Tuesday March 31st, 2015:
Thanks to everyone who joined the call today! A really fascinating discussion - I definitely recommend listening to the podcast if you have a spare moment. Some ideas that were batted around:
- Is it possible to control TB and DR TB without fixing underlying weak health systems? Is $1 million better spent as adding 1% to a primary health care system, 10% to a TB control program, or 50% to a TB research budget?
- There may be a role for some sort of generalizable model in settings of highly mobile, hotspot-oriented populations, to evaluate in a generic sense what types of interventions are likely to be most effective.
- We took an informal poll of the question, "among people who have prevalent MDR TB in Somalia, how many were primarily infected with a Rif-resistant strain?" Despite having a fair amount of collective TB knowledge in the room, there was not consensus on this...though most people felt that the number was probably over 50%. If this is true, then simply treating drug-susceptible TB (as is the #1 priority from the WHO) is unlikely to rapidly fix an MDR epidemic.
- Coming up with the data that would be necessary to populate a better model of MDR transmission in a place like Somalia...is incredibly hard!! We had a hard time even thinking of where one would start with that. It's not just the tricky natural history of TB, but also the lack of infrastructure and routine data in such settings.
- But despite all this, it's clear that TB epidemics, and especially MDR epidemics, are concentrated in areas of poor infrastructure. It's fascinating to think of why this is - a couple of hypotheses that were advanced during our discussion were that people in places like Somalia are more likely to have access to drugs like INH/RIF only (since simply interrupting treatment is not as likely to generate resistance), or that MDR requires a certain threshold of malnutrition/crowding/etc. to transmit - and that such a threshold may be much easier to reach in a place like Somalia.
Thanks again to everyone for a really fascinating discussion!! Our next call will be at the same time (2pm Eastern) on April 21st - I look forward to it!!
Tuesday December 16th, 2014:
Thanks to those who joined us for a very nice discussion today, and to Andrew for leading the charge! A couple of quick summary points:
- While tools that can help TB officials convert "measurable" quantities (like cost per case detected/treated) into "meaningful" ones (like cost per DALY) might be useful, it is often hard for health economists to see how they can be used, without linking them to a specific intervention. Meaning often hard to get published as well.
- It's difficult to measure the population impact of TB case finding - which may not be fully realized for a large number of years afterward.
- If we could measure the impact of case-finding activities over 10 years, our best guess is that spending $9000 per additional case treated would be similar in cost-effectiveness to ART. But it's hard to get policymakers to make this comparison, as empirical evidence for the population-level impact of ACF is limited (hard to do such studies), and certainly doesn't have the "sex appeal" of ART bringing people back from a wasted state. Instead, the decision is implicitly made to be willing to pay over 10 times more per DALY on ART than on our best guess for TB case finding.
As usual, we'll work to get the podcast up soon on the website (link here), under past discussions. I will be in touch in January with plans for the new year - we have to discuss internally here at Hopkins how we want to integrate these into our own team meetings (and Tuesday at noon will fall on the time that I have instructors' meetings for the course I am teaching).
Thanks to each of you for making this a great year for our modeling calls!! I certainly have learned a lot from each of you. I hope you all have a wonderful holiday!!
Tuesday November 25th, 2014:
Thanks to everyone who participated in our call yesterday! While I'm not sure there were any groundbreaking conclusions, as one person said, sometimes it's nice just to listen. I was struck by the convergence of themes from so many people - the difficulty of staying up-to-date in diverse fields, the challenge of maintaining respect across quantitative experts and policy experts (so that one doesn't feel either dumb or out-of-touch), the importance of maintaining fora where people can come together and learn from each other, the friction between personal aspirations and a system that is set up with different values in mind. I certainly enjoyed hearing people's perspectives.
The podcast is up here, for those who are interested to listen.
I hope that everyone has a wonderful Thanksgiving, and I will be in touch next month! Our next call will be on December 16, again at noon Eastern time.
Tuesday October 14th, 2014:
Thanks for another very interesting call yesterday! And a special thanks to Rosa for joining our group and answering our questions - and doing so very well!!
The podcast of the meeting is avaliable here
A few brief summary points:
- In this study, co-prevalent TB (7% among adults with documented TB infection) was found at a proportion not dissimilar from historical rates, which was an unexpected finding (as one might expect this proportion to be lower in modern Amsterdam, given expected low levels of transmission there).
- By contrast, the number of true incident cases beyond 2 years (2 cases among over about 3,000 person-years of follow-up, so < 0.075% per person-year, among people with documented TB infection) was quite small, lower than what one would expect on the basis of historical data even in developed countries.
- This suggests that recent infection may be playing more of a role in propagating TB epidemics in low-burden settings than we may think.
- It also has implications for how we model the process of progression from TB transmission/infection to active TB, at least in low-burden settings - most of this progression may occur in the first two years.
- There are a lot of practical considerations to incorporate when looking at these data, including how to use surveillance data, how these contacts were identified, the role of IPT (about half of first-ring contacts received IPT), the meaning of "nationality" as a variable, and possible confounding by indication (closest contacts might have been most likely to be included, and most likely to receive IPT). Putting this kind of data together is not a straightforward task!!
- A couple of interesting ancillary findings are that it was much worse, in terms of future TB risk, to be a contact of a younger adult (15-44) than an older one (45+), and that the population of contacts themselves was a pretty young one, with children - even those between 5-14 years old - having a very high risk of progression (despite the generally assumed overall low risk of TB disease in children).
I look forward to our next call in November! Thanks to everyone for your participation - great to hear many of you on the line today!!