Schizophrenia Research 109 (2009) 46–51

Contents lists available at ScienceDirect

Schizophrenia Research

j ourna l homepage: www.e lsev ie r.com/ locate /schres

Psychopathology and cognition in divergent functional outcomesin schizophrenia

R. Walter Heinrichs a,⁎, Narmeen Ammari a, Ashley Miles a,Stephanie McDermid Vaz b, Boyko Chopov a

a Department of Psychology, York University, 4700 Keele Street, Toronto, Ontario, Canada M3J 1P3b Cleghorn Early Intervention in Psychosis Program, St Joseph's Healthcare Hamilton 25 Charlton Ave East, Suite 703, Hamilton, ON, Canada L8N 1Y2

a r t i c l e i n f o

⁎ Corresponding author. Tel.:+1416 736 2100x66205E-mail addresses: walterh@yorku.ca (R.W. Heinric

ammarina@yorku.ca (N. Ammari), amiles@yorku.ca (Asmcdermi@stjosham.on.ca (S. McDermid Vaz), boyko@

0920-9964/$ – see front matter © 2009 Elsevier B.V.doi:10.1016/j.schres.2009.01.001

a b s t r a c t

Article history:Received 13 November 2008Received in revised form 1 January 2009Accepted 5 January 2009Available online 31 January 2009

Cognitive performance rather than symptoms, especially positive symptoms, is regarded as theprimary predictor of functional outcome in schizophrenia. However, contradictory evidenceexists and many studies fail to sample from the extremes of outcome measures. This studytested whether the differential importance assigned to symptoms and cognitive impairment issupportable in patients with high and low levels of community independence. Schizophreniapatients with highly unfavorable (n=24) and highly favorable (n=28) functional outcomes asdefined by community support requirements were studied. Standard cognitive andpsychopathology measures were analyzed using independent groups comparisons andoutcome prediction with logistic regression methods. Symptom severity and cognitive dataseparately accounted for significant amounts of variance in community independence. Positiveas well as negative symptoms, non-psychotic psychopathology and cognition generated largeeffect sizes between highly unfavorable and favorable outcome groups. The conditional validityof both overall psychopathology and positive symptoms was significant over and above thecontribution of cognition to outcome prediction. Results suggest researchers may haveunderestimated the role of psychopathology in general and positive symptoms in particular aspotential determinants of functional status in schizophrenia.

© 2009 Elsevier B.V. All rights reserved.

Keywords:Functional outcomeSymptomsCognition

1. Introduction

Over the last decade the persisting functional disabilitiesand dependencies observed in many schizophrenia patientshave been seen increasingly as consequences of cognitiveimpairment. Considerably less importance has been assigned topsychopathology, especially psychotic (positive) symptoms, interms of accounting for functional outcome variance. Green(2007), for example, has stated that, in contrast with cognition,psychotic symptoms are weak determinants of communityfunctioning. This view of cognitive impairment and symptomsis influential and grounded in several lines of evidence. First, the

; fax:+1416736 5814hs),. Miles),yorku.ca (B. Chopov)

All rights reserved.

.

.

cognition–functional outcome relationship has been confirmedin numerous studies, proving fairly consistent across commu-nity and psychosocial functioning and some aspects of voca-tional skill acquisition (Green et al., 2000; Matza et al., 2006;Green et al., 2004; Green, 2006). At the same time, it has beennoted that poor functional outcome often persists despitesignificant symptomatic improvement (San et al., 2007;Lindstrom et al., 2007). Anti-psychotic medication is relativelysuccessful in reducing symptom severity, but at bestmarginallysuccessful in improving cognition (Mishara and Goldberg,2004; Goldberg et al., 2007; Remillard et al., 2008). Moreover,scores on standard cognitive measures and symptom severityscales are only weakly related or even independent, implying apartial dissociation of these two facets of schizophrenic illness(Lucas et al., 2004; Rocca et al., 2006). Thus psychopathologydefines the clinical presentation anddiagnosis of schizophrenia,but cognitive performancemediates functional status. It follows

47R.W. Heinrichs et al. / Schizophrenia Research 109 (2009) 46–51

that treatment success is no longer reducible to symptomcontrol and improving cognitive performance has emerged as aprincipal target for the next generation ofmedications (Marder,2006).

Nonetheless, evidence contradicting this account of func-tional status has also emerged. In addition to data supportingassociations between negative and disorganization symptomsand functional variables (Milev et al., 2005; Kurtz et al., 2005),there are recent reports that positive symptoms predictcommunity outcome (Wittorf et al., 2008). Furthermore,longer treatment duration with antipsychotic medicationassociates not only with symptomatic improvements, but alsowith functional improvement (Dunayevich et al., 2007).Similarly, patients meeting criteria for symptom remissionfollowing treatment demonstrate significantly enhancedindependence in community living relative to patients notattaining remission (Helldin et al., 2007). Accordingly, thereare grounds for reconsidering the role of psychopathology inthe prediction and, potentially, determination of functionaloutcome.

It is noteworthy that most studies supporting the validityof cognitive test variables—and the relative invalidity ofpsychotic symptoms—as predictors of functional outcome usecorrelational methods and focus on accounting for significantamounts of outcome variance. While this approach has value,it is often difficult to determine whether an adequate range offunctional outcomes or symptom severity is being sampled. Intheory, restricted variance in predictor or criterion will limitvalidity. In practice, this restriction often occurs with respectto both symptoms and outcome. For example, Twamley et al.(2002) found no relationship between positive symptomsand level of residential independence in patients withschizophrenia. However, extremely low mean symptomscores suggest that the study sample comprised primarilyremitted patients. In other cases, interpretation of dataobtained from continuous outcome scales may be proble-matic because of floor or ceiling effects. Thus Bowie et al.(2006) reported descriptive statistics from the Specific Levelof Function Scale consistent with restricted variance andprobable ceiling effects on 4 out of 6 subscales. Even withmore adequate score distributions, the meaning of multi-variate outcome prediction may be unclear. A clinical sampleof patients with mild-to-moderate functional disability mayyield adequate variance for statistical purposes, but theresulting validities may not apply to patients with severe orno disability. These limitations are seldom noted or discussed,possibly encouraging unwarranted generalization of predic-tion validities across the complete range of clinical states andoutcome in schizophrenia.

In the present study, we were influenced by the remissioncriterion approach and applied a similar rationale to the studyof functional outcome. Given a sufficiently broad range ofclinical settings for recruitment, clear and meaningful criteriacan be applied to select patients with both highly favourableand unfavourable functional outcomes (Emsley et al., 2008;Mattsson et al., 2008). Hence we recruited patients withschizophrenia and schizoaffective disorder from a variety ofsettings including those with demanding rehabilitationprograms as well as from settings that require minimalinvolvement. The reasoning underpinning this approach wasthat it should yield adequate numbers of patients with

extreme functional outcomes, both favorable and unfavor-able, who could then be evaluated withmeasures of symptomseverity and cognition. The basic question motivating ourresearch was: does the comparison of schizophrenia patientswith markedly divergent functional status support or contra-dict the importance assigned to cognitive performancerelative to symptom severity, with special reference topositive symptoms?

2. Materials and methods

2.1. Participants

Patients were recruited from ambulatory outpatientsettings that required active program attendance and com-prised vocational and/or social rehabilitation. Settingsincluded the Hamilton Program for Schizophrenia, theCommunity Schizophrenia Service (St. Joseph's HealthcareHamilton), the Cleghorn Program (St. Joseph's HealthcareHamilton), the Canadian Mental Health Association (Torontobranch) and the Challenging Directions program (WhitbyMental Health Center). Male and female subjects whomet thefollowing criteria were included: 1) diagnosis of schizophre-nia or schizoaffective disorder by DSM-IV (American Psychia-tric Association Task Force on DSM-IV, 2000) criteria based onthe Structured Clinical Interview for DSM-IV (SCID; First et al.,1996), 2) age 18–65 years; 3) no history of seriousneurological or endocrine disorder; 4) no concurrent DSM-IV diagnosis of substance abuse; 5) no history of develop-mental disability; 6) minimum English reading level of grade4; 7) willingness and ability to sign informed consent; and8) normal or corrected vision.

A total of 156 patients, including 99 males and 57 femalesmet inclusion criteria. One hundred and fifty patients werereceiving anti-psychotic medication at the time of datacollection, with 129 patients treated with second-generationdrugs. Patients ranged from 21 to 65 years of age, with a meanof 41.10 (SD=9.34). To establish normal levels of communitysupport and independence, healthy comparison participants(n=71) were recruited by postings and advertisements forpaid research participation in community newspapers.Potential participants were screened for medical and psy-chiatric illness and history of drug abuse. This yielded 56males and 18 females, ranging from 18 to 65 years of age, witha mean of 40.59 (SD=13.68). All participants signedinformed consent and were paid for their time. The projectwas approved by the institutional review board at eachresearch site and by York University.

2.2. Psychopathology and cognitive performance measures

Current symptoms were evaluated with the Positive,Negative and General Psychopathology subscales of thePositive and Negative Syndrome Scale (PANSS; Opler et al.,1999). Neuropsychological tests measuring several aspects ofcognitive performance were administered including theVocabulary, Matrix Reasoning, Letter-Number Sequencingand Symbol Search subtests of the Wechsler Adult Intelli-gence Scale (WAIS-III; Wechsler, 1997). Vocabulary indexedverbal ability, whereas Matrix Reasoning indexed non-verbalability, Letter-Number Sequencing reflected auditory working

48 R.W. Heinrichs et al. / Schizophrenia Research 109 (2009) 46–51

memory and Symbol Search tapped processing speed. These 4tests represent the ability factors underlying the WAIS-III.Verbal memory acquisition and learning were measured withthe total number of words recalled over 5 trials on theCalifornia Verbal Learning Test (CVLT-II; Delis et al., 2000).Visual attention and response inhibition was assessed withthe signal detection score from the Continuous PerformanceTest (CPT-II; Conners, 2000), and word generation andfluency with phonemic trials of the Controlled Oral WordAssociation Test (COWAT; Benton et al., 1994). The selection oftests was based on considerations of efficiency and the natureand number of separable ability factors underpinningcognitive performance in schizophrenia patients (Nuechter-lein et al., 2004). However, social cognition and visualmemory functions were not represented in the battery.

Table 1Demographic and treatment characteristics of schizophrenia patients.

Outcome category

Highlyfavorable

Highlyunfavorable

χ2/t

Age (M, SD) 41.64 11.32 40.83 8.54 − .29Sex (n males/n total) 14/28 18/24 3.41High school completion (%) 86 71 1.72Parental SES (%) 54 61 .97Employed (%) 64 58 6.27Married (%) 25 0 9.98 a

Foreign born (%) 29 8 3.41First language (English, %) 71 83 1.03Medication (atypical, %) 89 74 4.61Medication (anti-Parkinsonian, %) 11 43 7.14 a

Note: SES = socioeconomic status, percent professional+white collar;employed = full and part time.

a Pb.05.

2.3. Classification of patients with divergent functional outcomesand statistical analysis

Functional outcome and community independence weremeasured with the Global Support rating from the Multi-dimensional Scale of Independent Functioning (MSIF; Jaegeret al., 2003) in both patients and healthy communityparticipants. The MSIF comprises a structured interview andself-report measure, with verification of information providedby history, chart abstractions, proxy reports and informantinterviews. The Global Support rating reflects the amount ofassistance a person receives across residential, work-relatedand educational settings. A rating of 1indicates no significantsupport beyond that considered a community norm forhealthy people. A rating of 2 indicates small amounts ofsupport in one or two environments, which could still be atthe low end of the normal range. In contrast, a rating of 4indicates moderate levels of support in all settings or compre-chensive support in one setting. A rating of 7 indicates that aperson receives complete services or “total support” in allsettings and even a 6 reflects comprehensive oversight andsupport. Ratings are based on detailed “anchors” provided foreach environment. The Global Support rating has been shownto have good inter-rater reliability (intraclass correla-tion=.75) and its component ratings vary significantly andin expected directions with differences in patients' residen-tial, work and educational situations. Although MSIF data arebased largely on self-report, previous studies show extremelylarge differences between patient and healthy controlsamples as well as stability over time periods up to 12months(Heinrichs et al., 2008a,b; Heinrichs et al. in press). The MSIFGlobal Support rating was used to assign patients to highlyfavorable (HF) and highly unfavorable (HU) outcome groups.The criterion for HF assignment was an MSIF score≤2 and forHU assignment a score ≥6 was required.

Independent groups t tests, with P values adjusted formultiple comparisons, were used to evaluate group differ-ences in mean demographic variables, symptom ratings andcognitive test scores. Frequency data were evaluated with theChi-squared statistic. Effect sizes (Cohen's d), defined as thedifference in group means divided by the pooled standarddeviation, were calculated for each symptomatic, cognitiveand functional group comparison. Binary logistic regressionwith groupmembership (HF/HU) as the criterionwas used to

assess the ability of symptomatic and cognitive data to predictoutcome extremes.

3. Results

Mean Global Support MSIF scores in patients (3.81, S.D.=1.42) and healthy comparison samples (1.15, S.D.=.36) differedsignificantly (t(228)=22.03, Pb.001) and corresponded to aneffect size (Cohen's d) of 2.24. This confirmed the validity of theMSIF as a highly discriminating measure of communityindependence. Group assignment criteria applied to the poolof 156 patients yielded 28 HF and 24 HU assignments.Accordingly, 18% of the clinical sample demonstrated commu-nity independence levels that approached normal function and15% were extremely dependent on support services. Within theHF group, 5 patients had ratings of 1 and 23 had ratings of 2 onthe 7-point MSIF support scale. In contrast, in the HU group, 22patients had a rating of 6 and 2 had ratings of 7. Demographicand treatment statistics for HF and HU patient groups arepresented in Table 1. The two outcome groups were statisticallyequivalent in the proportion of males, number of patients withEnglish as their primary language, foreign relative to domesticbirth status, employment rate, educational achievement, anddistribution of parental socioeconomic status. However, HFoutcomepatients hadhigher rates ofmarriage and lower rates oftreatment with anti-Parkinsonianmedication than HU outcomepatients.

Symptomatic and cognitive statistics are presented inTable 2. Relative to published norms for the schizophreniapopulation, HF outcome patients had low average levels ofpositive and negative symptoms and average levels of moregeneral psychopathology (i.e. anxiety and depression) symp-toms (Opler et al., 1999). In contrast, HU patients experienced,on average, positive and general symptom severities approxi-mately 1 standard deviation above population averages aswell as mildly elevated negative symptoms. All symptomaticdifferences between HU and HF groups were statisticallysignificant and associated with large (dN .8) effect sizes.Levels of cognitive performance were in average rangesrelative to healthy standardization samples (Wechsler, 1997;Delis et al., 2000; Conners, 2000) for HF outcome patients on

Table 2Psychopathology, cognitive performance and functional outcome inschizophrenia patients.

HF outcome HU outcome Statistic

n=28 n=24

M SD M SD t(63) d

PANSSPositive 16.46 4.57 23.83 5.87 −5.10a −1.39Negative 15.79 6.37 21.67 5.43 − 3.55a − .97General 37.07 7.83 47.04 8.88 −4.30a −1.18

WAIS-IIIVocabulary 11.43 3.07 6.54 2.74 6.35a 1.65Matrix reasoning 10.57 3.12 6.42 2.32 −5.37a 1.47Symbol search 9.71 2.91 5.79 2.75 −4.96a 1.36Letter-Number sequencing 10.25 2.98 6.13 2.92 −5.02a 1.37

Total words recalled (CVLT-II)) 44.71 12.73 34.00 8.64 −3.49a .96Detectability T score (CPT-II) 44.56 10.80 53.00 7.12 3.37a − .89Words generated (COWAT) 35.32 14.91 29.75 11.89 −1.47 .40

Note: HF= highly favorable; HU= highly unfavorable; PANSS= Positive andNegative Syndrome Scale (raw scores); WAIS-III = Wechsler AdultIntelligence Scale; CVLT-II = California Verbal learning Test (trials 1–5);CPT-II = Conners' Continuous Performance Test; COWAT = Controlled OralWord Association Test.aPb.004.

49R.W. Heinrichs et al. / Schizophrenia Research 109 (2009) 46–51

all measures. In contrast, central tendencies for HU patientswere within low average or impaired ranges. The HF groupdemonstrated significantly higher cognitive performance interms of WAIS-III Vocabulary, Letter-Number Sequencing,Matrix Reasoning and Symbol Search, CPT-II signal detectionand CVLT-II recall. No group differences were found in wordgeneration. Values of Cohen's d indicate large effect sizes forall cognitive comparisons except word generation.

To assess the relative and joint validity of symptomseverity and cognitive performance as predictors of functionaloutcome categories, a logistic multiple regression approachwas used. In the first analysis, the dichotomous outcomevariable (highly favorable/unfavorable) was regressed on acomposite symptom severity variable comprising the sum ofthe 3 PANSS scores (Positive, Negative, General). This modelwas statistically significant (χ2 (1)=22.98, Pb.001) andyielded an estimated R2 of .48 and an overall outcomeclassification accuracy of 77%. The accuracy was 71% for HUand 82% for HF outcome patients. The same procedure wasthen used for a composite measure of all 7 cognitive variables,also yielding a significant regression model (χ2 (1)=23.43,Pb.001). The estimated R2 for cognition was .48, and overalloutcome classification accuracy was 81%, with 79% of the HUand 82% of the HF patients correctly classified. With bothsymptomatic and cognitive data in a joint model (χ2 (2)=33.62, Pb.001), the R2 increased to .64 and overall accuracyimproved to 83%. Statistics for individual coefficients weresignificant for both cognition (Wald=7.37, P=.007) andsymptoms (Wald=6.99, P=.008) in the two-term equation.The increment in validity obtained by adding the cognitivedata to existing symptomatic data was statistically significant(χ2 (1)=10.63, P=.001). Moreover, the conditional validityof symptoms given the prior entry of the composite cognitivevariablewas also significant (χ2 (1)=10.18, P=.001). Finally,given the specific interest in positive symptoms, the sameregression procedure was carried out by first entering the

cognition term and then the Positive scale value from thePANSS. The joint cognition-Positive model was significant(χ2 (2)=37.73, Pb.001) and yielded an estimated R2 of .69with an outcome group classification accuracy of 85%. Inaddition, the incremental validity of the Positive scale datawas significant (χ2 (1)=14.30, Pb.001).

4. Discussion

The results of this investigation of highly divergent func-tional outcomes in schizophrenia suggest that researchers haveunderestimated the importance of psychopathology includingpositive symptoms for understanding and, possibly, influencingcommunity independence. We studied patients with levels ofindependence approaching normal functionality as well asthose receiving very comprehensive oversight and supportacross life settings. Psychopathology on its own accounted foralmost half of the outcome variance. Patients with optimallevels of independence have relatively mild symptoms, whilethose highly dependent on services and supports experiencesymptom severities more typical for the schizophrenia popula-tion. These severities include positive and negative as well asnon-psychotic symptoms associated with anxiety and depres-sion. Cognitive performance also differs very substantiallybetween outcome extremes and yields effect sizes similar inmagnitude to those produced by symptoms. Moreover, bothsymptoms and cognition contribute independently and sig-nificantly to the validity of functional outcome prediction. Wealso found evidence thatpositive symptomsadduniquevalidityto outcome prediction, over and above the contribution ofcognitive performance. Indeed, the prediction model compris-ing cognitive performance and positive symptom terms yieldedthe highest absolute validity of several regressionmodels underexamination.

Aunique feature of this studywas its focus onpatientswithunequivocally good and poor functional outcomes. Thefavorable outcome group demonstrated levels of indepen-dence that overlap with those shown by healthy communitysubjects. These high-functioning patients tend to performwithin average ranges on measures of general intellectualability and reasoning, working memory, processing speed,attention and response inhibition and verbal memory acquisi-tion. Meta-analytic findings indicate that average cognitiveperformance levels are markedly atypical for schizophreniapatients, who usually perform a standard deviation or morebelow healthy control values (Heinrichs & Zakzanis, 1998). Inaddition, the favorable outcome group had positive andnegative symptom levels .5–1.0 standard deviations belowthose typically reported for treated patients (Opler et al.,1999). Symptoms associated with anxiety and depressionoccur in favorable outcome patients at levels that are fairlytypical for the schizophrenia patient population as a whole.Relative to patients with poor outcomes and high supportrequirements, very large group differences were observedacross all symptomatic as well as cognitive variables. Forexample, the positive symptom effect size corresponds to ajoint distribution overlap of only 31%. This implies that asubstantial majority of poor outcome patients experiencelevels of psychosis completely outside the range experiencedby high-functioning, more independent patients.

50 R.W. Heinrichs et al. / Schizophrenia Research 109 (2009) 46–51

The importance of symptom severity in relation to func-tional outcome extremes was confirmed by the logisticmultiple regression results. Symptomatic data separatelycapture a significant amount of outcome variance and addvalidity to cognitive predictors of outcome extremes. Overall,it is likely that schizophrenia patients with themost favorableoutcomes are distinguished by a very broad set of advantagesencompassing clinical state and cognition. Relatively mildpsychopathology across a range of symptoms within andwithout the schizophrenia diagnostic criteria, as well asrelatively preserved cognitive abilities, define very successfuloutcome. It is also noteworthy that rates of anti-Parkinsonianmedication use are almost 4 times higher in poor relative togood outcome patients. This discrepancy occurs in theabsence of other medication differences, at least as coarselymeasured by proportions of first and second generationantipsychotic treatment. It may be that very low functionalityis part of a schizophrenia syndrome that includes poorresponse to anti-psychotic medication along with increasedlikelihood of side-effects and complications.

The results of this study are consistent with and extendrecent arguments and data supporting the importance ofpsychopathology in relation to important aspects of func-tional outcome (Perlick et al., 2008; Mohamed et al., 2008).The data challenge the view that psychotic symptoms are, atbest, weak determinants of community functioning. Thisfrequently advanced view is based partly on the observationthat medication is effective in controlling the severity ofpsychopathology, but not the severity of functional impair-ment (Green, 2007). However, it is possible that residualsymptoms undermine or prevent independent functioning inthe community. Our data suggest that high functioningpatients have a successful treatment response that includesnot only effective reduction in psychopathology, but also lessneed for supplementary medication to address side effects.When coupled with relatively preserved cognitive perfor-mance many of these patients are able to function with no orminimal supports in the community. It is noteworthy,however, that although a quarter of high functioning patientsare married, compared to none in the poor outcome group,employment rates are statistically similar and remain farbelow healthy population values.

Our findings depart from the literature in terms of themagnitude of predictor-outcome relationships. For example,cognitive predictors typically demonstrate only moderatevalidity and a ceiling of approximately 30% in terms ofaccounting for functional outcome variance (Bowie et al.,2006; Twamley et al., 2002; Heinrichs et al. in press). In thepresent study, a relatively brief set of cognitive measuresaccounted for 48% of the variance in outcome extremes andjoint cognition-psychopathology models accounted for up to69%. These are very substantial validities that may reflect themethod of extreme sampling. They also imply that the muchlower validities reported in the literature may in turn reflectsampling methods that lead to under-representation ofextremes. One possibility is that such under-representationcreates restricted variance in functionality and thereforeattenuates validity coefficients. Another, and perhaps moreinteresting possibility, is that cognitive performance andsymptom severity are essential mediators of extremely goodor poor, but not modal or more typical outcomes. The

functionality of the large group of patients “in the middle”with low average cognition and partially successful symptomcontrol, may be mediated by other variables. Assessing therelative merits of both psychometric and illness-relatedexplanations for high and low validity coefficients will bethe subject of future investigations.

The primary limitation of our approach to understandingfunctional outcome in schizophrenia is the relatively smallnumber of patients that demonstrate extremes in adjust-ment. Maximizing sample size and reliability require notonly large patient pools, but careful selection of recruitmentsites to ensure representation of both highly favorableand unfavorable outcomes. The main advantages of thisapproach include enhanced validity of prediction and theclarity and meaningful nature of the outcome contrasts,which provide a perspective not available in reports usingpoorly defined continuous outcome variables (see Jaegeret al., 2003) or those based on restricted variance. However,we focused exclusively on patient support and indepen-dence across residential, occupational and educational set-tings. Research on the cognition–functional outcomerelationship has found differential relationships betweencognitive performance and specific outcome indicators(McGurk and Mueser, 2004; Evans et al., 2004). Accordingly,we are unable to address the possibility that differentpatterns of cognition and psychopathology exist in relationto, say, vocational placement, program participation or so-cial functioning. It is noteworthy, in this regard, that recentfindings show psychopathology and cognition both relatingto quality of life measures (Perlick et al., 2008; Mohamedet al., 2008).

Another consideration is the heterogeneity of clinicalstates and phase of illness found in schizophrenia. Our samplerepresented a reasonable range of symptom severities basedon the PANSS norms. However, whether symptoms andcognitive performance predict extremes in functionalityacross phase of illness or medication history and responsecannot be determined with these data. In addition, weindexed psychopathology with standard sub-scales availablein the PANSS. It is possible that alternate instruments orsymptom typologies might generate different relationshipswith functionality or overlap more or less with cognition(Dollfus and Everitt, 1998). Finally, the present study islimited by the fact that, like most research on symptoms,cognition and outcome, it is correlational in nature. This kindof evidence does not allow for strong causal inferences aboutrelations between these variables, although it does supportthe formulation of hypotheses and conjectures that mayultimately lead to such inferences.

Given the success of previous generations of anti-psychotic medication in reducing symptom severity and theimportance of symptoms in functional outcome, it would be amistake to target primarily cognitive performance andnegative symptoms in the search for new schizophreniatreatments. Indeed, the data from this investigation suggestthat both psychopathology, including positive symptoms, andcognitive impairments contribute to poor community out-come. Accordingly, behavioral and medical therapies addres-sing these two aspects of schizophrenia jointly hold the mostpromise for ensuring that larger numbers of patientsexperience favorable outcomes.

51R.W. Heinrichs et al. / Schizophrenia Research 109 (2009) 46–51

Role of funding sourceFunding for this study was provided by the Ontario Mental Health

Foundation (OMHF) and by the Community Schizophrenia VocationalRehabilitation Foundation (CSVR). The OMHF and CSVR Foundation had norole in study design; in the collection, analysis and interpretation of data; inthe writing of the report; and in the decision to submit the paper forpublication.

ContributorsW. Heinrichs designed the study and wrote the protocol. N. Ammari

wrote the protocol. A. Miles and S. McDermid Vaz collected the data.W. Heinrichs and B. Chopov conducted the analyses. All authors contributedto and approved the final manuscript.

Conflicts of interestThere are no conflicts of interest with respect to this manuscript.

AcknowledgementsWe express our appreciation to Susan Strong, MHSc, Peter Prendergast,

MB, Suzanne Archie MD, Diana Smith, B.A., Andrew Miki, MA, ElizabethFaraone, B.Sc., Frances Carullo, B.A. Ashley Oman, B.A., Joel Goldberg, Ph.D.and the staff of the Community Schizophrenia Service, the Cleghorn Program,St. Joseph's Healthcare Hamilton, The Hamilton Program for Schizophrenia,The Canadian Mental Health Association Toronto Branch, ChallengingDirections and the Whitby Mental Health Center for their assistance.

References

American Psychiatric Association Task Force on DSM -IV, 2000. Diagnosticand Statistical Manual of Mental Disorders, 4th ed. APA,Washington, D.C.

Benton, A.L., Hamsher, K.D., Rey, G.J., Sivan, A.B., 1994. Multilingual AphasiaExamination. AJA Associates, Iowa City, Iowa.

Bowie, C.R., Reichenberg, A., Patterson, T.L., Heaton, R.K., Harvey, P.D., 2006.Determinants of real-world functional performance in schizophreniasubjects: correlations with cognition, functional capacity, and symptoms.Am. J. Psychiatry 163 (3), 418–425.

Conners, C.K., 2000. Conners' CPT-II for Windows. Multi-Health Systems,North Tonawanda, New York.

Delis, D.C., Kramer, J.H., Kaplan, E., Ober, B.A., 2000. California Verbal LearningTest-2nd ed (CVLT-II). Psychological Corporation, San Antonio, Texas.

Dollfus, S., Everitt, B., 1998. Symptom structure in schizophrenia: two-three-or four-factor models? Psychopath 31 (3), 120–130.

Dunayevich, E., Ascher-Svanum, H., Zhao, F., Jacobson, J.G., Phillips, G.A.,Dellva, M.A., Green, A.I., 2007. Longer time to antipsychotic treatmentdiscontinuation for any cause is associated with better functionaloutcomes for patients with schizophrenia, schizophreniform disorder,or schizoaffective disorder. J. Clin. Psychiatry 68 (8), 1163–1171.

Emsley, R., Chiliza, B., Schoeman, R., 2008. Predictors of long-term outcome inschizophrenia. Curr. Opin. Psychiatry 21 (2), 173–177.

Evans, J.D., Bond, G.R., Meyer, P.S., Kim, H.W., Lysaker, P.H., Gibson, P.J., Tunis, S.,2004. Cognitive and clinical predictors of success in vocational rehabilita-tion in schizophrenia. Schizophr. Res. 70 (2–3), 331–342.

First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 1996. Structured ClinicalInterview for DSM-IV Axis I Disorders, Non-Patient Edition (SCID-I/NP).New York State Psychiatric Institute, Biometrics Research, New York.

Goldberg, T.E., Goldman, R.S., Burdick, K.E., Malhotra, A.K., Lencz, T., Patel, R.C.,Woerner, M.G., Schooler, N.R., Kane, J.M., Robinson, D.G., 2007. Cognitiveimprovement after treatment with second-generation antipsychoticmedications in first-episode schizophrenia: is it a practice effect? Arch.Gen. Psychiatry 64 (10), 1115–1122.

Green, M.F., 2006. Cognitive impairment and functional outcome in schizo-phrenia and bipolar disorder. J. Clin. Psychiatry 67 (suppl 9), 3–8.

Green, M.F., 2007. Stimulating the development of drug treatments toimprove cognition in schizophrenia. Annu. Rev. Clin. Psychol. 3, 159–180.

Green, M.F., Kern, R.S., Braff, D.L., Mintz, J., 2000. Neurocognitive deficits andfunctional outcome in schizophrenia: arewemeasuring the “right stuff”?Schizophr. Bull. 26 (1), 119–136.

Green, M.F., Kern, R.S., Heaton, R.K., 2004. Longitudinal studies of cognitionand functional outcome in schizophrenia: implications for MATRICS.Schizophr. Bull. 72 (1), 41–51.

Heinrichs, R.W., Zakzanis, K.K., 1998. Neurocognitive deficit in schizophrenia:a quantitative review of the evidence. Neuropsychology 12 (3), 426–445.

Heinrichs, R.W., Ammari, N., McDermid Vaz, S., Miles, A., 2008a. Areschizophrenia and schizoaffective disorder neuropsychologically distin-guishable? Schizophr. Res. 99 (1–3), 149–154.

Heinrichs, R.W., Miles, A.A., Smith, D., Zargarian, T., Goldberg, J.O., Ammari, N.,McDermid Vaz, S., 2008b. Cognitive, clinical and functional character-istics of verbally superior schizophrenia patients. Neuropsychology 22(3), 321–328.

Heinrichs, R.W., Ammari, N., Miles, A., McDermid Vaz, S., in press. Cognitiveperformance and functional competence as predictors of communityindependence in schizophrenia. Schizophr Bull. (advance access pub-lished July 30, 2008).

Helldin, L., Kane, J.M., Karilampi, U., Norlander, T., Archer, T., 2007. Remissionin prognosis of functional outcome: a new dimension in the treatment ofpatients with psychotic disorders. Schizophr. Res. 93 (1–3), 160–168.

Jaeger, J., Berns, S.M., Czobor, P., 2003. The multidimensional scale ofindependent functioning: a new instrument for measuring functionaldisability in psychiatric populations. Schizophr. Bull. 29 (1), 153–168.

Kurtz, M.M., Moberg, P.J., Ragland, J.D., Gur, R.C., Gur, R.E., 2005. Symptomsversus neurocognitive test performance as predictors of psychosocialstatus in schizophrenia: a 1- and 4-year prospective study. Schizophr.Bull. 31 (1), 167–174.

Lindstrom, E., Eberhard, J., Levander, S., 2007. Five-year follow-up duringantipsychotic treatment: efficacy, safety, functional and social outcome.Acta Psychiatr. Scand. Suppl. 435, 5–16.

Lucas, S., Fitzgerald, D., Redoblado-Hodge, M.A., Anderson, J., Sanbrook, M.,Harris, A., Brennan, J., 2004. Neuropsychological correlates of symptomprofiles in first episode schizophrenia. Schizophr. Res. 71 (2–3), 323–330.

Marder, S.R., 2006. Initiatives to promote the discovery of drugs to improvecognitive function in severe mental illness. J. Clin. Psychiatry 67 (7), e03.

Matza, L.S., Buchanan, R., Purdon, S., Brewster-Jordan, J., Zhao, Y., Revicki, D.A.,2006. Measuring changes in functional status among patients withschizophrenia: the link with cognitive impairment. Schizophr. Bull. 32(4), 666–678.

Mattsson, M., Topor, A., Cullberg, J., Forsell, Y., 2008. Association betweenfinancial strain, social network and five-year recovery from first episodepsychosis. Soc Psychiatry Psychiatr Epidemiol. 43 (12), 947–952 (Dec).

McGurk, S.R., Mueser, K.T., 2004. Cognitive functioning, symptoms, and workin supported employment: a review and heuristic model. Schizophr. Res.70 (2–3), 147–173.

Milev, P., Ho, B.C., Arndt, S., Andreasen, N.C., 2005. Predictive values ofneurocognition and negative symptoms in functional outcome inschizophrenia: a longitudinal first-episode study with 7-year follow-up. Am. J. Psychiatry 162 (3), 495–506.

Mishara, A.L., Goldberg, T.E., 2004. A meta-analysis and critical review of theeffects of conventional neuroleptic treatment on cognition in schizo-phrenia: opening a closed book. Biol. Psychiatry 55 (10), 1013–1022.

Mohamed, S., Rosenheck, R., Swartz, M., Stroup, S., Lieberman, J.A., Keefe, R.S.,2008. Relationship of cognition and psychopathology to functionalimpairment in schizophrenia. Am. J. Psychiatry 165 (8), 978–987.

Nuechterlein, K.H., Barch,D.M., Gold, J.M.,Goldberg,T.E., Green,M.F.,Heaton,R.K.,2004. Identification of separable cognitive factors in schizophrenia.Schizophr. Res. 72 (1), 29–39.

Opler, L.A., Kay, S.R., Lindenmayer, J.P., Fiszbein, A., 1999. Structured ClinicalInterview for the Positive and Negative Syndrome Scale (SCI-PANSS).Multi-Health Systems, North Tonawanda, New York.

Perlick, D.A., Rosenheck, R.A., Kaczynski, R., Bingham, S., Collins, J., 2008.Association of symptomatology and cognitive deficits to functionalcapacity in schizophrenia. Schizophr. Res. 99 (1–3), 192–199.

Remillard, S., Pourcher, E., Cohen, H., 2008. Long-term effects of risperidoneversus haloperidol on verbal memory, attention, and symptomatology inschizophrenia. J. Int. Neuropsychol. Soc. 14 (1), 110–118.

Rocca, P., Castagna, F., Marchiaro, L., Rasetti, R., Rivoira, E., Bogetto, F., 2006.Neuropsychological correlates of reality distortion in schizophrenicpatients. Psychiatry Res 145 (1), 49–60.

San, L., Ciudad, A., Alvarez, E., Bobes, J., Gilaberte, I., 2007. Symptomaticremission and social/vocational functioning in outpatients with schizo-phrenia: prevalence and associations in a cross-sectional study. Eur.Psychiatry 22 (8), 490–498.

Twamley, E.W., Doshi, R.R., Nayak, G.V., Palmer, B.W., Golshan, S., Heaton, R.K.,Patterson, T.L., Jeste, D.V., 2002. Generalized cognitive impairments, abilityto perform everyday tasks, and level of independence in community livingsituations of older patients with psychosis. Am. J. Psychiatry 159 (12),2013–2020.

Wechsler, D., 1997.Wechsler Adult Intelligence Scale, third edition (WAIS-III).Psychological Corporation, Harcourt Brace Jovanovich, San Antonio,Texas.

Wittorf, A., Wiedemann, G., Buchkremer, G., Klingberg, S., 2008. Prediction ofcommunity outcome in schizophrenia 1 year after discharge fromimpatient treatment. Eur. Arch. Psychiatry Clin. Neurosci. 258 (1), 48–58.