Multiple linear regression analysis of the association between the PSG results and health care utilization
. | Model 1 (age), estimate (P value) . | Model 2 (genotype), estimate (P value) . | Model 3 (Hb), estimate (P value) . | Model 4 (all), estimate (P value) . |
---|---|---|---|---|
Mean SpO2 | ||||
ED visits | –0.21(<.05) | –0.30(<.001) | –0.22(<.05) | –0.11 (.29) |
Admissions | –0.24(<.01) | –0.39(<.0001) | –0.31(<.01) | –0.16 (.12) |
All encounters | –0.44(<.01) | –0.68(<.0001) | –0.51(<.01) | –0.26 (.17) |
%TST90 | ||||
ED visits | 0.05(<.0001) | 0.06(<.0001) | 0.05(<.001) | 0.04(<.01) |
Admissions | 0.04(<.01) | 0.05(<.0001) | 0.04(<.01) | 0.03(<.05) |
All encounters | 0.09(<.001) | 0.11(<.0001) | 0.10(<.001) | 0.07(<.01) |
PLMI | ||||
ED visits | 0.02 (.23) | 0.02(<.05) | 0.02 (.11) | 0.02 (.11) |
Admissions | 0.02 (.08) | 0.04(<.01) | 0.03(<.05) | 0.01 (.33) |
All encounters | 0.04 (.09) | 0.06(<.01) | 0.05(<.05) | 0.02 (.33) |
. | Model 1 (age), estimate (P value) . | Model 2 (genotype), estimate (P value) . | Model 3 (Hb), estimate (P value) . | Model 4 (all), estimate (P value) . |
---|---|---|---|---|
Mean SpO2 | ||||
ED visits | –0.21(<.05) | –0.30(<.001) | –0.22(<.05) | –0.11 (.29) |
Admissions | –0.24(<.01) | –0.39(<.0001) | –0.31(<.01) | –0.16 (.12) |
All encounters | –0.44(<.01) | –0.68(<.0001) | –0.51(<.01) | –0.26 (.17) |
%TST90 | ||||
ED visits | 0.05(<.0001) | 0.06(<.0001) | 0.05(<.001) | 0.04(<.01) |
Admissions | 0.04(<.01) | 0.05(<.0001) | 0.04(<.01) | 0.03(<.05) |
All encounters | 0.09(<.001) | 0.11(<.0001) | 0.10(<.001) | 0.07(<.01) |
PLMI | ||||
ED visits | 0.02 (.23) | 0.02(<.05) | 0.02 (.11) | 0.02 (.11) |
Admissions | 0.02 (.08) | 0.04(<.01) | 0.03(<.05) | 0.01 (.33) |
All encounters | 0.04 (.09) | 0.06(<.01) | 0.05(<.05) | 0.02 (.33) |
Model 1 includes age; model 2 includes genotype; model 3 includes Hb; and model 4 includes all significant covariates (age, genotype, and Hb).
Boldface indicates statistical significance.