Background South Asians with asthma experience much higher morbidity than whites with asthma in the UK. They also experience worse follow-up in primary care after hospitalisation compared to the white population. Previous trials in mixed ethnicity populations in the US and Europe suggest that: (a) an asthma specialist nurse intervention can reduce unscheduled asthma care; (b) education for physicians can lead to improvements in their patients' asthma control; (c) lay-led self-management programmes can improve asthma. We used adaptations of these interventions in an effort to reduce unscheduled care for South Asians with asthma living in London.

Objectives To test a multifaceted educational intervention for South Asians with asthma and their primary and secondary care physicians aimed at reducing unscheduled care.

Setting Eighty-seven general practices from two adjacent inner London boroughs were recruited and randomised to education (45 practices) or usual care (42 practices).

Participants We recruited South Asian patients with asthma from these practices attending two large nearby acute hospitals with unscheduled care.

Methods We adapted the Michigan Physician Asthma Care Education (PACE) Asthma Programme to educate specialist nurses and physicians from intervention practices in better consulting skills for ethnic minority patients. Patients recruited to the study from intervention practices were seen in outpatient care by PACE-trained specialist nurses who provided education and self-management advice, referred them for follow-up by their PACE-trained practices, and invited them to participate in an asthma adaptation of an anglicised version of the Stanford Chronic Disease Self Management Programme. Patients from control practices were seen in outpatient care by specialist nurses providing usual care. Patients were followed up for 1 year.

Main outcomes The primary outcome variable was time to first exacerbation with cough resulting in unscheduled primary and secondary care, and the secondary outcome variable was time to first asthma review in general practice.

Statistical method Cox's proportional hazards regression.

Results 375 patients participated: 183 from intervention, 192 from control practices. There was no evidence that the intervention increased time to first exacerbation with cough: median 171 days versus 189 days, adjusted HR (AHR)=1.08, 95% CI: 0.85 to 1.36. There was evidence that it reduced time to first asthma review: median 72 days versus 339 days in controls, AHR=2.04, 95% CI: 1.55 to 2.68.

Conclusion This multifaceted intervention failed to reduce unscheduled care for South Asians with asthma but did improve follow-up in primary care. Further efforts are needed to reduce unscheduled care in South Asians with asthma.