# Table 3 Multivariable analyses for the association between the hypothesis variables and guideline concordance*#

Hypothesis variables

LDL concordance (95% CI; p-value)

BP concordance (95% CI; p-value)

Patient factors

Age (per 10 year increment)

0.94 (0.91â€“0.97; p < 0.01)

0.86 (0.83â€“0.89; p < 0.01)

Distance to medical center (per 25 mile increment)

0.99 (0.99â€“1.00; p = 0.19)

0.99 (0.99â€“1.00; p = 0.30)

Diabetes

1.44 (1.35â€“1.54; p < 0.01)

1.01 (0.94â€“1.08; p = 0.89)

Vascular disease&

0.99 (0.91â€“1.07; p = 0.76)

0.91 (0.84â€“0.99; p = 0.03)

Processes of care

Total medications (per 1 medication increment)

1.01 (1.00â€“1.01; p < 0.01)

1.01 (1.00â€“1.01; p = 0.03)

Total Visits (per 1 visit increment)

1.02 (1.01â€“1.02; p < 0.01)

1.02 (1.01â€“1.02; p < 0.01)

Recent cardiac hospitalization

1.28 (1.11â€“1.49; p < 0.01)

1.34 (1.14â€“1.57; p < 0.01)

Structures of care

Clinic volume

1.01 (1.00â€“1.02; p = 0.76)

0.82 (0.68â€“0.99; p < 0.01)

On-site cardiologist

1.14 (1.00â€“1.30; p = 0.04)

1.10 (0.99â€“1.22; p = 0.10)

1. *The cumulative logit model estimates the probability that each subject is in a lower (i.e., better) category compared to a higher (i.e., worse) category of guideline concordance. The reported odds ratio is the probability of being in a lower category (i.e., better concordance) compared to a higher category (i.e., worse concordance). The outcome categories for LDL were the following: 0) LDL measurement and LDL <100 mg/dl, 1) LDL measurement, but LDLâ‰¥100 mg/dl, and 2) no LDL measurement. The outcome categories for BP were the following: 0) SBP<140 and DBP<90 mm Hg; 1) SBPâ‰¥140 or DBPâ‰¥90 mm Hg, and 2) no BP measurement.
2. #Candidate covariates included all Table 1 and 2 variables.
3. &Presence of cerebrovascular and/or peripheral vascular disease