Osman 2002.

Methods	STUDY DESIGN: Parallel group trial. LOCATION, NUMBER OF CENTRES: Single centre in Scotland, UK. DURATION OF STUDY: 12 months COMPLIANCE: Assessed via questionnaire report (81% returned at 1 month). CONFOUNDERS: At 12 months the differences between the 2 groups of patients remained greater for those for whom this had been a first admission. At one month return of questionnaire may be motivated by satisfaction with treatment.
Participants	N SCREENED not reported (study population represented 60% of all eligible patients over the study period) N RANDOMISED: 280 N COMPLETED : 226 questionnaires returned at one month (data on readmission at 12 months collected from patient records) M = 94 F = 186 CHARACTERISTICS: 22‐43 years old, 43% current smokers, 79% treated with ICS prior to admission, 23% outpatient appointment at chest clinic outpatient in previous 12 months INCLUSION CRITERIA: 14‐64 years old and admitted to the hospital with acute asthma EXCLUSION CRITERIA: Not reported
Interventions	Education group Self‐management programme (SMP) in 2 visits of 30 minutes each by a trained respiratory nurse. Discussion about asthma, booklet, self‐management plan (symptom and peak flow based). Control group Usual care. Only one visit by the respiratory nurse, two questionnaires sent after they leave hospital. All participants received a postal questionnaire FOLLOW‐UP PERIOD: 12 months
Outcomes	Re‐admission for acute asthma within 12 months; readmission 1 month after discharge; patient satisfaction with asthma explanation; written management plan; management at discharge
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	'...they were randomised by selecting from random numbers held in sealed envelopes.'
Allocation concealment (selection bias)	Low risk	Sealed envelopes
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Participants aware as to treatment group assignment. Independent assessor blinded to patient assignment collected data on readmission within 12 months from hospital records.
Incomplete outcome data (attrition bias) All outcomes	Low risk	97% patients contributed to primary outcome (readmission information). For the remainder of the outcomes: Data collected from patient notes (follow‐up at clinic only for initial admission; questionnaires returned by post), and analysed as available case (based on questionnaire at one month post‐discharge).
Selective reporting (reporting bias)	Low risk	Review primary outcome measured, analysed and disclosed in full.
Other bias	Low risk	No other sources of bias could be identified.