Table 2 Risk of dying on days with a mean apparent temperature of 30°C (lag 0–1) versus days with a temperature of 20°C among people age 65+ years dying in hospital and already in hospital two days before death: effect modification by age, gender, socioeconomic characteristics, hospital ward and type of hospital
No (%)OR (95% CI)* (unadjusted)OR (95% CI)* (age-adjusted)REM index† (p value)
Total (65+ years)94944 (100)1.32 (1.25 to 1.39)1.32 (1.25 to 1.39)(–)
Age (years)
    65–7426763 (28)1.23 (1.12 to 1.35)1.23 (1.12 to 1.35)1.00 (–)
    75–8438938 (41)1.26 (1.13 to 1.39)1.26 (1.13 to 1.39)1.02 (0.784)
    85+29243 (31)1.50 (1.35 to 1.66)1.50 (1.35 to 1.66)1.22 (0.005)
Gender
    Men47476 (50)1.26 (1.14 to 1.39)1.30 (1.21 to 1.41)1.00 (–)
    Women47468 (50)1.38 (1.28 to 1.48)1.38 (1.28 to 1.48)1.06 (0.307)
Marital status‡
    Married17083 (47)1.16 (1.01 to 1.34)1.16 (1.02 to 1.32)1.00 (–)
    Not married/widowed/divorced18974 (53)1.42 (1.27 to 1.57)1.38 (1.24 to 1.54)1.19 (0.040)
Income (area level)
    20th percentile20756 (22)1.33 (1.09 to 1.63)1.35 (1.10 to 1.66)1.00 (–)
    20th–50th percentile29915 (32)1.35 (1.20 to 1.52)1.36 (1.23 to 1.51)1.01 (0.926)
    50th–80th percentile27533 (29)1.37 (1.24 to 1.51)1.37 (1.22 to 1.53)1.01 (0.920)
    80th–100th percentile16205 (17)1.22 (1.08 to 1.38)1.24 (1.10 to 1.40)0.92 (0.475)
Type of hospital
    Public53512 (72)1.29 (1.21 to 1.38)1.33 (1.24 to 1.43)1.00 (–)
    Semi-public11742 (16)1.29 (1.11 to 1.49)1.31 (1.13 to 1.51)0.98 (0.802)
    Private§9570 (13)1.35 (1.14 to 1.60)1.35 (1.13 to 1.60)1.01 (0.917)
Ward of the hospital
    General medicine53481 (56)1.40 (1.31 to 1.51)1.40 (1.30 to 1.51)1.00 (–)
    Low-care load24299 (26)1.30 (1.17 to 1.45)1.35 (1.21 to 1.50)0.96 (0.531)
    High-care load5791 (6)1.10 (0.89 to 1.35)1.13 (0.90 to 1.41)0.80 (0.067)
    Intensive care10573 (11)1.08 (0.92 to 1.26)1.24 (1.03 to 1.50)0.88 (0.236)

  • *Odds ratio (OR) and 95% confidence intervals (CI). Results in italics are from random effects models.

  • †REM, relative effect modification index is calculated as the ratio between the specific OR and the OR from the reference category (from the age-adjusted model).

  • ‡Milan and Turin only. §Bologna, Milan and Rome only.