellors terms of age, gender, drugs and alcohol composite score,
deliveredtheILCintervention. adult ASPD criteria, substitution treatment at baseline,
and in terms of intake unit. No differences between
patients interviewed and patients lost to follow-up
Manualadherence
attained statistical significance, except that there was a
Of the 80 audiotaped sessions rated, 20.3 % were rated
difference between clinics in terms of follow-up rate
as a 5 on the adherence scale, 72.2 % were rated as a 4,
(χ2(12)=25.7, p=0.012). Rates of patients interviewed at
and7.6%wererated asa2or3.
least once ranged from 63 % in some clinics to 100 % in
others.
Follow-up
The follow-up rate at 3 months was 79 % (81 % in the Substanceuseoutcomes
TAU group and 77 % in the ILC group) and 69 % at Mean values for substance use variables in both groups
9 months (71 % in the TAU group and 68 % in the ILC at all assessment waves are summarized in Table 1. Each
group). Of all follow-up interviews, 75 % were con- row contains the mean and standard deviation at each
ducted at substance abuse treatment clinics, 9 % in the assessment wave for each group, as well as the standard
interviewees’ homes, 3 % in prisons, 12 % in other mean difference as an effect size indicating within-group
places, 1 % were conducted as telephone interviews, and change. In addition to means and standard deviations
12 % were conducted in various other places, including for the dependent variables and standardized mean dif-
cafés,andpublic libraries. ferences, Table 1 shows the percentage of patients with
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page7of12
Table1Descriptivestatisticsofsubstanceuseatallassessmentwaves(meansandstandarddeviations)
TAU ILC
Baseline 3months SMD 9months SMD Baseline 3months SMD 9months SMD
(n=74) (n=61) (n=55) (n=93) (n=72) (n=63)
ASIdrugsCS 0.19(0.13) 0.21(0.12) −0.12 0.16(0.13) 0.27** 0.20(0.13) 0.17(0.12) 0.23* 0.15(0.12) 0.42**
ASIalcoholCS 0.15(0.22) 0.12(0.22) 0.14 0.10(0.18) 0.22* 0.14(0.22) 0.12(0.22) 0.22 0.12(0.21) 0.10
DaysAbstinent 11.8(11.7) 10.8(11.2) −0.09 13.7(12.7) 0.17 9.07(10.9) 13.2(12.7) 0.38* 15.3(13.3) 0.58**
Abstinent 12% 13% 13% 3% 17% 21%
Dailyuse 36% 36% 33% 42% 37% 31%
Abstinentindicateszerodaysofillicitdrugoralcoholuseoutofthepast30.Significantdifferencesinfixed-effectsregression:*:p<0.05.**p<0.01.Nineparticipants
hadnotcompleteddataonsubstanceuseatbaseline
TAUTreatmentasusual,ILCImpulsivelifestylecounselling,ASIAddictionSeverityIndex,CSCompositescore,SMDstandardmeandifference
past-month abstinence (i.e., percent reporting zero days the ILC group had significantly less drug use at the 3-
ofuse) andthepercentage ofdaily users (percent report- monthfollow-upcomparedtotheTAUgroup(β=−0.041,
ing30daysofuseinthepast30days). p=0.018). In addition to the ILC and time variables,
The ILC group had reduced their drug use composite patients who received opioid substitution treatment at
score by a small effect size at 3 months (SMD=0.23, p= baselinehadmoreseveredruguse(β=0.11,p<0.001).
0.042) and at 9 months (SMD=0.42, p=0.000). The For the alcohol composite score, the Wald χ2 was
TAU group increased their drug use composite score 16.24 (df=8, p=0.039), and there was no significant
by a small effect size at 3 months (SMD=−0.12, ns) effect of randomization at either follow-up wave. The
and reduced itby a small effect sizeat9 months (SMD= whole group decreased their level of alcohol problems at
0.27,p=0.001). 9months(p=0.011).
The ILC group reduced their alcohol use composite For days abstinent, the Wald χ2 was 31.17, (df=8, p<
score by a small effect size (SMD=0.22, ns) at 3 months 0.001),andpatientsrandomizedtotheILCgroupreported
and at 9 months (SMD=0.11, ns). The TAU group 4.3 more days abstinent at 3 month follow-up (CI: 0.18 to
reduced theiralcohol use severity with a small effect size 8.46). Additionally, patients who received substitution
(SMD=0.14, ns) at 3 months and at 9 months (SMD= treatmenthadfewerdaysabstinent(β=−5.64,p=0.003).
0.22.p=0.020).
The ILC group increased their days abstinent with a Aggressionoutcomes
small effect size at 3months(SMD=0.38, p=0.042) and Mean values for aggression variables in both groups at
a moderate effect size at 9 months (SMD=0.58, p< all assessment waves are summarized in Table 3. Table 3
0.001). TheTAUgroup reducedtheir days abstinentbya ispresentedsimilarlytoTable1.
small effect size at 3 months (-0.09, ns) and increased it The ILC group reduced their general aggression (BPAQ)
bya smalleffect sizebyat9months(0.17,ns). with a moderate effect size at 3 months (SMD=0.51, p<
The proportion of patients who reported being com- 0.001)andalargeeffectsizeat9months(SMD=0.76,p<
pletely abstinent was stable in theTAU group, at 12 % at 0.001). The TAU group reduced their general aggression
baseline,and13%at3monthsfollow-upandat9months with a moderate effect size at 3 months (SMD=0.50, p<
follow-up. In the ILC group, 3 % were abstinent at base- 0.001)andincreaseditbyalargeeffectsizebyat9months
line,17%at3monthsand21%at9months.Asfordaily (0.76,p<0.001).
use, the TAU group reported 36 % at baseline and 36 % The ILC group reduced their interpersonal aggression
and 33 % at the two follow waves. In the ILC group 42 % (SRASBM) with a small effect size at 3 months (SMD=
reported daily use at baseline and 37 and 31 % did so at 0.57, p<0.001), and at 9 months (SMD=0.72, p<0.01).
thefollowupwaves. The TAU group reduced their interpersonal aggression
The results of mixed effects regression on substance with a moderate effect size at 3 months (SMD=0.57, p<
use outcomes are summarized in Table 2. For each 0.001) and a moderate effect size at 9 months (SMD=
dependent variable, Table 2 reports the effects of 0.61,p<0.001).
assessment wave for 3 and 9 months follow-up, and the The results of mixed effects regression on outcomes
interaction between randomization and assessment are summarized in Table 4. Similar to Table 2, for each
wave for both waves. Additionally, the table contains dependent variable,Table 4 reports the effects of assess-
the intraclass correlations for randomization site and ment wave for 3 and 9 months follow-up, and the inter-
patient with 95 % confidence intervals, and the model action between randomization and assessment wave for
Wald χ2. For the ASI drugs composite score, the Wald both waves, the intraclass correlations for randomization
χ2 was 93.90 (df=8, p<0.001). Patients randomized to site and patient with 95 % confidence intervals, and the
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page8of12
Table2Resultsofmixedeffectsregressiononsubstanceuseoutcomes(n=142)
Dependentvariable Coefficient 95%CI P-value
ASIDrugsCS ILCintercept 0.001 −0.026to0.046 0.588
3months 0.015 −0.017to0.047 0.362
9months −0.052 −0.086to−0.019 0.002
ILCX3months −0.052 −0.096to−0.009 0.018
ILCX9months −0.004 −0.049to0.042 0.872
ICCsite 0.017 0.000to0.361
ICCPatient 0.335 0.234to0.455
Waldχ2(8) 93.90 0.000
ASIAlcoholCS ILCintercept −0.012 −0.077to0.052 0.708
3months −0.041 −0.092to0.010 0.134
9months −0.066 −0.120to−0.112 0.016
ILCX3months 0.008 −0.061to0.077 0.814
ILCX9months 0.049 −0.023to0.121 0.182
ICCSite 0.072 0.016to0.272
ICCPatient 0.506 0.402to0.610
Waldχ2(8) 16.24 0.039
Daysabstinent ILCintercept −2.300 −6.179to1.580 0.245
3months 0.970 −4.027to2.088 0.534
9months 2.359 −0.863to5.810 0.151
ILCX3months 4.319 0.183to8.456 0.041
ILCX9months 3.584 −0.751to7.919 0.105
ICCSite 0.005 0.000to1.000
ICCPatient 0.471 0.369to0.576
Waldχ2(8) 31.17 0.000
Allanalysesadjustedforsiteandindividual,gender,age,andsubstitutionatbaseline
TAUTreatmentasusual,ILCImpulsivelifestylecounselling,ASIAddictionSeverityIndex,CSCompositescore,ICCIntraclasscorrelation
model Wald χ2. For the BPAQ, the Wald χ2 was 101.82 Discussion
(df=8, p<0.001). No differences were found between This trial provides the first evidence that the ILC pro-
ILC and TAU at any point, but across both groups, con- gram, a short-term, highly structured psychoeducational
siderable reductions in general aggression were observed intervention, increases the efficacy of treatment for sub-
atbothfollow-upwaves. stance use disorders for patients with comorbid ASPD
For the SRASBM, the Wald χ2 was 124.43 (df=8, p< andsubstanceusedisorder.
0.001),andtherewasnosignificanteffectofrandomization At the 3-month follow-up, patients who had been ran-
at either follow-up wave. No differences were found domized to the ILC group had increased days abstinent
betweenILCandTAUatanypoint,butacrossbothgroups compared to patients randomized to TAU and had less
considerable reductions in interpersonal aggression were severe drug use. As is typical in community substance
observedatbothfollow-upwaves. abuse treatment contexts, attendance to the intervention
Table3Descriptivestatisticsofaggressionatallassessmentwavesforpatients(meansandstandarddeviations)
TAU ILC
Baseline 3months SMD 9months SMD Baseline 3months SMD 9months SMD
(n=74) (n=61) (n=55) (n=93) (n=70) (n=63)
BPAQ 4.44(1.20) 3.83(1.16) 0.50** 3.52(1.25) 0.76** 4.38(1.10) 4.01(1.16) 0.34** 3.59(1.05) 0.72**
SRASBM 1.00(0.63) 0.64(0.46) 0.57** 0.61(0.52) 0.61** 0.93(0.61) 0.64(0.49) 0.47** 0.47(0.39) 0.75**
Nineparticipantshadnotcompleteddataonaggressionatbaseline
TAUTreatmentasusual,ILCImpulsivelifestylecounselling,BPAQBuss-PerryAggressionQuestionnaire,SRASBMSelf-ReportofAggressionandSocialBehavior
Measure,SMDstandardmeandifference**p<0.01
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page9of12
Table4Resultsofmixedeffectsregressiononaggressionoutcomes(n=142)
Dependentvariable Coefficient 95%CI P-value
Buss-PerryAggressionQuestionnaire ILCintercept −0.121 −0.500to0.258 0.533
3months −0.693 −0.946to−0.440 0.000
9months −0.967 −1.234to−0.700 0.000
ILCX3months 0.334 −0.001to0.677 0.056
ILCX9months 0.199 −0.161to0.558 0.279
ICCsite 0.000 0.000to0.000
ICCPatient 0.620 0.532to0.702
Waldχ2(8) 99.24 0.000
SRASBM ILCintercept −0.084 −0.256to0.088 0.339
3months −0.392 −0.522to−0.261 0.000
9months −0.460 −0.597to−0.322 0.000
ILCX3months 0.083 −0.092to0.260 0.351
ILCX9months 0.026 −0.158to0.210 0.782
ICCSite 0.000 0.000to0.000
ICCPatient 0.513 0.413to0.612
Waldχ2(8) 116.95 0.000
was less than perfect. In spite of this, the intention-to- Considering the results in Table 1, differences between
treat analysis supported thebenefitsofthetreatment. the two groups were more pronounced in terms of
No statistically significant effects were observed for patients remaining completely abstinent, than in terms
self-reported aggression. Across both groups, substantial ofpatientsusingnon-prescription drugs or alcohol daily;
decreases in self-reported aggression were observed at in both the ILC and the TAU condition, the proportion
both follow-up waves,indicating either thatparticipation of patients using substances daily was stable over time,
in substance abuse treatment reduced aggression, or whereas the proportion of patients that were abstinent
alternatively a regression to the mean effect [7], or a increasedintheILCgroup.
retest artefact [45]. It is plausible that the intervention The ILC program did not have a significant impact on
was too brief to have an impact on aggressive behavior. the alcohol composite score. In fact, the control group
However,thefindingsaresimilartothose ofa somewhat had reduced their level of alcohol severity by a small
more intensive intervention with cognitive behavioral effect size at 9 months, whereas the ILC group was vir-
therapy for psychiatric outpatients with ASPD, where tually unchanged. However, given that it did not attain
some effect on substance use was observed, but none on statistical significance, speculating about the reasons for
self-reportedaggression[16]. thisnegativefinding isunwarranted.
The brief non-intensive intervention delivered in this The current trial supported the findings by Davidson
study is not a cure for ASPD, but can constitute one of and colleagues that substance use could be influenced by
many small steps towards improving treatment for this targetingASPDusingapsychosocialintervention[16].
under-served population. Substance use is a factor that An additional finding from this study was that the
complicatesthetreatmentofanypsychiatricdisorder,and group ofpatientswho received opioidsubstitution medi-
if substance use is reduced, it opens up the possibility of cation at baseline had more severe drug problems and
further interventions and support, potentially increasing fewer days abstinent over the course of the trial. This
thepatient’ssocialandpsychologicalstability. finding does not necessarily indicate that substitution
As in several previous studies, we found evidence that treatment is ineffective, as it may just as well reflect pre-
individuals with ASPD reduced substance use during existing higher severity of drug problems in the opioid
standard substance abuse treatment [1, 39]; in the entire substitution group.
sample, alcohol and drugs composite scores were signifi-
cantly reduced at the 9-month follow-up. These findings Limitationsandstrengths
add further support to the view that individuals with Several limitations for this study must be acknowledged.
ASPD should not be excluded from substance abuse First, we had only self-reported data on substance use
treatment [28, 34, 39]. On the other hand, reductions in with no biological verification available. It is no longer
substanceusegenerallyrepresentedsmalleffect sizes. standard to use biological data for collecting data on
Thylstrupetal.BMCPsychiatry (2015) 15:283 Page10of12
drug use (e.g. [17, 30]), and a number of studies have Additional file
indicated that biological verification underestimates sub-
stance use and does not provide more valid data than Additionalfile1:TableS1. Descriptivestatisticsatbaseline(means
andstandarddeviations).(DOCX14kb)
self-reporting (e.g. [12, 51, 52]). Another limitation is
that the patient population in this study was heteroge-
neous, including patients using a wide range of sub- Competinginterests
Theauthorshavenocompetinginterests.
stances, and with a wide age range. On the other hand,
this means that most patients in substance abuse treat- Authors’contributions
ment with ASPD would meet inclusion criteria and BTandMHconceivedofthestudyandadaptedthemanual.BTcarriedoutthe
trainingofthecounsellors.Allthreeauthorswereinvolvedinallstepsofthe
could thus/potentially increase the generalizability of the
dataacquisition.MHdraftedthemanuscriptandcarriedoutthestatistical
findings[29,57]. analyses,andBTandSSrevisedthemanuscriptforcriticalintellectualcontent.
A further limitation is that we were not able to imple- Allauthorsreadandapprovedofthefinalmanuscript.
ment a standard of treatment as usual. The participating
Acknowledgements
clinics represented a wide range of rural and urban
ThisstudywassupportedbygrantsfromTrygfonden(j.nr.7-10-0705and
settings, and the service level depended on the local Reckitt-Benckiser(unnumbered),bothtoMortenHesse.Theauthorswishto
authorities and the division of labor between local psy- thankAbduKedir,JakobSanderHansenandOleJakobStorebøforuseful
commentsonpreviousversionsofthismanuscript.Allauthorswishtothank
chiatric, social, and substance abuse treatment services,
PamelaStarbirdforherlanguagecorrectionsforthismanuscript.
making it infeasible to standardize the treatment. Since
the intention of ILC is to increase help-seeking behavior Received:23July2015Accepted:26October2015
and compliance with treatment, it did not seem pertinent
tocontrolforoverallamountofservicesreceived. References
Further, the absence of an attention placebo condition 1. AltermanAI,RutherfordMJ,CacciolaJS,McKayJR,WoodyGE.Responseto
methadonemaintenanceandcounselinginantisocialpatientswithand
meant that the efficacy of the study may be overesti- withoutmajordepression.JNervMentDis.1996;184(11):695–702.
mated inthepresent analyses. doi:10.1097/00005053-199611000-00007.
The principle strength of this study lies mainly in its 2. APA.Diagnosticandstatisticalmanualofmentaldisorders.5thed.
Arlington,VA:AmericanPsychiatricAssociation;2013.
pragmatic approach to implementing screening and
3. BallSA.Comparingindividualtherapiesforpersonalitydisorderedopioid
intervention in a way that is practically feasible at dependentpatients.JPersDisord.2007;21(3):305–21.
busy clinics [17]. The intervention was specifically se- doi:10.1521/pedi.2007.21.3.305.
4. BallSA,Cobb-RichardsonP,ConnollyAJ,BujosaCT,O'NeallTW.Substanceabuse
lected to require minimal training of clinicians, yet
andpersonalitydisordersinhomelessdrop-incenterclients:symptomseverity
had a notable impact on the reduction of substance andpsychotherapyretentioninarandomizedclinicaltrial.ComprPsychiatry.
use. 2005;46(5):371–9.doi:10.1016/j.comppsych.2004.11.003.
5. BallSA,MaccarelliLM,LaPagliaDM,OstrowskiMJ.Randomizedtrialof
Although this study adds to the development of
dual-focusedvs.single-focusedindividualtherapyforpersonality
evidence-based targeted interventions for personality disordersandsubstancedependence.JNervMentDis.
disorders and ASPD, further research is needed to assess 2011;199(5):319–28.doi:10.1097/NMD.0b013e3182174e6f.
6. BanerjeeP,DugganC,HubandN,WatsonN.Briefpsychoeducationfor
the relative usefulness and credibility of psychoeducative
peoplewithpersonalitydisorder:apilotstudy.PsycholPsychother.
programs such as Impulsive Lifestyle Counselling or the 2006;79:385–94.doi:10.1348/147608305x57987.
models used by Banerjee et al. [6] or Zanarini and Fran- 7. BarnettAG,vanderPolsJC,DobsonAJ.Regressiontothemean:whatitisand
howtodealwithit.IntJEpidemiol.2005;34(1):215–20.doi:10.1093/ije/dyh29.
kenburg[56]forpatientswithASPD.
8. BatemanA,FonagyP.Comorbidantisocialandborderlinepersonality
Of particular interest would be studies with a focus disorders:mentalization-basedtreatment.JClinPsychol.2008;64(2):181–94.
on ASPD that examine different approaches at doi:10.1002/jclp.20451.
9. BlackDW,GunterT,LovelessP,AllenJ,SieleniB.Antisocialpersonality
substance abuse treatment services and compare them
disorderinincarceratedoffenders:psychiatriccomorbidityandqualityof
to other services offered to patients with ASPD in life.AnnClinPsychiatry.2010;22(2):113–20.Retrievedfrom<GotoISI>://
terms of their efficacy, their costs and the ease of WOS:000277846400006.
10. CohenJ.StatisticalPowerAnalysisfortheBehavioralSciences.2nded.
implementation.
Hillsdale,NJ:Erlbaum;1988.
11. CoidJ,YangM,TyrerP,RobertsA,UllrichS.Prevalenceandcorrelatesof
personalitydisorderinGreatBritain.BrJPsychiatry.2006;188:423–31.
doi:10.1192/bjp.188.5.423.
Conclusion
12. ColonHM,PerezCM,MelendezM,MarreroE,OrtizAP,SuarezE.Thevalidity
Impulsive lifestyle counselling, a brief psychoeducative ofdruguseresponsesinahouseholdsurveyinPuertoRico:comparisonof
intervention, had an impact on substance use in terms surveyresponseswithurinalysis.AddictBehav.2010;35(7):667–72.
doi:10.1016/j.addbeh.2010.02.006.
of drug use problems and days abstinent for outpatients
13. ComptonWM,ConwayKP,StinsonFS,ColliverJD,GrantBF.Prevalence,
with substance use disorders and antisocial personality correlates,andcomorbidityofDSM-IVantisocialpersonalitysyndromesand
disorder. More research is needed on how to provide alcoholandspecificdrugusedisordersintheUnitedStates:Resultsfrom
thenationalepidemiologicsurveyonalcoholandrelatedconditions.
optimal treatment for people with antisocial personality JClinPsychiatry.2005;66(6):677–85.Retrievedfrom<GotoISI>://WOS:
disorder. 000229989000002.