Prevalence issue.

[jackolney]

It appears as though my dynamic incidence weighting algorithm is really pumping up incidence post 2002. Maybe I need it to cap incidence not weight it >1?

[jackolney]

Consider:

    if(theTrans->GetBeta() && theTrans->GetIncidenceWeight() < 1 )
        serum = theRng->Sample( hivInc [Sex] [i] [j] * theTrans->GetIncidenceWeight() );
    else
        serum = theRng->Sample( hivInc [Sex] [i] [j] );

[jackolney]

Implemented and it appears to fix prevalence. Needs to be validated against UNAIDS estimates of prevalence though.

[jackolney]

UNAIDS estimates place the model output well below country-level estimates. Checks:

1. Output HIV incidence and check against UNAIDS
2. Check out natural mortality... I have a feeling that is limiting everything.

[jackolney]

Still not fixed, checked:

• [x] Natural Death Dates.
• [x] Initial Ages.

Still need to look at incidence more carefully. Plan:

• [x] Remove HIV-related death and look at HIV-prevalence (are people dying too soon?)
• [x] Else, compare incidence in the model to UNAIDS.

[jackolney]

• [x] HIV-related deaths are fine. Removing the age distribution at model start (Kenya age distribution in 1970) then we see a significant improvement in prevalence.
• [x] Initial Ages are fine - need to look into whether I was doing something different with natural mortality early on.

[jackolney]

UNAIDS estimates are for age between 15-49... I need to match that. Go go go. :sweat:

[jackolney]

Fixed! it was the age issue, duh.