v0.1.0.

• Processed sexual behaviour from 47 surveys (AIS, BAIS, DHS, PHIA)
• Multinomial logistic regression fit using Poisson trick via R-INLA for three risk categories
  • Fit to each country separately as well as all countries jointly
• Logistic regression used to differentiate highest risk category by estimating the proportion of those with non-regular sexual partners who report transactional sex
• Adjustment of transactional sex risk category to match external female sex worker estimates at a national level
• Calculation of incidence and new infections by risk category
• Analysis of the new infections that could be reached taking each possible risk stratification strategy