Table 10Risk of bias–haloperidol versus clozapine

StudyItemJudgmentDescription
Breier et al. 200256Adequate sequence generation?UNCLEARReported as a randomized trial, with no further details regarding sequence generation.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment in the trial report.
Blinding?YESReported as a DB trial, with investigators, raters, clinical staff, and patients blinded by using identical unmarked syringes administered by trained, unblinded, third-party personnel, who played no role in evaluating patients.
Incomplete outcome data addressed?YESITT principle used in analyses. Attrition rate was low (0.7%).
Free of selective reporting?YESProtocol was not available, but outcomes in methods and results are similar.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Citrome et al. 200162Adequate sequence generation?UNCLEARReported as a randomized trial, with no futher details regarding sequence generation.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment in the trial report.
Blinding?UNCLEARReported as a DB trial, with blinded raters performed all the clinical assessments and dosage changes requested by blinded psychiatrists, but no description of how the patients were blinded.
Incomplete outcome data addressed?NOITT principle was used during analyses, but attrition rate was high with only 58% of participants completing the trial.
Free of selective reporting?NOTwo outcomes (ESRS and NOSIE) are listed in the method section, but no results are presented.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Covington et al. 200070Adequate sequence generation?UNCLEARReported as a randomized trial, with no futher details regarding sequence generation.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment in the trial report.
Blinding?UNCLEARNo reporting of trial blinding.
Incomplete outcome data addressed?YESITT principle used during analyses and no dropouts were reproted.
Free of selective reporting?YESProtocol is not available, but outcomes in methods and results are similar.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Itoh et al. 1977155Adequate sequence generation?UNCLEARReported as a randomized trial, with no further details regarding sequence generation in trial report.
Allocation concealment?UNCLEARReported that those physicians who had conducted the trial or those who were related to the pharmaceutical company were excluded from the controllers who coded the DB trial and supervised the entire experiment.
Blinding?UNCLEARReported that the ‘coders’ were independent of physcians supervising the trial. So it seems that they were blinded, but no report of the subjects’ status.
Incomplete outcome data addressed?YESAll patients were analyzed.
Free of selective reporting?YESProtocol is not available, but outcomes in the methods and results match.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Kane et al. 200195Adequate sequence generation?YESReported as a randomized trial. Computer-generated randomization schedules (blocked by site) were provided to each site.
Allocation concealment?UNCLEARUse of sealed envelopes with treatment assignment were available to clinical personnel if needed to break the blind. Unsure whether they used sequentially numbered or opaque envelopes.
Blinding?UNCLEARReported as a DB trial: medication was administered under DB conditions. To maintain the blind, all subjects had a weekly blood draw. No further details regarding blinding.
Incomplete outcome data addressed?NOMore than 30% were droped out of the study, and no ITT was conducted.
Free of selective reporting?YESOutcomes in the method section match to the results.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Kleiser et al. 1994103Adequate sequence generation?UNCLEARReported as a randomized trial, with no further details regarding sequence generation in trial report.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment.
Blinding?UNCLEARReported as a DB trial, with no further details regarding blinding.
Incomplete outcome data addressed?YES51 patients were randomized, 51 analyzed. No report of ITT.
Free of selective reporting?YESProtocol is not available, but outcomes in the methods and results match.
Free of other bias?UNCLEARNo significant differences between treatment groups. No declaration of funding source.
Krakowski et al. 2006105Adequate sequence generation?UNCLEARReported as a block randomization with no further detail.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment.
Blinding?UNCLEARReported as a DB trial, with no further details regarding blinding.
Incomplete outcome data addressed?YESITT performed on all participants.
Free of selective reporting?YESProtocol is not available, but outcomes in the methods and results are similar.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Rosenheck et al. 1997126Adequate sequence generation?UNCLEARReported as a randomized trial, with no further details regarding sequence generation in trial report.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment in trial report.
Blinding?YESReported as a DB trial, with matching benztropine placebo. To maintain blinding, haloperidol-treated patients also received benztropine mesylate (2–10 mg/d) for extrapyramidal syndrome; clozapine patients received a matching benztropine placebo. Haloperidol patients participated in weekly blood counts as required for clozapine treatment.
Incomplete outcome data addressed?UNCLEARITT principle not used in analyses and attrition rate was moderate, with 346/423 participants not included in analyses.
Free of selective reporting?YESProtocol is not available, but outcomes in the methods and results are similar.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.
Volakva et al. 2002145Adequate sequence generation?UNCLEARReported as a randomized trial, with no further details regarding sequence generation.
Allocation concealment?UNCLEARNo information reported regarding allocation concealment in trial report.
Blinding?UNCLEARReported as a DB trial; clinicians who adjusted the dosing and outcome assessors were blinded. Patients were blinded to BDZ, not to the antipsychotics.
Incomplete outcome data addressed?YES101 patients randomized, and 101 patients analyzed. No clear description of how many randomized.
Free of selective reporting?UNCLEARNOSIE and ESRS are in method section. But, no numeric data is reported in results, just F-values, degrees of freedom, and p-values are reported.
Free of other bias?YESNo significant differences in baseline characteristics or other sources of bias detected.

BDZ = benzodiazepine; DB = double-blind; ESRS =Extrapyramidal Syndrome Rating Scale; ITT = intention-to-treat analysis; NOSIE = Nurses’ Observation Scale for Inpatient Evaluation

From: Appendix E, Risk of Bias Assessment for Randomized Controlled Trials and Nonrandomized Controlled Trials

Cover of First-Generation Versus Second-Generation Antipsychotics in Adults: Comparative Effectiveness
First-Generation Versus Second-Generation Antipsychotics in Adults: Comparative Effectiveness [Internet].
Comparative Effectiveness Reviews, No. 63.
Abou-Setta AM, Mousavi SS, Spooner C, et al.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.