LAUSR

Background There is a growing use of electronic databases for asthma studies, many of which use exacerbations to assess asthma control, severity and determine eligibility for biologic treatment.1 Databases identify severe exacerbations based on prednisolone prescriptions, however its accuracy in counting exacerbations is not validated. We determined whether healthcare databases accurately identify asthma exacerbation numbers and the true proportion of UK primary care patients eligible for treatment with mepolizumab. Methods Demographic and treatment information were collected from adults with asthma from ten UK general practices over a 12 months period. Frequent exacerbators (FE) were defined as ≥2 exacerbations in the past year. Oral corticosteroid (OCS) prescriptions were manually checked against clinical records for FE identified by database search. Prescriptions for non-asthma indications, dose <30mg daily, or within 7 days of a previous course were removed from the total exacerbation count. Blood eosinophil counts (BEC) in the previous year were obtained for FE from NHS Safe Haven. Results Of 2639 patients with active asthma, 254 (10%) FE were identified using electronic database searches. Whereas 185 (7%) FE were confirmed after manually reviewing OCS prescriptions. Database search overestimated FE by 37% and has a positive predictive value of 73%. Of 1000 prescriptions examined from eight practices, 302 (30%) prescriptions were discounted as an asthma exacerbation. The most common reason for overcounting of exacerbations was consecutive prescriptions given within 7 days. Less common reasons included OCS prescribed for other conditions, low dose maintenance OCS or emergency supply prescriptions. 30 patients had ≥4 exacerbations in the past year and/or were on maintenance OCS dose ≥5mg daily. 22 (73%) had BEC recorded in the past year. Twelve patients were eligible for mepolizumab according to NICE criteria. Conclusion Primary care electronic database searches overestimated FE by 37%. Accuracy can be improved by adding the date, indication and daily dose of OCS in the search algorithm. Using confirmed exacerbation numbers and applying blood eosinophil criteria where available, 0.5% of patients with active asthma attending primary care in an urban area could be considered for mepolizumab. References Kerkhof, et al. Thorax 2018;73:116–124.