European Psychiatry xxx (2014) xxx–xxx

G Model

EURPSY-3190; No. of Pages 8

Original article

Reasons for cannabis use in first-episode psychosis: Does strength ofendorsement change over 12 months?

A. Kolliakou a,*, D. Castle b, H. Sallis c, C. Joseph d, J. O’Connor d, B. Wiffen d,C. Gayer-Anderson e, G. McQueen d, H. Taylor d, S. Bonaccorso d, F. Gaughran d,S. Smith f, K. Greenwood g, R.M. Murray d, M. Di Forti d, Z. Atakan h, K. Ismail a

a Department of Psychological Medicine, PO92, Institute of Psychiatry, King’s College London, De Crespigny Park, SE5 8AF London, UKb Department of Psychiatry, University of Melbourne, Melbourne, Australiac MRC Integrative Epidemiology Unit, School of Social and Community Medicine, University of Bristol, Bristol, UKd Department of Psychosis Studies, Institute of Psychiatry, King’s College London, London, UKe Department of Health Service and Population Research, Institute of Psychiatry, King’s College London, London, UKf Department of Forensic and Neurodevelopmental Sciences, Institute of Psychiatry, King’s College London, London, UKg School of Psychology, University of Sussex, Sussex, UKh Department of Neuroimaging, Institute of Psychiatry, King’s College London, London, UK

A R T I C L E I N F O

Article history:

Received 4 August 2014

Received in revised form 29 October 2014

Accepted 29 October 2014

Available online xxx

Keywords:

Psychosis

Schizophrenia

Cannabis use

Reasons for cannabis use

Self-medication

Alleviation of dysphoria

A B S T R A C T

Background: Why patients with psychosis use cannabis remains debated. The self-medication hypothesis

has received some support but other evidence points towards an alleviation of dysphoria model. This

study investigated the reasons for cannabis use in first-episode psychosis (FEP) and whether strength in

their endorsement changed over time.

Methods: FEP inpatients and outpatients at the South London and Maudsley, Oxleas and Sussex NHS

Trusts UK, who used cannabis, rated their motives at baseline (n = 69), 3 months (n = 29) and 12 months

(n = 36). A random intercept model was used to test the change in strength of endorsement over the

12 months. Paired-sample t-tests assessed the differences in mean scores between the five subscales on

the Reasons for Use Scale (enhancement, social motive, coping with unpleasant affect, conformity and

acceptance and relief of positive symptoms and side effects), at each time-point.

Results: Time had a significant effect on scores when controlling for reason; average scores on each

subscale were higher at baseline than at 3 months and 12 months. At each time-point, patients endorsed

‘enhancement’ followed by ‘coping with unpleasant affect’ and ‘social motive’ more highly for their

cannabis use than any other reason. ‘Conformity and acceptance’ followed closely. ‘Relief of positive

symptoms and side effects’ was the least endorsed motive.

Conclusions: Patients endorsed their reasons for use at 3 months and 12 months less strongly than at

baseline. Little support for the self-medication or alleviation of dysphoria models was found. Rather,

patients rated ‘enhancement’ most highly for their cannabis use.

� 2014 Elsevier Masson SAS. All rights reserved.

Contents lists available at ScienceDirect

European Psychiatry

jo u rn al h om epag e: h t tp : / /ww w.eu ro p s y- jo ur n al .co m

1. Introduction

High rates of substance use disorder have been reported amongpatients with severe mental illness [27,34] with cannabis the mostpopular [5,13,16,26,40] and at much greater extent than in thegeneral population [5]. Persistent cannabis use by patients with anestablished psychosis has been associated with increased symp-toms, hospital readmissions and absence of remission [41], morepositive symptoms and a more continuous illness at follow-up

* Corresponding author. Tel.: +44 0 2032288561.

E-mail address: anna.kolliakou@kcl.ac.uk (A. Kolliakou).

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

http://dx.doi.org/10.1016/j.eurpsy.2014.10.007

0924-9338/� 2014 Elsevier Masson SAS. All rights reserved.

[11]. It has also been shown to make patients more likely to relapse[26] and experience difficulties leading to violence and criminalbehaviour [29]. In a systematic review by Zammit et al. [50],persistent cannabis use was also associated with increased relapseand non-adherence to treatment. Therefore, understanding whypatients with psychosis use cannabis is greatly important for thedevelopment of successful interventions.

Several hypotheses have been proposed to explain why patientsuse substances [30]:

� secondary substance use disorder models, which postulate thatsubstance use disorders are secondary to severe mental illness;

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx2

G Model

EURPSY-3190; No. of Pages 8

� secondary psychiatric illness models, which assume thatsubstance use disorders precede psychiatric disorders;� common factor models, which propose that psychiatric and

substance use disorders co-occur due to underlying sharedfactors;� bi-directional models, which assume that psychiatric and

substance use disorders activate and perpetuate each other;� multiple-risk factor models.

The most popular of the secondary substance use disordermodels, the self-medication hypothesis [19,20], proposes thatpatients take specific substances to relieve particular symptoms;substances are not chosen at random but have ‘‘psychopharma-cological specificity.’’ Khantzian [20] also argued that patients withschizophrenia might start taking substances, prior to the onset ofthe disorder, to self-medicate prodromal symptoms. Self-medica-tion of symptoms caused by antipsychotic medication may also beoccurring [36]. A variation of the traditional self-medicationhypothesis, the alleviation of dysphoria model, proposes thatsubstance misuse occurs to reduce unpleasant states such asboredom, depression and loneliness to which patients with severemental illness may be particularly susceptible. This model assumesthat patients do not choose specific substances to medicate specificundesired psychological states.

Findings from studies exploring self-reported reasons forsubstance use in patients with psychosis suggest that they largelyuse substances for their mood-enhancing properties and socialeffects with a small minority endorsing self-medication motivesfor their substance use. Spencer et al. [42] found that patientsmainly reported using alcohol and cannabis for enhancement,social and conformity reasons and to cope with dysphoricexperiences. Over half of the patients reported wanting to fit inwith peers as the main reason for initiating substance use in astudy by Laudet et al. [24]. Test et al. [46] found that the majority ofpatients said they used substances to relieve boredom followed byusing as a social activity and to feel less anxious and more relaxed.Most patients in a study by Dixon et al. [7] reported usingsubstances to get high and reduce feelings of depression followedclosely by using substances to relax and to increase feelings ofpleasure. Gregg et al. [15] also found that almost all patients withco-morbid alcohol or substance abuse/dependence reported usingsubstances to ‘‘chill out or relax.’’ Social reasons were also veryimportant followed by relief of boredom. Finally, in a review byGregg et al. [14], the ‘alleviation of dysphoria’ model was shown tobe largely supported by the self-report literature with littleevidence for the self-medication, common factor or bi-directionalmodels of co-morbid substance use. Another review by Perez et al.[33] also found that the 3 most popular reasons for substance usefrom self-report studies were, in order of preference:

� improvement in positive sensations;� relieving dysphoria;� social.

It has been shown that reasons for use differ betweensubstances and disorders [47] and so it is difficult to generalisefrom these results as none of the studies reported findings fromcannabis-specific analyses.

However, research on self-reported reasons for cannabis use, inparticular, has provided similar evidence. Addington and Duchak[1] reported the most frequent reasons for cannabis use in apatient-group with schizophrenia were to increase pleasure and toget high, to relax and reduce depression and to be more sociable. In2004, results from a study by Green et al. [12] showed that patientswith psychosis most commonly reported using cannabis forpositive mood alteration, coping with negative affect and for

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

social activity reasons. Relaxation was the least popular reason forcannabis use together with general coping with negative mood andcognitive enhancement. Availability of cannabis seemed to be avery important reason for its use. Schofield et al. [37] reported thatpatients diagnosed with a schizophrenia-spectrum disorderreported using cannabis to relax, to have as an activity withfriends and to relieve boredom. Less than a quarter of patients usedcannabis to reduce the side effects of antipsychotic medication, orto reduce positive symptoms. Fowler et al. [9] found that usingcannabis to reduce dysphoria was the most common reason,succeeded by social reasons and intoxication effects in their study.In Switzerland, Schaub et al. [35] reported that the majority ofpatients used cannabis to relax, get high and increase pleasure.Finally, in a study by Goswami et al. [10] all patients stated theyused cannabis to increase pleasure and the majority of patientsstated they used cannabis to get high, relax and satisfy curiosity.However, none of these studies were conducted with patientspresenting with a first-episode of psychosis. To date, only 3 studieshave explored the reasons for cannabis use in patients with a first-episode of psychosis [2,32,38]. Findings showed enhancement andsocial reasons to be mostly reported as reasons for use but allstudies focused primarily on adolescents and two of them [2,38]utilized qualitative methods only. A systematic review byKolliakou et al. [21] found little support for the self-medicationhypothesis with studies mostly reporting enhancement, ‘gettinghigh’ and alleviating negative states as the main reasons forcannabis use by patients with psychosis. No study, to date, hasinvestigated reasons for cannabis use longitudinally. It isimportant to explore longitudinal changes in reasons as theymight be malleable constructs susceptive to intervention and differwith time and circumstance. Finally, research so far has not utilizedUrinary Drug Screen (UDS) to corroborate use and explore withregards to different reasons.

The aims of this study are to:

� investigate how strength in endorsement of reasons for cannabisuse may vary over 12 months;� explore the specific differences between self-reported reasons

for cannabis use at study entry and then 3 and 12 months later;� examine how meaningful alternative confirmation of cannabis

use is by collecting UDSs at all 3 time-points, in a sample ofpatients presenting with a first-episode of psychosis.

We expected that the strength of endorsement wouldweaken over time with patients assigning lesser importanceto reasons at 3 and 12 months compared to baseline. Based onthe relevant literature, we anticipated that, at all 3 time-points,alleviation of negative affect would be the most stronglyendorsed reason for cannabis use. We did not anticipatedifferences in reasons for cannabis use between patients witha positive or negative UDS.

2. Methods

This study was nested within the Physical health and substanceUse Measures in first-onset Psychosis (PUMP) study, part of theIMPaCT (Improving Health and Reducing Substance Use inPsychosis) research programme. PUMP is a prospective observa-tional cohort, of patients with a FEP, which was followed up for12 months to test the association between lifestyle (diet, exerciseand substance use) and antipsychotic medication factors with (a)psychotic symptoms and (b) components of the metabolicsyndrome. Written informed consent was obtained from allpatients after study had been explained and ethics approval wasgranted by the Ethics Committee at the Institute of Psychiatry,King’s College London, UK.

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx 3

G Model

EURPSY-3190; No. of Pages 8

2.1. Subjects

Patients were consecutively approached for participation in thestudy if they:

� were aged 18–65 years;� had presented to statutory mental health services with a first-

onset psychotic disorder (ICD-10 definition of functionalpsychosis);� had experienced positive psychotic symptoms of at least 7 days

duration.

Patients were excluded from the study if they met criteria for:

� psychotic symptoms precipitated by an organic cause;� moderate or severe learning disability;� major medical or neurological illness;� insufficient command of English to complete assessments;� contact with health services for psychosis longer than 6 months

from identification;� pregnancy.

2.2. Recruitment

A sample of 321 patients was recruited to the PUMP study fromservices in the South London and Maudsley (SLaM), the Oxleas andthe Sussex NHS Trust in the UK between May 2008 and July 2011. Asummary of the recruitment and completion rates at the 3 time-points is provided in Fig. 1.

Sixty-nine of the 90 current users completed the Reasons forUse Scale (RFUS) [43] at baseline. However, 2 patients did notcomplete the item for ‘coping with unpleasant affect’ and 1 patientfor items ‘conformity and acceptance’ and ‘relief of positivesymptoms and side effects’ hence data from 67 and 68 users,respectively, are available for these variables. The 3-month follow-up assessments were conducted on only those patients recruited

Fig. 1. Recruitment and

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

from May 2008 to December 2010 as primary aim was to completethe 12-month follow-up assessments. Sixty-five patients weretherefore not re-contacted at 3 months. Twenty-nine of the30 current users completed the RFUS at 3 months. At 12 months,36 of the 37 current cannabis users completed the RFUS.

At baseline, 60% of participants were recruited from inpatientservices whereas at 3- and 12-month follow-up, 10% and 7% ofparticipants were hospitalised, respectively.

2.3. Measures

At each time-point, current cannabis use was ascertained byself-report; study participants were asked if they were presentlyusing cannabis.

The RFUS (Table 1) was used to assess reasons for cannabis use.This is a 26-item self-report instrument that includes items fromthe Drinking Motives Questionnaire [6] and additional motivesspecific to symptoms of mental illness. Internal reliability of thescale has been demonstrated in patients with psychotic disorders[43]. The instrument comprises 5 subscales that are believed toreflect a participant’s reasons for drug use and utilises a 5-pointscale that reflects how often cannabis is used for each reason:never/almost never (0), some of the time (1), half of the time (2),often (3), and almost always/always (4). Participants are invited torate how much of the time each subscale item is a reason for theircannabis use. The mean score is then calculated for each subscale.

For the Urine Drug Test (UDS), the Concateno Drug Screen TestCup (urine) was used as to corroborate self-reported cannabis use.The test is a lateral flow chromatographic immunoassay for thequalitative detection of drugs and drug metabolites in urine for in

vitro diagnostic use [48].

2.4. Statistical analysis

A random intercept model was used to longitudinally testwhether mean subscale scores on the RFUS changed over

completion rates.

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

Table 1Subscales and subscale items of the RFUS.

Subscales Subscale items

Enhancement Because it makes you feel good

Because it’s fun

To get high

Social motives Because it’s what most of your friends do

when you get together

Because it makes a social gathering more enjoyable

As a way to celebrate

To be sociable

Coping with

unpleasant affect

Because it helps when you are feeling nervous

Because it helps when you are feeling depressed

To forget your worries

To feel more motivated

To make it easier to sleep

To help me concentrate

Because you feel more self-confident or

sure of yourself

To relieve boredom

To decrease restlessness

To slow down racing thoughts

To relax

Conformity and

acceptance

So you won’t feel left out

To be liked

To help you talk to others

To be part of a group

Because your friends pressure you to do it

Relief of positive

symptoms and

side effects

To get away from the voices

To reduce side effects of medication

To feel less suspicious/paranoid

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx4

G Model

EURPSY-3190; No. of Pages 8

12 months. This model takes into account the fact that observa-tions from the same individual will not be independent. Fixedeffects of reason (5 subscales) and time (baseline, 3 months and12 months) were included in the model, in addition to a randomeffect of participant. This model allowed cannabis users who haddata at any of the time-points to contribute to the analysis – unlikea repeated-measures ANOVA test, which excludes subjects withany missing values.

A series of paired-sample t-tests was carried out to test thespecific differences between mean scores on each subscale of theRFUS at baseline, 3 months and 12 months. Ten pairs were createdfrom the 5 subscales at each time-point. A Bonferroni correctionwas applied to adjust for multiple testing.

Mann-Whitney tests were carried out to test differences oneach subscale at each time-point between patients with positiveand negative UDSs.

All analyses were conducted using the Statistical Package forthe Social Sciences Version 20 (SPSS 20; [31]).

3. Results

Age of participants at baseline ranged from 18–60 years(median = 25 years). At baseline, 53% (36) of participants reportedtheir ethnic background as white, 34% (23) as black and 13% (9) asother. At 3 and 12 months, respectively, 55% (16) and 58% (21)reported their ethnic background as white, 28% (8) and 39% (14) asblack and 17% (5) and 3% (1) as other. The majority of patients weremale at all three time-points; 75% (51) at baseline, 90% (26) at3 months and 61% (16) at 12 months.

Since the random intercept model assumes that observationsfrom the same individual are not independent, cannabis userswhose RFUS scores were available at one time-point but reportednot using cannabis at another, were excluded from the analysisaltogether. Thus, 50 patients’ scores at baseline, 26 at 3 months and25 at 12 months were utilised for this model. Relief of positive

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

symptoms and side effects’ and ‘12 months’ were used as thebaseline values in the random intercept model. We found strongevidence of an effect of both time (F(2, 591) = 3.7, P = 0.024) andreason (F(4, 559) = 119, P < 0.001). Mean (M) subscale scores wereexamined and ‘enhancement’ reasons (M = 3.1) were morestrongly endorsed than ‘relief of positive symptoms and sideeffects’ (M = 1.5; t = 18.2, P < 0.001). ‘Social motives’ were alsorated more highly (M = 2.3) than ‘relief of positive symptoms andside effects’ (M = 1.5; t = 7.9, P < 0.001). Similarly, ‘coping withunpleasant affect’ was rated more highly (M = 2.5) than ‘relief ofpositive symptoms and side effects’ (M = 1.5), (t = 8.7, P < 0.001)(Fig. 2).

We found evidence of an effect of time when controlling forreason. On average, scores on any subscale were .23 points higherat baseline than 12 months (t = 2.6, P = 0.010) and .21 points higherat 3 months than 12 months (t = 2.1, P = 0.036) (Table 2).

After a Bonferroni correction adjusting the statistical signifi-cance level (0.05/30 = 0.0017), 22 t-tests remained significant. Atbaseline, enhancement was rated by participants much morehighly than social use (t(68) = 6.57, P < 0.001), coping withunpleasant affect (t(66) = 6.61, P < 0.001), conformity and accep-tance (t(67) = 11.93, P < 0.001) and relief of positive symptomsand side effects (t(67) = 11.17, P < 0.001), while social motive wasrated more highly than conformity and acceptance (t(67) = 7.90,P < 0.001) and relief of positive symptoms and side effects(t(67) = 5.76, P < 0.001). Coping with unpleasant affect wasendorsed more highly than conformity and acceptance(t(66) = 7.80, P < 0.001) and relief of positive symptoms and sideeffects (t(66) = 9.03 P < 0.001). These results confirm the baselineobservations from the random intercept model.

A similar pattern was also shown at 3 and 12 months,respectively, with enhancement again endorsed more stronglythan social use (t(28) = 3.82, P < 0.001/t(35) = 4.74, P < 0.001),coping with unpleasant affect (t(35) = 4.81, P < 0.001), conformityand acceptance (t(28) = 7.49, P < 0.001/t(35) = 10.37, P < 0.001)and relief of positive symptoms and side effects (t(28) = 5.40,P < 0.001/t(35) = 7.57, P < 0.001). Social motive was rated morehighly than conformity/acceptance (t(28) = 4.68, P < 0.001/t(35) = 6.10, P < 0.001) and relief of positive symptoms and sideeffects (t(35) = 3.47, P = 0.001). Coping with unpleasant affect wasendorsed more strongly than conformity and acceptance(t(28) = 8.06, P < 0.001/t(35) = 9.02, P < 0.001) and relief of posi-tive symptoms and side effects (t(28) = 7.10, P < 0.001/t(35) = 5.96, P < 0.001). Table 3 summarises the mean scores onthe RFUS subscales at each time-point.

Of the 37 cannabis users who provided a UDS at baseline, 40%(15) were positive and 60% (22) were negative. At 3 months,24 screens were available: 70% (17) positive and 30% (7) negative.At 12 months there were 82% (18) positive and 18% (4) negativeresults from 22 available screens. At baseline, patients with apositive UDS rated ‘coping with unpleasant affect’ more highlythan those with a negative UDS (U = 244.5, P = 0.013). No otherdifferences at any time-point were observed between the twogroups.

4. Discussion

Much empirical research suggests that patients with psychosisuse cannabis for the same reasons as the general population[12,33] – for its ‘enhancing’ effects e.g. ‘to get high’ and also toreduce negative states such as depression and dysphoria[15,21,33,47]. Social reasons are also important but relief ofpositive symptoms and medication side effects have usually beenfound to be the least common motives [15,21,33,47]. Although theidea that patients take cannabis for relief of positive symptoms ormedication side effects cannot be dismissed, the evidence points

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

Fig. 2. Box-plot of changes in reasons for cannabis use at the 3 time-points.

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx 5

G Model

EURPSY-3190; No. of Pages 8

towards enhancement and alleviation of dysphoria as the mostfrequent reasons for use of cannabis by both patients withpsychosis and healthy subjects [1,2,9,10,12,21,32,33,35,37,38,47].

However, studies have employed different criteria for diagnosisof psychosis or schizophrenia [35,37] and others have beenunderpowered [10]. A variety of techniques and instruments havebeen used with both interviews and questionnaires being utilizedwith reasons identified through free response, open-endedquestions or fixed lists [1,9,10,12,32,33,35,37]. Studies have alsomostly reported results from outpatient populations with schizo-phrenia-spectrum disorders only [1,9,10,12,35,37] and those whorecruited patients with a first-episode of psychosis primarilyfocused on adolescents [2,32,38]. In the current study, weaddressed the methodological issues above by applying standar-dized diagnostic criteria, included adult inpatients and outpatientswith a first presentation of any psychotic disorder and utilized avalidated instrument to assess reasons for use.

Our results showed that patients with FEP who used cannabisfelt more strongly about their reasons for use at baseline than at

Table 2Random intercept model testing the differences in reasons for cannabis use while

controlling for time.

Effect size 95% CI Sig

Reason

Relief of positive symptoms

and side effects

Enhancement 1.75 1.56–1.94 < 0.001

Social motive 0.76 0.57–0.95 < 0.001

Coping with unpleasant affect 0.85 0.66–1.04 < 0.001

Conformity and acceptance �0.67 �0.26–0.12 0.482

Time

12 months

Baseline 0.23 0.06–0.41 0.010

3 months 0.21 0.01–0.41 0.036

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

3 months and 12 months follow-up. Although previous research onchanges in the strength of reason endorsement over time isminimal, we anticipated that contact with services and treatmentfor psychosis over 12 months from first presentation wouldincorporate some intervention targeting cannabis use. Over timepatients might have had the opportunity to explore the reasons fortheir cannabis use and possible effects it had on their illnesstrajectory, which may have reduced the level of certainty they feltover what supported their use. It has also been previouslysuggested that the reasons patients endorse for their cannabisuse do not always appear to be justified by the subsequentperceived effects of the actual use [1,12,45]. For instance, patientshave reported using cannabis to relieve anxiety or depression or torelax but did not describe relief from these symptoms or achievingrelaxation after use [1,12,45]. In our opinion, similarly withencouraging patients to explore their reasons, supporting them inrecognizing the perceived and actual effects of cannabis use andemphasizing any discrepancies is likely to have lessened theconfidence in the reasons they had previously consideredimportant for their use.

Patients, at each time-point, endorsed ‘enhancement’ morehighly as a reason for their use than any other motive. ‘Coping withunpleasant affect’ followed closely but was not rated more highlythan ‘social motive’. Relief of positive symptoms or medication side

Table 3Mean RFUS subscale scores at each time.

Subscale Baseline 3 months 12 months

Enhancement 3.2 3.2 3.2

Social motive 2.3 2.2 2.0

Coping with unpleasant

affect

2.3 2.5 2.1

Conformity and acceptance 1.5 1.3 1.1

Relief of positive symptoms

and side effects

1.4 1.6 1.4

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx6

G Model

EURPSY-3190; No. of Pages 8

effects was the least endorsed motive. These data support previousstudies, which showed no evidence for the self-medicationhypothesis [1,9,10,12,35,37]. However, only two studies [10,12]have reported that enhancing effects were rated most highly forcannabis use among patients with psychosis. The rest of the studies[1,9,35,37] reported that coping with unpleasant affect – mainlyrelief of boredom and anxiety – were much more stronglyendorsed than any other reason followed by enhancement andsocial use.

Finally, results from UDS analysis in our study challenge thenotion that UDS to corroborate cannabis use is meaningful incertain settings. At baseline, we were not surprised at the low rateof positive UDS from patients who self-reported cannabis use asthe majority of patients were hospitalised and unlikely to haveused cannabis around the time of assessment. Although chronicuse may result in positive UDS for an extended period, thedetection window for cannabis is typically between 2–4 days [8]and our results were very much expected. It is notable that thesepatients still considered themselves cannabis users and lack ofaccess was not confused with voluntary abstinence. Swartz et al.[44] also reported that although 16.3% of patients with schizo-phrenia reported substance use, only 12.4% produced a positiveurine screen. However, a high level of agreement between patients’self-reported cannabis use and UDS was reported by Hides et al.[17] but only 8 of their 49 UDS samples were collected duringhospital admission. The higher agreement between self-report andurine screening at 3 and 12 months in our study also validatesthese findings. Our results highlight the variation in frequency andquantity of use in patients routinely classified as self-reportedcannabis users and further add to existing evidence that self-reported cannabis use in inpatient settings might be sufficient forclassification without the need for biomedical screening[39,49]. However, depending on the study design and samplecharacteristics, relying solely on self-report measures for drug usein patients with schizophrenia should be treated with caution asneurocognitive impairments may restrict accurate reporting[3]. Gathering information from multiple sources such as collateralreports and biological testing might in some cases provide morerobust results [28]. At baseline, users with positive UDS endorsed‘coping with unpleasant affect’ more strongly than those withnegative tests. This difference is largely justified in a population ofyoung adults who must have been unquestionably distressed bythe emergence of their first psychotic symptoms and experience ofhospital admission. As the majority of participants at 3 and12 months were outpatients – and likely to be more stable – copingwith unpleasant affect was not rated more highly by users with apositive UDS at these subsequent time-points.

The main strength of this study is that it utilised a validated andreliable measure for assessing reasons for cannabis use in patientswith psychosis. Patients with schizophrenia are known to appearmore suggestible [23] so using a Likert scale rather than yes/noendorsement statements yields more nuanced responses,increases variability and strengthens significant observations.Another strength is the sample size at baseline, which is the largestto date; the smaller sample sizes at 3 and 12 months as well as theone utilised for the random intercept model are still relatively largecompared to previous research [10]. The consistency of ourfindings observed at the 3 time-points suggests that results areunlikely to be due to chance even with the given sample sizevariability.

Inclusion of only outpatients has occasionally been intentionalto avoid bias from severely ill patients who might have alreadyfailed in an attempt to minimize their symptoms by self-medication and would be less likely to choose this reason fortheir cannabis use [10]. Although a majority of inpatientsparticipated at baseline and most of the patients at 3 and

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

12 months were treated in the community, neither inpatientsnor outpatients rated ‘relief of positive symptoms and medicationside effects’ highly. The inclusion of hospitalised patients did notproduce different results to those reported from research onoutpatient populations only. The inclusion of patients withpsychosis spectrum diagnoses and FEP in particular is anotherstrength as findings contribute to the evidence-base by showingthat reasons for cannabis use remain fairly similar in psychoticpopulations and are consistent with results from research inschizophrenia-only patients [1,9,10,12,35,37].

4.1. Limitations

Our study did not take into account the different classes ofmedication by which patients were treated. The only study thatconsidered this difference found that fewer patients treated withatypical rather than typical antipsychotics were likely to report thatcannabis improved the feeling of being ‘slowed down’ caused by themedication [1]. However, Schaub et al. [35] showed that patientstreated with atypical antipsychotics chose only one reason moreoften than healthy controls and that was using cannabis to ‘reduceboredom’. There is a possibility that different classes of antipsycho-tics as well as treatment with antidepressant or anxiolyticmedication or a combination of both might lead to over- orunder-stating the importance that certain reasons have in sustainingcannabis use. This would also obscure findings when investigatingreasons for cannabis use in treated and medication-free patients,although no research has examined this distinction so far.

Although the sample size was adequately large for thisinvestigation, the number of patients completing the RFUS wasnot sufficient to allow for further analysis based on extent ofcannabis exposure in terms of frequency or quantity. Patientsclassified as ‘abusing’ cannabis have been shown to endorse illnessand medication-related reasons for use more often than cannabisusers [9]. Similarly, in the study by Schofield et al. [37], greateramount and frequency of cannabis use was observed when sideeffects were motivators. On the contrary, daily cannabis users weremore likely to state they used cannabis to increase pleasure and goalong with the group [35]. None of the studies observed anincreased report of self-medication reasons for use across thedifferent cannabis groups. Lack of endorsement of self-medicationmotivators in our study might have been due to infrequentcannabis use. This was not possible to consider although studiesthat examined different patterns of use have not consistentlyreported a greater incentive to relieve positive symptoms ormedication side effects by patients using cannabis at higherfrequency or greater amounts.

According to the Positive and Negative Syndrome Scale (PANSS;[18]), psychopathology levels were low in our sample. The PANSS isa measure of current psychopathology and is interpreted as eithera total score (range 30–210) or grouped into severity categories ofmildly ill (< 58), moderately ill (58–74), markedly ill (75–95) andseverely ill (> 96) [25]. It also comprises of positive (score 7 to 49),negative (score 7 to 49), and general (score 16 to 112) symptomsubscales. Of patients who completed the RFUS, 84%, 91% and 92%were assessed as mildly or moderately ill at baseline, 3 months and12 months, respectively. No patient scored on the severe end of thePANSS spectrum (> 96) at 3 months and only one was assessed asseverely ill at baseline and 12 months. For ethical reasons, patientswere only approached when judged well enough to be providedwith information about the study and invited to consent. It is onlynatural that within this time patients’ mental health improved. Thelow positive symptomatology in our sample could have affectedthe results, similarly. Median scores in the positive symptomssubscale were 15 (IQR = 21-11), 13 (IQR = 18-9) and 11 (IQR = 15-8) at baseline, 3 months and 12 months, respectively. It should be

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx 7

G Model

EURPSY-3190; No. of Pages 8

argued that low total and positive symptom severity indicatesrelative recovery which could have prevented patients fromchoosing self-medication motives and allowed them to enjoycannabis for other reasons, which they were more likely to endorseas important. Low scores on the negative symptoms subscale [14(IQR = 20-11), 15 (IQR = 21-10) and 14 (IQR = 20-10)] make itunlikely that endorsement of social motives were under-reporteddue to demotivation or impairments in social functioning.

4.2. Research and clinical implications

Our study has shown that strength in the endorsement ofreasons for cannabis use may change in response to social,emotional and environmental factors and should be examined aspart of a continuous process to identify triggers for relapse.Similarly to the difference between initiating and sustainingmotives for use [4], reasons for cannabis use may vary at differentstages of treatment and recovery. It is very promising that self-reported reasons for use appear to validate an existing clinicalpicture [22] as in a clinical setting where standardised measuresare not routinely used to assess reasons for use, observation andactive listening can help identify what sustains cannabis use. Thiscan also be an effective strategy with patients who are reluctant toopenly discuss cannabis use per se.

Patients in our study widely endorsed reasons for cannabis useto effect psychological enhancement and social enjoyment and torelieve unpleasant states unrelated to their illness. Exploringperceived effects of cannabis use and challenging positiveexpectations might be advantageous to patients whose primarymotive is psychological enhancement. Patients, who use cannabisfor social reasons, may benefit from assertiveness training toencourage exploration of their patterns, identify high-risk situa-tions and feel socially comfortable in the absence of cannabis. Forthe minority, who use cannabis to relieve positive symptoms andmedication side effects, psycho-education could be a valuablestarting point of intervention. As Gregg et al. [14] proposed,patients who use drugs primarily for social and enhancementreasons should benefit from interventions such as MotivationalInterviewing, whereas those who use drugs primarily for emo-tional reasons (e.g. negative states) might find Cognitive Beha-vioural Therapy more beneficial. It is likely that such patient-centred strategies would be helpful to patients with psychosisattempting to reduce or give up their cannabis use.

Future research should aim to utilize validated instruments anddiagnostic criteria for psychosis and cannabis exposure classifica-tion in order to facilitate the generalizability of results. Studiescomparing reasons given by in- and outpatients with varyingdegrees of psychopathology would be beneficial in establishingwhether severity of symptoms influences the endorsement ofreasons. Inclusion of patients with any psychotic diagnosis shouldalso be used. Cannabis exposure classification would highlight anyeffects that frequency or quantity of use might independently haveon reasons for use. Finally, corroborating self-reported cannabisuse with biomedical testing could be an advantage to therobustness of results but should be interpreted with cautionwhen studying inpatient populations.

Awareness of patients’ reasons for cannabis use would allowclinicians to plan psychotherapeutic and pharmacological inter-ventions to target widely endorsed incentives such as relief fromdysphoria, psychological enhancement and social motives as wellas positive psychotic symptoms and medication side effects, whichfor some patients might still be strong motivators in sustainingtheir cannabis use. Taking into consideration worldwide legisla-tion changes in the last decade, cannabis will inadvertentlybecome even more accepted and widely available as a recreationaldrug of use. Identifying which reasons maintain use in patients

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

with an early onset of psychosis is the first step in protecting thosemost vulnerable from its harmful effects.

Disclosure of interest

The authors declare that they have no conflicts of interestconcerning this article.

Funding: This paper summarises independent research fundedby the National Institute for Health Research (NIHR) under itsIMPACT Programme (Grant Reference Number RP-PG-0606-1049).The views expressed are those of the authors and not necessarilythose of the National Health Service, the NIHR or the Department ofHealth.

References

[1] Addington J, Duchak V. Reasons for substance use in schizophrenia. ActaPsychiatr Scand 1997;96:329–33.

[2] Archie S, Boydell KM, Stasiulis E, Volpe T, Gladstone BM. Reflections of youngpeople who have had a first episode of psychosis: what attracted them to usealcohol and illicit drugs? Early Interv Psychiatry 2013;7(2):193–9.

[3] Bahorik AL, Newhill CE, Queen CC, Eack SM. Under-reporting of drug useamong individuals with schizophrenia: prevalence and predictors. PsycholMed 2014;44(1):61–9.

[4] Baigent M, Holme G, Hafner RJ. Self-reports of the interaction betweensubstance abuse and schizophrenia. Aust N Z J Psychiatry 1995;29(1):69–74.

[5] Barnett JH, Werners U, Secher SM, Hill KE, Brazil R, Masson K, et al. Substanceuse in a population-based clinic sample of people with first-episode psychosis.Br J Psychiatry 2007;190:515–20.

[6] Cooper ML. Motivations for alcohol use among adolescents: development andvalidation of a four-factor model. Psychol Assess 1994;6:117–28.

[7] Dixon L, Haas G, Weiden PJ, Sweeney J, Frances AJ. Drug abuse in schizophrenicpatients: clinical correlates and reasons for use. Am J Psychiatry 1991;148:224–30.

[8] Donovan DM, Bigelow GE, Brigham GS, Carroll KM, Cohen AJ, Gardin JG, et al.Primary outcome indices in illicit drug dependence treatment research:systematic approach to selection and measurement of drug use end-pointsin clinical trials. Addiction 2012;107(4):694–708.

[9] Fowler IL, Carr VJ, Carter NT, Lewin TJ. Patterns of current and lifetimesubstance use in schizophrenia. Schizophr Bull 1998;24:443–5.

[10] Goswami S, Mattoo SK, Basu D, Sing G. Substance-abusing schizophrenics: dothey self-medicate? Am J Addict 2004;13:139–50.

[11] Grech A, Van Os J, Jones PB, Lewis SW, Murray RM. Cannabis use and outcomeof recent onset psychosis. Eur Psychiatry 2005;20(4):349–53.

[12] Green B, Kavanagh DJ, Young RMCD. Reasons for cannabis use in men with orwithout psychosis. Drug Alcohol Rev 2004;23:445–53.

[13] Green B, Young R, Kavanagh D. Cannabis use and misuse prevalence amongpeople with psychosis. Br J Psychiatry 2005;187:306–13.

[14] Gregg L, Barrowclough C, Haddock G. Reasons for increased substance use inpsychosis. Clin Psychol Rev 2007;27:494–510.

[15] Gregg L, Haddock G, Barrowclough C. Self-reported reasons for substance usein schizophrenia: a Q methodological investigation. Ment Health Subst Use2009;2(1):24–39.

[16] Harrison I, Joyce EM, Mutsatsa SH, Hutton SB, Huddy V, Kapasi M, et al.Naturalistic follow-up of co-morbid substance use in schizophrenia: the WestLondon first-episode study. Psychol Med 2007;38:79–88.

[17] Hides L, Dawe S, Kavanagh DJ, Young RM. Psychotic symptom and cannabisrelapse in recent-onset psychosis. Prospective study. Br J Psychiatry2006;189:137–43.

[18] Kay SR, Fiszbein A, Opler LA. The Positive and Negative Syndrome Scale(PANSS) for schizophrenia. Schizophr Bull 1987;13:261–76.

[19] Khantzian EJ. The self-medication hypothesis of addictive disorders. Am JPsychiatry 1985;142:1259–64.

[20] Khantzian EJ. The self-medication hypothesis of substance use disorders: areconsideration and recent applications. Harv Rev Psychiatry 1997;4:231–44.

[21] Kolliakou A, Joseph C, Atakan A, Ismail K, Murray RM. Why do patients withpsychosis use cannabis and are they ready to change their use? Int J DevNeurosci 2011;29:335–46.

[22] Kolliakou A, Ismail K, Fusar-Poli P, Atakan Z. Why do psychotic patients usecannabis? Case series. Curr Pharm Des 2012;18:4950–9.

[23] Kot T, Serper M. Increased susceptibility to auditory conditioning in halluci-nating schizophrenic patients: a preliminary investigation. J Nerv Ment Dis2002;190:282–8.

[24] Laudet AB, Magura S, Vogel HS, Knight EL. Perceived reasons for substancemisuse among persons with a psychiatric disorder. Am J Orthopsychiatry2004;74(3):365–75.

[25] Leucht S, Kane JM, Kissling W, Hamann J, Etscel E, Engel RR. What does thePANSS mean? Schizophr Res 2005;79(2–3):231–8.

[26] Linszen DH, Dingemans PM, Lenior ME. Cannabis abuse and the course ofrecent-onset schizophrenic disorders. Arch Gen Psychiatry 1994;51(4):273–9.

[27] McCreadie RG. Use of drugs, alcohol and tobacco by people with schizophre-nia: case-control study. Br J Psychiatry 2002;181:321–5.

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007

A. Kolliakou et al. / European Psychiatry xxx (2014) xxx–xxx8

G Model

EURPSY-3190; No. of Pages 8

[28] McHugo GJ, Drake RE, Brunette MF, Haiyi Xie, Essok SM, Green AI. Enhancingvalidity in co-occurring disorders treatment research. Schizophr Bull 2006;32(4):655–65.

[29] Miles H, Johnson S, Amponsah-Afuwape S, Finch E, Leese M, Thornicroft G.Characteristics of subgroups of individuals with psychotic illness and acomorbid substance use disorder. Psychiatr Serv 2003;54:554–61.

[30] Mueser KT, Drake RE, Wallach MA. Dual diagnosis: a review of etiologicaltheories. Addict Behav 1998;23(6):717–34.

[31] Statistical Package for the Social Sciences. Version 20 for Windows. Chicago:SPSS Inc; 2011.

[32] Pencer A, Addington J. Reasons for using substances in adolescents with andwithout psychosis. Early Interv Psychiatr 2008;2:42–4.

[33] Perez LG, Santacana AM, Baquero DB, Perez-Sola V. Reasons and subjectiveeffects of cannabis use among people with psychotic disorders: a systematicreview. Actas Esp Psiquiatr 2014;42(2):83–90.

[34] Regier DA, Farmer ME, Rae DS, Locke BZ, Keith SJ, Judd LL, et al. Comorbidity ofmental disorders with alcohol and other drug abuse. Results from the Epide-miologic Catchment Area (ECA) study. JAMA 1990;264:2511–8.

[35] Schaub M, Fanghaenel K, Stohler R. Reasons for cannabis use: patients withschizophrenia versus matched healthy controls. Aust N Z J Psychiatry 2008;42:1060–5.

[36] Schneier FR, Siris SG. A review of psychoactive substance use and abuse inschizophrenia patterns of drug choice. J Nerv Ment Dis 1987;175(11):641–52.

[37] Schofield D, Tennant C, Nash L, Degenhardt L, Cornish A, Hobbs C, et al. Reasonsfor cannabis use in psychosis. Aust N Z J Psychiatry 2006;40:570–4.

[38] Seddon LJ, Copello A, Birchwood M. Cannabis use and abstention in first-episodepsychosis: the participants’ view. Mental Health Subst Use 2013;6(1):47–58.

[39] Selten JP, Bosman IJ, de Boer D, Veen ND, van der Graaf Y, Maes RA, et al. Hairanalysis for cannabinoids and amphetamines in a psychosis incidence study.Eur Neuropsychopharmacol 2002;12:27–30.

Please cite this article in press as: Kolliakou A, et al. Reasons for cannachange over 12 months? European Psychiatry (2014), http://dx.doi.

[40] Sevy S, Robinson DG, Solloway S, Alvir JM, Woerner MG, Bilder R, et al.Correlates of substance misuse in patients with first-episode schizophreniaand schizoaffective disorder. Acta Psychiatr Scand 2001;104:367–74.

[41] Sorbara F, Liraud F, Assens F, Abalan F, Verdoux H. Substance use and thecourse of early psychosis: a 2-year follow-up of first-admitted subjects. EurPsychiatry 2003;18:133–6.

[42] Spencer C, Castle D, Michie PT. Motivations that maintain substance useamong individuals with psychotic disorders. Schizophr Bull 2002;28(2):233–47.

[43] Spencer C, Castle D, Michie PT. Motivations that maintain substance useamong individuals with psychotic disorders. Schizophr Bull 2002;28:233–47.

[44] Swartz MS, Swanson JW, Hannon MJ. Detection of illicit substance use amongpersons with schizophrenia by radioimmunoassay of hair. Psychiatr Serv2003;54(6):891–5.

[45] Swendsen J, Ben-Zeev D, Granholm E. Real-time electronic ambulatory moni-toring of substance use and symptom expression in schizophrenia. Am JPsychiatry 2011;168(2):202–9.

[46] Test MA, Wallisch LS, Allness DJ, Ripp K. Substance use in young adults withschizophrenic disorders. Schizophr Bull 1989;15(3):465–76.

[47] Thornton LK, Baker AL, Lewin TJ, Kay-Lambkin FJ, Kavanagh D, Richmond R,et al. Reasons for substance use among people with mental disorders. AddictBehav 2012;37:427–34.

[48] Tietz NW. Textbook of Clinical Chemistry. W.B. Saunders Company; 1986.[49] Wolford GL, Rosenberg SD, Drake RE, Mueser KT, Oxman TE, Hoffman D, et al.

Evaluation of methods for detecting substance use disorder in persons withsevere mental illness. Psychol Addict Behav 1999;13:313–26.

[50] Zammit S, Moore TH, Lingford-Huges A, Barnes TR, Jones PB, Burke M, et al.Effects of cannabis use on outcomes of psychotic disorders: systematic review.Br J Psychiatry 2008;193(5):357–63.

bis use in first-episode psychosis: Does strength of endorsementorg/10.1016/j.eurpsy.2014.10.007