4. Suggested design for future trial.

Methods	Allocation: centralised sequence generation with table of random numbers or computer‐generated code, stratified by severity of illness, sequence concealed till interventions assigned. Blinding: could be optional, depending on choice of outcome. Duration: 12 months.
Participants	Diagnosis: schizophrenia and co‐occurring ongoing substance misuse (clinical criteria). N = 300*. Age: adults. Sex: men and women. Setting: any.
Interventions	1. Risperidone: clinically indicated dose. N = 150. 2. Olanzapine: clinically indicated dose. N = 150.
Outcomes	Global state: CGI‐I and CGI‐S. Substance use: pragmatic binary/continuous measure. Well‐being: pragmatic binary/continuous measure. Craving: pragmatic binary/continuous measure. Service outcomes: re‐hospitalisation, days in hospital, time attending psychiatric outpatient clinic. Quality of life: important change. Adverse effects: including mortality, weight change and extrapyramidal symptoms. Satisfaction with care: patients/carers. Leaving the study early. Economic data. Other routine data, such as incidents with the police,
Notes	* size of study to detect a 10% difference in improvement with 80% certainty. For all outcomes there should be binary cut‐off points of clinically important improvement, defined before the study starts.