Univariate binary logistic regression
Univariate binary logistic regression was performed using the enter
method to determine contribution of independent variables towards
patient readmission (Table 3 ). Neither age nor sex were
significant contributing factors to patient reattendance. Increasing
BOOST score was a significant contributor to patient readmission with an
odds ratio of 1.5 (95% CI 1.1 – 2.0, p = 0.006). Components of the
BOOST score which significantly contributed to risk of readmission
include diabetes mellitus diagnosis (OR 2.8, p = 0.014), >9
prescribed medications (OR 2.4, p = 0.045), prescribed insulin (OR 4.7,
p = 0.036), and recent hospitalisation (OR 4.8, p = <0.001).
Increasing LACE index was also significantly associated with readmission
with an odds ratio of 1.5 (95% CI 1.2 – 2.1, p = 0.003). The only
statistically significant component of the LACE index was number of
hospitalisations in the prior 6 months which had an odds ratio of 2.0
(95% CI 1.3 – 3.1, p = 0.001). Odds for emergency admission and
comorbidity index were skewed as all indexed patients were classified as
an emergency attendance, and all but one patient obtained a comorbidity
index score of 5.