Skip to Main Content
Table 2—

Multivariate regression analysis of 3-year cost (n = 1,694)

VariableCoefficientSEZ-statP95% CI
Female 0.0802 0.0900 0.89 0.373 −0.0963 to 0.2567 
Age 62 years 0.0174 0.0031 5.57 0.000 0.0113–0.0235 
(Age 62 years) ∧2 −0.0003 0.0001 −2.22 0.026 −0.0006 to 0.0000 
Female × (age 62 years) −0.0097 0.0041 −2.37 0.018 −0.0177 to −0.0017 
[Female × (age 62 years)]∧2 0.0003 0.0002 1.29 0.196 −0.0002 to 0.0008 
A1c 7.5% and A1c >7.5% 0.0534 0.0220 2.43 0.015 0.0103–0.0966 
A1c 7.5% and A1c <7.5% −0.0218 0.0431 −0.51 0.613 −0.1063 to 0.0627 
Duration of diabetes (12 years) 0.0019 0.0027 0.71 0.478 −0.0034 to 0.0072 
CHD 0.7639 0.0646 11.83 0.000 0.6374–0.8905 
Hypertension 0.2233 0.0471 4.74 0.000 0.1310–0.3156 
Lipids 0.0951 0.0597 1.59 0.111 −0.0219 to 0.2122 
Depression 0.3043 0.0709 4.29 0.000 0.1654–0.4432 
Pharmacy coverage 0.3995 0.0714 5.59 0.000 0.2594–0.5395 
Low income 0.1196 0.0699 1.71 0.087 −0.0175 to 0.2566 
Low education −0.0265 0.0917 −0.29 0.773 −0.2062 to 0.1532 
Constant 9.1013 0.0985 92.37 0.000 8.9081–9.2944 
VariableCoefficientSEZ-statP95% CI
Female 0.0802 0.0900 0.89 0.373 −0.0963 to 0.2567 
Age 62 years 0.0174 0.0031 5.57 0.000 0.0113–0.0235 
(Age 62 years) ∧2 −0.0003 0.0001 −2.22 0.026 −0.0006 to 0.0000 
Female × (age 62 years) −0.0097 0.0041 −2.37 0.018 −0.0177 to −0.0017 
[Female × (age 62 years)]∧2 0.0003 0.0002 1.29 0.196 −0.0002 to 0.0008 
A1c 7.5% and A1c >7.5% 0.0534 0.0220 2.43 0.015 0.0103–0.0966 
A1c 7.5% and A1c <7.5% −0.0218 0.0431 −0.51 0.613 −0.1063 to 0.0627 
Duration of diabetes (12 years) 0.0019 0.0027 0.71 0.478 −0.0034 to 0.0072 
CHD 0.7639 0.0646 11.83 0.000 0.6374–0.8905 
Hypertension 0.2233 0.0471 4.74 0.000 0.1310–0.3156 
Lipids 0.0951 0.0597 1.59 0.111 −0.0219 to 0.2122 
Depression 0.3043 0.0709 4.29 0.000 0.1654–0.4432 
Pharmacy coverage 0.3995 0.0714 5.59 0.000 0.2594–0.5395 
Low income 0.1196 0.0699 1.71 0.087 −0.0175 to 0.2566 
Low education −0.0265 0.0917 −0.29 0.773 −0.2062 to 0.1532 
Constant 9.1013 0.0985 92.37 0.000 8.9081–9.2944 

The dependent variable is an estimate of 3 years of health care costs. The generalized linear regression uses a γ distribution with a log-link function and a covariance matrix that is robust to heterscedasticity and allows for clustering of variance by primary care clinic. Observations are weighted by length of enrollment.

Close Modal

or Create an Account

Close Modal
Close Modal