Exposure to persons with symptoms of respiratory or gastrointestinal infection and relative risk of disease: self-reported observations by controls in a randomized intervention trial

Table 2 Association of reported homologous exposure with reported respiratory tract symptoms during the same or the following week including comparisons between sites of exposure

Reported exposures		Reported symptoms: Person-week in relation to exposure/Number and proportion of weeks with symptoms/Statistical analysis
Site	Number of weeks	Same week		Following week		Either week
Site	Number of weeks	Number	Proportion/predictive margin^a	Number	Proportion/predictive margin^a	Number	Proportion/predictive margin^a
Only at work (W)	1,068	262	0.245/0.236 (CI^b 0.198 to 0.274) RR(W/0)^c = 4.0 (CI 3.2 to 4.9), P < 0.001	221	0.207/0.203 (CI 0.169 to 0.237) RR(W/0) = 2.7 (CI 2.2 to 3.2), P < 0.001	359	0.336/0.334 (CI 0.287 to 0.382) RR(W/0) = 2.90 (CI 2.5 to 3.4), P < 0.001
Only elsewhere (E)	1,448	434	0.300/0.291 (CI 0.251 to 0.331) RR(E/0) = 4.9 (CI 4.0 to 6.0), P < 0.001 RR(E/W) = 1.231 (CI 1.1 to 1.4)	366	0.253/0.221 (CI 0.191 to 0.251) RR(E/0) = 2.9 (CI 2.5 to 3.4), P < 0.001 RR(E/W) = 1.1 (CI 0.90 to 1.28)	584	0.403/0.404 (CI 0.358 to 0.449) RR(E/0) = 3.5 (CI 3.2 to 4.1), P < 0.001 RR(E/W) = 1.2 (CI 1.0 to 1.4)
Both (WE)	763	280	0.367/0.369 (CI 0.302 to 0.435) RR(WE/0) = 6.2 (CI 4.9 to 7.8), P < 0.001	218	0.286/0.296 (CI 0.231 to 0.360) RR(WE/0) = 3.9 (CI 3.2 to 5.0), P < 0.001	339	0.444/0.453 (CI 0.384 to 0.521) RR(WE/0) = 3.9 (CI 3.3 to 4.7), P < 0.001

^aGEE logistic regression model using Stata was fitted for RTI outcome with seasonality, triplet and homologous reported exposure (same or previous week) as covariates. The working correlation matrix was ‘independent’ and chosen by the quasi-likelihood model criterion (QIC).
^b95% confidence interval.
^cRR, risk ratio, where 0 = No exposure from Table 1.

ISSN: 1745-6215