Covariate adjustment of spirometric and smoking phenotypes: The potential of neural network models.

To increase power and minimize bias in statistical analyses, quantitative outcomes are often adjusted for precision and confounding variables using standard regression approaches. The outcome is modeled as a linear function of the precision variables and confounders; however, for many complex phenotypes, the assumptions of the linear regression models are not always met. As an alternative, we used neural networks for the modeling of complex phenotypes and covariate adjustments. We compared the prediction accuracy of the neural network models to that of classical approaches based on linear regression. Using data from the UK Biobank, COPDGene study, and Childhood Asthma Management Program (CAMP), we examined the features of neural networks in this context and compared them with traditional regression approaches for prediction of three outcomes: forced expiratory volume in one second (FEV1), age at smoking cessation, and log transformation of age at smoking cessation (due to age at smoking cessation being right-skewed). We used mean squared error to compare neural network and regression models, and found the models performed similarly unless the observed distribution of the phenotype was skewed, in which case the neural network had smaller mean squared error. Our results suggest neural network models have an advantage over standard regression approaches when the phenotypic distribution is skewed. However, when the distribution is not skewed, the approaches performed similarly. Our findings are relevant to studies that analyze phenotypes that are skewed by nature or where the phenotype of interest is skewed as a result of the ascertainment condition.

Asthma Treatments and Mental Health Visits After a Food and Drug Administration Label Change for Leukotriene Inhibitors.

PURPOSE: In 2009, the US Food and Drug Administration (FDA) mandated a label change for leukotriene inhibitors (LTIs) to include neuropsychiatric adverse events (eg, depression and suicidality) as a precaution. This study investigated how this label change affected the use of LTIs and other asthma controller medications, mental health visits, and suicide attempts. METHODS: We analyzed data (2005-2010) from 5 large health plans in the US Population-Based Effectiveness in Asthma and Lung Diseases (PEAL) Network. The study cohort included children and adolescents (n = 30,000), young adults (n = 20,000), and adults (n = 90,000) with asthma. We used interrupted time series to examine changes in rates of LTI dispensings, non-LTI dispensings, mental health visits, and suicide attempts (using a validated algorithm based on a combination of diagnoses of injury or poisoning and psychiatric conditions). FINDINGS: The label change was associated with abrupt reductions in LTI use among all age groups (relative reductions of 8.3%, 15.1%, and 6.0% among adolescents, young adults, and adults, respectively, compared with expected rates at 1 year after the warnings). Although we detected immediate offset increases in non-LTI asthma medication use, these increases were not sustained among adolescents and young adults. There were small increases in mental health visits among LTI users. IMPLICATIONS: The FDA label change for LTIs communicated possible risk of neuropsychiatric events. Communication and enhanced awareness may have increased reporting of mental health symptoms among young adults and adults. It is important to assess intended and unintended consequences of FDA warnings and label changes.