Temperature and particulate matter as environmental factors associated with seasonality of influenza incidence – an approach using Earth observation-based modeling in a health insurance cohort study from Baden-Württemberg (Germany)

Table 4 Attributable cases

	Model	Restricted	Restricted	Full	Full
	Seasons	All	Recent	All	Recent
Temperature	Prediction 5-Percentile	3113 (2814–3443)	3423 (3046–3846)	3315 (2994–3671)	3615 (3214–4067)
	Prediction 95-Percentile	474 (435—516)	396 (360—437)	187 (136—256)	116 (81—166)
	Gain	-2639	-3027	-3128	-3499
	Relative Risk	0.152	0.116	0.0564	0.0320
Precipitation	Prediction 5-Percentile	553 (500–612)	469 (415–529)	601 (520–696)	389 (325–467)
	Prediction 95-Percentile	760 (689—838)	638 (570—714)	743 (638—866)	476 (398—569)
	Gain	207	169	142	87
	Relative Risk	1.37	1.36	1.23	1.22
PM_2.5	Prediction 5-Percentile	550 (497–608)	440 (390–497)	601 (516–699)	329 (271–398)
	Prediction 95-Percentile	984 (888—1091)	942 (829—1070)	897 (779—1034)	739 (617—886)
	Gain	435	502	297	411
	Relative Risk	1.79	2.14	1.49	2.25
NO₂	Prediction 5-Percentile	950 (843–1069)	735 (638–847)	892 (761–1045)	535 (439–652)
	Prediction 95-Percentile	656 (600—717)	585 (529—647)	722 (628—829)	464 (397—542)
	Gain	-293	-150	-170	-71
	Relative Risk	0.691	0.796	0.810	0.867

Enlistment of estimated cases per 100,000 persons per year for variation in environmental stressors. For each of the four models, the estimated number of cases is given for the 5- and 95 percentiles of the environmental stressors: temperature; precipitation; PM2.5 and NO2, with the 95% confidence interval in brackets. For each stressor, the gain in estimated cases was computed as the difference between estimations for the 5- and 95-percentiles. Similarly, the relative risk is the 95%-to-5% ratio of case predictions. Full models include quarter as independent variable, and restricted models do not. Models with all seasons include all seasons from 2010/11 and later, whilst models with recent seasons include only the seasons 2012/13 and later

ISSN: 1476-069X