Trained intensivist coverage and survival outcomes in critically ill patients: a nationwide cohort study in South Korea

Annals of Intensive Care

Table 3 Mixed effects Cox regression model for 1-year mortality

Variable	HR (95% CI)	P value
Intensivist group (non-intensivist group)	0.85 (0.83, 0.89)	< 0.001
Age, year	1.03 (1.03, 1.04)	< 0.001
Sex, men (vs women)	1.18 (1.17, 1.19)	< 0.001
Having a job (vs unemployment)	0.96 (0.94, 0.97)	< 0.001
Household income level
Q1 (Lowest) (vs medical aid program)	0.84 (0.82, 0.86)	< 0.001
Q2 (vs medical aid program)	0.84 (0.83, 0.85)	< 0.001
Q3 (vs medical aid program)	0.80 (0.78, 0.82)	< 0.001
Q4 (Highest) (vs medical aid program)	0.78 (0.76, 0.80)	< 0.001
Unknown (vs medical aid program)	0.88 (0.85, 0.90)	< 0.001
Residence
Rural area (vs Urban area)	1.04 (1.03, 1.05)	< 0.001
Unknown (vs Urban area)	1.52 (0.47, 1.56)	< 0.001
CCI, point	1.12 (1.11, 1.13)	< 0.001
Admitting department
IM (vs non-IM)	1.67 (1.64, 1.69)	< 0.001
Hospital admission through ER	1.34 (1.33, 1.35)	< 0.001
Hospital level^a
B (vs A)	1.32 (1.31, 1.33)	< 0.001
C (vs A)	0.70 (0.68, 0.73)	< 0.001
Isolated ICU admission	1.30 (1.28, 1.32)	< 0.001
Surgery associated hospital admission	0.76 (0.76, 0.77)	< 0.001
Year of admission
2017 (vs 2016)	0.96 (0.95, 0.97)	< 0.001
2018 (vs 2016)	0.85 (0.84, 0.86)	< 0.001
2019 (vs 2016)	0.84 (0.83, 0.85)	< 0.001

HR hazard ratio; CI confidence interval; ICU intensive care unit; CCI Charlson comorbidity index; IM internal medicine; ED emergency department; ECMO extracorporeal membrane oxygenation; CRRT continuous renal replacement therapy
^aHospital location, type of hospital, total number of doctors, specialist doctors, nurses, and pharmacists, total number of hospital beds, and total number of operating rooms were used for hierarchical approach to account for clustering at the level of the hospital. Detailed information was presented in Additional file 1: Table S1