Table 2.

Distributions used as priors for model parameters. Test performance parameters are beta-distributed based on literature studies with with parameters (α, β) as given, equal to one plus the numbers of positive and negative results in the study, respectively. The rates of spontaneous recovery and treatment seeking were sampled using Markov Chain Monte Carlo methods (see main text). The proportion of the population who were sexually active is a beta distribution with parameters such that its central 95% credible interval matches the 95% confidence interval for the proportion estimated from Natsal-3. The proportion of incident infections which are asymptomatic was calibrated so that the sampled prevalence in 15-24-year-olds matched the Natsal estimates (see main text).

Parameter Description	Population		α	β	Mean (95% range: (2.5th, 97.5th) centiles)	Reference
Test performance: sensitivity; probability of a positive result, given infected	men		33	1	0.971 (0.894, 0.999)	6
	women		130	13	0.909 (0.857, 0.950)	7
Test performance: (1-specificity); probability of a positive result, given uninfected	men		3	951	0.00314 (0.00065, 0.00756)	6
	women		5	2324	0.00215 (0.00070, 0.00439)	7
Rate of spontaneous recovery (year-1)	men		Distribution sampled using Markov Chain Monte Carlo methods.		0.470 (0.059, 1.274)	8,10–17
	women				0.728 (0.591, 0.874)	8,12–15,18–22
Rate of treatment seeking (year-1)	men		Distribution sampled using Markov Chain Monte Carlo methods.		14.4 (8.5, 22.9)	23
	women				14.4 (8.5, 22.9)	23
Proportion of population sexually active	men	16-19 years	506	208	0.709 (0.675, 0.741)	25
		20-24 years	467	45.8	0.911 (0.885, 0.934)
		16-24 years	1090	234	0.823 (0.802, 0.843)
	women	16-19 years	532	238	0.691 (0.658, 0.723)	25
		20-24 years	569	53.8	0.914 (0.890, 0.934)
		16-24 years	1280	283	0.819 (0.799, 0.838)
Proportion of incident infections that are asymptomatic	men		Calibrated to reproduce Natsal-3 prevalence		0.511 (0.264, 0.759)	2
	women				0.615 (0.468, 0.752)