TABLE 3 ] .

Multivariable Analysis of the Cross-sectional Associations of Individual QOL Outcomes on CB-Only Status as Predictor (Referent: CAO Only)

Indexes	LSC (n = 643)	P Value	COPDGene (n = 2,731)	P Value
SGRQ
Total score	0.26 ± 0.02	< .001	0.67 ± 0.01	< .001
Impacts score	0.45 ± 0.03	< .001	0.81 ± 0.01	< .001
Activity score	0.24 ± 0.02	< .001	0.54 ± 0.01	< .001
Symptom score	0.43 ± 0.02	< .001	0.71 ± 0.01	< .001
Cough	3.42 (2.16-5.40)	< .001	…	…
Phlegm	6.80 (4.25-10.86)	< .001	…	…
Wheeze	3.49 (1.81-6.72)	< .001	…	…
Attack-free days	-0.08 ± 0.05	.13	…	…
Good days	-0.05 ± 0.05	.31	…	…
am wheeze	1.74 (1.03-2.93)	.04	…	…
Item interrupted sleep	2.47 (1.48-4.14)	< .001	…	…
SF-36
General Health Perception	−0.11 ± 0.01	< .001	−0.25 ± 0.01	< .001
Bodily Pain	−0.11 ± 0.01	< .001	−0.22 ± 0.01	< .001
Vitality	−0.14 ± 0.01	< .001	−0.24 ± 0.01	< .001
Mental Health	−0.06 ± 0.01	< .001	−0.13 ± 0.01	< .001
Role Emotional	−0.07 ± 0.01	< .001	−0.19 ± 0.01	< .001
Social Functioning	−0.09 ± 0.01	< .001	−0.19 ± 0.01	< .001
Derived Depressive Mood	1.61 (1.06-2.44)	.03	…	…

Data are presented as PE ± SE or OR (95% CI). CAO was defined by an absolute postbronchodilator value of FEV₁/FVC < 70%. CB was defined as the presence of at least two of the following three criteria: (1) doctor diagnosis of CB; (2) phlegm production over the previous 4 wk; and (3) phlegm production for most days for ≥ 3 consecutive months during the year. After 18 comparisons, the Bonferroni correction is estimated to be 0.003. Data prefaced with “SGRQ” were obtained from the SGRQ, including its component activity, impact, and symptom dimension, or from its individual items that were coded for presence or absence and significant in the univariate analysis. A positive parameter estimate suggests a worse QOL for the CB group. Data prefaced with “SF-36” were obtained from the SF-36 questionnaire, including its component dimensions, or from its individual items that were coded for presence or absence. A negative parameter estimate suggests a worse QOL for the CB group. Exclusion criterion was a history of asthma; inclusion criteria were ≥ 20 pack-y of smoking and age 40-75 y. A priori covariates included self-reported age, sex, Hispanic ethnicity, pack-years of smoking, current smoking status, and baseline FEV₁ % predicted. P values for categorical variables were based on logistic regression (OR = OR as estimate of strength of association) and for continuous variables were based on linear regression methods (PE = PE as estimate of strength of association). PE = parameter estimate; QOL = quality of life. See Table 1 for expansion of other abbreviations.