Table 3.

Assessment of whether number and level of contacts mediate the effect of quarantine on secondary cases.

Term in adjusted model* (contrasts with crude model B, in Table 2)	Secondary Case Count Ratio (SCCR)	p-value under large sample assumption and using bootstrapped variance estimate (in italics)	95% confidence interval under large sample assumption and using bootstrapped variance estimate (in italics)
			Lower limit	Upper limit
Quarantine (v. no)	0.3524498	0.046 0.550	0.1266103 0.0048919	0.9811279 20.14694
Total contacts (continuous†)	0.9999988	0.657 0.090	0.9999937 0.9999816	1.000004 1.000017
Total close contacts (continuous)	1.061969	0.005 0.009	1.018544 1.015351	1.107246 1.107246

* Poisson regression for secondary case count, adjusting for total and close contacts.

†Continuous term transformed by taking the square root of raw data; used to achieve normally-distributed residuals.