Table 4 Inputs and summary outputs for all the model runs and fits for the five model fitting scenarios
ScenarioModel inputs for each model fitting scenarioProjections from all model runsProjections from all the model fits†Number of validated model fits‡
Proportion in risk groupNumber IDUs share with per monthNumber of syringe sharing incidents per IDU per monthOverall HCV prevalence:mean (range)Number of model fits to data†Overall HIV prevalence:mean (range)§HCV prevalence: mean (range)
Low¶High¶Low¶High¶Low¶High¶Overall⩽2 years injecting>2 years injectingIn HIV infected subjects
Data summaryNANANANANANANANA2.7% (1–5%)17.3% (13–22%)14.5% (8–24%)18.2% (13–24%)100% (63–100%)NA
Scenario 122–82%13–26%6–912–201–40–866.7% (32–95%)0NANANANANA0
Scenario 222–82%13–26%1–2*2–4*4–160–4666.7% (32–95%)0NANANANANA0
Scenario 322–82%13–26%0.02–0.72–61119.5% (7–50%)1111.5% (1–5%)13.7% (13–17%)9.1% (8–10%)16.4% (15–20%)28.0% (22–35%)0
Scenario 422–82%13–26%0.03–1.42–111124.6% (8–64%)1001.7% (1–5%)13.9% (13–19%)9.2% (8–11%)16.7% (15–23%)26.3% (18–34%)0
Scenario 529–89%6–20%0.25–4.05.6–10.01155.6% (8–92%)7431.3% (1–5%)17.9% (13–21%)12.0% (9–16%)21.3% (15–24%)51.6% (22–77%)125
  • HCV, hepatitis C virus; IDU, intravenous drug user; NA, not applicable.

  • Ranges in brackets are 95% confidence intervals for data and 95% credibility intervals for model estimates which are the 2.5–97.5% percentiles of the model projections.

  • *The low-risk sharers also shared once with 0.02–0.74 IDUs per month with whom they had not shared before (“strangers”) and the high-risk sharers shared once with 1.9–6.0 “strangers” per month.

  • †A model fit was defined as any model simulation that agreed with the observed overall prevalence of HIV and HCV among all IDUs, and the prevalence of HCV among IDUs who have been injecting for ⩽2 years and >2 years, and for scenarios 1–4 the HCV prevalence among IDUs who have obtained most syringes from a pharmacy, shop, hospital, health worker or elsewhere.

  • ‡A validated model fit is any model fit that also agreed with the observed HCV prevalence among HIV-infected individuals.

  • §The model overall HIV prevalence is the projected HIV prevalence when the model simulation first agrees with all the fitting data, and so it is expected to be towards the lower bound of the survey HIV prevalence confidence bounds.

  • ¶In all model fitting scenarios except scenario 5, low-risk IDUs share syringes and acquire their syringes from a reliable source (pharmacy, shop, hospital, health worker) whereas high-risk IDUs acquire their syringes from an unreliable source (found, IDU, drug dealer, etc). In scenario 5, low-risk IDUs share syringes but have not shared with a “stranger” in the last week, whereas high-risk IDUs have shared with a “stranger” in the last week.