A proposed case-control framework to probabilistically classify individual deaths as expected or excess during extreme hot weather events

Table 1 Summary of variants for each combination and standard deviations (SD) for case ranks between trials

Combination	Significance tally	Coefficient in expected direction	Alpha	Mean of rank SD	SD of rank SD	Median of rank SD	Interquartile range of rank SD
Univariate Combinations
#1	Summed across variables	T	0.10	36.94	12.09	40.37	17.50
#2	Summed across variables	F	0.10	36.61	11.67	40.10	15.81
#3	Summed across variables	T	0.05	41.12	12.68	44.44	17.72
#4	Summed across variables	F	0.05	41.05	12.71	44.40	18.18
Multivariate Combinations
#1	Count of significant variables	T	0.10	50.56	12.77	54.23	17.52
#2	Count of significant variables	F	0.10	50.56	12.92	54.43	17.76
#3	Count of significant variables	T	0.05	60.07	14.76	64.25	18.56
#4	Count of significant variables	F	0.05	60.31	14.81	63.89	22.29
#5	At least one significant variable	T	0.10	89.74	13.62	93.17	15.21
#6	At least one significant variable	F	0.10	89.83	13.44	94.09	14.82
#7	At least one significant variable	T	0.05	66.83	18.16	74.09	27.79
#8	At least one significant variable	F	0.05	67.07	17.95	74.75	23.72

Bold highlighting indicates the univariate and multivariate combinations selected as the most promising combinations based on minimized variability in the ranking results

ISSN: 1476-069X