Skip to main content
. 2019 May 30;6(1):e000425. doi: 10.1136/bmjresp-2019-000425

Table 3.

Regression model results for primary and secondary outcomes using individual-level data (level 2)

Outcome Model COPD admissions in analysis (n) Change postindex date in comparator sites, estimate (95% CI) Change postindex date in implementation sites, estimate (95% CI) Group–time interaction, estimate (95% CI) P value for interaction term
COPD readmission within 28 days Unadjusted 19 097 OR=0.97 (0.84 to 1.11) OR=0.94 (0.81 to 1.10) OR=0.97 (0.79 to 1.20) 0.804
Adjusted* 18 324 OR=0.93 (0.81 to 1.07) OR=0.95 (0.81 to 1.11) OR=1.02 (0.83 to 1.26) 0.865
All-cause readmission within 28 days Unadjusted 17 742 OR=1.02 (0.91 to 1.14) OR=0.86 (0.76 to 0.98) OR=0.85 (0.71 to 1.01) 0.060
Adjusted* 16 981 OR=1.00 (0.88 to 1.12) OR=0.88 (0.77 to 1.00) OR=0.88 (0.74 to 1.05) 0.156
COPD readmission within 90 days† Unadjusted 18 381 IRR=0.99 (0.95 to 1.03) IRR=0.98 (0.84 to 1.15) IRR=0.99 (0.84 to 1.16) 0.899
Adjusted* 17 634 IRR=0.98 (0.95 to 1.01) IRR=0.99 (0.86 to 1.13) IRR=1.00 (0.87 to 1.16) 0.971
Length of stay for COPD admission Unadjusted 19 343 IRR=0.99 (0.94 to 1.04) IRR=0.93 (0.88 to 0.97) IRR=0.94 (0.88 to 1.01) 0.071
Adjusted* 18 565 IRR=0.98 (0.94 to 1.03) IRR=0.93 (0.88 to 0.97) IRR=0.94 (0.88 to 1.01) 0.100
Number of emergency department attendances Unadjusted 15 646 IRR=1.14 (1.04 to 1.26) IRR=0.63 (0.56 to 0.70) IRR=0.55 (0.47 to 0.63) <0.001
Adjusted* 14 900 IRR=1.12 (1.02, 1.24) IRR=0.63 (0.56 to 0.71) IRR=0.56 (0.48 to 0.65) <0.001
In-hospital mortality for COPD admissions Unadjusted 19 343 OR=0.79 (0.64 to 0.98) OR=0.79 (0.63 to 0.99) OR=1.00 (0.74 to 1.36) 0.992
Adjusted* 18 565 OR=0.78 (0.63 to 0.97) OR=0.82 (0.65 to 1.03) OR=1.05 (0.77 to 1.44) 0.741
90-day mortality† Unadjusted 17 664 IRR=0.93 (0.78 to 1.10) IRR=0.89 (0.73 to 1.08) IRR=0.96 (0.74 to 1.24) 0.734
Adjusted 16 944 IRR=0.93 (0.80 to 1.08) IRR=0.93 (0.79 to 1.11) IRR=1.01 (0.80 to 1.26) 0.946

*Adjusted for age, sex, ethnicity and IMD quintile.

†CI calculated using robust SEs; the Wald test was used to generate the p value for interaction.

COPD, chronic obstructive pulmonary disease; IMD, Index of Multiple Deprivation; IRR, incidence rate ratio.