Table 1 lists the pathogen, host, behavioural and environmental/building related factors significantly associated with indoor air bioaerosol load after backward elimination in (logistic)generalised linear models

From: Indoor air surveillance and factors associated with respiratory pathogen detection in community settings in Belgium

Remaining variables

p-value

Adjusted odds ratio and 95% CI

(a) Pathogen detection (all pathogens) in a logistic regression model

Pathogen

<0.0001

 

Age group

<0.0001

Month

0.0024

CO2

0.0015

1.09 (CI 1.03–1.15) per 100 ppm increase in CO2

Natural ventilation

0.0097

0.88 (CI 0.80–0.97) per step increase (Likert scale)

Remaining variables

p-value

Coefficient and 95% CI

(b) Pathogen concentration (qPCR Ct of positive samples, all pathogens) in a linear regression model

Pathogen

<0.0001

 

Age group

<0.0001

Month

0.0020

CO2

<0.0001

−0.08 (CI −0.12 to −0.04) per increase of 100 ppm

Portable air filtration

0.0005

0.58 (CI 0.25–0.91)

  1. Panel a lists the factors significantly associated with pathogen detection in a logistic regression model. It also shows effect sizes (odds ratios and 95% CI) for CO2 and natural ventilation, after adjustment for pathogen, age group and month. See Supplementary Table 4 for unadjusted estimates. Panel b) lists the factors significantly associated with pathogen concentration (measured in qPCR Ct values) in a linear regression model. It also shows effect sizes (change in Ct value and 95% CI) for CO2 and portable air filtration, after adjustment for pathogen, age group and month. See Supplementary Table 4 for unadjusted estimates. P-values are two-sided. They were estimated using the Chi squared method (no adjustment for multiple comparisons). Almost identical results of alternative models are shown in Supplementary Table 4.
Back to article page