Table IV.

Mixed-model regression analysis to evaluate COVID-19 mortality with government interventions and other factors globally and by income categorization.

	Global analysis
Items	β-coefficient	95% CI
Post intermediate (vs. pre intermediate) interventions (Model I)	2.55	2.30, 2.79
Post high (vs. pre high) interventions (Model II)	1.24	1.09, 1.39
Post very high (vs. pre very high) interventions (Model III)	0.65	0.49, 0.82
	High income
Post intermediate (vs. pre intermediate) interventions (Model I)	3.10	2.78, 3.43
Post high (vs. pre high) interventions (Model II)	0.88	0.65, 1.11
Post very high (vs. pre very high) interventions (Model III)	0.32	0.11, 0.53
	Upper middle income
Post intermediate (vs. pre intermediate) interventions (Model I)	1.83	1.51, 2.15
Post high (vs. pre high) interventions (Model II)	2.07	1.79, 2.35
Post very high (vs. pre very high) interventions (Model III)	1.16	0.74, 1.58
	Lower middle income
Post intermediate (vs. pre intermediate) interventions (Model I)	3.44	3.00, 3.88
Post high (vs. pre high) interventions (Model II)	1.37	0.99, 1.75
Post very high (vs. pre very high) interventions (Model III)	1.49	1.20, 1.78
	Low income
Post intermediate (vs. pre intermediate) interventions (Model I)	1.54	0.95, 2.13
Post high (vs. pre high) interventions (Model II)	0.87	0.35, 1.39
Post very high (vs. pre very high) interventions (Model III)	1.43	0.66, 2.20

All Stringency Index levels were compared to the low level in separate models. The Stringency Index was categorized in four quartiles, reflecting low to very high government intervention. All three models were adjusted for Index for Risk Management (INFORM) Epidemic Risk Index, proportion of population aged ⩾70 years, neoplasms disability-adjusted life years (DALYs), respiratory infections and tuberculosis DALYs and diabetes and kidney diseases DALYs.