Table 3.

Multivariate binary logistic regression results using the 2010 derivation cohort (n = 8,700) and with all variables maintained in the model. A patient’s RRS score is calculated by multiplying the variable value by its beta coefficient. However, for the categorical variables that are not normally distributed (mCCI, ED visits, inpatient stays, ambulatory mediations, and length of stay), the raw value identifies a beta coefficient by virtue of its category. For these categorical variables, the beta coefficient is multiplied by 1 in the RRS calculation.

Variable	Beta Coefficient	Std. Error	P-value
Acute Admission (i.e., not scheduled)	0.26	0.063	<0.0001
Age	0.007	0.002	<0.0001
Male Sex	0.28	0.048	<0.0001
Married	-0.13	0.051	0.008
ED Visits within one year
• 0	–
• 1–2	-0.25	0.053	<0.0001
• 3–5	0.82	0.073	<0.0001
• ≥6	1.16	0.113	<0.0001
Inpatient Visits within one year
• 0	–
• 1–2	0.39	0.055	0.39
• 3–5	0.82	0.073	<0.0001
• ≥6	1.16	0.113	<0.0001
Uninsured	0.22	0.10	0.03
Lives alone	0.017	0.067	0.80
Length of Stay
• 0–1	–
• 2–3	-0.16	0.069	0.019
• 4–8	0.23	0.072	0.002
• ≥9	0.39	0.099	<0.0001
Medications on the Patient’s Inpatient Medication List Two Days Prior to Discharge	0.03	0.004	<0.0001
Medications on the Patient’s At nbulatory M edication List
• 0–1	–
• 2–5	-0.10	0.056	0.087
• 6–14	-0.23	0.061	<0.0001
• ≥15	-0.48	0.105	<0.0001
Modified Charlson Comorbidity Index
• 0–3	–
• 4–8	0.68	0.075	<0.0001
• ≥9	0.72	0.096	<0.0001