van Nimwegen 2008.

Methods	Allocation: randomised Blindness: double‐blind Duration: 6 weeks Design: subgroup with substance (cannabis) reported, superiority, parallel group, randomised trial. Setting: outpatients. Multisite across 4 sites in the Netherlands. (Academic Medical Centre University of Amsterdam, Erasmus Medical Centre Rotterdam, Panassia Psychomedical Centre in the Hague, Mediant in Enschede).
Participants	Diagnosis: Structured Clinical Interview for DSM (SCID‐I) diagnosis of schizophreniform disorder, schizophrenia or schizoaffective disorder, cannabis self‐report and urine testing for cannabis. N = 138 (subgroup of 41 (29.7%) used cannabis). Age: 18 to 30 years, mean age ˜25 years Sex: 80% male. Ethnicity: not reported Exclusion criteria: pregnant or lactating, no adequate contraception, known hypersensitivity to any ingredient of olanzapine or risperidone. Concomitant use of any other antipsychotic drug than olanzapine or risperidone. Use of depot anti‐psychotics for a period of at least three months prior to the study or the use of other psychotropic medication other than oxazepam or biperiden. Narrow‐angle glaucoma, neurological or endocrine disease.
Interventions	Risperidone: flexible dosing, 1.25 mg, 2.5 mg, 3.75 mg, 5 mg, titrated to a fixed dose within the first week. N = 21 Olanzapine: flexible dosing, 5 mg, 10 mg, 15 mg, 20 mg, titrated to fixed dose within first week. N = 20
Outcomes	Substance use: cannabis use self‐report scores ‒ change data (joints per week) Craving for substances: Obsessive Compulsive Drug Use Scale (OCDUS), Desires for Drug Questionairre (DDQ) ‒ endpoint data. Leaving the study early Unable to use: Subjective Well‐being: Subjective Well‐being Under Neuroleptics (SWN) score (no subgroup mean, SD or N)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description of how sequence was generated.
Allocation concealment (selection bias)	Unclear risk	No description of where sequence was kept and who allocated participants. Nevertheless tablets were described as identical‐looking.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Study described as "double blind" with identically appearing capsules, although no description is given as to how blinding was achieved. Different side‐effect profiles of the two medications could have lead to unblinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Described as "double blind" with identically appearing capsules, although no description is given as to how outcome assessors were kept masked from treatment. Different side‐effect profiles of the two medications could have lead to unblinding.
Incomplete outcome data (attrition bias) All outcomes	High risk	ITT analysis with single imputation method (LOCF). Although attrition was comparable across groups it is unclear if groups differed with regards to other factors such as symptoms severity and other baseline measures.
Selective reporting (reporting bias)	Unclear risk	Only some outcomes stated in protocol are reported. Unclear how some factors, such as symptom severity that was not reported, could have impacted on reported outcomes of SWN and craving.
Other bias	High risk	Study funded by Eli‐Lilly