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 4 Association of reported homologous exposure with reported gastrointestinal tract symptoms during the same or the following week including comparisons between sites of exposure

Reported exposure		Reported symptoms: Person-week in relation to exposure/Number and proportion of weeks with GTI symptoms/Statistical analysis
Site	Number	Same week		Following week		Either week
Site	Number	Number	Proportion/predictive margin	Number	Proportion/predictive margin	Number	Proportion/predictive margin
Only at work (W)	261	41	0.157/0.162 (CI^a 0.116 to 0.208) RR(W/0)^b = 11.4 (CI 7.5 to 17.5), P < 0.001	14	0.054/0.572 (CI 0.0290 to 0.0854) RR(W/0) = 2.6 (CI 1.6 to 4.4), P < 0.001	51	0.195/0.201 (CI 0.144 to 0.258) RR(W/0) = 6.0 (CI 4.2 to 8.6), P < 0.001
Only elsewhere (E)	437	109	0.249/0.246 (CI 0.193 to 0.298) RR(E/0) = 17.4 (CI 11.8 to 25.7), P < 0.001 RR(E/W) = 1.5 (CI 1.0 to 2.0)	48	0.110/0.0674 (CI 0.046 to 0.089) RR(E/0) = 3.1 (CI 2.2 to 4.3), P < 0.001 RR(E/W) = 1.18 (CI 0.52 to 1.84)	131	0.300/0.311 (CI 0.248 to 0.375) RR(E/0) = 9.0 (CI 6.8 to 11.9), P < 0.001 RR(E/W) = 1.5 (CI 1.0 to 2.0)
Both W and E	69	9	0.130/0.176 CI 0.053 to 0.299) RR(WE/0) = 12.4 (CI 6.3 to 24.5), P < 0.001	5	0.072/0.101 (CI 0.013 to 0.188) RR(WE/0) = 4.6 (CI 1.9 to 11.0), P < 0.001	12	0.174/0.186 (CI 0.049 to 0.322) RR(WE/0) = 5.5 (CI 1.5 to 9.7), P < 0.001

GEE logistic regression model using Stata was fitted for GTI 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).
^a95% confidence interval.
^bRR, risk ratio, where 0 = No exposure from Table 3.

ISSN: 1745-6215