Table 4

Multivariate regressions for predictors of hospital mortality, length of stay and 30-day readmission
Model Clinical variable Odds ratio (95% CI) p value
Logistic regression for hospital mortality (n =34523) Provider Group
Traditional-FP Reference
FP-Hospitalist 0.881 (0.779 – 0.996) 0.043
GIM-Hospitalist 0.355 (0.288 – 0.436) <0.001
Other-IM 0.490 (0.426 – 0.562) <0.001
Age 1.025 (1.022 – 1.028) <0.001
Charlson score
0 Reference
1–2 3.300 (2.816 – 3.868) <0.001
3 or more 9.417 (8.072 – 10.988) <0.001
Length of stay
0 – 1 days 2.847 (2.563 – 3.161) <0.001
2 – 9 days Reference
10 or more days 1.282 (1.174 – 1.399) <0.001
Number of interventions
0 Reference
1 0.909 (0.829 – 0.997) 0.044
2 or more 1.278 (1.154 – 1.414) <0.001
Clinical variable Odds ratio (95% CI) p value
Logistic regression for 30-day readmission (n = 30902) Provider Group
Traditional-FP Reference
FP-Hospitalist 0.766 (0.678 – 0.867) 0.010
GIM-Hospitalist 0.800 (0.675 – 0.948) 0.004
Other-IM 0.831 (0.734 – 0.941) <0.001
Age 1.005 (1.003 – 1.007) <0.001
Initial discharge to long-term care facility 0.443 (0.393 – 0.500) <0.001
Charlson score
0 Reference
1–2 1.474 (1.344 – 1.617) <0.001
3 or more 2.442 (2.220 – 2.687) <0.001
Length of stay
0 – 1 days 1.349 (1.228 – 1.482) <0.001
2 – 9 days Reference
10 or more days 1.149 (1.051 – 1.257) 0.002
Clinical variable Ods Ratio (95% CI) p value
Linear regression for length of stay (n = 34005) Constant 2.590 (2.077 – 3.103) <0.000
Provider Group
Traditional-FP Reference
FP-Hospitalist 0.096 (−0.136 – 0.329) 0.417
GIM-Hospitalist −2.975 (−3.302 – -2.647) <0.001
Other IM −3.592 (−3.829 - -3.355) <0.001
Charlson score
0 Reference
1–2 0.672 (0.512 – 0.832) <0.001
3 or more 1.986 (1.816 – 2.157) <0.001
Age in years 0.058 (0.054 – 0.061) <0.001
Transfer in from other acute care hospital 0.901 (0.570 – 1.232) <0.001
Number of interventions 1.997 (1.937 – 2.056) <0.001
Emergent admit −0.583 (−0.996 - -0.170) 0.006

Legend: FP = family physician, GIM = general internal medicine, IM = internal medicine, CI = confidence interval.

Yousefi and Chong

Yousefi and Chong BMC Health Services Research 2013 13:204   doi:10.1186/1472-6963-13-204

Open Data