Table 2.

Logistic regression analyses for the study populations with BD I, BD II, BD NOS and the combined population, adjusted for age, sex, and psychiatric comorbidity

Logistic regression analyses	n	aOR	CI, 95%	P	R2
BD I
Any episode	5618	0.78	0.70–0.87	< 0.001	0.06–0.08
Depressive episode	5607	0.81	0.72–0.90	< 0.001	0.05–0.07
Manic episode	5592	0.85	0.72–0.97	0.02	0.02–0.03
Hypomanic episode	5598	0.90	0.79–0.99	0.05	0.03–0.04
Mixed episode	5556	0.93	0.78–1.07		0.03–0.06
BD II
Any episode	5574	0.90	0.84–0.96	0.001	0.07–0.1
Depressive episode	5566	0.93	0.87–0.99	0.02	0.06–0.08
Hypomanic episode	5557	0.87	0.80–0.93	< 0.001	0.06–0.09
Mixed episode	5547	0.96	0.87–1.04		0.03–0.06
BD NOS
Any episode	3424	0.82	0.74–0.90	< 0.001	0.1–0.14
Depressive episode	3412	0.82	0.74–0.90	< 0.001	0.09–0.12
Hypomanic episode	3406	0.93	0.83–1.02		0.06–0.09
Mixed episode	3412	1.01	0.89–1.13		0.05–0.09
Total, all subpopulations
Any episode	14,616	0.84	0.80–0.88	< 0.001	0.09–0.12
Depressive episode	14,585	0.87	0.83–0.92	< 0.001	0.08–0.10
Hypomanic episode	14,561	0.89	0.84–0.94	< 0.001	0.05–0.08
Mixed episode	14,515	0.94	0.88–1.01	0.097	0.04–0.07

Adjusted odds ratios (aOR) estimating differences for patients in counties with high lithium prevalence compared with counties with low lithium prevalence. The aOR can be interpreted as the decreased odds for a mood episode in the population when the lithium prescription rate is greater by 10% points