Schizophrenia Research 151 (2013) 191–196
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Psychiatric treatment following participation in the CapOpus randomized trial for patients with comorbid cannabis use disorder and psychosis Carsten Rygaard Hjorthøj ⁎, Sonja Orlovska, Allan Fohlmann, Merete Nordentoft Copenhagen University Hospital, Mental Health Center Copenhagen, Bispebjerg Bakke 23, DK-2400 Copenhagen, Denmark
a r t i c l e
i n f o
Article history: Received 8 May 2013 Received in revised form 23 September 2013 Accepted 13 October 2013 Available online 1 November 2013 Keywords: Schizophrenia Dual diagnosis Substance use disorder Psychosocial interventions
a b s t r a c t Background: Randomized trials targeting cannabis use disorders in patients with psychosis have generally been unsuccessful. One of the largest such trials was the CapOpus trial, which had an impact on the number of monthly joints used, but not on the number of days with cannabis use or positive or negative symptoms. Objective: To investigate the effects of CapOpus on psychiatric treatment. Methods: Six-month randomized trial on participants meeting ICD-10 criteria for cannabis use disorder and schizophrenia-spectrum psychosis. Participants were randomized to treatment as usual (TAU, n = 51) alone versus TAU plus CapOpus (n = 52) consisting of motivational interviewing and cognitive behavior therapy. Data regarding psychiatric treatment was obtained from complete nationwide registers. Analyses were intention-to-treat. Cox and poisson regression were used as appropriate. Results: Compared with treatment as usual, participants in the CapOpus group had an overall higher risk of having a psychiatric emergency room contact (hazard ratio 2.02, 95% confidence interval 1.22–3.34). Participants in CapOpus also had more contacts with psychiatric emergency rooms (incidence rate ratio 3.47 (2.64–4.57)) and more admissions to psychiatric hospitals (incidence rate ratio 2.24 (1.65–3.03)); conversely, CapOpusparticipants spent fewer days admitted to psychiatric hospitals than treatment-as-usual participants (incidence rate ratio 0.72 (0.68–0.75)). Conclusions: CapOpus led to earlier and more psychiatric emergency room contacts and admissions that, however, were of fewer days. This pattern could indicate that participants receiving treatment as usual were inadequately treated. However, it cannot be excluded that the differences might be an adverse reaction to the psychosocial intervention. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Cannabis use has been linked to a range of poor outcomes in patients with schizophrenia and other types of psychoses, e.g. more severe psychotic symptoms, poorer adherence to medication, increased risk of homelessness, and an increased risk of re-hospitalizations (Fenton et al., 1997; Caspari, 1999; Coldham et al., 2002; Grech et al., 2005; Hides et al., 2006; Miller et al., 2009). Most trials seeking to treat cannabis use disorders in psychosis have been largely unsuccessful (Hjorthøj et al., 2009). Most recently, several trials have reported that psychosocial interventions such as motivational interviewing and/or cognitive behavior therapy may be better than treatment as usual at reducing quantity, but not frequency, of cannabis use in patients with psychosis (Barrowclough et al., 2010; Bonsack et al., 2011). The CapOpus trial is to date one of the largest trials including patients with cannabis use disorder (Hjorthøj et al., 2013). The primary results of this trial showed a possible reduction in number ⁎ Corresponding author. Tel.: +45 2613 6290. E-mail address: [email protected] (C.R. Hjorthøj). 0920-9964/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.schres.2013.10.014
of monthly joints, but not days, with cannabis use. The intervention did not reduce psychotic symptoms. The trial applied intention-to-treat analyses but suffered from (expectedly) high loss to follow-up. The research protocol thus detailed secondary analyses based on Danish registers, in which there is complete follow-up of all patients. Nationwide registers of contact with psychiatric treatment services offer an exciting possibility of tracking patients over time. Within trials, researchers are generally limited to a few time-points when assessing patients with the risk of seeing patients only at certain times, e.g. when they feel well enough to participate in the assessments. Registers thus allow for a potentially better assessment outside these periods. With the present secondary analysis, we sought to estimate the impact of specialized addiction intervention on contact to other psychiatric treatment. 2. Methods The present investigation is a protocol-pre-planned analysis of the CapOpus randomized trial, including patients from September 2007 to December 2010. The protocol of the trial is available online at www.
192
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
capopus.dk and the design and primary results have been published in two separate papers (Hjorthøj et al., 2008; Hjorthøj et al., 2013). The primary results were focused on self-reported outcome measures. In the present investigation, we report results on register-based outcomes. 2.1. Patients Patients were recruited by referral from outpatient clinics in the Copenhagen area. Participants met ICD-10 criteria for schizophreniaspectrum psychosis (chapter F2) and cannabis use disorder (chapter F12). Participants were 17 to 42 years. Dependence to other substances was allowed if cannabis was the primary substance of abuse. Following informed consent and the baseline assessment, patients were randomized 1:1 to either CapOpus plus treatment as usual (TAU) or TAU alone. Randomization was performed centrally by the Copenhagen Trial Unit with block size (varying between 6, 8, and 10) unknown to the investigators, ensuring complete allocation concealment. Randomization was computerized, and stratified by severity of self-reported cannabis use (0–14 days or 15–30 days in the preceding month) and by type of referring clinic (used as a proxy for the severity of psychosis). Neither patients nor staff was blind to treatment allocation. 2.2. Intervention CapOpus was offered as an add-on treatment to TAU. The intervention was fully manual-based and set to last six months from first contact with the patient. The intervention used the Stages of Change model with different approaches for each stage (i.e. precontemplation, contemplation, preparation, action, maintenance, and relapse)(Prochaska and DiClemente, 1992). Motivational interviewing (Miller and Rollnick, 2002) was applied in combination with cognitive behavior therapy, with focus on assisting participants themselves to identify goals and problems related to cannabis use: Advantages and
disadvantages of continued use, development of personal strategies for craving and “trigger situations”, etc. One or two weekly sessions were planned with an assertive approach that entailed text messages, phone calls, and visiting patients in their homes. Caseload was intended to be around 1:10 but was usually between 1:6 and 1:7. Group sessions were planned but never implemented because patients did not wish to participate in them. CapOpus consultants also incorporated case managers from TAU in the treatment, as well as patients' families if the patients consented. 2.3. Control All patients in the trial received TAU and most patients were already receiving TAU at inclusion. In these cases, TAU simply consisted of continuing the existing treatment. For patients who were not receiving TAU at inclusion into the trial, the CapOpus addiction consultant facilitated referral to either Opus (if the patient met criteria for treatment in these teams) or community mental health centers. TAU could also take place in Assertive Community Treatment teams (Nordentoft et al., 1996; Vendsborg et al., 1999; Petersen et al., 2005). TAU was not organizationally part of the CapOpus trial and did not terminate when the six-month trial period ended. Caseload was around 1:10 in Opus and Assertive Community Treatment teams and 1:20 to 1:30 in Community Mental Health Centers. No manual exists regarding treatment of cannabis use disorders in TAU and thus the content of interventions targeted at such disorders probably ranged from nearly nothing to various systematic approaches adopted by case managers. 2.4. Outcomes The outcome measures in the present investigation are psychiatric contacts registered in the Danish Psychiatric Central Research Register (Mors et al., 2011). The register is nationwide and covers all inpatient
Participants referred to CapOpus (n=129)
Excluded (n=26) Not meeting inclusion criteria (n=4) Referred but never showed up (n=6) Not interested (n=16)
Interviewed for baseline assessment (n=103)
Randomized to CapOpus plus treatment as usual (n=52)
Randomized to treatment as usual (n=51)
Not found in registers due to invalid identification number (n=1)
Not found in registers due to invalid identification number (n=1)
Included in intention to treat analyses (n=51)
Included in intention to treat analyses (n=50)
Fig. 1. Flow chart of the CapOpus trial.
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
2.5. Sample size
Table 1 Selected baseline characteristics of the CapOpus trial.
Cannabis using days last month Standard joints last month Age Male Diagnosis Schizophrenia Schizotypal disorder Other/unclear diagnosis Age at first psychiatric contacta No. of psychiatric admissions per yearb No. of psychiatric emergency room visit per yearb No. of days admitted to psychiatric hospitals per yearb
193
TAU (n = 51)
CapOpus (n = 52)
15.4 (11.5) 47.1 (55.2) 27.1 (6.3) 40 (78.4%)
14.5 (11.1) 45.7 (48.2) 26.6 (6.3) 38 (73.1%)
p = 0.69 p = 0.89 p = 0.66 p = 0.53
22 (43.1%) 19 (37.3%) 10 (19.6%) 23.2 (5.2) 1.4 (2.2) 1.8 (2.3)
31 (59.6%) 13 (25.0%) 8 (15.4%) 21.1 (5.8) 1.9 (4.5) 2.3 (2.6)
p = 0.24
p = 0.07 p = 0.45 p = 0.32
73.4 (97.3)
70.1 (85.5)
p = 0.86
a and b Comparisons based on n = 50 for TAU and n = 51 for CapOpus due to invalid identification numbers leading to missing register-based information a Outpatient data only available after 1995. Difference drives by two outliers with first contact at age 5. P = 0.18 when excluding these two outliers. b Per year since first contact with psychiatric services
contacts with psychiatric treatment facilities from 1969 and outpatient contacts from 1995. Outpatient data was not considered as an outcome as patients were already in outpatient treatment upon inclusion into the trial. Since the register is complete, there is no missing data. For all participants, the following variables were constructed using data from the register starting the day each participant was included into the trial: Time at risk of first psychiatric hospital admission; time at risk of first psychiatric emergency room visit; time at risk of either admission or emergency room visit; number of psychiatric hospital admissions; number of psychiatric emergency room visits; and number of days admitted to psychiatric hospitals. Time-to-event data were analyzed using Cox proportional hazards regression with censoring occurring at date of death, date of emigration, or July 31st 2011 (the last date for which we had register-based information), whichever came first. We estimated hazard ratios (HR) for the intervention with corresponding 95% confidence intervals. In addition, we drew Kaplan–Meier curves comparing the intervention group with the control group. For the count outcomes, we estimated incidence rate ratios (IRR) and 95% confidence intervals using Poisson regression with the logarithm of time from inclusion to date of censoring as offset. Analyses were done according to the intention-totreat principle, analyzing all included participants. We conducted further analyses adjusted for the number of TAU-sessions received as part of the intervention, self-reported frequency of cannabis use in the preceding month (assessed using the Timeline Follow-Back technique) and assessor-rated severity of psychotic symptoms, assessed using the Positive and Negative Syndrome Scale for Schizophrenia (PANSS) (Kay et al., 1987; Sobell and Sobell, 1992). Since this would potentially lead to outcomes preceding exposures, we excluded psychiatric treatment received in the first six months after randomization. Analyses were carried out in SPSS version 20.
The sample size calculation for the CapOpus trial was based on the primary outcome measure, self-reported days with cannabis-use in the preceding month. For this outcome we needed 22 participants per group, but aimed at 60 to 70 in order to account for attrition and subgroup analyses. Slow recruitment of patients forced us to stop inclusion at 103 patients, which was still well above what was necessary for the primary analysis. We did not perform calculations related to power for the outcomes reported in the present secondary analysis. 2.6. Ethics and approvals The study was approved by the ethics committee for the Danish Capital Region (H-D-2007-0028) and the Danish Data Protection Agency (2007-41-0616). The trial was registered at ClinicalTrials.gov (NCT00484302). 3. Results Participant flow relevant to the present investigation is depicted in Fig. 1. A total of 103 patients were included and randomized to either CapOpus (n = 52) or TAU (n = 51). Two patients (one in each group) were registered under invalid identification numbers and could thus not be traced in the registers. Consequently, 101 patients (51 in CapOpus, 50 in TAU) were included in the analyses. Table 1 shows selected baseline characteristics. There was no indication of baseline differences except for a tendency to earlier age at first psychiatric contact in the CapOpus group compared with the TAU group. The difference was primarily driven by two patients in the CapOpus group having their first contact at age five, whereas the lowest age in the TAU group was 12years. Excluding these outliers, the p-value between the two groups was 0.18. Two persons, both in the CapOpus group, were censored prior to end of study, one of which had died and the other emigrated. Table 2 shows the relative use of psychiatric services in CapOpus compared to TAU. Patients were followed for a mean of 1001.9 (SD 258.8) days from inclusion to end of follow-up. The total amount of follow-up time was 101,190 days. Patients were at risk of emergency room contacts for 52,062 days, at risk of admission for 51,739 days, and at risk for any psychiatric contact for 46,270days. The mean number of emergency room contacts was 2.9 (SD 5.1, range 0 to 28) and Poisson regression showed that the incidence was highly and significantly increased in CapOpus compared with TAU by a factor 3.47 (95% confidence interval (C.I.) 2.64 to 4.57). Cox regression also showed a hazard ratio of 2.02 (95% C.I. 1.22 to 3.34) for time to first emergency room contact. The mean number of admissions was 2.0 (SD 2.8, range 0 to 18) and higher in the CapOpus group than in the TAU group. There was no statistically significant difference in time to first admission between the groups but a tendency towards earlier admissions in the CapOpus group (p = 0.12). Conversely, the mean number of days admitted to psychiatric hospitals was 68.7 (SD 114.5, range 0 to 562)
Table 2 Effects of CapOpus on level of psychiatric treatment. TAU (n = 50)
CapOpus (n = 51)
Cox proportional hazards regression analyses; hazard ratios and 95% confidence intervals Time to psychiatric emergency room contact 1 (ref.) Time to psychiatric hospital admission 1 (ref.) Time to any psychiatric contact 1 (ref.)
2.02 (1.22 to 3.34) 1.48 (0.91 to 2.41) 1.54 (0.96 to 2.46)
p = 0.006 p = 0.12 p = 0.07
Poisson regression analyses; incidence rate ratios and 95% confidence intervals Number of emergency room contacts Number of psychiatric hospital admissions Number of days admitted to psychiatric hospitals
3.47 (2.64 to 4.57) 2.24 (1.65 to 3.03) 0.72 (0.68 to 0.75)
p b 0.001 p b 0.001 p b 0.001
1 (ref.) 1 (ref.) 1 (ref.)
194
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
Intervention CapOpus Treatment as usual CapOpus-censored TAU-censored
Percentage admitted to psychiatric hospitals
100
80
60
40
20
0
Log rank: p=0.11 0
1
2
3
4
Years since inclusion into the trial Fig. 2. Time to first psychiatric hospital contact by intervention group.
with significantly fewer days in the CapOpus group than in the TAU group (incidence rate ratio 0.72 (95% C.I. 0.68 to 0.75)). Time to first emergency room contact and first psychiatric admission following inclusion into the trial are illustrated in Figs. 2 and 3, respectively. Participants in the CapOpus group had a much faster contact with both types of psychiatric services in the period immediately following inclusion. Chart reviews revealed that this was often because CapOpus consultants had facilitated contact with emergency rooms and inpatient psychiatric wards. However, when
starting the evaluation at six months after the inclusion (when the intervention was set to finish), the same tendencies were evident. We conducted further analyses including number of TAU-sessions received during the trial as well as information regarding consumption of cannabis and severity of psychotic symptoms. This reduced the sample size to between 59 and 65 due to missing values. In these adjusted analyses, the effect of CapOpus regarding time to psychiatric emergency room contact or time to psychiatric hospital admission was virtually unchanged. The association between CapOpus and
Intervention CapOpus Treatment as usual) (TAU) CapOpus-censored TAU-censored
Percentage contacted psychiatric emergency rooms
100
80
60
40
20
0
Log rank: p=0.005 0
1
2
3
4
Years since inclusion into the trial Fig. 3. Time to first psychiatric emergency room contact by intervention group.
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
number of psychiatric emergency room visits was reduced to IRR=2.18 (1.76–2.72), p b 0.001, a reduction of around a third of the unadjusted association. The association between CapOpus and number of psychiatric hospital admissions disappeared entirely, at IRR = 1.38 (0.83–2.31), p = 0.22. Conversely, the apparent protective effect of CapOpus on number of days admitted to psychiatric hospitals was even more pronounced in this adjusted model, at IRR = 0.55 (0.51– 0.60), p b 0.001. 4. Discussion In the present pre-planned secondary analysis of data from the CapOpus trial, we observed diverging tendencies regarding the impact of CapOpus specialized addiction intervention on contact with psychiatric services following inclusion. Participants randomized to receive CapOpus had more admissions and emergency room visits within the psychiatric system compared with participants randomized to receive TAU. Time-to-first psychiatric contact was also shorter in CapOpus than in TAU. On the other hand, participants in the CapOpus group were admitted to psychiatric hospitals significantly fewer days than participants in the TAU group. Increased rates of contact with psychiatric treatment facilities may have several causes and the association with the intervention (if causal) may be both beneficial and harmful: The increase in contact may be a positive finding and reflect a more appropriate treatment. It is possible that CapOpus addiction consultants facilitated contact to psychiatric treatment, which would mean that the treatment was appropriate but would not have occurred in the absence of being randomized into the CapOpus group. Similarly, if patients were taught, through CapOpus, to contact psychiatric treatment facilities when experiencing serious symptoms, an increase in contact would be expected and considered positive. Chart reviews revealed that CapOpus consultants did facilitate contact with treatment facilities making this a likely explanation. The increase in psychiatric contact remained well after the intervention had ceased, possibly indicating a teaching aspect of CapOpus. Secondary adjusted analyses showed largely similar results, indicating that symptoms, amount of TAU received, and frequency of cannabis use were not the true causes of the changes in psychiatric treatment associated with CapOpus. The exception to this was that, in adjusted analyses, there no longer appeared to be an increase in psychiatric hospital admissions in the CapOpus group. This is perhaps more likely to be a power issue than a case of confounding, since the original CapOpus investigation did not find an association between treatment allocation and level of psychotic symptoms. The possibility of the increase in contact being a negative consequence of CapOpus cannot be fully rejected. Starting in a new intervention with new staff may be highly stressful and stressors have repeatedly been shown to exacerbate psychotic symptoms (Norman and Malla, 1993; Myin-Germeys and Van, 2007). Since the increase in psychiatric contact was evident even when ignoring the intervention period, this does not seem probable. The increase could also be related to withdrawal symptoms if the intervention had a rapid effect on cannabis use. Nonetheless, the primary analysis showed that the odds of becoming abstinent did not differ between the two interventions and it was not the addiction consultants' experience that participants decreased their cannabis use rapidly. The fact that the CapOpus group actually had fewer days admitted to psychiatric hospitals following randomization also makes these two explanations less probable. A third explanation could be that, at first, receiving a new psychosocial intervention is stressful leading to increased contact, but over time the intervention is beneficial and leads to shorter stays in psychiatric hospitals. Previous trials have not found statistically significant differences between the intervention and control groups, although often with a tendency to slightly higher rates of psychiatric treatment contact in groups receiving Motivational Interviewing (Edwards et al., 2006;
195
Craig et al., 2008; Barrowclough et al., 2010; Bonsack et al., 2011). One trial reported lower hospitalization rates related to receipt of Motivational Interviewing (Bonsack et al., 2011). This tendency towards either no difference in psychiatric treatment or an increase in the intervention group may be exclusive to cannabis use. Trials targeting both cannabis use disorders and other types of alcohol or substance use disorders have showed a significant decrease in psychiatric treatment use among participants randomized to Motivational Interviewing (alone or along with other treatment elements) (Barrowclough, 2001; Haddock, 2003; James, 2004; Bellack et al., 2006). This is consistent with an observation that efficacy of psychosocial interventions appears higher in trials including all types of substance use disorders than in trials reporting separately for patients with cannabis use disorders (Hjorthøj et al., 2009). If engagement in new treatments is generally stressful with the stress leading to an increase in psychiatric contact, one would not expect to observe these differences in results depending on the type of substances used. As such, the more likely explanation appears to be that patients in the intervention groups are taught to seek treatment when the need arises. This is not reflected in trials targeting all substances as the intervention leads to a reduction in symptom levels, something neither achieved in CapOpus nor in other trials targeting cannabis use disorders separately. It remains unclear whether this reflects that trials should target all substances or that trials doing so miss potential differential effects depending on the types of primarily used substances. 4.1. Strengths and limitations Danish registers regarding psychiatric treatment allow complete follow-up of all participants as randomized according to the intentionto-treat principle. A further strength of the present analysis is that it was pre-planned, minimizing the risk of selective outcome reporting bias from the trial. Furthermore, the trial included more sessions and more participants with cannabis use disorder and psychosis than previous trials, reducing the risk of problems relating to statistical power. Unfortunately, however, we are limited in the information we have regarding the adequacy of psychiatric treatment. The present study also has certain limitations inherited from the overall design of the study. Patients were recruited by referral and as such will probably not be representative of all patients meeting the inclusion criteria. Neither treatment staff nor patients were blind to the allocated treatment, increasing the risk of collateral intervention bias. It seems unlikely, however, that participants receiving CapOpus should actively seek out collateral interventions that increased their degree of psychiatric contact. As such, we do not believe that this potential bias significantly influences our results. 4.2. Implications The findings in this secondary analysis do not change the conclusions presented in the primary analyses: that there is not sufficient evidence to introduce CapOpus style interventions (Motivational Interviewing and cognitive behavior therapy for six months) alone with regard to treating cannabis use disorders in patients with comorbid psychosis. Further research is needed in order to determine whether the differences in psychiatric contact observed between the intervention groups is a positive treatment effect or an adverse reaction. 4.3. Conclusion CapOpus specialized addiction intervention led to an increase in psychiatric admissions and emergency room visits but also to fewer inpatient days. It appears that this reflects a more adequate level of treatment and consequently that the control group was under-treated. However, it cannot be excluded that the association can be considered
196
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
an adverse reaction to the intervention. As such, future trials on psychosocial interventions in similar and other populations should still be monitored for safety. Role of funding source None. Contributors CRH and MN designed the original CapOpus trial. AF performed a large part of the intervention. CRH collected the data for the trial. CRH and SO performed the statistical analyses. CRH wrote the first draft of the manuscript. SO, AF, and MN critically revised the manuscript. All authors approved the final version of the manuscript. Conflicts of interest All authors declare that they have no conflicts of interest.
Acknowledgment This work was supported by The Lundbeck Foundation; The Municipality of Copenhagen; The Egmont Foundation; The Health Insurance Foundation; The Ministry of Social Welfare; Aase and Ejnar Danielsen's Foundation; and the Wørzner Foundation. The funding bodies of the trial had no influence on design, analysis, interpretation, drafting of the manuscript, decision to publish, or any other areas other than funding.
References Barrowclough, C., 2001. Randomized controlled trial of motivational interviewing, cognitive behavior therapy, and family intervention for patients with comorbid schizophrenia and substance use disorders. Barrowclough, C., Haddock, G., Wykes, T., Beardmore, R., Conrod, P., Craig, T., Davies, L., Dunn, G., Eisner, E., Lewis, S., Moring, J., Steel, C., Tarrier, N., 2010. Integrated motivational interviewing and cognitive behavioural therapy for people with psychosis and comorbid substance misuse: randomised controlled trial. BMJ 341, c6325. Bellack, A.S., Bennett, M.E., Gearon, J.S., Brown, C.H., Yang, Y., 2006. A randomized clinical trial of a new behavioral treatment for drug abuse in people with severe and persistent mental illness. Arch. Gen. Psychiatry 63, 426–432. Bonsack, C., Gibellini, M.S., Favrod, J., Montagrin, Y., Besson, J., Bovet, P., Conus, P., 2011. Motivational intervention to reduce cannabis use in young people with psychosis: a randomized controlled trial. Psychother. Psychosom. 80, 287–297. Caspari, D., 1999. Cannabis and schizophrenia: results of a follow-up study. Eur. Arch. Psychiatry Clin. Neurosci. 249, 45–49. Coldham, E.L., Addington, J., Addington, D., 2002. Medication adherence of individuals with a first episode of psychosis. Acta Psychiatr. Scand. 106, 286–290. Craig, T.K., Johnson, S., McCrone, P., Afuwape, S., Hughes, E., Gournay, K., White, I., Wanigaratne, S., Leese, M., Thornicroft, G., 2008. Integrated care for co-occurring disorders: psychiatric symptoms, social functioning, and service costs at 18 months. Psychiatr. Serv. 59, 276–282. Edwards, J., Elkins, K., Hinton, M., Harrigan, S.M., Donovan, K., Athanasopoulos, O., McGorry, P.D., 2006. Randomized controlled trial of a cannabis-focused
intervention for young people with first-episode psychosis. Acta Psychiatr. Scand. 114, 109–117. Fenton, W.S., Blyler, C.R., Heinssen, R.K., 1997. Determinants of medication compliance in schizophrenia: empirical and clinical findings. Schizophr. Bull. 23, 637–651. Grech, A., Van Os, J., Jones, P.B., Lewis, S.W., Murray, R.M., 2005. Cannabis use and outcome of recent onset psychosis. Eur. Psychiatry 20, 349–353. Haddock, G., 2003. Cognitive-behavioural therapy and motivational intervention for schizophrenia and substance misuse. 18-month outcomes of a randomised controlled trial. Hides, L., Dawe, S., Kavanagh, D.J., Young, R.M., 2006. Psychotic symptom and cannabis relapse in recent-onset psychosis. Prospective study. Br. J. Psychiatry 189, 137–143. Hjorthøj, C., Fohlmann, A., Larsen, A.-M., Madsen, M.T., Vesterager, L., Gluud, C., Arendt, M.C., Nordentoft, M., 2008. Design paper: The CapOpus trial: a randomized, parallel-group, observer-blinded clinical trial of specialized addiction treatment versus treatment as usual for young patients with cannabis abuse and psychosis. Trials 9, 42. Hjorthøj, C., Fohlmann, A., Nordentoft, M., 2009. Treatment of cannabis use disorders in people with schizophrenia spectrum disorders — a systematic review. Addict. Behav. 34, 520–525. Hjorthøj, C.R., Fohlmann, A., Larsen, A.-M., Gluud, C., Arendt, M.C., Nordentoft, M., 2013. Specialized psychosocial treatment plus treatment as usual versus treatment as usual for patients with cannabis use disorder and psychosis — the CapOpus randomised trial. Psychol. Med. 43, 1499–1510. James, W., 2004. A group intervention which assists patients with dual diagnosis reduce their drug use: a randomized controlled trial. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. Miller, W.R., Rollnick, S., 2002. Motivational Interviewing: Preparing People for Change, 2nd ed. Guilford, New York. Miller, R., Ream, G., McCormack, J., Gunduz-Bruce, H., Sevy, S., Robinson, D., 2009. A prospective study of cannabis use as a risk factor for non-adherence and treatment dropout in first-episode schizophrenia. Schizophr. Res. 113, 138–144. Mors, O., Perto, G.P., Mortensen, P.B., 2011. The Danish psychiatric central research register. Scand. J. Public Health 39, 54–57. Myin-Germeys, I., van, O.J., 2007. Stress-reactivity in psychosis: evidence for an affective pathway to psychosis. Clin. Psychol. Rev. 27, 409–424. Nordentoft, M., Knudsen, H.C., Jessen-Petersen, B., Krasnik, A., Sælan, H., Treufeldt, P., Wetcher, B., 1996. CCPP-Copenhagen Community Psychiatric Project. Implementation of community mental health centres in Copenhagen: effects of service utilization, social integration, quality of life and positive and negative symptoms. Soc. Psychiatry Psychiatr. Epidemiol. 31, 336–344. Norman, R.M., Malla, A.K., 1993. Stressful life events and schizophrenia. I: a review of the research. Br. J. Psychiatry 162, 161–166. Petersen, L., Jeppesen, P., Thorup, A., Abel, M.B., Ohlenschlaeger, J., Christensen, T.O., Krarup, G., Jorgensen, P., Nordentoft, M., 2005. A randomised multicentre trial of integrated versus standard treatment for patients with a first episode of psychotic illness. Br. Med. J. 331, 602. Prochaska, J.O., DiClemente, C.C., 1992. Stages of change in the modification of problem behaviors. Prog. Behav. Modif. 28, 183–218. Sobell, L.C., Sobell, M.B., 1992. Timeline follow-back: a technique for assessing self-reported alcohol consumption. In: Litten, R.Z., Allen, J. (Eds.), Measuring Alcohol Consumption: Psychosocial and Biological Methods. Humana Press, New Jersey, pp. 41–72. Vendsborg, P., Nordentoft, M., Hvenegaard, A., Søgaard, J., 1999. Opsøgende psykoseteam: Assertive Community Treatment Institut for sundhedsvæsen, Copenhagen.
Contents lists available at ScienceDirect
Schizophrenia Research journal homepage: www.elsevier.com/locate/schres
Psychiatric treatment following participation in the CapOpus randomized trial for patients with comorbid cannabis use disorder and psychosis Carsten Rygaard Hjorthøj ⁎, Sonja Orlovska, Allan Fohlmann, Merete Nordentoft Copenhagen University Hospital, Mental Health Center Copenhagen, Bispebjerg Bakke 23, DK-2400 Copenhagen, Denmark
a r t i c l e
i n f o
Article history: Received 8 May 2013 Received in revised form 23 September 2013 Accepted 13 October 2013 Available online 1 November 2013 Keywords: Schizophrenia Dual diagnosis Substance use disorder Psychosocial interventions
a b s t r a c t Background: Randomized trials targeting cannabis use disorders in patients with psychosis have generally been unsuccessful. One of the largest such trials was the CapOpus trial, which had an impact on the number of monthly joints used, but not on the number of days with cannabis use or positive or negative symptoms. Objective: To investigate the effects of CapOpus on psychiatric treatment. Methods: Six-month randomized trial on participants meeting ICD-10 criteria for cannabis use disorder and schizophrenia-spectrum psychosis. Participants were randomized to treatment as usual (TAU, n = 51) alone versus TAU plus CapOpus (n = 52) consisting of motivational interviewing and cognitive behavior therapy. Data regarding psychiatric treatment was obtained from complete nationwide registers. Analyses were intention-to-treat. Cox and poisson regression were used as appropriate. Results: Compared with treatment as usual, participants in the CapOpus group had an overall higher risk of having a psychiatric emergency room contact (hazard ratio 2.02, 95% confidence interval 1.22–3.34). Participants in CapOpus also had more contacts with psychiatric emergency rooms (incidence rate ratio 3.47 (2.64–4.57)) and more admissions to psychiatric hospitals (incidence rate ratio 2.24 (1.65–3.03)); conversely, CapOpusparticipants spent fewer days admitted to psychiatric hospitals than treatment-as-usual participants (incidence rate ratio 0.72 (0.68–0.75)). Conclusions: CapOpus led to earlier and more psychiatric emergency room contacts and admissions that, however, were of fewer days. This pattern could indicate that participants receiving treatment as usual were inadequately treated. However, it cannot be excluded that the differences might be an adverse reaction to the psychosocial intervention. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Cannabis use has been linked to a range of poor outcomes in patients with schizophrenia and other types of psychoses, e.g. more severe psychotic symptoms, poorer adherence to medication, increased risk of homelessness, and an increased risk of re-hospitalizations (Fenton et al., 1997; Caspari, 1999; Coldham et al., 2002; Grech et al., 2005; Hides et al., 2006; Miller et al., 2009). Most trials seeking to treat cannabis use disorders in psychosis have been largely unsuccessful (Hjorthøj et al., 2009). Most recently, several trials have reported that psychosocial interventions such as motivational interviewing and/or cognitive behavior therapy may be better than treatment as usual at reducing quantity, but not frequency, of cannabis use in patients with psychosis (Barrowclough et al., 2010; Bonsack et al., 2011). The CapOpus trial is to date one of the largest trials including patients with cannabis use disorder (Hjorthøj et al., 2013). The primary results of this trial showed a possible reduction in number ⁎ Corresponding author. Tel.: +45 2613 6290. E-mail address: [email protected] (C.R. Hjorthøj). 0920-9964/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.schres.2013.10.014
of monthly joints, but not days, with cannabis use. The intervention did not reduce psychotic symptoms. The trial applied intention-to-treat analyses but suffered from (expectedly) high loss to follow-up. The research protocol thus detailed secondary analyses based on Danish registers, in which there is complete follow-up of all patients. Nationwide registers of contact with psychiatric treatment services offer an exciting possibility of tracking patients over time. Within trials, researchers are generally limited to a few time-points when assessing patients with the risk of seeing patients only at certain times, e.g. when they feel well enough to participate in the assessments. Registers thus allow for a potentially better assessment outside these periods. With the present secondary analysis, we sought to estimate the impact of specialized addiction intervention on contact to other psychiatric treatment. 2. Methods The present investigation is a protocol-pre-planned analysis of the CapOpus randomized trial, including patients from September 2007 to December 2010. The protocol of the trial is available online at www.
192
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
capopus.dk and the design and primary results have been published in two separate papers (Hjorthøj et al., 2008; Hjorthøj et al., 2013). The primary results were focused on self-reported outcome measures. In the present investigation, we report results on register-based outcomes. 2.1. Patients Patients were recruited by referral from outpatient clinics in the Copenhagen area. Participants met ICD-10 criteria for schizophreniaspectrum psychosis (chapter F2) and cannabis use disorder (chapter F12). Participants were 17 to 42 years. Dependence to other substances was allowed if cannabis was the primary substance of abuse. Following informed consent and the baseline assessment, patients were randomized 1:1 to either CapOpus plus treatment as usual (TAU) or TAU alone. Randomization was performed centrally by the Copenhagen Trial Unit with block size (varying between 6, 8, and 10) unknown to the investigators, ensuring complete allocation concealment. Randomization was computerized, and stratified by severity of self-reported cannabis use (0–14 days or 15–30 days in the preceding month) and by type of referring clinic (used as a proxy for the severity of psychosis). Neither patients nor staff was blind to treatment allocation. 2.2. Intervention CapOpus was offered as an add-on treatment to TAU. The intervention was fully manual-based and set to last six months from first contact with the patient. The intervention used the Stages of Change model with different approaches for each stage (i.e. precontemplation, contemplation, preparation, action, maintenance, and relapse)(Prochaska and DiClemente, 1992). Motivational interviewing (Miller and Rollnick, 2002) was applied in combination with cognitive behavior therapy, with focus on assisting participants themselves to identify goals and problems related to cannabis use: Advantages and
disadvantages of continued use, development of personal strategies for craving and “trigger situations”, etc. One or two weekly sessions were planned with an assertive approach that entailed text messages, phone calls, and visiting patients in their homes. Caseload was intended to be around 1:10 but was usually between 1:6 and 1:7. Group sessions were planned but never implemented because patients did not wish to participate in them. CapOpus consultants also incorporated case managers from TAU in the treatment, as well as patients' families if the patients consented. 2.3. Control All patients in the trial received TAU and most patients were already receiving TAU at inclusion. In these cases, TAU simply consisted of continuing the existing treatment. For patients who were not receiving TAU at inclusion into the trial, the CapOpus addiction consultant facilitated referral to either Opus (if the patient met criteria for treatment in these teams) or community mental health centers. TAU could also take place in Assertive Community Treatment teams (Nordentoft et al., 1996; Vendsborg et al., 1999; Petersen et al., 2005). TAU was not organizationally part of the CapOpus trial and did not terminate when the six-month trial period ended. Caseload was around 1:10 in Opus and Assertive Community Treatment teams and 1:20 to 1:30 in Community Mental Health Centers. No manual exists regarding treatment of cannabis use disorders in TAU and thus the content of interventions targeted at such disorders probably ranged from nearly nothing to various systematic approaches adopted by case managers. 2.4. Outcomes The outcome measures in the present investigation are psychiatric contacts registered in the Danish Psychiatric Central Research Register (Mors et al., 2011). The register is nationwide and covers all inpatient
Participants referred to CapOpus (n=129)
Excluded (n=26) Not meeting inclusion criteria (n=4) Referred but never showed up (n=6) Not interested (n=16)
Interviewed for baseline assessment (n=103)
Randomized to CapOpus plus treatment as usual (n=52)
Randomized to treatment as usual (n=51)
Not found in registers due to invalid identification number (n=1)
Not found in registers due to invalid identification number (n=1)
Included in intention to treat analyses (n=51)
Included in intention to treat analyses (n=50)
Fig. 1. Flow chart of the CapOpus trial.
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
2.5. Sample size
Table 1 Selected baseline characteristics of the CapOpus trial.
Cannabis using days last month Standard joints last month Age Male Diagnosis Schizophrenia Schizotypal disorder Other/unclear diagnosis Age at first psychiatric contacta No. of psychiatric admissions per yearb No. of psychiatric emergency room visit per yearb No. of days admitted to psychiatric hospitals per yearb
193
TAU (n = 51)
CapOpus (n = 52)
15.4 (11.5) 47.1 (55.2) 27.1 (6.3) 40 (78.4%)
14.5 (11.1) 45.7 (48.2) 26.6 (6.3) 38 (73.1%)
p = 0.69 p = 0.89 p = 0.66 p = 0.53
22 (43.1%) 19 (37.3%) 10 (19.6%) 23.2 (5.2) 1.4 (2.2) 1.8 (2.3)
31 (59.6%) 13 (25.0%) 8 (15.4%) 21.1 (5.8) 1.9 (4.5) 2.3 (2.6)
p = 0.24
p = 0.07 p = 0.45 p = 0.32
73.4 (97.3)
70.1 (85.5)
p = 0.86
a and b Comparisons based on n = 50 for TAU and n = 51 for CapOpus due to invalid identification numbers leading to missing register-based information a Outpatient data only available after 1995. Difference drives by two outliers with first contact at age 5. P = 0.18 when excluding these two outliers. b Per year since first contact with psychiatric services
contacts with psychiatric treatment facilities from 1969 and outpatient contacts from 1995. Outpatient data was not considered as an outcome as patients were already in outpatient treatment upon inclusion into the trial. Since the register is complete, there is no missing data. For all participants, the following variables were constructed using data from the register starting the day each participant was included into the trial: Time at risk of first psychiatric hospital admission; time at risk of first psychiatric emergency room visit; time at risk of either admission or emergency room visit; number of psychiatric hospital admissions; number of psychiatric emergency room visits; and number of days admitted to psychiatric hospitals. Time-to-event data were analyzed using Cox proportional hazards regression with censoring occurring at date of death, date of emigration, or July 31st 2011 (the last date for which we had register-based information), whichever came first. We estimated hazard ratios (HR) for the intervention with corresponding 95% confidence intervals. In addition, we drew Kaplan–Meier curves comparing the intervention group with the control group. For the count outcomes, we estimated incidence rate ratios (IRR) and 95% confidence intervals using Poisson regression with the logarithm of time from inclusion to date of censoring as offset. Analyses were done according to the intention-totreat principle, analyzing all included participants. We conducted further analyses adjusted for the number of TAU-sessions received as part of the intervention, self-reported frequency of cannabis use in the preceding month (assessed using the Timeline Follow-Back technique) and assessor-rated severity of psychotic symptoms, assessed using the Positive and Negative Syndrome Scale for Schizophrenia (PANSS) (Kay et al., 1987; Sobell and Sobell, 1992). Since this would potentially lead to outcomes preceding exposures, we excluded psychiatric treatment received in the first six months after randomization. Analyses were carried out in SPSS version 20.
The sample size calculation for the CapOpus trial was based on the primary outcome measure, self-reported days with cannabis-use in the preceding month. For this outcome we needed 22 participants per group, but aimed at 60 to 70 in order to account for attrition and subgroup analyses. Slow recruitment of patients forced us to stop inclusion at 103 patients, which was still well above what was necessary for the primary analysis. We did not perform calculations related to power for the outcomes reported in the present secondary analysis. 2.6. Ethics and approvals The study was approved by the ethics committee for the Danish Capital Region (H-D-2007-0028) and the Danish Data Protection Agency (2007-41-0616). The trial was registered at ClinicalTrials.gov (NCT00484302). 3. Results Participant flow relevant to the present investigation is depicted in Fig. 1. A total of 103 patients were included and randomized to either CapOpus (n = 52) or TAU (n = 51). Two patients (one in each group) were registered under invalid identification numbers and could thus not be traced in the registers. Consequently, 101 patients (51 in CapOpus, 50 in TAU) were included in the analyses. Table 1 shows selected baseline characteristics. There was no indication of baseline differences except for a tendency to earlier age at first psychiatric contact in the CapOpus group compared with the TAU group. The difference was primarily driven by two patients in the CapOpus group having their first contact at age five, whereas the lowest age in the TAU group was 12years. Excluding these outliers, the p-value between the two groups was 0.18. Two persons, both in the CapOpus group, were censored prior to end of study, one of which had died and the other emigrated. Table 2 shows the relative use of psychiatric services in CapOpus compared to TAU. Patients were followed for a mean of 1001.9 (SD 258.8) days from inclusion to end of follow-up. The total amount of follow-up time was 101,190 days. Patients were at risk of emergency room contacts for 52,062 days, at risk of admission for 51,739 days, and at risk for any psychiatric contact for 46,270days. The mean number of emergency room contacts was 2.9 (SD 5.1, range 0 to 28) and Poisson regression showed that the incidence was highly and significantly increased in CapOpus compared with TAU by a factor 3.47 (95% confidence interval (C.I.) 2.64 to 4.57). Cox regression also showed a hazard ratio of 2.02 (95% C.I. 1.22 to 3.34) for time to first emergency room contact. The mean number of admissions was 2.0 (SD 2.8, range 0 to 18) and higher in the CapOpus group than in the TAU group. There was no statistically significant difference in time to first admission between the groups but a tendency towards earlier admissions in the CapOpus group (p = 0.12). Conversely, the mean number of days admitted to psychiatric hospitals was 68.7 (SD 114.5, range 0 to 562)
Table 2 Effects of CapOpus on level of psychiatric treatment. TAU (n = 50)
CapOpus (n = 51)
Cox proportional hazards regression analyses; hazard ratios and 95% confidence intervals Time to psychiatric emergency room contact 1 (ref.) Time to psychiatric hospital admission 1 (ref.) Time to any psychiatric contact 1 (ref.)
2.02 (1.22 to 3.34) 1.48 (0.91 to 2.41) 1.54 (0.96 to 2.46)
p = 0.006 p = 0.12 p = 0.07
Poisson regression analyses; incidence rate ratios and 95% confidence intervals Number of emergency room contacts Number of psychiatric hospital admissions Number of days admitted to psychiatric hospitals
3.47 (2.64 to 4.57) 2.24 (1.65 to 3.03) 0.72 (0.68 to 0.75)
p b 0.001 p b 0.001 p b 0.001
1 (ref.) 1 (ref.) 1 (ref.)
194
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
Intervention CapOpus Treatment as usual CapOpus-censored TAU-censored
Percentage admitted to psychiatric hospitals
100
80
60
40
20
0
Log rank: p=0.11 0
1
2
3
4
Years since inclusion into the trial Fig. 2. Time to first psychiatric hospital contact by intervention group.
with significantly fewer days in the CapOpus group than in the TAU group (incidence rate ratio 0.72 (95% C.I. 0.68 to 0.75)). Time to first emergency room contact and first psychiatric admission following inclusion into the trial are illustrated in Figs. 2 and 3, respectively. Participants in the CapOpus group had a much faster contact with both types of psychiatric services in the period immediately following inclusion. Chart reviews revealed that this was often because CapOpus consultants had facilitated contact with emergency rooms and inpatient psychiatric wards. However, when
starting the evaluation at six months after the inclusion (when the intervention was set to finish), the same tendencies were evident. We conducted further analyses including number of TAU-sessions received during the trial as well as information regarding consumption of cannabis and severity of psychotic symptoms. This reduced the sample size to between 59 and 65 due to missing values. In these adjusted analyses, the effect of CapOpus regarding time to psychiatric emergency room contact or time to psychiatric hospital admission was virtually unchanged. The association between CapOpus and
Intervention CapOpus Treatment as usual) (TAU) CapOpus-censored TAU-censored
Percentage contacted psychiatric emergency rooms
100
80
60
40
20
0
Log rank: p=0.005 0
1
2
3
4
Years since inclusion into the trial Fig. 3. Time to first psychiatric emergency room contact by intervention group.
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
number of psychiatric emergency room visits was reduced to IRR=2.18 (1.76–2.72), p b 0.001, a reduction of around a third of the unadjusted association. The association between CapOpus and number of psychiatric hospital admissions disappeared entirely, at IRR = 1.38 (0.83–2.31), p = 0.22. Conversely, the apparent protective effect of CapOpus on number of days admitted to psychiatric hospitals was even more pronounced in this adjusted model, at IRR = 0.55 (0.51– 0.60), p b 0.001. 4. Discussion In the present pre-planned secondary analysis of data from the CapOpus trial, we observed diverging tendencies regarding the impact of CapOpus specialized addiction intervention on contact with psychiatric services following inclusion. Participants randomized to receive CapOpus had more admissions and emergency room visits within the psychiatric system compared with participants randomized to receive TAU. Time-to-first psychiatric contact was also shorter in CapOpus than in TAU. On the other hand, participants in the CapOpus group were admitted to psychiatric hospitals significantly fewer days than participants in the TAU group. Increased rates of contact with psychiatric treatment facilities may have several causes and the association with the intervention (if causal) may be both beneficial and harmful: The increase in contact may be a positive finding and reflect a more appropriate treatment. It is possible that CapOpus addiction consultants facilitated contact to psychiatric treatment, which would mean that the treatment was appropriate but would not have occurred in the absence of being randomized into the CapOpus group. Similarly, if patients were taught, through CapOpus, to contact psychiatric treatment facilities when experiencing serious symptoms, an increase in contact would be expected and considered positive. Chart reviews revealed that CapOpus consultants did facilitate contact with treatment facilities making this a likely explanation. The increase in psychiatric contact remained well after the intervention had ceased, possibly indicating a teaching aspect of CapOpus. Secondary adjusted analyses showed largely similar results, indicating that symptoms, amount of TAU received, and frequency of cannabis use were not the true causes of the changes in psychiatric treatment associated with CapOpus. The exception to this was that, in adjusted analyses, there no longer appeared to be an increase in psychiatric hospital admissions in the CapOpus group. This is perhaps more likely to be a power issue than a case of confounding, since the original CapOpus investigation did not find an association between treatment allocation and level of psychotic symptoms. The possibility of the increase in contact being a negative consequence of CapOpus cannot be fully rejected. Starting in a new intervention with new staff may be highly stressful and stressors have repeatedly been shown to exacerbate psychotic symptoms (Norman and Malla, 1993; Myin-Germeys and Van, 2007). Since the increase in psychiatric contact was evident even when ignoring the intervention period, this does not seem probable. The increase could also be related to withdrawal symptoms if the intervention had a rapid effect on cannabis use. Nonetheless, the primary analysis showed that the odds of becoming abstinent did not differ between the two interventions and it was not the addiction consultants' experience that participants decreased their cannabis use rapidly. The fact that the CapOpus group actually had fewer days admitted to psychiatric hospitals following randomization also makes these two explanations less probable. A third explanation could be that, at first, receiving a new psychosocial intervention is stressful leading to increased contact, but over time the intervention is beneficial and leads to shorter stays in psychiatric hospitals. Previous trials have not found statistically significant differences between the intervention and control groups, although often with a tendency to slightly higher rates of psychiatric treatment contact in groups receiving Motivational Interviewing (Edwards et al., 2006;
195
Craig et al., 2008; Barrowclough et al., 2010; Bonsack et al., 2011). One trial reported lower hospitalization rates related to receipt of Motivational Interviewing (Bonsack et al., 2011). This tendency towards either no difference in psychiatric treatment or an increase in the intervention group may be exclusive to cannabis use. Trials targeting both cannabis use disorders and other types of alcohol or substance use disorders have showed a significant decrease in psychiatric treatment use among participants randomized to Motivational Interviewing (alone or along with other treatment elements) (Barrowclough, 2001; Haddock, 2003; James, 2004; Bellack et al., 2006). This is consistent with an observation that efficacy of psychosocial interventions appears higher in trials including all types of substance use disorders than in trials reporting separately for patients with cannabis use disorders (Hjorthøj et al., 2009). If engagement in new treatments is generally stressful with the stress leading to an increase in psychiatric contact, one would not expect to observe these differences in results depending on the type of substances used. As such, the more likely explanation appears to be that patients in the intervention groups are taught to seek treatment when the need arises. This is not reflected in trials targeting all substances as the intervention leads to a reduction in symptom levels, something neither achieved in CapOpus nor in other trials targeting cannabis use disorders separately. It remains unclear whether this reflects that trials should target all substances or that trials doing so miss potential differential effects depending on the types of primarily used substances. 4.1. Strengths and limitations Danish registers regarding psychiatric treatment allow complete follow-up of all participants as randomized according to the intentionto-treat principle. A further strength of the present analysis is that it was pre-planned, minimizing the risk of selective outcome reporting bias from the trial. Furthermore, the trial included more sessions and more participants with cannabis use disorder and psychosis than previous trials, reducing the risk of problems relating to statistical power. Unfortunately, however, we are limited in the information we have regarding the adequacy of psychiatric treatment. The present study also has certain limitations inherited from the overall design of the study. Patients were recruited by referral and as such will probably not be representative of all patients meeting the inclusion criteria. Neither treatment staff nor patients were blind to the allocated treatment, increasing the risk of collateral intervention bias. It seems unlikely, however, that participants receiving CapOpus should actively seek out collateral interventions that increased their degree of psychiatric contact. As such, we do not believe that this potential bias significantly influences our results. 4.2. Implications The findings in this secondary analysis do not change the conclusions presented in the primary analyses: that there is not sufficient evidence to introduce CapOpus style interventions (Motivational Interviewing and cognitive behavior therapy for six months) alone with regard to treating cannabis use disorders in patients with comorbid psychosis. Further research is needed in order to determine whether the differences in psychiatric contact observed between the intervention groups is a positive treatment effect or an adverse reaction. 4.3. Conclusion CapOpus specialized addiction intervention led to an increase in psychiatric admissions and emergency room visits but also to fewer inpatient days. It appears that this reflects a more adequate level of treatment and consequently that the control group was under-treated. However, it cannot be excluded that the association can be considered
196
C.R. Hjorthøj et al. / Schizophrenia Research 151 (2013) 191–196
an adverse reaction to the intervention. As such, future trials on psychosocial interventions in similar and other populations should still be monitored for safety. Role of funding source None. Contributors CRH and MN designed the original CapOpus trial. AF performed a large part of the intervention. CRH collected the data for the trial. CRH and SO performed the statistical analyses. CRH wrote the first draft of the manuscript. SO, AF, and MN critically revised the manuscript. All authors approved the final version of the manuscript. Conflicts of interest All authors declare that they have no conflicts of interest.
Acknowledgment This work was supported by The Lundbeck Foundation; The Municipality of Copenhagen; The Egmont Foundation; The Health Insurance Foundation; The Ministry of Social Welfare; Aase and Ejnar Danielsen's Foundation; and the Wørzner Foundation. The funding bodies of the trial had no influence on design, analysis, interpretation, drafting of the manuscript, decision to publish, or any other areas other than funding.
References Barrowclough, C., 2001. Randomized controlled trial of motivational interviewing, cognitive behavior therapy, and family intervention for patients with comorbid schizophrenia and substance use disorders. Barrowclough, C., Haddock, G., Wykes, T., Beardmore, R., Conrod, P., Craig, T., Davies, L., Dunn, G., Eisner, E., Lewis, S., Moring, J., Steel, C., Tarrier, N., 2010. Integrated motivational interviewing and cognitive behavioural therapy for people with psychosis and comorbid substance misuse: randomised controlled trial. BMJ 341, c6325. Bellack, A.S., Bennett, M.E., Gearon, J.S., Brown, C.H., Yang, Y., 2006. A randomized clinical trial of a new behavioral treatment for drug abuse in people with severe and persistent mental illness. Arch. Gen. Psychiatry 63, 426–432. Bonsack, C., Gibellini, M.S., Favrod, J., Montagrin, Y., Besson, J., Bovet, P., Conus, P., 2011. Motivational intervention to reduce cannabis use in young people with psychosis: a randomized controlled trial. Psychother. Psychosom. 80, 287–297. Caspari, D., 1999. Cannabis and schizophrenia: results of a follow-up study. Eur. Arch. Psychiatry Clin. Neurosci. 249, 45–49. Coldham, E.L., Addington, J., Addington, D., 2002. Medication adherence of individuals with a first episode of psychosis. Acta Psychiatr. Scand. 106, 286–290. Craig, T.K., Johnson, S., McCrone, P., Afuwape, S., Hughes, E., Gournay, K., White, I., Wanigaratne, S., Leese, M., Thornicroft, G., 2008. Integrated care for co-occurring disorders: psychiatric symptoms, social functioning, and service costs at 18 months. Psychiatr. Serv. 59, 276–282. Edwards, J., Elkins, K., Hinton, M., Harrigan, S.M., Donovan, K., Athanasopoulos, O., McGorry, P.D., 2006. Randomized controlled trial of a cannabis-focused
intervention for young people with first-episode psychosis. Acta Psychiatr. Scand. 114, 109–117. Fenton, W.S., Blyler, C.R., Heinssen, R.K., 1997. Determinants of medication compliance in schizophrenia: empirical and clinical findings. Schizophr. Bull. 23, 637–651. Grech, A., Van Os, J., Jones, P.B., Lewis, S.W., Murray, R.M., 2005. Cannabis use and outcome of recent onset psychosis. Eur. Psychiatry 20, 349–353. Haddock, G., 2003. Cognitive-behavioural therapy and motivational intervention for schizophrenia and substance misuse. 18-month outcomes of a randomised controlled trial. Hides, L., Dawe, S., Kavanagh, D.J., Young, R.M., 2006. Psychotic symptom and cannabis relapse in recent-onset psychosis. Prospective study. Br. J. Psychiatry 189, 137–143. Hjorthøj, C., Fohlmann, A., Larsen, A.-M., Madsen, M.T., Vesterager, L., Gluud, C., Arendt, M.C., Nordentoft, M., 2008. Design paper: The CapOpus trial: a randomized, parallel-group, observer-blinded clinical trial of specialized addiction treatment versus treatment as usual for young patients with cannabis abuse and psychosis. Trials 9, 42. Hjorthøj, C., Fohlmann, A., Nordentoft, M., 2009. Treatment of cannabis use disorders in people with schizophrenia spectrum disorders — a systematic review. Addict. Behav. 34, 520–525. Hjorthøj, C.R., Fohlmann, A., Larsen, A.-M., Gluud, C., Arendt, M.C., Nordentoft, M., 2013. Specialized psychosocial treatment plus treatment as usual versus treatment as usual for patients with cannabis use disorder and psychosis — the CapOpus randomised trial. Psychol. Med. 43, 1499–1510. James, W., 2004. A group intervention which assists patients with dual diagnosis reduce their drug use: a randomized controlled trial. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261–276. Miller, W.R., Rollnick, S., 2002. Motivational Interviewing: Preparing People for Change, 2nd ed. Guilford, New York. Miller, R., Ream, G., McCormack, J., Gunduz-Bruce, H., Sevy, S., Robinson, D., 2009. A prospective study of cannabis use as a risk factor for non-adherence and treatment dropout in first-episode schizophrenia. Schizophr. Res. 113, 138–144. Mors, O., Perto, G.P., Mortensen, P.B., 2011. The Danish psychiatric central research register. Scand. J. Public Health 39, 54–57. Myin-Germeys, I., van, O.J., 2007. Stress-reactivity in psychosis: evidence for an affective pathway to psychosis. Clin. Psychol. Rev. 27, 409–424. Nordentoft, M., Knudsen, H.C., Jessen-Petersen, B., Krasnik, A., Sælan, H., Treufeldt, P., Wetcher, B., 1996. CCPP-Copenhagen Community Psychiatric Project. Implementation of community mental health centres in Copenhagen: effects of service utilization, social integration, quality of life and positive and negative symptoms. Soc. Psychiatry Psychiatr. Epidemiol. 31, 336–344. Norman, R.M., Malla, A.K., 1993. Stressful life events and schizophrenia. I: a review of the research. Br. J. Psychiatry 162, 161–166. Petersen, L., Jeppesen, P., Thorup, A., Abel, M.B., Ohlenschlaeger, J., Christensen, T.O., Krarup, G., Jorgensen, P., Nordentoft, M., 2005. A randomised multicentre trial of integrated versus standard treatment for patients with a first episode of psychotic illness. Br. Med. J. 331, 602. Prochaska, J.O., DiClemente, C.C., 1992. Stages of change in the modification of problem behaviors. Prog. Behav. Modif. 28, 183–218. Sobell, L.C., Sobell, M.B., 1992. Timeline follow-back: a technique for assessing self-reported alcohol consumption. In: Litten, R.Z., Allen, J. (Eds.), Measuring Alcohol Consumption: Psychosocial and Biological Methods. Humana Press, New Jersey, pp. 41–72. Vendsborg, P., Nordentoft, M., Hvenegaard, A., Søgaard, J., 1999. Opsøgende psykoseteam: Assertive Community Treatment Institut for sundhedsvæsen, Copenhagen.
