Figure 1: Model Structure

Figure 2:Model calibration for the HIV MSM population in the United Kingdom. Plot (A)shows on the x-axis the case runs arranged in order from the case run that produces the lowest calibration error compared to the observed data to the largest. On the y-axis, we show the overall error distance. Plots (B), (D), and (F) are for the dynamics of the prevalence of the MSM HIV-, HIV+ aware and the HIV+ unaware. Plots (C) and (E), respectively show the dynamics of incidence of diagnosed HIV infection among MSM in either the E or J state and the A state. In all plots, the blue trajectory shows results from the best case run of the model, the red point shows the calibration target and the green ribbons or envelopes going from the darker inner ones to the outer lighter ones show the range of cases that belong to the top (i.e. best) 0.2%, 1% and 10% of the case runs. These ribbons do not show trajectories but rather help show that the top trajectories are bounded to lie within these regions. Notice the range shown by the green ribbons for the year 2002, illustrates the effect of having sampled the initial prevalence conditions.

Figure 3: Model calibration for the HIV MSM population in France Plot (A) shows on the x-axis the case runs arranged in order from the case run that produces the lowest calibration error compared to the observed realists to the largest. On the y-axis, we show the overall error distance. Plots (B), (D), and (F) for the dynamics of the prevalence of the MSM HIV-, HIV+ aware, and the HIV+ unaware. Plots (C) and (E) respectively show the dynamics of the incidence of diagnosed HIV infection among MSM in either the E or J state and the A state. In all plots, the blue trajectory shows results from the best case run of the model, the red point shows the calibration target and the green ribbons or envelopes going from the darker inner ones to the outer lighter ones show the range of cases that belong to the top (i.e. best) 0.2%, 1% and 10% of the case runs. These ribbons do not show trajectories but rather help show that the top trajectories are bounded to lie within these regions. Notice the range shown by the green ribbons for the year 2002, illustrates the effect of having sampled the initial prevalence conditions.

Model calibration fo the HIV MSM population in Poland Plot (A) shows on the x-axis the case runs arranged in order from the case run that produces the lowest calibration error compared to the observed realists to the largest. On the y-axis, we show the overall error distance. Plots (B), (D), and (F) for the dynamics of the prevalence of the MSM HIV-, HIV+ aware, and the HIV+ unaware. Plots (C) and (E) respectively show the dynamics of the incidence of diagnosed HIV infection among MSM in either the E or J state and the A state. In all plots, the blue trajectory shows results from the best case run of the model, the red point shows the calibration target and the green ribbons or envelopes going from the darker inner ones to the outer lighter ones show the range of cases that belong to the top (i.e. best) 0.2%, 1% and 10% of the case runs. These ribbons do not show trajectories but rather help show that the top trajectories are bounded to lie within these regions. Notice the range shown by the green ribbons for the year 2002, illustrates the effect of having sampled the initial prevalence conditions

Figure 5: Incidence of HIV-positive MSM individuals who are diagnosed prior to the AIDS stage following a change in testing rate in the year 2020 of (A-C) 50%, (D-F) no change, (G-I) 150% and (J-L) 200%. Results for the UK (right), France (center), and Poland (left) are respectively shown along the different columns. The lines show the median trajectory of 200 calibrated case runs. The bands around the medians show the inter-quartile ranges across the set of calibrated cases.

Figure 6:Prevalence of HIV+ unaware MSM (red line) and HIV-POSITIVE MSM on treatment (blue line) following a change in testing rate in the year 2020 of (A-C) 50%, (D-F) no change, (G-I) 150% and (J-L) 200%. Results for the UK (right), France (center), and Poland (left) are respectively shown along with the different columns. The lines show the median trajectory of 200 calibrated case runs. The bands around the medians show the inter-quartile ranges across the set of calibrated cases.

Figure 7: Changes to the prevalence of the HIV unaware MSM population (A-C), the number of HIV-POSITIVE MSM on treatment (D-F), an annual incidence of HIV-POSITIVE MSM that gets diagnosed (G-I) for the year 2030 following different increases in diagnosis rate. Results for the UK (right), France (centre), and Poland (left) are respectively shown along with the different columns. The lines show the median trajectory of 200 calibrated case runs. The band shows the inter-quartile range and the out-band spans the 5 to the 95th percentile.