Table 2.
Simple Logistic
Regression |
Multiple Logistic
Regression (N = 508) |
||||||
---|---|---|---|---|---|---|---|
Variable | n | Odds Ratio | 95% CI | P | Odds Ratio | 95% CI | P |
Female vs male | 535 | 1.6 | 1.1, 2.2 | .0069 | 1.4 | 1.0, 2.1 | .053 |
African American vs white | 535 | 0.91 | 0.62, 1.3 | .64 | 0.76 | 0.50, 1.2 | .22 |
Native American vs white | 535 | 0.78 | 0.53, 1.2 | .23 | 0.74 | 0.47, 1.2 | .19 |
Pills only vs no diabetes medication | 535 | 3.2 | 1.7, 6.2 | .0004 | 2.3 | 1.1, 4.6 | .028 |
Insulin and pills vs no diabetes medication | 535 | 6.5 | 3.2, 13 | <.0001 | 3.3 | 1.5, 7.2 | .0033 |
Log duration of diabetes | 535 | 1.5 | 1.3, 1.7 | <.0001 | 1.3 | 1.1, 1.6 | .0047 |
SF-12 Physical Health/10 | 509 | 0.89 | 0.77, 1.0 | .12 | 0.98 | 0.82, 1.2 | .86 |
SF-12 Mental Health/10† | 509 | 0.83 | 0.71, 0.97 | .018 | 0.85 | 0.71, 1.0 | .062 |
Shown how to test | 535 | 0.86 | 0.61, 1.2 | .40 | 1.0 | 0.66, 1.6 | .90 |
Helped to test | 535 | 0.59 | 0.43, 0.83 | .0021 | 0.58 | 0.39, 0.86 | .0068 |
Helped to understand results | 535 | 0.98 | 0.69, 1.4 | .90 | 0.83 | 0.53, 1.3 | .39 |
Health care provider recommended | 535 | 5.0 | 3.3, 7.7 | <.0001 | 4.3 | 2.7, 6.8 | <.0001 |
Attended diabetes class | 534 | 1.8 | 1.2, 2.8 | .0095 | 1.5 | 0.93, 2.5 | .092 |
Self-monitoring of blood glucose (SMBG) outcome was analyzed using logistic regression and the proportional odds model among participants who reported SMBG in the previous week.
The proportional odds assumption was not satisfied (P < .05) for the simple regression model only.