Table 2

Logistic regression results: risk of AME among older adults associated with beneficiary attributes, health attributes and region of residence (n=428 373 person-months)

	OR	Lower 95% CL	Upper 95% CL	p Value
Beneficiary attributes
Age at baseline	1.01	1.00	1.01	*
Male^a	1.40	1.27	1.54	***
Never married^b	0.95	0.73	1.24
Widowed^b	1.02	0.91	1.14
Separated/divorced^b	1.12	0.94	1.35
Hispanic^c	0.92	0.72	1.18
African-American^c	1.19	1.01	1.40	*
Other ethnicity/race^c	0.75	0.52	1.08
Annual income <$10 000^d	0.90	0.78	1.04
Grade school only^e	0.88	0.76	1.02
High school degree^e	0.90	0.78	1.04
Some college, no degree^e	0.89	0.77	1.04
Beneficiary health attributes
Number of chronic conditions	1.27	1.24	1.29	***
ADL limitations	1.04	1.00	1.09	*
IADL limitations	1.01	0.97	1.05
Self-reported fair to poor health^f	1.32	1.18	1.48	***
Underweight^g	1.12	0.88	1.43
Overweight^g	0.95	0.85	1.05
Obese^g	0.93	0.82	1.05
Rural–urban location
Urban location^h	0.91	0.81	1.02
Rural location^h	1.07	0.90	1.28
Northeastⁱ	1.06	0.93	1.21
Midwestⁱ	1.13	1.01	1.27	*
Westⁱ	0.99	0.86	1.14

Estimates were from random-effects multinomial logistic regression and adjusted for insurance status, death and cohort year. Robust estimates of the SEs were used. The following reference categories apply: ^afemale, ^bmarried, ^cwhite/Caucasian, ^dincome ≥10 001, ^ecollege degree or higher, ^fhealth good to excellent, ^gnormal weight, ^hsuburban location, and ⁱsouth. *p<0.05; ***p<0.001.
ADL, activities of daily living; AME, adverse medical event; CL, confidence limit; IADL, instrumental activities of daily living.