An issue has been identified with the estimates reported in Table 5 of our published article.
Table 5.
Multivariate analysisa of diabetes medication adherence.
| Multivariate analysis (N = 1678) | Estimate | Odds Ratio (95% CI) | p-value |
|---|---|---|---|
| Age Group | 0.097 | ||
| Age Group (70–74 vs. 65–69) | 0.216 | 1.241 (0.768, 2.007) | 0.374 |
| Age Group (75+ vs. 65–69) | −0.301 | 0.740 (0.438, 1.251) | 0.258 |
| Disability (No vs. Yes) | −0.447 | 0.640 (0.446, 0.918) | 0.016* |
| Gender (Male vs. Female) | 0.145 | 1.155 (0.817, 1.634) | 0.409 |
| Health Literacy | 0.003 | 1.003 (0.986, 1.021) | 0.718 |
| Household Couldn't Afford Balanced Meals (No vs. Yes) | 0.361 | 1.435 (1.031, 1.996) | 0.033* |
| Interview Language (Spanish vs. English) | 0.606 | 1.832 (1.173, 2.862) | 0.008* |
| Race | 0.722 | ||
| Race (Black vs. White) | −0.131 | 0.877 (0.567, 1.358) | 0.552 |
| Race (Other vs. White) | 0.016 | 1.016 (0.674, 1.533) | 0.938 |
| Usual Place for Health Care (No vs. Yes) | −0.527 | 0.590 (0.280, 1.246) | 0.164 |
Statistical Methods
The multivariable analysis utilized logistic regression with applicable reference categories to examine the effect of age group, disability, gender, health literacy score, affordability of balanced meals, interview language, race, and usual place of health care on diabetes medication adherence.
A 5% significance level was used for all tests. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Inc., Cary, NC). Adjustments for the NHANES survey sampling design were included in each of the analyses with SAS’ PROC SURVEYLOGISTIC procedure.
The authors would like to apologize for any inconvenience caused.