search

optimamodel / optima

Optima HIV software tool
http://optimamodel.com
GNU Lesser General Public License v3.0
7 stars 1 forks source link

Low viral suppression with huge budgets? #1605

Closed Rowanmh closed 7 years ago

Rowanmh commented 7 years ago

So running GA and generating BOCs noticed that outcomes were getting worse with the larger budgets (10x and quite unnecessarily 100x in particular), rather than a flat line (most countries).

Investigation of how outcomes could possibly be getting worse despite huge funding seem to be explained in the cascades...

Default funding is the current budget Large is 10x current budget Huge is 100x current budget

It seems like while more people are diagnosed and a higher percentage of people are diagnosed, less people are virally suppressed.

This may be linked specifically to testing rates as running the higher budget BOCs does generate warnings that HIV testing is making the situation worse.

Rowanmh commented 7 years ago

Something wrong in the settings of the projects?

Guess I could work around by setting a constraint on hiv testing programs to not overfund beyond saturation?

(it's still the case with lower increases in budgets, just not noticeable because the other gains from increased budget up to a point are so significant)

Rowanmh commented 7 years ago

This could also relate to Issue #1596

Rowanmh commented 7 years ago

@robynstuart discussed with Cliff already, seems like with no programs effecting numvlmon that I need to be setting a fixed propsupp (you might have said this before but I didn't realise it applied to all projects not just with fixed proptx), but I'm not sure what value to use, 0.9 or even higher Cliff suggested 0.95 to 0.98 even. Waiting on your advice before rerunning.

That will fix the issue for the GM, but it still seems worth investigating in general, as having more spare money and the same number of people treated because everyone is diagnosed and treated shouldn't result in decreasing viral monitoring/suppression.