Table 3—

Medication adherence and use of preventive diabetes services*

Unadjusted means/proportions
Model-based (adjusted) estimates
No major depressionMajor depressionMean differenceOdds ratio95% CIP (n)
Nonadherence to medication
    Oral hypoglycemic18.8%24.5%3.621.18–6.06<0.005
    Lipid lowering19.3%27.2%6.793.11–10.46<0.0005
    ACE inhibitors21.6%27.9%5.591.41–9.780.01
Preventive diabetes services
    No HbA1c test within the last year4.0%6.3%1.91.27–2.87<0.005 (4,347)
    Fewer than three HbA1c tests per year among patients with HbA1c >8%55.1%55.5%0.90.70–1.21NS (1,922)
    No retinal examination within the last year19.4%17.2%1.00.74–1.24NS (4,347)
    Less than two examinations within 2 years among patients with retinopathy24.8%22.5%0.80.55, 1.31NS (1,179)
    No microalbumin urine test within the last year among patients not taking ACE inhibitors47.0%45.6%1.00.80–1.33NS (1,831)
  • *

    * Nonadherence to medication outcomes are continuous measures of the percentage of nonadherent days for each individual. Preventive diabetes services outcomes are dichotomous, indicating use or no use. Model-based estimates of percentage point difference in adherence and odds ratios with corresponding CIs and P values are adjusted for age, sex, marital status, education, race/ethnicity, medication risk (a pharmacy-based comorbidity measure), diabetes complications, and treatment intensity. Models also adjust for clustering by clinic and physician.