The association between hospital length of stay before rapid response system activation and clinical outcomes: a retrospective multicenter cohort study

Table 4 Analysis in the propensity score-based 1:1 matched patients

Variables	Early deterioration (N = 4454)	Late deterioration (N = 4454)	P value^c
Primary outcome
28-day mortality	859 (19.3%)	1108 (24.9%)
Unadjusted OR	1.00	1.39 (1.26–1.54)	< 0.001
Adjusted OR¹	1.00	1.66 (1.45–1.91)	< 0.001
Secondary outcome
In-hospital mortality	999 (22.5%)	1309 (29.4%)
Unadjusted OR	1.00	1.45 (1.31–1.59)	< 0.001
Adjusted OR^a	1.00	1.72 (1.51–1.95)	< 0.001
ICU admission	1325 (29.8%)	1362 (30.6%)
Unadjusted OR	1.00	1.04 (0.95–1.14)	0.392
Adjusted OR^a	1.00	1.10 (0.99–1.22)	0.080
Hospital LOS after RRS activation (days)	11 (5–21)	14 (6–27)	< 0.001
LOS in ICU^b (days)	4 (2–9)	5 (3–10)	0.305

LOS length of stay, RRS rapid response system, OR odds ratio, IRR incidence rate ratio, ICU intensive care unit
^aTo minimize effects of remaining confounding factors after matching, adjusted analyses were also performed. Following variables were adjusted: age, department of admission (medical vs. surgical), whether patients were in postoperative state, comorbidities of solid cancer, hematological malignancy, chronic lung disease, hepatobiliary disease, or cerebrovascular disease, whether DNR (do-not-resuscitate) discussion was made after RRS activation, and NEWS2 (national early warning score 2)
^bAnalysis on LOS in ICU included only patients who were admitted to the ICU
^cP values were calculated by conditional logistic regression for binary outcomes and by Wilcoxon signed-rank test for continuous outcomes

ISSN: 1465-993X