Table 2

Patient characteristics and outcomes for hospital spells included in the study

Control site, post-intervention (N=2553)Control site, pre-intervention (N=1262)Intervention site, post-intervention (N=5753)Intervention site, pre-intervention (N=6050)Difference-in-difference OR/mean (95% CI)
Age
 Missing0 (0.0%)0 (0.0%)0 (0.0%)0 (0.0%)
 Median (IQR): band midpoints82.5 (72.5–87.5)82.5 (72.5–87.5)82.5 (72.5–87.5)82.5 (72.5–87.5)0.91 (0.80 to 1.04)
Gender
 Missing0 (0.0%)0 (0.0%)0 (0.0%)0 (0.0%)
 Female1293 (50.7%)642 (50.9%)2816 (49.0%)3084 (51.0%)1.07 (0.92 to 1.25)
 Male1260 (49.4%)620 (49.1%)2937 (51.1%)2966 (49.0%)
Ethnicity
 Missing39 (1.5%)10 (0.8%)9 (0.2%)17 (0.3%)
 Black or black British193/2514 (7.7%)106/1252 (8.5%)526/5744 (9.2%)480/6033 (8.0%)1.30 (0.98 to 1.73)
 Mixed/other ethnic groups/not stated227/2514 (9.0%)112/1252 (9.0%)524/5744 (9.1%)486/6033 (8.1%)1.47 (1.10 to 1.97)
 White1136/2514 (45.2%)544/1252 (43.5%)2565/5744 (44.7%)3007/6033 (49.8%)
 Asian or Asian British958/2514 (38.1%)490/1252 (39.1%)2129/5744 (37.1%)2060/6033 (34.2%)1.29 (1.09 to 1.53)
Index of Multiple Deprivation
 Missing138 (5.4%)49 (3.9%)775 (13.5%)508 (8.4%)
 n; median (IQR)2415; 25.6 (17.1–33.2)1213; 25.85 (17.2–32.4)4978; 17.7 (12.3–24.9)5542; 16.7 (11.7–23.7)
 n; mean (SD)2415; 25.6 (10.7)1213; 25.8 (10.1)4978; 19.3 (10.1)5542; 18.4 (9.3)1.20 (1.05 to 1.36)
Any-cause readmission category
 No readmission <90 days1588 (62.2%)774 (61.3%)3908 (67.9%)3976 (65.7%)
 <7 days211/965 (21.9%)96/488 (19.7%)359/1845 (19.5%)410/2074 (19.8%)
 ≥7–<30 days409/965 (42.4%)198/488 (40.6%)698/1845 (37.8%)811/2074 (39.1%)
 ≥30–<90 days345/965 (35.8%)194/488 (40.0%)788/1845 (42.7%)853/2074 (41.1%)
Mortality
 Missing0 (0.0%)0 (0.0%)0 (0.0%)0 (0.0%)
 In-hospital mortality243 (9.5%)139 (11.0%)550 (9.6%)893 (14.8%)
Length of stay
 Missing0 (0.0%)0 (0.0%)0 (0.0%)0 (0.0%)
 Median (IQR) days7.5 (3.2–15.2)8.8 (4.0–17.6)6.7 (2.8–12.9)7.1 (3.0–14.4)
  • Demographic characteristics and unadjusted outcomes for patients discharged following an admission with heart failure. ORs are for difference-in-difference in each characteristic between sites across pre-intervention and post-intervention periods, calculated using logistic regression for gender, ordered logistic regression for age band and multinomial regression for ethnicity.