2.4 | Statistical Analysis
This study was a retrospective comparison of two main groups of
patients: those who tested positive or negative for an RVI pre-admission
and/or upon admission. Patients who tested positive were further
categorized as having a high- or low-risk infection and as to whether
the patient presented with LR and/or UR symptoms.
We reported demographic and other characteristics as well as outcomes
overall and by RVI group: RVI negative; RVI positive, low-risk,
asymptomatic; RVI positive, low-risk, symptomatic; RVI positive,
high-risk, asymptomatic; and RVI positive, high-risk, symptomatic. We
also presented results collapsing all RVI positive groups to a single
group. Summary statistics include the frequency of each categorical
characteristic and five number summaries (mean, standard deviation,
median, interquartile range, minimum, and maximum) for each continuous
or ordinal characteristic.
We used contingency tables and logistic regression models to evaluate
differences in the primary outcomes among the RVI groups. Categorical
data were presented with their counts and frequencies and depending on
cell counts were statistically tested using either the Chi-Square Test
or the Fisher’s Exact Test. We used similar methods to evaluate all
secondary outcomes. The overall model was first evaluated, with the
overall model p-value adjusted for multiple tests using the Holm
procedure. Further comparison of RVI groups was only conducted if the
overall model test was significant.
3 | Results :
Table 1-4 present detailed data for all patients. Figures 1-3 present
findings based on data and statistical analyses.