Table 2

Incidence rate ratios (IRRs) obtained with a GLM based on the Poisson distribution for both datasets

PredictorUnivariable modelMultivariable model*
IRR (95% CI)P valueIRR (95% CI)P value
Meteorological variables
 Mean values within a day
  Ambient temperature (per 5°C)0.872 (0.870 to 0.873)<0.0010.874 (0.872 to 0.875)<0.001
  Relative humidity (per 10%)0.977 (0.975 to 0.980)<0.0010.985 (0.980 to 0.990)<0.001
  Precipitation during the previous hour (per 1 mm)0.913 (0.907 to 0.918)<0.0010.986 (0.972 to 1.000)0.053
  Snowfall (per 1 mm)1.088 (1.085 to 1.091)<0.0011.004 (0.998 to 1.010)0.184
  Cloud coverage (per 10%)0.992 (0.991 to 0.993)<0.0010.99 (0.989 to 0.991)<0.001
  Wind speed (per 10 kph)1.061 (1.057 to 1.065)<0.0011.000 (0.993 to 1.007)0.984
  Atmospheric pressure (per 1 hPa)0.994 (0.993 to 0.995)<0.001NA†
 Differences between maximum and minimum values within a day
  Ambient temperature (per 5°C)1.012 (1.009 to 1.016)<0.0010.982 (0.976 to 0.988)<0.001
  Relative humidity (per 10%)0.996 (0.994 to 0.998)<0.0010.996 (0.993 to 1.000)0.035
  Precipitation during the previous hour (per 1 mm)0.98 (0.979 to 0.981)<0.0011.006 (1.003 to 1.009)<0.001
  Snowfall (per 1 mm)1.335 (1.325 to 1.345)<0.0011.004 (1.003 to 1.006)<0.001
  Cloud coverage (per 10%)1.006 (1.005 to 1.006)<0.0011.003 (1.002 to 1.003)<0.001
  Wind speed (per 10 kph)1.053 (1.050 to 1.057)<0.0011.013 (1.008 to 1.018)<0.001
  Atmospheric pressure (per 1 hPa)1.175 (1.169 to 1.181)<0.001NA†
 Difference in mean value from the previous day
  Ambient temperature (per 5°C)1.003 (0.997 to 1.010)0.2801.075 (1.068 to 1.082)<0.001
  Relative humidity (per 10%)0.990 (0.987 to 0.992)<0.0011.006 (1.002 to 1.010)0.004
  Precipitation during the previous hour (per 1 mm)0.987 (0.983 to 0.991)<0.0010.984 (0.977 to 0.990)<0.001
  Snowfall (per 1 mm)0.994 (0.991 to 0.998)0.0020.993 (0.989 to 0.996)<0.001
  Cloud coverage (per 10%)0.997 (0.997 to 0.998)<0.0011.004 (1.003 to 1.005)<0.001
  Wind speed (per 10 kph)0.990 (0.985 to 0.994)<0.0011.007 (1.002 to 1.013)0.010
  Atmospheric pressure (per 1 hPa)1.017 (1.012 to 1.022)<0.001NA†
Chronological variables
 Year (per 3 years)‡,§ 1.059 (1.056 to 1.062)<0.0011.059 (1.056 to 1.062)<0.001
 Year squared§ 0.998 (0.998 to 0.998)<0.0010.998 (0.998 to 0.998)<0.001
 Japanese holiday1.154 (1.141 to 1.166)<0.0011.099 (1.087 to 1.111)<0.001
 Day of the week
  Sunday1 (Ref.)1 (Ref.)
  Monday0.98 (0.971 to 0.988)<0.0010.971 (0.963 to 0.980)<0.001
  Tuesday0.904 (0.896 to 0.913)<0.0010.905 (0.897 to 0.913)<0.001
  Wednesday0.879 (0.871 to 0.886)<0.0010.879 (0.871 to 0.887)<0.001
  Thursday0.88 (0.872 to 0.888)<0.0010.879 (0.872 to 0.887)<0.001
  Friday0.889 (0.881 to 0.897)<0.0010.888 (0.880 to 0.896)<0.001
  Saturday0.909 (0.901 to 0.917)<0.0010.909 (0.901 to 0.917)<0.001
 Season
  Winter1 (Ref.)1 (Ref.)
  Spring0.742 (0.737 to 0.747)<0.0010.875 (0.868 to 0.882)<0.001
  Summer0.56 (0.556 to 0.564)<0.0010.892 (0.880 to 0.904)<0.001
  Autumn0.67 (0.666 to 0.675)<0.0010.892 (0.883 to 0.901)<0.001
  • IRRs are exponentiated regression coefficients corresponding to multiplicative terms for the estimated number of cardiac arrests for each 1-unit, 5-unit or 10-unit increase in a meteorological variable.

  • *In the multivariable model, exponentiated regression coefficients were adjusted for prefecture and all of the variables listed in this table.

  • †Atmospheric pressure was excluded from the multiple Poisson regression analysis because of high multicollinearity.

  • ‡In the model, 2005 was considered year 0.

  • §In the univariable model, we included both year and year squared in the same model.

  • GLM, generalised linear model; IRR, incidence rate ratio; NA, not applicable.