Table 2. Results Of Model Building, Adjusted Ors Of Poor Inhaler Technique, Defined As Making An Error In 20% Of Steps For Any Inhaler, And For Each Device (2/11 Steps Mdi, 2/10 Advair Diskus, 3/14 Spiriva Handihaler).

	Error for any inhaler used (n=688)		MDI (n=616)		Diskus (n=190)		Spiriva (n=212)
Variable	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Education
< High School	referent		referent		referent		referent
High school	0.64 (0.21-1.92)	0.42	0.61 (0.22-1.72)	0.350	0.82 (0.36-1.85)	0.628	0.65 (0.40-2.95)	0.574
Trade school/some college	0.35 (0.17-0.70)	0.003	0.31 (0.14-0.67)	0.003	0.48 (0.21-1.10)	0.085	0.62 (0.24-1.52)	0.311
College or higher	0.25 (0.11-0.61)	0.002	0.24 (0.10-0.55)	0.001	0.36 (0.14-0.91)	0.031	0.59 (0.25-1.40)	0.230
Trend	0.62 (0.54-0.72)	<0.001	0.60 (0.52-0.69)	<0.001	0.68 (0.53-0.89)	0.004	0.84 (0.72-1.00)	0.057
Race
White	referent		referent		referent		referent
Black	3.25 (1.86-5.61)	<0.001	3.82 (2.47-5.92)	<0.001	2.24 (1.00-5.04)	0.051	2.60 (1.46-4.52)	0.001
Other	1.96 (1.14-3.35)	0.014	1.81 (0.91-3.62)	0.093	3.70 (0.33-41.1)	0.289	2.78 (0.39-19.61)	0.305
Smoker	1.46 (0.88-2.43)	0.145	1.33 (0.84-2.09)	0.225	2.05 (0.82-5.12)	0.124	1.10 (0.36-3.33)	0.864
Gold stage
No AFO	referent		referent		referent		referent
1 to 2	1.22 (0.82-1.32)	0.323	1.39 (0.98-1.96)	0.062	0.64 (0.21-2.01)	0.447	1.00 (0.41-2.42)	1.00
3 to 4	1.14 (0.68-1.92)	0.628	1.17 (0.66-2.10)	0.570	0.49 (0.13-1.88)	0.301	0.74 (0.38-1.46)	0.392
Count of Devices	1.20 (1.02-1.41)	0.029	0.97 (0.66-2.10)	0.795	1.40 (1.01-1.93)	0.043	1.05 (0.44-1.50)	0.773

Models adjusted for age in decades, income, and gender. Included variables are the primary predictor and any variable that attained p<0.1 in bivariate analyses for any inhaler, excluding variables that were highly collinear (Pearson's correlation coefficient>0.4). Bold typeface indicates statistical significance. Abbreviations: AFO=airflow obstruction