Statistical Analyses
Summary descriptive statistics included means and standard deviations or
median and range for continuous variables and frequencies and
percentages for categorical data. Characteristics such as body mass
index (BMI) that were repeated over the study period were averaged over
their individual baseline and follow-up time. Differences between airway
fungi groups were determined using ANOVAs or Chi-squared tests depending
on the distribution. Given the potential impact of ABPA on clinical
status, we also compared individuals with rare and frequent fungi by
ABPA status, comparing characteristics and outcomes across the 4 groups.
Multiple linear mixed effects regression models allowing for random
slope and intercept per patient were used to assess percent predicted
forced expiratory volume in one second (ppFEV1) over time by fungal
exposure group using maximum likelihood estimation. ppFEV1 over
follow-up time was standardized within patient by taking their average
ppFEV1 during their baseline. The first model only included time and
fungal exposure group, the second model adjusted for chronic P.
aeruginosa and ABPA status, and the third and fourth models compared
ppFEV1 by airway fungi and also adjusted for P. aeruginosastratified by ABPA status. Models were re-run adjusting for transition
from one group to another in the follow up period to assess any change
in association or inference. All three-way and two-way interactions
between time with fungal status, ABPA, chronic or ever P.
aeruginosa were assessed. Statistical significance was set at a Type I
error rate of 0.05. All data cleaning, summaries, and testing were done
using R (R Foundation for Statistical Computing, Vienna, Austria. URL
https://www.R-project.org/).