Iloperidone 2016.

Study characteristics
Methods	Randomisation: computer‐generated random sequence (1:1 ratio). Allocation: centralised, interactive voice response system. Blinding: double‐blind, identical appearing capsules. Duration: terminated early after 68 relapse events, patients were followed up for up to 26 weeks. Design: parallel. Location: multi‐centre. Setting: outpatients.
Participants	Diagnosis: Schizophrenia (DSM‐IV), not hospitalised ar the time of screening, then stabilised on iloperidone for at least 12 weeks before the relapse‐prevention phase. N = 303. Gender: 178 men, 125 women. Age: 38.3 years. History: duration stable‐ clinically stable for at least 12 weeks, no change in treatment for at least 4 weeks, duration ill‐ 12.4 years, mean duration of hospitalisation‐ n.i., number of previous hospitalisations‐ n.i., age at onset‐ 25.9 years, severity of illness‐ mean PANSS total score 55,4, mean CGI‐S total score 3,3, baseline antipsychotic dose‐ n.i., remission at baseline‐ n.i..
Interventions	1. Drug: Iloperidone. N = 153 Flexible dose. Mean dose: 15 mg/day. Allowed dose range: 8 mg/day to 24 mg/day. 2. Placebo: duration of taper: n.i. N = 150. Rescue medication: anticholinergics (antiparkinson medication), lorazepam (agitation, severe restlessness, insomnia), zolpidem (insomnia).
Outcomes	Examined Relapse (rating scale based and/or clinical judgement and/or need for hospitalisation or increase in the level of psychiatric care). Leaving the study early ‐ due to adverse events. Social functioning: Sheehan Disability Scale. Adverse effects Death Unable to use/Not included Leaving the study early ‐ due to any cause/inefficacy (no usable data, no crude numbers for relapse are available, subjects withdrawn due to early termination counted as dropouts). Global state‐ Number of participants improved (no usable data) Global state ‐ Number of participants in remission (no usable data, CGI‐I not reported). Mental state: PANSS change in total score, BPRS change in total score (no predefined outcomes of interest)
Notes	Sponsored by Vanda Pharma.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random sequence.
Allocation concealment (selection bias)	Low risk	Centralised, interactive voice response system.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blind, identical appearing capsules
Blinding (performance bias and detection bias) Subjective outcomes	Unclear risk	Double‐blind, identical appearing capsules
Blinding (performance bias and detection bias) Objective outcomes	Low risk	Double‐blind, identical appearing capsules
Incomplete outcome data (attrition bias) All outcomes	High risk	The overall attrition rate (57%) was high, and more participants in the placebo group left the study early due to relapse. This difference may have biased the results of outcomes other than leaving the study early and relapse, which was assessed using the Kaplan‐Meier survival curve analysis. Data for secondary outcomes were analysed on an ITT basis (not full ITT, because only participants who received at least one dose of study medication were included) .
Selective reporting (reporting bias)	Unclear risk	Some secondary efficacy outcomes are reported incompletely so that they cannot be entered in the meta‐analysis (e.g. CGI‐I scores).
Other bias	High risk	Terminated early, but it was pre‐planned.