O'Byrne 1996.

Methods	Study Design: Parallel RCT Randomisation: Computer generated code‐ stratified by allergy status and randomised in blocks of 3 at each centre Concealment of Allocation: coded identical containers Double Blinding: Yes Withdrawals / dropouts: Described. Adverse events: Described. Statistical analysis: Intention to treat Jadad Score: 4
Participants	Study site: multicentre trial in 7 cities, at primary care physicians No eligible: Not stated No randomised: 57 (20/17/20) No completed: 39 (12/13/14) Sex: Male 25 Female 32 Age: mean 37/32/36 Diagnostic criteria for asthma: Objective lung function:> 10% PEF diurnal variation at least 3 of the last 7 days. Inclusion criteria: > 18 years, mild asthma and on clinical evaluation not requiring ICS. Exclusion criteria: No ICS in past 3 months, Requiring ICS Baseline severity of asthma: PEF mean (SD): Intervention 1: 364 (115); Intervention 2: 381 (112); Intervention 3 : 403 (116)
Interventions	1: BUD 800mcg/day (400mcg bid) 2: BUD 400mcg /day (200mcg bid) 3: Placebo Delivery device: nebuhaler Duration of treatment: 16 weeks Run in phase: Other: Beta agonists as needed, theophylline if remained constant through out study, allowed budesonide and/or prednisone for exacerbation during study.
Outcomes	costs, hospitalisations, ER visits, PEF, symptoms, exacerbations, adverse events
Notes	For morning PEF it was assumed the variance estimate was SE.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated code‐ stratified by allergy status and randomised in blocks of 3 at each centre