N ontuberculous mycobacteria (NTM) are opportunistic environmental pathogens that can cause chronic lung disease (1,2). NTM are identified through laboratory testing for acid-fast bacilli, which test for all mycobacteria, including Mycobacterium tuberculosis. Multiple studies have described increasing NTM pulmonary disease (NTM PD) prevalence in the United States (1,3-6), a phenomenon that might be caused by true increase in disease rates, new efficient testing technologies, increased mycobacterial testing, or any combination of those. We assessed trends in mycobacterial testing rates and NTM PD prevalence from 2009 through 2015. We also analyzed factors associated with differential testing rates and prevalence across subpopulations. The Study The population for our study comprised persons whose medical encounters were represented in the Cerner HealthFacts Electronic Health Record database (https://sc-ctsi.org/resources/cerner-healthfacts). We extracted microbiological, demographic, and clinical data for all patient encounters at 31 facilities across the United States that continually reported microbiological data (Appendix, https://wwwnc. cdc.gov/EID/article/26/9/20-0749-App1.pdf) and that speciated mycobacterial culture results from 2009 through 2015. We included only microbiology data collected from pulmonary body sites and that used the words "AFB" and "culture" in the testing description (i.e., mycobacterial culture tests). For analyses of mycobacterial culture testing and pathogenic NTM culture positivity rates (Appendix), we used the number of unique inpatients and outpatients at the 31 facilities as the population denominator. To estimate mycobacterial culture testing trends, we used Poisson regression models fit through quasilikelihood methods, which enable overdispersion (7). We analyzed trends within the overall study population and subpopulations stratified by age, sex, race/ethnicity, concurrent conditions, facility size, region, and facility teaching status. To identify variables associated with the odds of mycobacterial culture testing per facility encounter and the odds of pathogenic NTM culture positivity per facility encounter, we fit 2 mixed-effect logistic regression models to the data. We adjusted these models for patient age, sex, interactions between age and sex, race/ethnicity, teaching facility status, facility census region, encounter year, and whether the patient had a pulmonary computed tomographic scan or radiograph during the study period. The following concurrent conditions have been associated with a higher risk for NTM PD and were included as predictors in the model: bronchiectasis (4,8), chronic obstructive pulmonary disease (4,8), cystic fibrosis (CF) (9), lung cancer (4,5,8), and rheumatoid arthritis (8). We included deidentified patient number as a random effect to account for clustering among an individual patient's multiple encounters. Persons with mycobacterial culture tests were older and had more concurrent conditions than the overall population in the 31 study facilities...