Table 3 Output of the logistic regression models developed by choosing different cut-offs to discriminate high- versus low-ABHR consumption groups within the central tertile of the distribution of predicted ABHR data (cut-offs outside this range are reported in Supplementary Table S2). For each cutoff i of ABHR consumption (dependent variable), hospitals were divided in 2 groups (high vs. low) and the logistic regression was built as described in the Methods. The table shows the odds ratio (OR₁₀) of belonging to the high-consumption group for every 10-point increase in the HHSAF score (independent variable). Values reported in bold indicate cut-off values for which statistical significance is approached (i = 58–59)

1.090 [0.956–1.244]

0.198

49

1.074 [0.944–1.222]

0.280

1.097 [0.963–1.249]

0.165

50

1.074 [0.944–1.222]

0.280

1.097 [0.963–1.249]

0.165

51

1.074 [0.944–1.222]

0.280

1.084 [0.953–1.232]

0.220

52

1.074 [0.944–1.222]

0.280

1.082 [0.952–1.230]

0.226

53

1.074 [0.944–1.222]

0.280

1.082 [0.952–1.230]

0.226

54

1.074 [0.944–1.222]

0.280

1.083 [0.953–1.230]

0.224

55

1.074 [0.944–1.222]

0.280

1.083 [0.953–1.230]

0.224

56

1.074 [0.944–1.222]

0.280

1.068 [0.940–1.212]

0.313

57

1.074 [0.944–1.222]

0.280

1.068 [0.940–1.212]

0.313

58

1.124 [0.981–1.288]

0.092

1.068 [0.940–1.212]

0.313

59

1.124 [0.981–1.288]

0.092

1.068 [0.940–1.212]

0.313

60

1.092 [0.955–1.249]

0.198

1.068 [0.940–1.212]

0.313

61

1.028 [0.902–1.172]

0.675

1.068 [0.940–1.212]

0.313

62

1.059 [0.925–1.213]

0.408

1.102 [0.967–1.256]

0.143

63

1.059 [0.925–1.213]

0.408

1.093 [0.959–1.245]

0.182

64

1.059 [0.925–1.213]

0.408

1.093 [0.959–1.245]

0.182

65

1.059 [0.925–1.213]

0.408

1.093 [0.959–1.245]

0.182

66

1.060 [0.924–1.216]

0.408

1.074 [0.944–1.222]

0.280

67

1.060 [0.924–1.216]

0.408