x (2007) xxx–xxx

+ MODEL

www.elsevier.com/locate/psychres

ARTICLE IN PRESS

Psychiatry Research x

Characteristics of smokers with a psychotic disorder and implicationsfor smoking interventions

Amanda Baker a,b,⁎, Robyn Richmond c, Melanie Haile a, Terry J. Lewin a,b,Vaughan J. Carr a,b, Rachel L. Taylor c, Paul M. Constable a, Sylvia Jansons c,

Kay Wilhelm d, Kristen Moeller-Saxone e

a Centre for Mental Health Studies, University of Newcastle, University Drive, Callaghan, NSW 2308, Australiab Neuroscience Institute of Schizophrenia and Allied Disorders, Darlinghurst, Sydney, NSW 2010, Australia

c School of Public Health and Community Medicine, University of New South Wales, Sydney, NSW 2052, Australiad School of Psychiatry, University of New South Wales, Sydney, NSW 2052, Australia

e SANE Australia, 153 Park Street, South Melbourne, Victoria 3205, Australia

Received 4 November 2005; received in revised form 31 March 2006; accepted 14 May 2006

Abstract

Despite high rates of smoking among people with psychotic disorders, and the associated health and financial burden, fewstudies have investigated the characteristics of this group of smokers. This paper reports data from 298 smokers with an ICD-10psychotic disorder residing in the community (56.7% with schizophrenia or schizoaffective disorder), including an examination oftheir demographic and clinical characteristics, smoking behaviours, severity of nicotine dependence, stage of change, and reasonsfor smoking and for quitting. Standardized self-report instruments were used, in conjunction with structured interviews, as part ofthe first phase of a randomized controlled trial. On average, participants smoked 30 cigarettes per day, commenced smoking dailyat about 18 years of age (5 years before illness onset), and had made 2–3 quit attempts in their lifetime. Higher levels of nicotinedependence and concurrent hazardous use of alcohol or cannabis were associated with a younger age at smoking initiation. Thepresent sample was also more likely to report stress reduction, stimulation and addiction as reasons for smoking, compared to ageneral sample of smokers. Males, precontemplators and participants with concurrent hazardous substance use cited fewer reasonsfor quitting smoking. These and other subgroup differences in smoking characteristics are used to illustrate potential implicationsfor the nature and timing of smoking interventions among people with a psychotic disorder.© 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Australia; Psychotic disorders; Smoking; Smoking cessation; Motivation; Nicotine

⁎ Corresponding author. Centre for Mental Health Studies, Univer-sity of Newcastle, University Drive, Callaghan, NSW 2308, Australia.Tel.: +61 2 4924 6610; fax: +61 2 4924 6608.

E-mail address: Amanda.Baker@newcastle.edu.au (A. Baker).

0165-1781/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights resedoi:10.1016/j.psychres.2006.05.021

1. Introduction

1.1. Smoking and psychosis

The prevalence of smoking among people withpsychotic disorders is very high. A recent meta-analysisof 42 studies across 20 countries (De Leon and Diaz,2005) reported that the odds of people with schizophrenia

PSY-05508; No of Pages 12

rved.

2 A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

or schizoaffective disorder being current smokers was 5.3times higher compared to the general population. Thecorresponding odds, relative to other severe mentaldisorders, were 1.9, even after controlling for potentialconfounding variables. Such high rates of smokingamong those with psychotic disorders may contribute tothe 20% reduction in life expectancy reported inschizophrenia (Wilhelm, 1998). Further evidence forsuch an association is the finding that ischaemic heartdisease is the most common cause of death among peoplewith schizophrenia (Lawrence et al., 2001).

Based on the large number of studies describing highrates of smoking among schizophrenia and schizoaffec-tive disorder samples, De Leon and Diaz (2005) haveposited that a biological factor may make these groupsmore prone to smoke. In reviewing the literature on the‘effects of nicotine in populations genotypically and/orphenotypically related to schizophrenia’, Kumari andPostma (2005) have suggested that smoking may serveas a form of self-medication to correct for sensory andcognitive deficits in schizophrenia. They further suggestthat the beneficial effects of nicotine may be explainedin terms of the drug's interaction with the dopaminergicand glutamatergic transmitter systems. In addition,Ziedonis and Williams (2003) have summarized thenumerous psychological and social factors that may actto increase the risk of nicotine addiction among peoplewith psychotic disorders. These include: limited educa-tion; poverty; unemployment; peer pressure; and thevalues held by those within the treatment system. Assome patients start daily smoking after the onset ofschizophrenia, there may be influences from otherpatients and the treatment environment on smokingbehaviours. Certainly, neither mental health nor generalmedical providers routinely diagnose nicotine depen-dence, which would better allow for the opportunity todiscuss smoking cessation (Peterson et al., 2003).

While many smokers with schizophrenia would liketo stop smoking, available data suggest that smokingcessation rates among people with schizophrenia arequite low. The cessation rates reported in De Leon andDiaz's (2005) meta-analyses were 9% for schizophreniaversus 14% to 49% for the general population. Ziedonisand George (1997) have suggested that this may relate tolower motivation to quit. Motivation for smokingcessation refers both to reasons why smokers wish toquit and to the strength of their desire to do so (Marlatt,1988). Little is known about the motivation to quitamong people with a psychotic disorder.

The stages of changemodel, whichmeasures smokers'readiness to quit, has been well validated in generalsamples (Prochaska et al., 2004). Diclemente et al. (1991)

demonstrated that smokers in the precontemplation andcontemplation stages of change were significantly lesslikely to make a quit attempt over a 6-month periodcompared to those in the preparation stages. Those in thepreparation stage were also more successful in their quitattempts. In a recent study of smoking among people whowere inpatients in a psychiatric hospital, Reichler et al.(2001) reported that nearly one-third of the sample(29.5%) were precontemplators, while almost half(45.9%) were either contemplating quitting or reducingtheir smoking and a small number were preparing forchange (13%) or taking action (11.6%). Addington et al.(1997) reported comparable levels of motivation among agroup of people with schizophrenia.

Curry et al. (1990) have posited a model of motiva-tion based on intrinsically versus extrinsically motivatedbehaviour. Intrinsically motivated behaviours are thosefor which the rewards are internal to the person (such asone's health), while extrinsically motivated behavioursare performed in response to external rewards (such asmoney) or punishment. Curry et al. found that success-ful quitters differentiated between intrinsic and extrinsicmotivation and had significantly higher levels of intrin-sic motivation and lower levels of extrinsic motivation.While no formal assessment has been made of the typeof motivation people with schizophrenia have forquitting smoking, one study by Kelly and McCreadie(1999) suggested that many patients would like to quitsmoking for health reasons.

1.2. The current study

Despite the magnitude of the public health burden ofsmoking among people with psychotic disorders,increasing identification of the factors that influencesmoking, and evidence that a sizeable proportion ofpeople with psychotic disorders are prepared to considerquitting smoking, few studies have investigated thecharacteristics of this group of smokers or comparedthem with other groups. Such information may beimportant when designing smoking interventions forpeople with psychotic disorders. Detailed data relatingto the behavioural domains of reasons for smoking,reasons for wanting to quit, type of motivation (intrinsicor extrinsic), level of dependence, and past history ofquit attempts has not typically been collected fromsmokers with psychotic disorders. Moreover, mostprevious studies have focused solely on schizophrenia,not the spectrum of psychotic disorders.

The present paper had three aims: (i) to describe thedemographic and clinical characteristics of smokerswith a psychotic disorder residing in the community; (ii)

3A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

to describe their smoking behaviours, including fre-quency of tobacco consumption, severity of nicotinedependence, stage of change and reasons for smokingand for quitting, using well validated assessmentinstruments; and (iii) where available, to comparethese behaviours with data reported for other samples.This study formed part of a longitudinal randomizedcontrolled trial (RCT) that examined the efficacy ofmotivational interviewing, cognitive behaviour therapy(CBT) and nicotine replacement therapy (NRT) com-pared to usual treatment at 3-, 6- and 12-monthson smoking cessation among people with a psychoticdisorder residing in the community. Baker et al.(in press) details the therapeutic interventions evaluatedin this trial, patterns of treatment engagement andsubject retention, and the treatment outcomes achieved.

2. Methods

2.1. Sample

A total of 473 people were referred to the study, 360met the study's inclusion criteria, of whom 298 (82.8%)attended the baseline assessment, forming the samplefor the present paper (recruitment location: Sydney, 152;Newcastle region, 146). Inclusion criteria were: Aged atleast 18 years; daily consumption of at least 15cigarettes; and fulfillment of ICD-10 diagnostic criteriafor a psychotic disorder. Participants were also requiredto express an interest in quitting smoking. Exclusioncriteria were: Medical conditions that would precludeuse of NRT; acute psychosis (in which case re-assessment 1 month post-screening was arranged); andevidence of cognitive impairment.

During the first 9 months of the recruitment periodanother treatment study among people with a psychoticdisorder was running concurrently in one of therecruitment locations (Newcastle region) aimed atreducing drug and alcohol consumption. Design andmethods for this project are reported elsewhere (Bakeret al., 2006). Potential participants who also met criteriafor weekly use of cannabis on the Opiate TreatmentIndex (Darke et al., 1991), or alcohol consumptionexceeding recommended National Health and MedicalResearch Council (NHMRC) guidelines for hazardoususe in the month prior to baseline interview, werepreferentially channelled into that study for the first9 months of the present study's recruitment period.Subsequently, potential participants from the Newcastleregion with similar characteristics were over-sampledfor the present study. Consequently, overall, 20.5% (30/146) of the Newcastle-recruited participants in the

current study were also using either alcohol or cannabisat hazardous levels, which is comparable to the regionalrate reported by Fowler et al. (1998); for the Sydney-recruited participants, the corresponding rate was lowerat 8.6% (13/152).

2.2. Procedure

Participants were recruited between March 2001 andOctober 2002 from Sydney and the Newcastle regionprimarily by means of notices placed within communitymental health centres. Notices stated that universityresearchers were conducting a project to help smokersquit. When a volunteer contacted the research team, thepurpose and design of the study was described. After aninitial telephone screening, participants were scheduledfor a baseline assessment appointment. Participantswere assured that all information was strictly confiden-tial to the research team who were independent of anytreatment agency and that refusal to participate wouldnot affect their relationship with the agency in any way.Participants were asked to provide written consent totake part in the study. The initial assessment tookapproximately 90 min to complete. Four trainedinterviewers (three psychologists and one psychiatricnurse) conducted the interviews.

2.3. Measures

Data were collected on demographic characteristics,past and present tobacco, alcohol and other drug use andmental health, treatment history, stage of change,nicotine dependence, reasons for smoking and motiva-tion to quit smoking.

2.3.1. Diagnostic measuresDiagnosis was determined using the Diagnostic

Interview for Psychosis (DIP) (Jablensky et al., 2000;Castle et al., 2006), a semi-structured interview thatconfirms diagnosis using the Operational Criteria forPsychosis (OPCRIT) (McGuffin et al., 1991) andproduces a diagnostic classification in accordance withthe International Classification of Diseases, 10thRevision (ICD-10). The DIP also provided informationon socio-demographic characteristics, premorbid ad-justment, current symptoms and medication, socialfunctioning and service utilization.

2.3.2. Measures of current psychopathology andfunctioning

Psychiatric symptomatology was rated using the 24-item Brief Psychiatric Rating Scale (Ventura et al., 1993)

4 A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

in which severity of symptoms is assessed on a series ofseven-point Likert scales, with higher scores indicatinggreater severity (global scores, range: 24–168). Severityof depression was assessed by the Beck DepressionInventory II (BDI-II) (Beck et al., 1988), a 21-item self-report measure pertaining to the previous 2 weeks (range:0–63). Symptoms are rated on four-point Likert scales (0–3), with higher scores indicating more severe depression.Anxiety was measured by the State-Trait Anxiety Inven-tory (Self-Evaluation Questionnaire) (STAI) (Spielberger,1983). This scale differentiates anxiety as a state (S-Anxiety scale) and as a trait (T-Anxiety scale). The State-Anxiety scale consists of 20 statements about how theparticipant feels “right at this moment”, while the Trait-Anxiety scale consists of 20 statements about how theparticipant feels “in general”. Scores are rated on four-point Likert scales, with higher scores indicating moresevere anxiety (range: 20–80 for both scales). Meanscores for neuropsychiatric patients have been reportedto be 47.74 (S.D.=13.24) for state anxiety and 46.62(S.D.=12.41) for trait anxiety (Spielberger, 1983).

General health functioning was measured using the12-item Short Form survey (SF-12), derived from theSF-36. The SF-12 has been shown to be reliable andvalid in clinical and population based applications in theUS and other countries, including Australia (Ware et al.,1996; Lim and Fisher, 1999). The scale has also beenvalidated for use with homeless persons, who, likepeople with mental illness, also tend to experiencebarriers in accessing health care and have a high preva-lence of physical and mental illnesses and substancedependence (Larson, 2002). The scale produces physicalcomponent scores (PCS) and mental health componentscores (MCS) (Ware et al., 1996; Andrews, 2002). Bothscores are designed to have a mean of 50 and a standarddeviation of 10. Lower scores on the PCS and MCSindicate greater disability (i.e., <30=severe disability;30–40=moderate disability; 40–50=mild disability;>50=no disability).

2.3.3. Measures of smokingReasons for smoking were assessed using the 12-

item Reasons for Smoking Questionnaire (RSQ)developed by Pederson et al. (1996) for use among ageneral population of smokers, with two additionalquestions relating specifically to mental illness. Partici-pants responded ‘yes’ or ‘no’ to each of the 14 possiblereasons presented. Following an examination of itemcharacteristics and relationships (e.g., factor analysis),five subscale scores were derived, namely addiction(items: Habit, craving; range: 0–2), stress reduction(items: Relaxation, to take a break, reduce stress; range:

0–3), arousal (items: Peps me up, weight control,enjoyment, to help concentration; range: 0–4), mentalillness (items: Symptoms of illness, medication sideeffects; range: 0–2) and partner smoking (range: 0–1).

Reasons for quitting were assessed using the 20-itemReasons for Quitting (RFQ) scale developed by Curryet al. (1990). This scale includes 10 intrinsic items thatdefine two 5-item sub-dimensions related to healthconcerns and self-control and 10 extrinsic items thatdefine two 5-item sub-dimensions related to immediatereinforcement and social pressure. Motivation to quitsmoking was measured using the 11-item Readiness andMotivation to Quit Smoking Questionnaire (RMQ)developed by Crittenden et al. (1994). This is an elabo-rated stages of readiness scale which subdividesProchaska and Diclemente's (1983) precontemplationstage into three lower stages. The lowest stage of readinessis defined as those people not contemplating quitting orcutting down (PC-1). The second stage is defined as thosepeople not contemplating quitting but seriously thinkingabout cutting down (PC-2) and the third stage is defined asthose who are contemplating quitting but not within6 months (PC-3) (Crittenden et al., 1994). Contemplationand preparation stages of change are the same as in theProchaska and Diclemente (1983) scheme and wereclassified as follows: Not planning to quit in the nextmonth, or planning to quit within 1 month with no 24-h quit attempt in the past year (Contemplator); andplanning to quit within the next month with a 24-h quitattempt in the past year (Preparation).

Nicotine dependence was measured by the Fagers-tom Test for Nicotine Dependence (FTND) (Fagerstromet al., 1996). The FTND is a widely used and validated6-item questionnaire for assessing severity of nicotinedependence, which yields scores ranging from 0 to 10(Heatherton et al., 1991). Typically, scores of 6 or highersignify nicotine dependence (Fagerstrom et al., 1996). Ithas been suggested that establishing a score on theFTND may help predict success at stopping smoking(Heatherton et al., 1991). Mean FTND scores forsmokers in the general population have typically beenfound to range between 3.0 and 4.3 (Fagerstrom et al.,1996). A recent study by De Leon et al. (2002a) reportedmean FTND scores of 6.3 (S.D. =2.2) for U.S.psychiatric patients compared to 3.9 (S.D.=2.6) forU.S. controls. For people with schizophrenia, themean FTND score was 6.2 (S.D.=2.2).

2.4. Statistical analyses

Data were analyzed using SPSS for Windows(Version 12.0). Categorical variables were analyzed

5A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

using chi-square tests, while the main analyses exam-ining subgroup differences on the continuous outcomevariables were analyses of co-variance (ANCOVAs),controlling for potential confounders (e.g., age, gender,diagnostic category), followed by Scheffé post-hoccomparisons. Single sample t-tests were also used tocompare overall profiles for our psychiatric group withselected normative data from other general populationstudies. As a partial control for the number of statisticaltests conducted, the threshold for significance was set atP<0.01.

3. Results

3.1. Sample characteristics

The demographic and clinical characteristics of thesample are displayed in Table 1. The typical participantwas aged in their late 30s, Australian born, single,receiving welfare support, and had not completed seniorhigh school. Schizophrenia was the most commonprimary psychiatric diagnosis. Service utilization andillness history profiles were generally consistent with

Table 1Demographic and clinical characteristics of the sample at pre-treatment asse

Demographic characteristics IllnessMean age (S.D., range) — years 37.24 (11.09, 18–64) FamMale 52.3% (n=156) PsycAustralian born 84.9% (n=253) MeaSingle, never married 65.8% (n=196) CouCompleted highest school year available 35.2% (n=105) –Age left school (S.D., range) 16.11 (1.74, 6–23) –Employed full or part time 23.2% (n=69) –

Receiving welfare support 95.6% (n=285) –ICD-10 primary diagnosis a Pattern

Severe depression with psychosis 6.4% (n=19) HazBipolar disorder, mania 9.1% (n=27) HazSchizophrenia 42.3% (n=126) MedicaSchizoaffective disorder 14.4% (n=43) UsaOther psychosis 27.9% (n=83) Help

Service utilisation (past 12 months)(S.D., range) a

Curren

Hospital admissions: BPR–At least one psychiatric admission 38.9% (n=116) STA–Mean number of hospital admissions 0.70 (1.01, 0–5) STA–Average length of hospitaladmission (days)

19.38 (42.55, 0–364) BDI

Mean number of visits toGeneral Practitioner

12.22 (16.46, 0–104) SF-1

Mean number of visits toCommunity Mental Health Team

17.16 (39.77, 0–365) SF-1

BPRS=Brief Psychiatric Rating Scale (Ventura et al., 1993); STAI=State TInventory (Beck et al., 1988); SF-12=Short Form Survey (Ware et al., 1996a Based on responses to the DIP (Jablensky et al., 2000; Castle et al., 20

those reported in the national Low Prevalence (psy-chotic) Disorders Study for participants recruitedthrough community mental health services (Carr et al.,2002). Similarly, antipsychotic medication was beingtaken in the month prior to interview by the majority ofthe sample, with most participants reporting thismedication to be helpful. On average, participantswere mildly symptomatic as measured by the BPRS,mildly depressed (BDI-II), moderately anxious (STAI),and mildly disabled (SF-12). In view of these samplecharacteristics, two additional simplified categoricalvariables were constructed for possible use in subse-quent analyses: Diagnostic category – schizophrenia orschizoaffective disorder (n=169, 56.7%) versus otherICD-10 psychotic disorders (n=129, 43.3%); andcomorbidity category – concurrent hazardous use ofalcohol or cannabis (n=43, 14.4%) versus no concurrenthazardous use of these substances (n=255, 85.6%).

3.2. Smoking profiles

Most of the sample (n=209, 70.1%) had been pre-viously advised to quit smoking by a health professional,

ssment (N=298)

factors a

ily history of schizophrenia 22.1% (n=66)hosocial stressor prior to onset of illness 75.2% (n=224)n age of illness onset (years) (S.D., range) 22.93 (7.28, 9–50)rse of psychotic disorder:Single episode, good or unknown recovery 7.4% (n=22)Multiple episodes, good recovery 26.8% (n=80)Multiple episodes, minimalrecovery or deterioration

39.6% (n=118)

Chronic, clear deterioration 26.2% (n=78)s of substance use (OTI past month)ardous use of alcohol 6.0% (n=18)ardous use of cannabis 10.1% (n=30)tion a

ge of anti-psychotic medication 82.9% (n=247)fulness of anti-psychotic medication (n=230) 71.5% (n=213)t psychopathology and functioning

S global score mean (S.D., range) 34.09 (9.98, 24–94)I State mean (S.D., range) 41.49 (12.54, 20–76)I Trait mean (S.D., range) 47.72 (12.21, 20–79)-II score (S.D., range) 16.24 (13.60, 0–60)

2 (PCS) (S.D., range) 47.05 (7.53, 27–61)

2 (MCS) (S.D., range) 46.17 (8.33, 22–62)

rait Anxiety Inventory (Spielberger, 1983); BDI-II=Beck Depression; Andrews, 2002).06).

6 A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

usually their general practitioner (GP) (185/209, 88.5%).Only 26.8% (80/209) and 14.1% (42/209) of thissubgroup had been advised to do so by their psychiatristor case managers respectively. A small percentage (44/298, 14.8%) had been advised by a health professionalthat smoking was an “acceptable activity for them”. Withrespect to stage of change for quitting smoking, 13.1%were precontemplators (PC-1: 8; PC-2: 12; and PC-3:19); 49.7% were contemplators (C: 148); and 37.2%were preparing to take action (PA: 111).

The left-hand data column of Table 2 summarizes theoverall sample's smoking characteristics and reasons forquitting, while the right-hand columns report means(and S.D.'s) for these variables by gender and stage ofchange. Reasons for smoking are not reported in Table 2as there were no gender or stage of change effects (seebelow). The overall means (N=298) for the reasons forsmoking subscales were: Stress reduction (0–3), 2.56(S.D.=0.76); stimulation (0–4), 2.01 (S.D.=1.11);mental illness (0–2), 0.36 (S.D.=0.57); addiction (0–2), 1.88 (S.D.=0.38); and partner smokes (0–1), 0.11(S.D.=0.31).

A series of analyses was conducted to examineassociations between the smoking variables and age,gender, diagnostic category, comorbidity category, andstage of change. Most of these analyses involved one-way ANCOVAs, examining associations between theselected smoking variable and the independent variableof interest, while controlling for age, gender, diagnostic

Table 2Smoking characteristics, reasons for smoking and reasons for quitting by ge

Smoking variables(initial assessment)

Overall Gender

Males Females

Means (S.D.)

Sample size 298 156 142Smoking characteristicsCigarettes per day 30.47 (13.24) 29.03 (11.50) 32.05 (14Age first smoked 14.98 (5.22) 14.54 (4.78) 15.45 (5.Age smoking daily 17.81 (5.16) 17.49 (4.64) 18.15 (5.Quit attempts (lifetime) 2.58 (1.59) 2.65 (1.57) 2.50 (1.Nicotine dependence score (0–10) 8.06 (2.08) 7.72 (2.16) 8.43 (1.

Reasons for quittingHealth concerns (0–4) 2.61 (1.14) 2.50 (1.18) 2.74 (1.Self-control (0–4) 2.64 (1.17) 2.45 (1.29) 2.85 (0.Immediate reinforcement (0–4) 2.38 (1.16) 2.08 (1.18) 2.71 (1.Social influence (0–4) 1.11 (1.05) 0.95 (0.97) 1.30 (1.Intrinsic minus extrinsic 0.88 (0.91) 0.96 (0.93) 0.80 (0.Overall scale score (0–4) 2.19 (0.84) 1.99 (0.85) 2.40 (0.a Comparisons between groups were based on one-way ANCOVAs (c

appropriate), with Scheffé follow-up comparisons, as required: ⁎P<0.01; ⁎⁎Pof significant differences (e.g., pc=differs significantly from pre-contemplat

category, and comorbidity category (as appropriate), andusing Scheffé follow-up comparisons, if required. Forage, which was coded as a continuous independentvariable, partial correlations (pr) were calculated (from aparallel series of regression analyses). Table 2 reportsthe statistically significant comparisons involvinggender and stage of change, while analyses relating tothe other independent variables are reported below.

On average, participants were heavy smokers, highlydependent on nicotine, they had begun smoking at anearly age and had been daily smokers for about 5 yearsbefore they were first diagnosed with a mental illness(see Table 2). Participants also reported having madeonly a relatively small number of quit attempts in theirlifetime. Older age was associated with higher dailyusage (pr=0.18, P<0.01), an older age at smokinginitiation (pr=0.21, P<0.001) and commencement ofdaily smoking (pr=0.21, P<0.01), and a higher FTNDscore (pr=0.18, P<0.01). There were no gender ordiagnostic category differences in smoking character-istics, however, participants with concurrent hazardoususe of alcohol or cannabis initiated smoking at a youngerage (mean: 12.51 versus 15.39 years, F(1293)=8.03,P<0.01) and commenced daily smoking at a youngerage (mean: 15.49 versus 18.20 years, F(1293)=7.03,P<0.01). As shown in Table 2, stage of change tended tohave a linear association with number of lifetime quitattempts, with participants classified as precontempla-tors (PC) or contemplators (C) having made significantly

nder and stage of change a

Stage of change

Pre-contemplators (PC) Contemplators (C) Preparing foraction (PA)

39 148 111

.79) 29.87 (14.44) 31.30 (12.15) 29.57 (14.21)65) 14.67 (4.14) 15.52 (5.64) 14.36 (4.95)68) 17.44 (3.80) 17.74 (5.25) 18.03 (5.47)61) 1.74 (1.60)⁎⁎pa 2.41 (1.62)⁎pa 3.09 (1.36)94) 7.87 (2.29) 8.35 (1.97) 7.73 (2.12)

08) 2.09 (1.29) 2.65 (1.12) 2.74 (1.06)99)⁎ 1.68 (1.33) 2.69 (1.10)⁎⁎pc 2.91 (1.04)⁎⁎pc

05)⁎⁎ 1.73 (1.27) 2.48 (1.15) 2.48 (1.07)⁎pc

10)⁎ 0.85 (0.97) 1.18 (1.06) 1.12 (1.06)88) 0.59 (0.92) 0.84 (0.90) 1.03 (0.89)78)⁎⁎ 1.59 (0.93) 2.25 (0.82)⁎⁎pc 2.31 (0.74)⁎⁎pc

ontrolling for age, gender, diagnostic and comorbidity category, as<0.001. For stage of change, superscripted letters indicate the pattern

ors).

Table 3Selected comparisons with other studies

Smoking variables Currentstudy

Comparisonsample

Statisticalcomparisons a

Means (S.D.'s)

Reason for smoking Study A b t(297)Stress reduction (0–3) 2.56 (0.76) 2.17 8.80⁎⁎

Stimulation (0–4) 2.01 (1.11) 1.53 7.49⁎⁎

Addiction (0–2) 1.88 (0.38) 1.55 14.97⁎⁎

Reason for quitting Study B c F(1,1511)

Health concerns (0–4) 2.61 (1.13) 2.63 (0.99) 0.09, NSSelf-control (0–4) 2.64 (2.80) 2.09 (1.06) 29.62⁎⁎

Immediatereinforcement (0–4)

2.38 (1.16) 1.56 (1.04) 141.96⁎⁎

Social influence (0–4) 1.11 (0.85) 0.53 (0.59) 190.90⁎⁎

Intrinsic minusextrinsic

0.89 (0.91) 1.31 (0.79) 63.56⁎⁎

Overall scalescore (0–4)

2.19 (0.84) 1.73 (0.65) 104.84⁎⁎

Nicotine dependence(Fagerstrom) score

Study C d F(1700)

Full sample 8.06 (2.08) 3.60 (2.60) 595.44⁎⁎

(n=169) (n=66) F(1233)

Schizophreniasubgroup

8.18 (2.10) 6.20 (2.20) 39.94⁎⁎

a Based on single sample t-tests (with Study A as the referencevalue) or one-way ANOVAs: NS, non-significant; ⁎P<0.01;⁎⁎P<0.001.b Study A: sample of smokers participating in a phone survey

(n=387) (Pederson et al., 1996).c Study B: volunteer sample of smokers participating in a

randomized evaluation of self-help interventions for smokingcessation (n=1215) (Curry et al., 1997).d Study C: general sample of smokers (US) (n=404) and

schizophrenia sample of smokers (n=66) (De Leon et al., 2002b).

7A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

fewer quit attempts since starting smoking than those inthe preparation stage (PA).

Reported reasons for smoking were not associatedwith any of the predictor variables examined (i.e., age,gender, diagnostic or comorbidity category, and stage ofchange), suggesting that they tend to represent a sharedset of beliefs among this particular group of smokers. Incontrast, reported reasons for quitting were significantlyassociated with age, gender, comorbidity category, andstage of change. Older participants were more likely tocite extrinsic reasons such as immediate reinforcement(e.g., from money saved, or reduced cleaning) asmotivating factors for quitting (pr=0.15, P<0.01). Asshown in Table 2, female participants reported signifi-cantly more reasons for quitting thanmales (i.e., they hadhigher average scores on the overall RFQ). They also hadsignificantly higher scores on three of the four sub-dimensions, being more likely to report as reasons forquitting a desire for self-control, immediate reinforce-ment motivation, and social influences. On the otherhand, participants with concurrent hazardous use ofalcohol or cannabis reported significantly fewer reasonsfor quitting than those without such comorbidity (meanoverall RQF score: 1.82 versus 2.25, F(1293)=8.60,P<0.01). They were also less likely to report immediatereinforcement motivation as a reason for quitting (mean:1.87 versus 2.47, F(1293)=7.93, P<0.01). Similarly,participants in the precontemplation stage reportedsignificantly lower average scores on the overall RFQcompared to those in the contemplation and preparationstages (see Table 2). Average scores on two of the foursub-dimensions also differed significantly according tostage of change. Participants in the precontemplationstage reported significantly lower desire for self-controlcompared to those in the contemplation and preparationstages. A similar pattern emerged for immediatereinforcement motivation, with a significant differencebetween precontemplators and those preparing foraction.

Following Aguilar et al. (2005), the current samplewas divided into two subgroups, those with ‘moderate tohigh’ nicotine dependence (FTND≤7, n=150, meanconsumption of 23.80 cigarettes per day) and those with‘very high’ nicotine dependence (FTND>7, n=148,mean consumption of 37.22 cigarettes per day). Withintheir large sample of schizophrenia outpatients, Aguilaret al. (2005) also found approximately equal proportionswith moderate to high (36.8%) and very high (32.4%)nicotine dependence. Similar ANCOVAs to thosedescribed above (i.e., controlling for age, gender,diagnostic and comorbidity category) were used toexamine differences between these nicotine dependence

subgroups. Participants with very high nicotine depen-dence initiated smoking at a younger age (mean: 14.36versus 15.59 years, F(1292)=11.30, P<0.001) andcommenced daily smoking at a younger age (mean:17.13 versus 18.47 years, F(1292)=12.61, P<0.001).However, there were no significant differences betweenthese two subgroups in the number of lifetime quitattempts, nor in their reported reasons for smoking orquitting, although there was a tendency for those withvery high nicotine dependence to cite health concernsamong their reasons for quitting (mean: 2.82 versus2.41, F(1292)=6.59, P=0.011).

3.3. Comparisons with other studies

Table 3 shows comparisons between our sample andother studies that had employed the same assessmentinstruments. Reasons for smoking were compared withthe general population group surveyed by Pedersonet al. (1996) using the RSQ. However, it was only

8 A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

possible to estimate aggregate levels from the originalstudy (and not S.D.'s) for three of our five factors.Single sample t-tests revealed significant differencesbetween the general population group and our psychi-atric sample, who were more likely to cite stressreduction, stimulation and addiction as reasons forsmoking compared to the general population.

Average scores on the RFQ, for each motivationdimension and for the intrinsic–extrinsic motivationdifference score, and overall, are summarized in Table 3for the present sample and a volunteer sample ofsmokers surveyed by Curry et al. (1997). Our samplereported significantly higher overall motivation scores.Average scores on three of the four sub-dimensions andthe intrinsic–extrinsic difference score also differedsignificantly between the two samples. While healthconcerns were equivalent between the two groups, oursample was more likely to report self-control, immediatereinforcement, and social influences as reasons forquitting. Level of extrinsic relative to intrinsic motiva-tion was proportionately higher among the presentsample (i.e., our psychiatric group was less likely toreport a preference for intrinsic factors as reasons forquitting, compared to the general population). We alsocompared our psychiatric group to the general popula-tion group surveyed by de Leon et al. (2002a) and foundthat our sample had a significantly higher level ofnicotine dependence compared to the general populationof smokers. Our diagnostic subgroup with schizophreniaor schizoaffective disorder also had a significantlyhigher level of nicotine dependence compared to deLeon's schizophrenia subgroup (P<0.01).

4. Discussion

Employing standardized assessment instruments, thepresent study identified several significant differences insmoking characteristics between subgroups within thepresent sample of smokers with a psychotic disorderliving in the community and between the present sampleand other studies. These findings may have importantimplications for smoking interventions among peoplewith a psychotic disorder.

With regard to differences in smoking characteristicswithin the present sample, participants who reported ayounger age at initiation to smoking (or an earliercommencement of daily consumption) tended to beyounger, to have concurrent hazardous use of alcohol orcannabis, and/or to have very high levels of nicotinedependence. Conversely, heavier cigarette use (based onconsumption per day) and higher nicotine dependencescores were reported by those who were older. The

absence of differences in smoking characteristicsbetween those with schizophrenia or schizoaffectivedisorder and the remaining participants (with otherpsychotic disorders) is inconsistent with some previousstudies, however, De Leon and Diaz (2005) suggest thatit is still unclear whether schizophrenia patients whosmoke are heavier users or more dependent thansmokers with other severe mental illnesses. Participantsat earlier stages of change also reported significantlyfewer quit attempts in the present study, which shouldencourage clinicians to consider each quit attempt aspotentially shoring up motivation for a further, perhapsmore successful, quit attempt. However, a recent studyshowed that psychiatrists tend to offer limited counsel-ling on smoking related health risks to patients withestablished mental illness and may miss many suchopportunities (Himelhoch and Daumit, 2003).

The current data indicate that, especially in theabsence of significant subgroup differences in reasonsfor smoking, it is important to ask people about theirreasons for quitting, preferably employing an instrumentsuch as the RFQ, in order to adequately assess intrinsicand extrinsic motivators. Typically, females were morelikely to cite a variety of reasons for quitting (e.g., self-control, factors providing immediate reinforcement,social influences), while participants with concurrenthazardous use of alcohol or cannabis, or at theprecontemplation stage of change, were less likely toidentify reasons for quitting. One of the implications ofthese findings, and following on from work by Curryet al. (1990, 1997), is that motivational interventionsamong precontemplative smokers with a psychoticdisorder may benefit from focusing more on enhancingintrinsic factors such as a desire for self-control, and lesson any immediate reinforcement associated withquitting. More generally, it may be appropriate toattempt to strengthen any RFQ domains that are under-represented, especially among males, beginning with theintrinsic factors (health concerns and self-control). Insupport of Kelly and McCreadie (1999), health concernswere among the most highly cited reasons for quitting inthe current study, particularly among those with thehighest levels of nicotine dependence.

The brief comparisons that were undertaken withother smoking studies (see Table 3) provided a valuablecontextual framework, the findings from which largelyreinforce the conclusions drawn from the within studyanalyses. For example, the present sample differed intheir reasons for quitting from a non-psychopathologicalvolunteer sample (Curry et al., 1997), having signifi-cantly higher overall motivation scores. However, theywere less likely to display a preference for intrinsic

9A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

factors relative to extrinsic motivators. Tipping thebalance in favour of intrinsic versus extrinsic motivatorsduring smoking cessation interventions, particularlyduring motivational interviewing, may be crucial.

In the current sample, approximately equal numbersof participants had moderate to high and very highnicotine dependence, which is comparable to the datareported by Aguilar et al. (2005). The substantiallyhigher levels of nicotine dependence among the presentsample compared to the schizophrenia sample of DeLeon et al. (2002b) may partially reflect differences insampling. Their data were from inpatients and out-patients with schizophrenia who were seeking mentalhealth treatment, whereas the present data were obtainedfrom a sample of people with psychotic disorders whowere responding to an invitation to be involved in asmoking cessation study. The current sample may alsobe demonstrating an increased awareness of healthrelated messages.

Our ability to recruit a large sample of peopleinterested in giving up smoking attests to the potentiallikelihood of attracting sufficient numbers of people ifsmoking cessation interventions were to be offered bymental health services. Only one-eighth (13.1%) of thepresent sample was classified as precontemplators,whereas other studies among non-treatment seekingsmokers with a long-term psychiatric illness have foundmuch higher rates of precontemplation, with 52%–79%of patients classified as precontemplators, 20%–25%classified as contemplators and 3%–15% classified aspreparing for action (Hall et al., 1995; Carosella et al.,1999; De Leon et al., 2002b), which is very similar tothe distribution of stages of change among the generalpopulation (Etter et al., 2004).

The current sample began smoking at a young ageand had become daily smokers on average 5 yearsbefore diagnosis with psychosis. This suggests that thedevelopment of early interventions among teenagers toprevent transition to daily smoking is an important areafor future research. Such interventions may also havewider benefits, as tobacco could function as a ‘gatewaydrug’, leading to involvement with other drugs (Kandel,2002). In addition, the finding that, compared to generalsamples of smokers, the present sample had highernicotine dependence and rated stress reduction, stimu-lation and addiction more highly as reasons for smokingmay have direct implications for interventions to reducestress and boredom and to address dependence ade-quately via pharmacological and psychological means.However, initial attributions for smoking may have beenformed prior to illness onset and may not reflect currentinfluences or circumstances. The lack of associations

between the predictor variables examined in this studyand the reported reasons for smoking tends to suggestthat these beliefs are long-standing and essentiallyshared among this group of smokers.

Recently, Dolan et al. (2004) reported that smokerswith schizophrenia who exhibited poorer performanceon tests of executive function and spatial memory priorto implementation of a smoking cessation interventionwere less able to quit smoking. The authors suggestedthat they might continue to smoke because of theneuropsychological benefits that they receive fromcigarette smoking. Future studies should includeneuropsychological measures.

GPs are well placed to conduct smoking interven-tions, given that approximately three-quarters ofAustralian GPs treat patients with a psychotic disorder(Lewin and Carr, 1998). Participants in the present studyreported visiting their GP around a dozen times per year,which is consistent with other Australian psychosisstudies (Carr et al., 2002). In addition, the present studyfound that among those offered quit advice, this hadmost often come from their GP. Interventions deliveredby GPs using the Smokescreen program have beenshown to be successful among general patient popula-tions (Richmond et al., 1993, 1998; Richmond andWodak, 1998; Richmond, 1999) and may be suited topeople with severe mental illnesses. The currentAustralian guidelines for best clinical practice forsmoking cessation in general practice include recom-mendations for people with a mental illness (Zwar et al.,2003). The longitudinal nature of GP care may bebeneficial in addressing unsuccessful cessation attemptsand in delivering longer-term pharmacological andcounselling interventions for smoking, taking fluctua-tions in psychiatric symptomatology into account.Training of mental health professionals in offeringbrief smoking advice might also be indicated.

4.1. Limitations

The main limitations of the current study relate tosampling issues and the lack of suitable studies withwhich to make appropriate comparisons. Firstly, thecurrent sample was not selected to be representative ofall smokers with a current diagnosis of psychosis, beingdrawn from community members willing to participatein a smoking RCT. For example, the high overall levelsof nicotine dependence and the absence of differences insmoking characteristics between the diagnostic sub-groups may be associated with recruitment to anintervention trial. Likewise, the observed differencebetween the Newcastle and Sydney samples in current

10 A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

hazardous consumption of alcohol or cannabis (20.5%vs. 8.6%) may reflect variations in recruitment sources,strategies or engagement rates across these locations orrepresent true regional differences. Secondly, there maybe psychometric limitations associated with using theFTND among smokers with schizophrenia (Dolan et al.,2004), necessitating some adjustments to this instru-ment. Thirdly, as evidenced by the relatively smallnumber of comparison studies in Table 3, it was difficultto find suitable research against which to comparereported reasons for smoking and reasons for quitting. Itis also problematic to compare samples recruited fordifferent purposes and from different countries, al-though this is standard practice in meta-analyses (e.g.,De Leon and Diaz, 2005). Similarly, it is difficult tomake universal service delivery or treatment recom-mendations with respect to smoking interventionsbecause of differences between countries in smokingbase rates, health promotion programs and serviceprovision models (e.g., 95.6% of the current samplereceived welfare support, which included access tosubsidized health care). A broader spectrum of disordersmay need to be included in future studies (i.e. not justpsychoses) to facilitate a more direct examination ofdifferences in smoking related behaviours and attribu-tions. Researchers may also need to study groups atdifferent time points or illness phases rather thanattempting to generalise about smokers as a whole.

4.2. Conclusions

This is one of the first large-scale, comprehensivestudies of smoking and smoking-related phenomena in asample of smokers with a psychotic disorder. Althoughit is difficult to make universal recommendations withrespect to smoking interventions, we suggest thefollowing: Expect keen interest from patients, withmany being at the contemplation and action stages;encourage precontemplators to consider quitting; askabout reasons for quitting (in addition to reasons forsmoking); focus motivational interventions on thedevelopment of intrinsic motivators; target youngpeople for smoking cessation interventions (beforetransition to daily smoking); target those at risk forhazardous use of other substances (before more severenicotine dependence has developed); address lifestylefactors, such as stress and boredom, in conjunction withappropriate pharmacological interventions for nicotinedependence; involve GPs, who are well placed toconduct smoking cessation interventions (and may findrecent guidelines for smoking cessation among peoplewith a mental illness useful); and consider training

mental health professionals in brief interventions forsmoking cessation.

Acknowledgment

This research was funded through grants from theAustralian Rotary Research Fund, the CommunityHealth and Tuberculosis Association, and the NationalHealth and Medical Research Council (grant number:141708). Thanks to Sarah Robertson and Jane Waltonfor assistance with initial stages of the project, RenayGreig for her assistance with the literature review forthis paper and her contribution to the follow-upinterviews, and Angeline Dalmau for assistance withthe follow-up interviews. We also wish to thank all ofthe participants and the various agencies and healthprofessionals who assisted with recruitment, includingthe NISAD Schizophrenia Research Register.

References

Addington, J., el-Guebaly, N., Addington, D., Hodgins, D., 1997.Readiness to stop smoking in schizophrenia. Canadian Journal ofPsychiatry 42, 49–52.

Aguilar, M.C., Gurpegui, M., Diaz, F.J., de Leon, J., 2005. Nicotinedependence and symptoms in schizophrenia: naturalistic study ofcomplex interactions. British Journal of Psychiatry 186, 215–221.

Andrews, G., 2002. A brief integer scorer for the SF-12: validity of thebrief scorer in Australian community and clinical settings.Australian and New Zealand Journal of Public Health 26,508–510.

Baker, A., Richmond, R., Haile, M., Lewin, T.J., Carr, V.J., Taylor, R.L., Jansons, S., Wilhelm, K., in press. Randomized controlled trialof a smoking cessation intervention among people with a psychoticdisorder. American Journal of Psychiatry.

Baker, A., Bucci, S., Lewin, T., Kay-Lambkin, F., Constable, P., Carr,V., 2006. Cognitive-behavioural therapy for substance usedisorders in people with psychotic disorders. Randomisedcontrolled trial. British Journal of Psychiatry 188, 439–448.

Beck, A.T., Steer, R.A., Garbin, M.G., 1988. Beck DepressionInventory—Second Edition Manual. Harcourt Brace and Com-pany, Orlando.

Carosella, A., Ossip-Klein, D.J., Owens, C.A., 1999. Smokingattitudes, beliefs, and readiness to change among acute and longterm care inpatients with psychiatric diagnoses. AddictiveBehaviors 24, 331–344.

Carr, V.J., Lewin, T.J., Barnard, R.E., Walton, J.M., Allen, J.L.,Constable, P.M., Chapman, J.L., 2002. Comparisons betweenschizophrenia patients recruited from Australian general practicesand public mental health services. Acta Psychiatrica Scandinavica105, 346–355.

Castle, D.J., Jablensky, A., McGrath, J.J., Carr, V., Morgan, V.,Waterreus, A., Valuri, G., Stain, H., McGuffin, P., Farmer, A.,2006. The diagnostic interview for psychoses (DIP): development,reliability and applications. Psychological Medicine 36, 69–80.

Crittenden, K.S., Manfredi, C., Lacey, L., Warnecke, R., Parsons, J.,1994. Measuring readiness and motivation to quit smoking amongwomen in public health clinics. Addictive Behaviors 19, 497–507.

11A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

Curry, S., Wagner, E.H., Grothaus, L.C., 1990. Intrinsic and extrinsicmotivation for smoking cessation. Journal of Consulting andClinical Psychology 58, 310–316.

Curry, S.J., Grothaus, L., McBride, C., 1997. Reasons for quitting:intrinsic and extrinsic motivation for smoking cessation in apopulation-based sample of smokers. Addictive Behaviors 22,727–739.

Darke, S., Heather, N., Hall, W., Ward, J., Wodak, A., 1991.Estimating drug consumption in opioid users: reliability andvalidity of a recent use episodes method. British Journal ofAddiction 86, 1311–1316.

De Leon, J., Diaz, F.J., 2005. A meta-analysis of worldwide studiesdemonstrates an association between schizophrenia and tobaccosmoking behaviours. Schizophrenia Research 76, 135–157.

De Leon, J., Becona, E., Gurpegui, M., Gonzalez-Pinto, A., Diaz, F.J.,2002a. The association between high nicotine dependence andsevere mental illness may be consistent across countries. Journal ofClinical Psychiatry 63, 812–816.

De Leon, J., Diaz, F.J., Rogers, T., Browne, D., Dinsmore, L., 2002b.Initiation of daily smoking and nicotine dependence in schizo-phrenia and mood disorders. Schizophrenia Research 56, 47–54.

Diclemente, C.C., Prochaska, J.O., Fairhurst, S.K., Velicer, W.F.,Velasquez, M.M., Rossi, J.S., 1991. The process of smokingcessation: an analysis of precontemplation, contemplation, andpreparation stages of change. Journal of Consulting and ClinicalPsychology 59, 295–304.

Dolan, S.L., Sacco, K.A., Termine, A., Seyal, A.A., Dudas, M.M.,Vessicchio, J.C., Wexler, B.E., George, T.P., 2004. Neuropsycho-logical deficits are associated with smoking cessation treatmentfailure in patients with schizophrenia. Schizophrenia Research 70,263–275.

Etter, M., Mohr, S., Garin, C., Etter, J., 2004. Stages of change insmokers with schizophrenia or schizoaffective disorder and in thegeneral population. Schizophrenia Bulletin 30, 459–468.

Fagerstrom, K.O., Kunze, M., Schoberberger, R., Breslau, N., Hughes,J.R., Hurt, R.D., Puska, P., Ramstrom, L., Zatonski, W., 1996.Nicotine dependence versus smoking prevalence: comparisonsamong countries and categories of smokers. Tobacco Control 5,52–56.

Fowler, I.L., Carr, V.J., Carter, N.T., Lewin, T.J., 1998. Patterns ofcurrent and lifetime substance use in schizophrenia. SchizophreniaBulletin 24, 443–455.

Hall, R.G., Duhamel, M., McClanahan, R., Miles, G., Nason, C.,Rosen, S., Schiller, P., Tao-Yonenaga, L., Hall, S.M., 1995. Levelof functioning, severity of illness, and smoking status amongchronic psychiatric patients. Journal of Nervous and MentalDisease 183, 468–471.

Heatherton, T.F., Kozlowski, L.T., Frecker, R.C., Fagerstrom, K.O.,1991. The Fagerstrom test for nicotine dependence: a revision ofthe Fagerstrom tolerance questionnaire. British Journal ofAddiction 86, 1119–1127.

Himelhoch, S., Daumit, G., 2003. To whom do psychiatrists offersmoking cessation counseling? American Journal of Psychiatry160, 2228–2230.

Jablensky, A., McGrath, J., Herrman, H., Castle, D., Gureje, O., Evans,M., Carr, V., Morgan, V., Korten, A., Harvey, C., 2000. Psychoticdisorders in urban areas: an overview of the study on LowPrevalence Disorders. Australian and New Zealand Journal ofPsychiatry 34, 221–236.

Kandel, D., 2002. Stages and Pathways of Drug Involvement.Examining the Gateway Hypothesis. Cambridge UniversityPress, Cambridge.

Kelly, C., McCreadie, R.G., 1999. Smoking habits, current symptoms,and premorbid characteristics of schizophrenic patients in Niths-dale, Scotland. American Journal of Psychiatry 156, 1751–1757.

Kumari, V., Postma, P., 2005. Nicotine use in schizophrenia: the self-medication hypotheses. Neuroscience and Biobehavioral Reviews29, 1021–1034.

Larson, C.O., 2002. Use of the SF-12 instrument for measuring thehealth of homeless persons. Health Services Research 37,733–749.

Lawrence, D., Holma, D., Jablensky, A., 2001. Preventable PhysicalIllness in People with Mental Illness. The University of WesternAustralian, Perth.

Lewin, T.J., Carr, V.J., 1998. Rates of treatment of schizophrenia bygeneral practitioners: a pilot study. Medical Journal of Australia168, 166–168.

Lim, L.L., Fisher, D., 1999. Use of the 12-item Short Form (SF-12)Health Survey in an Australian heart and stroke population.Quality of Life Research 8, 1–8.

Marlatt, G.A., 1988. Matching clients to treatment: treatment modelsand stages of change. In: Donovan, D.M., Marlatt, G.A. (Eds.),Assessment of Addictive Behaviors. The Guilford BehavioralAssessment Series. Guilford PressNew York, pp. 474–483.

McGuffin, P., Farmer, A., Harvey, I., 1991. A polydiagnosticapplication of operational criteria in studies of psychotic illness.Development and reliability of the OPCRIT system. Archives ofGeneral Psychiatry 48, 764–770.

Pederson, L., Bull, S., Ashley, M.J., MacDonald, J.K., 1996. Quittingsmoking: why, how, and what might help. Tobacco Control 5,209–214.

Peterson, A.L., Hryshko-Mullen, A.S., Cortez, Y., 2003. Assessmentand diagnosis of nicotine dependence in mental health settings.American Journal of Addiction 12, 192–197.

Prochaska, J.O., Diclemente, C.C., 1983. Stages and processes of self-change of smoking: toward an integrative model of change.Journal of Consulting and Clinical Psychology 51, 390–395.

Prochaska, J.O., Velicer, W.F., Prochaska, J.M., Johnson, J.L., 2004.Size, consistency, and stability of stage effects for smokingcessation. Addictive Behaviors 29, 207–213.

Reichler, H., Baker, A., Lewin, T., Carr, V., 2001. Smoking among in-patients with drug-related problems in an Australian psychiatrichospital. Drug and Alcohol Review 20, 231–237.

Richmond, R., 1999. The physician can make a difference withsmokers: evidence based clinical approaches. International Journalof Tuberculosis and Lung Disease 3, 100–112.

Richmond, R., Wodak, A., 1998. Where should alcohol and tobaccointerventions be provided? (Editorial) Australian and New ZealandJournal of Public Health 22, 745–746.

Richmond, R., Makinson, R., Kehoe, L., Giugni, A., Webster, I., 1993.One year evaluation of three smoking cessation programmesadministered by general practitioners. Addictive Behaviors 18,187–199.

Richmond, R., Mendelsohn, C., Kehoe, L., 1998. Family physicians'utilization of a brief smoking cessation program followingreinforcement contact after training: a randomized trial. PreventiveMedicine 27, 77–83.

Spielberger, C.D., 1983. Manual for the State-Trait Anxiety Inventory(Form Y). Consulting Psychologists Press, Palo Alto.

Ventura, J., Lukoff, D., Nuechterlein, K.H., Liberman, R.P., Green, M.,Shaner, A., 1993. Appendix 1: brief Psychiatric Rating Scale(BPRS) Expanded Version (4.0) scales, anchor points andadministration manual. International Journal of Methods inPsychiatric Research 3, 227–243.

12 A. Baker et al. / Psychiatry Research xx (2007) xxx–xxx

ARTICLE IN PRESS

Ware Jr., J.E., Kosinski, M., Keller, S.D., 1996. A 12-item Short FormHealth Survey: construction of scales and preliminary tests ofreliability and validity. Medical Care 34, 220–233.

Wilhelm, K., 1998. The relevance of smoking and nicotine to clinicalpsychiatry. Australasian Psychiatry 6, 130–132.

Ziedonis, D.M., George, T.P., 1997. Schizophrenia and nicotine use:report of a pilot smoking cessation program and review ofneurobiological and clinical issues. Schizophrenia Research 23,247–255.

Ziedonis, D.M., Williams, J.M., 2003. Addressing tobacco use inmental health and addiction settings. Psychiatric Annals 33,425–426.

Zwar, N., Richmond, R., Borland, R., Stillman, S., Cunningham, M.,Litt, J., 2003. Smoking Cessation Guidelines for AustralianGeneral Practice: Practice Handbook. Commonwealth Departmentof Health and Ageing, Canberra.

Minor corrections to published paper: References The two Baker et al. papers should read: Baker, A., Richmond, R., Haile, M., Lewin, T.J., Carr, V.J., Taylor, R.L., Jansons, S.,

Wilhelm, K., 2006a. A Randomized controlled trial of a smoking cessation intervention among people with a psychotic disorder. American Journal of Psychiatry 163, 1934-1942.

Baker, A., Bucci, S., Lewin, T.J., Kay-Lambkin, F., Constable, P.M., Carr, V.J.,

2006b. Cognitive-behavioural therapy for substance use disorders in people with psychotic disorders. Randomised controlled trial. British Journal of Psychiatry 188, 439-448.

Other corrections Section 1.2. The current study Last sentence: - change "Baker et al. (in press)" to "Baker et al. (2006a)" Section 2.1. Sample Second paragraph, second sentence: - change "(Baker et al., 2006)" to "(Baker et al., 2006b)" Section 3.2. Smoking profiles There are several minor errors throughout this section in the way that the degrees of freedom are reported (in the 4th, 5th and 6th paragraphs): - in each case, there should be a comma and space after “F(1” to show that there are two degrees of freedom parameters. Thus, “F(1293)” becomes “F(1, 293)”, and “F(1292)” becomes “F(1, 292)”. Section 3.3. Comparisons with other studies Likewise, in Table 3, the degrees of freedom should be reported as “F(1, 700)” and “F(1, 233)”.