S3-eu-west-1.amazonaws.com



Table 1: Long-Acting Injectable antipsychotics versus oral antipsychotics (review question 2)Patient or population: patients with schizophrenia and poor medication adherenceIntervention: Long-Acting Injectable antipsychoticsComparison: oral antipsychoticsOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskOral antipsychoticsLong-Acting Injectable antipsychoticsRelapse (longest follow-up, at least 6 months)283 per 1000263 per 1000(224 to 312)RR 0.93 (0.79 to 1.10)5329(21 studies)⊕???very low1,2,3,4NSKey outcomeHospitalisation (at least 1 hospitalisation within study duration, at least 6 months)258 per 1000224 per 1000(180 to 278)RR 0.87 (0.70 to 1.08)2390(10 studies)⊕⊕??low1,3NSAll-cause discontinuation414 per 1000401 per 1000(360 to 447)RR 0.97 (0.87 to 1.08)4978(19 studies)⊕⊕??low1,3NSMortality9 per 10005 per 1000(3 to 12)RR 0.60 (0.28 to 1.30)4302(8 studies)⊕???very low1,3,4NSQuality of lifeQLSThe mean quality of life in the intervention groups was0.64 standard deviations lower(1.99 lower to 0.72 higher)906(2 studies)⊕???very low1,3,4NSInjection site adverse events11 per 100089 per 1000(8 to 1000)RR 7.80 (0.68 to 89.73)1055(2 studies)⊕???very low1,3,4NSDiscontinuation due to adverse events33 per 100035 per 1000(26 to 47)RR 1.06 (0.78 to 1.45)4749(18 studies)⊕???very low1,3,4,7NSNumber of violent episodes per month during the studyThe mean number of violent episodes per month during the study in the intervention groups was1.19 lower(1.84 to 0.54 lower)46(1 study)⊕⊕??low3,8Statistically significant butbased on only one studyRehospitalisation, cohort designModerateHR 0.36 (0.17 to 0.75)5(1 study)⊕⊕??low5,6Statistically and clinically significantNOTE: cohort studyAll-cause discontinuation, cohort designModerateHR 0.41 (0.27 to 0.61)5(1 study)⊕⊕??low5,6Statistically and clinically significantNOTE: cohort studyRisk of hospitalisation, mirror image design(Follow-up: 12 months)ModerateRR 0.43 (0.35 to 0.53)(16 studies)⊕???very low9,10Statistically and clinically significantNOTE: mirror-image studiesNumber of hospitalisations, mirror image designModerateRR 0.38 (0.24 to 0.51)(15 studies)⊕???very low9,10Statistically and clinically significantNOTE: mirror-image studies*The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).CI: Confidence interval; RR: Risk ratio; HR: Hazard ratio; NS: Not statistically significant ScalesQLS: Heinrichs-Carpenter Quality of Life ScaleGRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Many studies with unclear randomisation sequence generation and allocation concealment and/or high risk of performance/detection bias.2 Studies before 2005 report positive findings compared with studies after 2005, but even in studies after 2005 there was some inconsistency between results.3 Indirectness.4 Either end of the CI would yield a different result.5 Adjusted for: age at diagnosis, sex, duration of first hospital episode, and current and previous use of anxiolytics, hypnotics and sedatives, antidepressants, drugs used in addictive disorders, analgesics, antiparkinsonian drugs, blood glucose-lowering drugs, lipid-modifying agents, previous use of antipsychotics, during the follow-up and the choice of initial antipsychotic (serving as a surrogate for the patient’s clinical status at baseline and thus reflecting the clinical correlates determining the selection of treatment).6 One single study.7 Inconsistent results across included studies8 Only 1 study9 Mirror-image studies are associated with high risk of bias, e.g. expectation bias, selection bias and regression to the mean.10 Estimates vary between studies (not all CIs overlap).Table 2: SSRI compared with placebo for negative symptoms (review question 3)Patient or population: schizophreniaIntervention: SSRI as add-on to antipsychotic therapyComparison: placeboOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskPlaceboAntidepressants (SSRI)Negative symptoms, end of treatment (duration: 4 weeks to 6 months) PANSS, SANS, BPRSThe mean negative symptoms at end of treatment in the intervention groups was0.31 standard deviations lower(0.51 to 0.10 lower)565(14 studies)⊕⊕??low1,2Small effect size of questionable clinical relevanceKey outcomePositive symptoms, end of treatment (duration: 4 weeks to 6 month) PANSS, SAPS, BPRSThe mean positive symptoms at end of treatment in the intervention groups was0.07 standard deviations lower(0.25 lower to 0.11 higher)492(12 studies)⊕⊕⊕?moderate1NSAll-cause discontinuation (duration: 4 weeks to 6 months)128 per 1000168 per 1000(107 to 263)RR 1.38 (0.88 to 2.16)473(11 studies)⊕???very low1,2,3NSNeurological side effects, end of treatment The mean neurological side effects at end of treatment in the intervention groups was0.02 standard deviations lower(0.32 lower to 0.28 higher)336(8 studies)⊕⊕⊕?moderate1NSAgitation, end of treatment Number of events308 per 100058 per 1000(6 to 609)RR 0.19 (0.02 to 1.98)26(1 study)⊕⊕??low4NSQuality of life, end of treatment QLSThe mean quality of life at end of treatment in the intervention groups was6.30 standard deviations lower(17.22 lower to 4.62 higher)47(1 study)⊕⊕⊕?moderate4NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).SSRI: Selective serotonin reuptake inhibitor; CI: Confidence interval; RR: Risk ratio; NS: Not statistically significant-152408318500ScalesPANSS: Positive and Negative Syndrome ScaleSANS: Scale for Assessment of Negative SymptomsBPRS: Brief Psychiatric Rating ScaleSAPS: Scale for Assessment of Positive SymptomsQLS: Quality of Life ScaleGRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Considerable number of risk-of-bias items judged as 'unclear'.2 Asymmetric funnel plot.3 Different ends of CI yields different conclusion.s4 Small sample size.Table 3: SNRI compared with placebo for negative symptoms Patient or population: schizophreniaIntervention: SNRI as add-on to antipsychotic therapyComparison: placeboOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskPlaceboAntidepressants (SNRI)Negative symptoms , end of treatment (duration: 4 weeks to 6 months) PANSSThe mean negative symptoms at end of treatment in the intervention groups was1.38 standard deviations lower(2.07 to 0.68 lower)40(1 study)⊕???very low1,2,3Large effect size, but based on only one studyKey outcomePositive symptoms , end of treatment (duration: 4 weeks to 6 month) PANSSThe mean positive symptoms at end of treatment in the intervention groups was0.00 standard deviations higher(0.62 lower to 0.62 higher)40(1 study)⊕???very low1,2,3,4NSAll-cause discontinuation ( duration: 4 weeks to 6 months)200 per 1000150 per 1000(38 to 586)RR 0.75 (0.19 to 2.93)40(1 study)⊕???very low1,2,3,4NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).SNRI: serotonin and noradrenaline reuptake inhibitor; CI: Confidence interval; RR: Risk ratio; NS: Not statistically significant38106985000ScalesPANSS: Positive and Negative Syndrome ScaleGRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Risk of performance bias (insufficient blinding).2 Small sample size.3 Only one study.4 Different ends of CI yield different conclusions.Table 4: Maintenance antipsychotic treatment for non-remitted schizophrenia patients (review question 4)Patient or population: patients with non-remitted schizophreniaIntervention: maintenance antipsychotic treatmentComparison: placeboOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskPlaceboMaintenance antipsychotic treatmentRelapse up to 3 months335 per 1000147 per 1000(124 to 178)RR 0.44 (0.37 to 0.53)1737(10 studies)⊕⊕⊕?moderate1,2Relapse risk more than halved in the short termRelapse from 7 months to 1 year552 per 1000210 per 1000(177 to 254)RR 0.38 (0.32 to 0.46)3038(18 studies)⊕⊕⊕?moderate1,2,3Relapse risk more than halved in the medium termKey outcomeNumber of participants hospitalised (> 7 months)285 per 1000145 per 1000(114 to 188)RR 0.51 (0.40 to 0.66)1402(8 studies)⊕⊕⊕?moderate1,2Risk of hospitalisation halvedKey outcomeAdverse effects: weight gain >= 7% (7 to 12 months)30 per 100084 per 1000(38 to 184)RR 2.83 (1.29 to 6.20)1145(4 studies)⊕⊕⊕?moderate2,3Risk of substantial weight gain increased 2.8 timesAdverse effects: at least one adverse event (7 to 12 months)438 per 1000424 per 1000(385 to 464)RR 0.97 (0.88 to 1.06)1826(6 studies)⊕⊕??low2,4,5NSLeaving the study early due to adverse events (> 7 months)36 per 100027 per 1000(17 to 45)RR 0.76 (0.46 to 1.26)1782(11 studies)⊕⊕??low1,2,5NSSuicide (7 to 12 months)2 per 10001 per 1000(0 to 17)RR 0.32 (0.01 to 7.86)1055(4 studies)⊕⊕⊕?moderate2,5NSSuicide attempt4 per 10003 per 1000(0 to 28)RR 0.7 (0.07 to 6.65)610(2 studies)⊕⊕??low2,5,6NSQuality of life (7 to 12 months)QoLThe mean quality of life (7 to 12 months) in the intervention groups was0.01 standard deviations lower(0.29 lower to 0.26 higher)205(1 study)⊕⊕??low3,7,8NSSocial functioningGAF, PSPThe mean social functioning in the intervention groups was0.12 standard deviations higher(0.46 lower to 0.7 higher)346(2 studies)⊕⊕??low2,5,9NSViolent/aggressive behaviour (7 to 12 months)190 per 100057 per 1000(29 to 114)RR 0.30 (0.15 to 0.60)288(2 studies)⊕⊕⊕?moderate2,9Substantial reductions in violent/aggressive behaviour, but based on few studies* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).CI: Confidence interval; RR: Risk ratio; NS: Not statistically significant-571563500ScalesQoL: Schizophrenia Quality-of-Life ScaleGAF: Global Assessment of Functioning ScalePSP: Personal and Social Performance Scale GRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Many studies with unclear randomisation sequence generation and allocation concealment.2 Only maintenance trials of not remitted patients were included, however, all the trials only recruited patients previously stabilised on antipsychotic drug treatment and many trials required a fairly low burden of symptoms to be included in the maintenance versus placebo phase.3 High risk of attrition bias.4 All included studies high risk of attrition bias.5 Either end of the CI would give a different results.6 Half of items either unclear or low risk of attrition bias7 High risk of performance bias, attrition bias and other bias (study was stopped after interim analysis and showed clear advantage of AP).8 Included patients had a low symptom score.9 Of few studies available many items with high risk of bias.Table 5: Family intervention (review question 5)Patient or population: patients with schizophrenia and functional impairmentIntervention: family interventionComparison: TAUOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskTAUFamily interventionFamily burden, end of treatmentFBIS, SBAS, SDSS, FBS, ZBI, CFIThe mean family burden at end of treatment in the intervention groups was0.56 standard deviations lower(1.13 lower to 0.01 higher)386(8 studies)⊕⊕??low1,2,3,4NSKey outcomeClinical relapse, end of treatment377 per 1000208 per 1000(177 to 245)RR 0.55 (0.47 to 0.65)2760(34 studies)⊕⊕??low1,2,3,5,6,7Risk of relapse halved in the short termKey outcomeClinical relapse, longest follow-up (min 4-6 months)487 per 1000375 per 1000(292 to 477)RR 0.77 (0.60 to 0.98)634(11 studies)⊕⊕⊕?moderate1,2,3,6Risk of relapse significantly reduced in the longer termDays at hospital, end of treatmentThe mean days at hospital at end of treatment in the intervention groups was3.20 lower(4.54 to 1.86 lower)533(8 studies)⊕⊕⊕?moderate1,2Clinically relevant reduction of inpatient durationCarer satisfaction, end of treatmentSSQ6, VSSS, PSQThe mean carer satisfaction at end of treatment in the intervention groups was0.34 standard deviations higher(0.05 to 0.63 higher)275(4 studies)⊕⊕⊕?moderate1,2,6Moderate and clinically relevant effect sizeQuality of life, end of treatmentQLSThe mean quality of life at end of treatment in the intervention groups was0.50 standard deviations higher(0.25 to 0.75 higher)263(2 studies)⊕⊕⊕?moderate1,2,3,6Moderate and clinically relevant effect sizeSocial functioning, end of treatmentSFS, SLFS, SOFAS. SDSS, HoNOS(A lower value reflects better functioning)The mean social functioning at end of treatment in the intervention groups was0.42 standard deviations lower(0.70 to 0.15 lower)772(10 studies)⊕⊕??low1,2,3,4Small effect sizeCrime (imprisonment), longest follow-up (min 4-6 months)158 per 1000150 per 1000(35 to 654)RR 0.95 (0.22 to 4.14)39(1 study)⊕⊕??low1,2,3,6,8NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).1333511557000TAU: Treatment as usual; CI: Confidence interval; RR: Risk ratio; NS: Not statistically significant Scales:FBIS: Family Burden Interview ScheduleSBAS: Social Behaviour Assessment ScheduleSDSS: Social Disability Screening Schedule FBD: Family Burden ScaleZBI: Zarit Burden InterviewCFI: Camberwell Family InterviewSSQ6: The Six-item Social Support Questionnaire VSSS: Verona Service Satisfaction ScalePSQ: Patient Satisfaction QuestionnaireSFS: Social Functioning ScaleSLFS: Specific Level of Functioning ScaleSOFAS: Social and Occupational Functioning Assessment ScaleHoNOS: Health of the Nation Outcome ScaleQLS: Quality of Life ScaleGRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Risk of selection bias1 Risk of selection bias.2 Risk of performance bias.3 Risk of detection bias.4 High degree of heterogeneity among studies.5 Risk of attrition bias.6 Risk of reporting bias.7 Funnel plot suggests risk of publication bias.8 95% CI could be in favour of both intervention and control.Table 6: Neurocognitive training (review question 6)Patient or population: patients with schizophrenia and functional impairmentIntervention: neurocognitive trainingComparison: TAUOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskNeurocognitive trainingGlobal cognition score (Z score), end of treatmentThe mean global cognition score (z score) at end of treatment in the intervention groups was0.28 standard deviations higher(0.7 lower to 0.13 higher)118(2 studies)⊕⊕??low1,2,3NSKey outcomeSocial functioning, end of treatmentSBS, SFS, SSSI, WHODAS, SOFASThe mean social functioning at end of treatment in the intervention groups was0.56 standard deviations higher(0.16 to 0.96 higher)479(6 studies)⊕⊕??low1,2,4Moderate effect sizeSocial functioning, longest follow-up (min 4-6 months)SBS, SFS, SoFASThe mean social functioning at longest follow-up in the intervention groups was0.26 standard deviations higher(0.01 to 0.51 higher)261(4 studies)⊕⊕??low1,2,3,5Small effect sizeWorking memory, end of treatmentANS, ACT, BACS, WAIS, WAIS II, WAIS III, WAIS-R The mean working memory at end of treatment in the intervention groups was0.66 standard deviations higher(0.27 to 1.04 higher)574(9 studies)⊕⊕??low1,2,4,5Moderate effect sizeVerbal learning and memory (total score), end of treatmentHVLT, RAVLTThe mean verbal learning and memory at end of treatment in the intervention groups was0.50 standard deviations higher(1.37 lower to 2.37 higher)97(2 studies)⊕⊕??low2,4,5,6NSVerbal learning, end of treatmentRAVLT, CVLT, WLM, WMS-ST, HVLTThe mean verbal learning at end of treatment in the intervention groups was0.23 standard deviations higher(0.09 to 0.55 higher)330(6 studies)⊕???very low1,2,3,4Small effect sizeVerbal memory, end of treatmentCVLT, HVLT, RAVLT, Cognistat, Groebe DfR16, BACS, WMS-LT, HVLT-RThe mean verbal memory at end of treatment in the intervention groups was0.34 standard deviations higher(0.04 lower to 0,71 higher)578(10 studies)⊕⊕??low1,2,4,5NSSymptoms, End of treatmentPANSS, BPRSThe mean symptoms at end of treatment in the intervention groups was0.12 standard deviations lower(0.32 lower to 0.08 higher)367(6 studies)⊕⊕⊕?moderate1,2NSQuality of life, end of treatmentQOLI, OLS, SQoLThe mean quality of life at end of treatment in the intervention groups was0.85 standard deviations higher(0.34 lower to 2.03 higher)257(4 studies)⊕???very low1,2,3,4,5NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).TAU: Treatment as usual; CI: Confidence interval; NS: Not statistically significant38106985000Scales:SBS: Social Behaviour ScheduleSFS: Social Functioning ScaleSSSI: Scale of Social Skills of chronic schizophrenia InpatientsWHODAS: Disability Assessment Schedule from World Health OrganizationSOFAS: Social and Occupational Functioning Assessment ScaleANS: Auditory Number SequencingACT: Auditory Consonant TrigramsBACS: Brief Assessment of Cognition in SchizophreniaWAIS: Wechsler Adult Intelligence ScaleWAIS II: Wechsler Adult Intelligence Scale IIWAIS III: Wechsler Adult Intelligence Scale III WAIS-R: Wechsler Adult Intelligence Scale–RevisedRAVLT: Rey Verbal Learning and Memory TestCVLT: California Verbal Learning TestWLM: Word List Memory testWMS-ST: Wechsler Memory Scale STHVLT: Hopkins Verbal Learning TestCognistat: Neurobehavioral Cognitive Status ExaminationGroebe DfR16: Grober and Buschke test (differed free recall)WMS-LT: Wechsler Memory Scale LTHVLT-R: Hopkins Verbal Learning Test-RevisedPANNS: Positive and Negative Syndrome ScaleBPRS: Brief Psychiatric Rating Scale GRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Risk of selection bias.2 Risk of performance bias.3 95% CI could be in favour of both intervention and control.4 Considerable inconsistency between studies.5 Risk of attrition bias.6 Risk of reporting bias.Table 7: Social cognitive training (review question 7)Patient or population: patients with schizophrenia and functional impairmentIntervention: social cognitive trainingComparison: TAUOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskTAUSocial cognitive trainingTheory of mind, end of treatmentPST, Hinting task, Attribution of intentions1The mean theory of mind at end of treatment in the intervention groups was0.29 standard deviations higher(0.40 lower to 0.98 higher)126(3 studies)⊕⊕??low2,3,4,5,6NSKey outcomeTheory of mind, longest follow-up (min 4-6 months)Eyes task, hinting taskThe mean theory of mind at longest follow-up in the intervention groups was0.45 standard deviations higher(0.67 lower to 1.57 higher)99(2 studies)⊕⊕??low2,3,5,6NSEmotion processing/emotion perception, end of treatmentPFA, ERT, POFA, Emotion discrimination taskThe mean emotion processing/emotion perception at end of treatment in the intervention groups was0.81 standard deviations higher(0.50 to 1.12 higher)178(5 studies)⊕⊕⊕?moderate2,3,4,5Large effect sizeEmotion processing/emotion perception, longest follow-up (min 4-6 months)FEITThe mean emotion processing/emotion perception at longest follow-up in the intervention groups was2.65 higher(0.78 to 4.52 higher)39(1 study)⊕⊕??low3,5,7Large effect sizeKey outcomeSocial function, end of treatmentSFS, VSSS, GSFS, Whodas21The mean social function at end of treatment in the intervention groups was0.02 standard deviations lower(0.32 lower to 0.27 higher)178(4 studies)⊕⊕⊕?moderate2,3,4NSSocial function, longest follow-up (min 4-6 months)SFS, VSSS, GSFS, PSPThe mean social function at longest follow-up in the intervention groups was0.54 standard deviations higher(0.04 to 1.04 higher)200(4 studies)⊕⊕⊕?moderate2,3,4Moderate effect sizeSocial perception, end of treatmentEPS, TASITThe mean social perception at end of treatment in the intervention groups was0.06 standard deviations higher(0,51 lower to 0,38 higher)77(2 studies)⊕⊕⊕?moderate2,3,4NSSocial perception, longest follow-up (min 4-6 months)TASITThe mean social perception at longest follow-up in the intervention groups was0.00 standard deviations higher(0.51 lower to 0.50 higher)60(1 study)⊕⊕??low2,3,4,7NSSymptomatic relapse, end of treatment259 per 1000194 per 1000(117 to 321)RR 0.75 (0.45 to 1.24)238(3 studies)⊕⊕??low2,3,8NSSymptoms, end of treatmentPANSS, BPRSThe mean symptoms, end of treatment in the intervention groups was0.08 standard deviations lower(0.39 lower to 0.22 higher)266(6 studies)⊕⊕⊕?moderate2,3,4NSQuality of life, end of treatment QLS-S, WHOQoL Social, SF-36 Mental Health, QWBThe mean quality of life at end of treatment in the intervention groups was0.49 standard deviations higher(0.01 lower to 0.98 higher)204(5 studies)⊕⊕??low2,3,4,6NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).TAU: Treatment as usual; CI: Confidence interval; RR: Risk ratio; NS: Not statistically significant-1524010985500Scales:PST: Theory of Mind Picture Sequencing TaskHinting task: Corcoran, Mercer, & Frith, 1995Attribution of intentions: Number of correct answers. Eyes task: Number of correct answers.PFA: Pictures of Facial AffecsERT: Emotion Recognition TestPOFA: Pictures of Facial AffectEmotion discrimination task: Number of correct answers.FEIT: Face Emotion Identification TaskSFS: Social Functioning ScaleVSSS: Verona Service Satisfaction ScaleGSFS: Global Social Functioning ScaleWhodas: Disability Assessment Schedule from World Health OrganizationPSP: Personal and Social Performance Scale EPS: “Escala de Percepción Social” (translated: Social Perception ScaleTASIT: The Awareness of Social Inference TaskPANNS: Positive and Negative Syndrome ScaleBPRS: Brief Psychiatric Rating ScaleQLS-S: The Quality of Life Scale – SocialWHOQoL Social: WHO-Quality of Life—short version (Social domain)SF-36 Mental Health: SF-36 Health Survey (Mental health domain) QWB: Quality of Well-Being ScaleGRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Scales reversed.2 Risk of selection bias.3 Risk of performance bias.4 Risk of detection bias.5 Risk of reporting bias.6 Considerable inconsistency between studies.7 Small sample size.8 95% CI could be in favour of both social cognition and TAU.Table 8: Cognitive behavioural therapy for persisting symptoms (review question 8)Patient or population: patients with schizophrenia and functional impairmentIntervention: cognitive behavioural therapyComparison: TAUOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskTAUCBTPsychotic symptoms, end of treatmentPANSS positive, SAPS, BPRS positiveThe mean psychotic symptoms, end of treatment in the intervention groups was0.34 standard deviations lower(0.58 to 0.10 lower)1078(15 studies)⊕⊕⊕?moderate1,2,3Small effect size but clinically relevantKey outcomeNegative symptoms, end of treatmentPANSS negative, SANS, BPRS negative, BRIANSThe mean negative symptoms, end of treatment in the intervention groups was0.32 standard deviations lower(0.59 to 0.04 lower)1214(18 studies)⊕⊕⊕?moderate1,2,3,4Small effect size but clinically relevantKey outcomePsychotic symptoms, longest follow-up ( min 4-6 month)PANSS positive, SAPS, BPRS positiveThe mean psychotic symptoms at longest follow-up in the intervention groups was0.09 standard deviations lower(0.38 lower to 0.19 higher)892(10 studies)⊕⊕??low1,2,5NSNegative symptoms, longest follow-up ( min 4-6 month)PANSS negative, SANS, BPRS negative, BRIANSThe mean negative symptoms at longest follow-up in the intervention groups was0.08 standard deviations lower(0.3 lower to 0.13 higher)1011(12 studies)⊕???very low1,2,4,5,6NSSocial function, end of treatmentSOFAS, SPS, SFS, GAS, GAFThe mean social function, end of treatment in the intervention groups was0.07 standard deviations higher(0.10 lower to 0.23 higher)575(8 studies)⊕⊕⊕?moderate1,2NSDistress, end of treatment PSYRATS(hallucinations)The mean distress, end of treatment in the intervention groups was0.22 lower(1.28 lower to 0.84 higher)202(5 studies)⊕⊕??low1,2,5,6NSRelapse, end of treatment213 per 1000171 per 1000(102 to 282)RR 0.80 (0.48 to 1.32)363(4 studies)⊕⊕??low1,2,6NSQuality of life, end of treatmentQLS, WHOQOL, QSQThe mean quality of life, end of treatment in the intervention groups was0.03 standard deviations higher(0.32 lower to 0.38 higher)297(4 studies)⊕⊕??low1,2,4,6NSDays in hospital, end of interventionThe mean days in hospital, end of intervention in the intervention groups was10.64 lower(32.14 lower to 10.86 higher)425(4 studies)⊕⊕??low1,2,6NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).TAU: treatment as usual; CI: confidence interval; RR: Risk ratio; NS: Not statistically significant3810-254000Scales:PANSS-Positive: Positive and Negative Syndrome Scale (Positive Domain)SAPS: Scale for Assessment of Positive SymptomsBPRS-Positive: Brief Psychiatric Rating Scale (Positive domain)PANSS-Negative: Positive and Negative Syndrome Scale (Negative domain)SANS: Scale for the Assessment of Negative SymptomsBPRS-Negative: Brief Psychiatric Rating Scale (Negative domain)BRIANS: Brief Rating Instrument for Assessment of Negative Symptoms ScaleSOFAS: Social and Occupational Functioning Assessment ScaleSPS: Social Provision Scale:SFS: Social Functioning ScaleGAS: Global Assessment ScaleGAF: Global Functioning ScalePSYRATS: Psychotic Symptom Rating Scales (Hallucinations)QLS: Quality of Life ScaleWHOQOL: World Health Organization’s Quality of Life scaleQSQ: Quality of Life Enjoyment and Satisfaction Questionnaire:GRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Risk of performance bias.2 Risk of reporting bias.3 Inconsistency is explained by the study by Grant et al. (low I? without it).4 Risk of selection bias.5 Considerable heterogeneity.6 95% CI could be in favour of both TAU and CBT with clinical relevance.Table 9: Cognitive behavioural therapy + motivational interviewing for cannabis and/or central stimulant abuse (review question 9)Patient or population: patients with schizophrenia and comorbid cannabis and/or central stimulant abuseIntervention: cognitive behavioural therapy + motivational interviewingComparison: TAUOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskTAUCognitive behavioural therapy + motivational interviewingCannabis use, end of treatmentFrequency per month, OTI scaleThe mean cannabis use at end of treatment in the intervention groups was0.06 standard deviations lower(0.42 lower to 0.29 higher)127(2 studies)⊕⊕⊕?moderate1,2,3,4NSKey outcomeAmphetamine, end of treatmentOTI scaleThe mean amphetamine at end of treatment in the intervention groups was0.16 higher(0.73 lower to 1.04 higher)20(1 study)⊕???very low1,4NSKey outcomeCannabis use, longest follow-up (min 4-6 months)Frequency per month, OTI scaleThe mean cannabis use at longest follow-up in the intervention groups was0.03 standard deviations higher(0.34 lower to 0.41 higher)168(3 studies)⊕⊕⊕?moderate5NSAmphetamine longest follow-up (min 4-6 months)OTI scaleThe mean amphetamine, estimated daily use, at 12 months follow-up in the intervention groups was0.13 higher(0.11 lower to 0.37 higher)17(1 study)⊕???very low2,3,4,5NSSymptoms, end of treatmentPANSS,SANSThe mean symptoms, end of treatment in the intervention groups was0.06 standard deviations higher(0.38 lower to 0.51 higher)158(3 studies)⊕⊕⊕?moderate6NSRelapse (mental state), end of treatment556 per 1000278 per 1000(117 to 650)RR 0.5 (0.21 to 1.17)36(1 study)⊕⊕??low4,7NSUse of alcohol, end of treatmentFrequency per month, OTI scaleThe mean use of alcohol at end of treatment in the intervention groups was0.32 standard deviations higher(0.17 lower to 0.81 higher)68(2 studies)⊕⊕??low2,4,6,8NSQuality of life, end of treatmentBQOL, WHOQOL, MANSAThe mean quality of life at end of treatment in the intervention groups was0.17 standard deviations lower(0.48 lower to 0.13 higher)190(3 studies)⊕⊕⊕?moderate5NSSocial functioning, end of treatment SFS, GAF average score The mean social functioning at end of treatment in the intervention groups was0.08 standard deviations lower(0.54 lower to 0.37 higher)209(3 studies)⊕⊕⊕?moderate9NSMortality, end of treatmentModerateRR 0.72 (0.22 to 2.41)493(3 studies)⊕⊕⊕?moderate7,10NS31 per 100022 per 1000(7 to 75)Number of imprisonments, longest follow-up (min 4-6 months)265 per 1000132 per 1000(55 to 286)OR 0.42 (0.16 to 1.11)110(1 study)⊕⊕??low1,7NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).TAU: Treatment as usual, CI: Confidence interval; RR: Risk ratio; OR: Odds ratio; NS: Not statistically significant -57156032500Scales:OTI scale: Opiate Treatment Index ScalePANSS-Negative: Positive and Negative Syndrome Scale (Negative domain)SANS: Scale for the Assessment of Negative SymptomsBQOL: Brief Quality of Life ScaleWHOQOL: World Health Organization’s Quality of Life scaleMANSA: Manchester Short Assessment of Quality of LifeSFS: Social Functioning ScaleGAF: average score: Global Assessment of Functioning (average score)GRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Risk of attrition bias.2 Risk of selection bias.3 Risk of reporting bias.4 Small sample size.5 Risk of performance bias.6 Risk of detection bias.7 Absolute effect contains both evidence for and against treatment.8 Skewed data.9 Sign of heterogeneity10 Risk of bias estimated as serious. All 3 studies had a significant dropout rate and with unclear risk of bias. Table 10: Assertive Community Treatment (review question 10)Patient or population: patients with schizophrenia and unable to retain contact with outpatient mental health servicesIntervention: Assertive Community TreatmentComparison: TAUOutcomesIllustrative comparative risks* (95% CI)Relative effect(95% CI)No of Participants(studies)Quality of the evidence(GRADE)CommentsAssumed riskCorresponding riskTAUAssertive Community TreatmentLoss of contact, longest follow-up(max 24 months)263 per 1000105 per 1000(71 to 160)RR 0.40 (0.27 to 0.61)1538(8 studies)⊕⊕⊕?moderate1,2,3,4Statistically significant and clinically relevant findingKey outcomeDays of hospital per month, longest follow-up(max 24 months)The mean days of hospital per month at longest follow-up in the intervention groups was0.86 lower(1.38 to 0.35 lower)3717(26 studies)⊕⊕??low1,2,3,4,5Statistically significant and clinically relevant findingOther health care costs, longest follow-up (max 24 months)Emergency room visits311 per 1000352 per 1000(224 to 548)RR 1.13 (0.72 to 1.76)178(1 study)⊕⊕⊕?moderate1,2,3,4NSSymptoms, longest follow-up (max 24 months)CSI, BPRS, SCL-90, PSE, CPRS, split-GAFThe mean symptoms at longest follow-up in the intervention groups was0.27 standard deviations lower(0.38 to 0.15 lower)1289(10 studies)⊕⊕⊕?moderate1,2,3,4Small effect sizeQuality of life, longest follow-up (max 24 months)QOLI, LQoLP, MANSA, The mean quality of life at longest follow-up in the intervention groups was0.10 lower(0.36 lower to 0.16 higher)453(6 studies)⊕⊕??low1,2,3,4,5NSPatient satisfaction, longest follow-up (max 24 months) UKU ConSat, CSQThe mean patient satisfaction at longest follow-up in the intervention groups was0.75 standard deviations higher(0.38 to 1.11 higher)127(2 studies)⊕⊕⊕?moderate2,4Moderate effect sizeMortality (all causes), longest follow-up (max 24 months)38 per 100032 per 1000(18 to 51)RR 0.89 (0.53 to 1.51)1742(12 studies)⊕⊕⊕?moderate1,2,3,4NSSocial functioning, longest follow-up (max 24 months)DAS, ISSI, RFS, SCSThe mean social functioning in the intervention groups was0.28 standard deviations higher(0.65 higher to 0.10 lower)260(3 studies)⊕⊕??low1,2,3,4,5NSCrime, longest follow-up (max 24 months)police contact, arrests , imprisoned158 per 1000133 per 1000(82 to 210)RR 0.84 (0.52 to 1.33)1404(10 studies)⊕⊕??low1,2,3,4,5NS* The basis for the assumed risk is the median control group risk across studies; if otherwise explanation is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).TAU: Treatment as usual; CI: Confidence interval; RR: Risk ratio; NS: Not statistically significant-342907874000Scales: CSI: Colorado Symptom IndexBPRS: Brief Psychiatric Rating ScaleSCL-90: Hopkins symptoms check listPSE: Present State ExaminationCPRS: Comprehensive Psychopathological Rating Scalesplit-GAF: Global Assessment of FunctioningQOLI: Lehman’s Quality of Life InterviewLQoLP: Lancashire Quality of Life ProfileMANSA: Manchester Short Assessment ofQuality of LifeUKU ConSat: UKU Consumer Satisfaction Rating ScaleCSQ: Client Satisfaction QuestionnaireDAS: WHO Psychiatric Disability Assessment ScheduleISSI: InterviewSchedule for Social InteractionRFS: Role Functioning ScaleSCS: Strauss-Carpenter ScaleGRADE Working Group grades of evidenceHigh quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.Very low quality: We are very uncertain about the estimate.1 Risk of attrition bias. 2 Risk of reporting bias.3 Risk of selection bias.4 Risk of performance bias.5 Wide variation across studies. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download