RT Journal Article
SR Electronic
T1 Usefulness of the asthma predictive index in clinical practice: A systematic review and clinical epidemiological analysis
JF European Respiratory Journal
JO Eur Respir J
FD European Respiratory Society
SP P4094
VO 40
IS Suppl 56
A1 Sotiris Fouzas
A1 Paul Brand
YR 2012
UL http://erj.ersjournals.com/content/40/Suppl_56/P4094.abstract
AB The Asthma Predictive Index (API) has been developed from follow-up data of the Tucson population-based study to predict whether preschool children with wheeze develop asthma by age 6 years. The stringent API (> 3 episodes of wheeze + one or more major criteria [eczema; parental asthma] or two or more minor criteria [allergic rhinitis; eosinophilia; wheezing apart from colds]) has been praised for its simplicity and its ability to rule out later asthma.We conducted a systematic review of studies using the API to predict the outcome of preschool wheeze, using likelihood ratios (LR) and Bayes' theorem to calculate the diagnostic value of the API to predict asthma in situations of different prevalence of disease.We found one study preliminary validating the API in the derivation cohort, two studies validating the API in different birth cohorts, and one study using the index in a clinical cohort. In the derivation cohort, the positive (LR+) and negative likelihood ratios (LR-) were 7.3 and 0.8, respectively, indicating poor to modest predictive value to rule asthma in or out at age 6. In the two population validation studies, the LR- was comparable (0.68-0.87), whilst the LR+ was lower (2.06-2.50). In the clinical study, LR+ was 2.1, LR- 0.72. In a clinical population with pretest likelihood (prevalence) of asthma at age 6 of 40%, using the API misclassifies 1 in 3 children with a negative, and 1 in 5 children with a positive API.We conclude that the API is insufficiently accurate to be used in predicting the outcome of preschool wheezing in clinical practice.