Table 2.
Results from GLM predicting SGA initiation
| Variable | Odds ratio | 95% CI – lower | 95% CI – upper | p |
|---|---|---|---|---|
| Age (decade) | 0.997 | 0.997 | 0.997 | <.0001 |
| Gender (female) | 0.846 | 0.815 | 0.878 | <.0001 |
| ≥ 50% disability status | 0.740 | 0.717 | 0.763 | <.0001 |
| Marital status (married) | 1.011 | 0.985 | 1.039 | .4062 |
| Race/ethnicity a | ||||
| African American | 1.068 | 1.024 | 1.114 | .0021 |
| Hispanic | 0.989 | 0.901 | 1.087 | .8222 |
| Other race/ethnicity | 1.096 | 0.964 | 1.246 | .1614 |
| Region b | ||||
| Midwest | 0.991 | 0.955 | 1.028 | .6221 |
| South | 1.288 | 1.245 | 1.332 | <.0001 |
| West | 0.919 | 0.886 | 0.953 | <.0001 |
| Clinical/treatment factors in prior year | ||||
| Psychotic features | 1.696 | 1.520 | 1.891 | <.0001 |
| Bipolar type I (versus type II/NOS)c | 0.968 | 0.935 | 1.003 | .0736 |
| Antidepressant prescription | 0.903 | 0.882 | 0.924 | <.0001 |
| Antimanic prescription | 0.620 | 0.606 | 0.634 | <.0001 |
| Comorbidities in prior year | ||||
| Diabetes | 0.919 | 0.882 | 0.957 | <.0001 |
| Obesity | 0.958 | 0.913 | 1.004 | .0739 |
| Hyperlipidemia | 0.991 | 0.960 | 1.023 | .5860 |
| Substance abuse | 0.965 | 0.933 | 0.999 | .0407 |
| Tobacco use disorder | 1.065 | 1.030 | 1.103 | .0003 |
| Anxiety disorder | 1.082 | 1.051 | 1.114 | <.0001 |
| Sleep disorder | 1.555 | 1.435 | 1.685 | <.0001 |
| Cardiac dysrhythmia | 1.096 | 0.999 | 1.202 | .0520 |
| Liver disorder | 1.145 | 1.079 | 1.214 | <.0001 |
| Kidney disorder | 1.187 | 1.069 | 1.318 | .0014 |
| Thyroid disorder | 1.021 | 0.956 | 1.089 | .5372 |
| Traumatic brain injury | 1.109 | 0.944 | 1.303 | .2068 |
| Change per year | 0.988 | 0.982 | 0.993 | <.0001 |
aComparison: White.
bComparison: Northeast.
cPatient was labeled as bipolar type I for a given month if at least 10% of bipolar encounters to date were for bipolar type I. See text for details.