Loading ...
Sorry, an error occurred while loading the content.
 

Malaria research

Expand Messages
  • Alexander Berger
    Hi all, We ve been doing some more research on insecticide-treated bednets (ITNs), in particular trying to better understand the case for universal
    Message 1 of 2 , Apr 23, 2012
      Hi all,

      We've been doing some more research on insecticide-treated bednets (ITNs), in particular trying to better understand the case for universal distribution (i.e., aiming to give nets to everyone who lives in a particular area regardless of age) vs. distribution targeted at children under five (who die from malaria most frequently) and how ITN effectiveness relates to malaria prevalence.

      To do this, we looked at the openmalaria malaria epidemiology simulator produced by the Swiss Tropical and Public Health Institute. We're not very confident in the results and haven't fully explored their implications, but they generally seem to comport with the theory that universal distribution is as good or better at preventing mortality than targeted distribution. It also appears that, contrary to intuition, the bednets are not necessarily most effective at preventing mortality in areas with the highest transmission rates.

      These tentative results are consistent with our recommendation of AMF and don't affect our bottom line at all.

      Background

      Most of AMF’s distribution partners, including the one we visited in Malawi, focus on universal distribution. We wondered about this because if the main benefit of ITNs is in reducing mortality for children under five, it seemed that targeting children under five could achieve most of the benefits of universal coverage, for a fraction of the cost. In the area served by AMF’s Malawi distribution partner, ITN coverage amongst children was already fairly high, so it seemed that reaching universal coverage might not be as helpful as somewhere else where the proportion of children already covered was lower. Our preliminary investigation of this question during the fall consisted of:

      • Speaking with Christian Lengeler, the author of the Cochrane review of the evidence regarding ITN distribution discussed in our ITN distribution intervention report. (This Cochrane review appears to us to be the mostwidely cited evidence of the efficacy of ITNs.) We have published our notesfrom this conversation.
      • Speaking with two scholars that Professor Lengeler referred us to on this question, Dave Smith and Thomas Smith. We have published notes from each of these conversations.
      • Reviewing literature that these scholars referred us to.

      In our main model of the cost-effectiveness of distributing ITNs (XLS), we assumed that half of the benefits of ITN distribution for saving children come from directly covering individual kids, while the other half comes from interrupting malaria transmission in the community. Roughly speaking, the greater proportion of the benefits of deworming that come from community effects, the more effective universal distribution will be relative to distribution targeted at children.

      What we did

      To continue our investigation of this question, we asked Dario Amodei to run a simulation comparing universal and targeted ITN distribution campaigns. Using software produced by the Swiss Tropical and Public Health Institute, Dario was able to compare the expected outcomes of a universal and a targeted ITN distribution campaign. The software runs probabilistic simulations, which means it might not get the same answer every time except with very large sample sizes. The models appear to sensitive to a number of assumptions in ways that we haven't fully worked out.

      Once we had the results of Dario's analysis, we asked Rob Mather at the Against Malaria Foundation about them. He consulted with scholars from AMF's Malaria Advisory Group, who went though some back and forth, and then sent us the record of their exchange.

      Thanks to Dario and Alex Dingle, who helped him set up the modeling software. 

      What we found

      Universal bednet distribution was generally as good or better than targeted distribution with respect to mortality in the modeling scenarios that Dario ran for us, though we're not sure to what extent this conclusion is sensitive to assumptions. We don't fully understand what drives the outcomes in these scenarios and have done any assessment of whether the modeling is reasonable.

      One of Dario's other tentative findings was that higher transmission of malaria does not necessarily translate into greater effectiveness for ITNs. We contacted Rob Mather to ask about this, and he sent the question to the Malaria Advisory Group. Thomas Smith, a scholar originally referred to us by Christian Lengeler who also serves on AMF’s Malaria Advisory Group, wrote in his reply:

      We would expect a higher coverage of ITNs to have a substantial effect on transmission measures whatever the starting level of transmission (>0), but the effect on measures of morbidity and mortality would also depend on the age groups considered and the time horizon. While it is clearly the case that there is no benefit in using ITNs if there is no malaria at all, the effects on morbidity and mortality are greatest at intermediate levels of transmission and the relationship between the public health benefits of given levels of scaling up and the initial EIR is complicated. It is certainly not the case that ITNs have least utility in low endemic areas, since it is precisely there that there is a chance they will interrupt transmission.

      We've posted additional files related to this analysis (e.g. charts from Dario's analysis, source code, and emails from the AMF Malaria Advisory Group) here (ZIP file; requires this software to run the malaria models).

    • Alexander Berger
      And, as Rob astutely pointed out, it is* LLIN distribution*, not *deworming*, that should appear in this sentence: Roughly speaking, the greater proportion of
      Message 2 of 2 , Apr 23, 2012
        And, as Rob astutely pointed out, it is LLIN distribution, not deworming, that should appear in this sentence: "Roughly speaking, the greater proportion of the benefits of deworming that come from community effects, the more effective universal distribution will be relative to distribution targeted at children."

        Sorry for the mistake and second email!

        On Mon, Apr 23, 2012 at 12:50 PM, Alexander Berger <alexander.is@...> wrote:
        Hi all,

        We've been doing some more research on insecticide-treated bednets (ITNs), in particular trying to better understand the case for universal distribution (i.e., aiming to give nets to everyone who lives in a particular area regardless of age) vs. distribution targeted at children under five (who die from malaria most frequently) and how ITN effectiveness relates to malaria prevalence.

        To do this, we looked at the openmalaria malaria epidemiology simulator produced by the Swiss Tropical and Public Health Institute. We're not very confident in the results and haven't fully explored their implications, but they generally seem to comport with the theory that universal distribution is as good or better at preventing mortality than targeted distribution. It also appears that, contrary to intuition, the bednets are not necessarily most effective at preventing mortality in areas with the highest transmission rates.

        These tentative results are consistent with our recommendation of AMF and don't affect our bottom line at all.

        Background

        Most of AMF’s distribution partners, including the one we visited in Malawi, focus on universal distribution. We wondered about this because if the main benefit of ITNs is in reducing mortality for children under five, it seemed that targeting children under five could achieve most of the benefits of universal coverage, for a fraction of the cost. In the area served by AMF’s Malawi distribution partner, ITN coverage amongst children was already fairly high, so it seemed that reaching universal coverage might not be as helpful as somewhere else where the proportion of children already covered was lower. Our preliminary investigation of this question during the fall consisted of:

        • Speaking with Christian Lengeler, the author of the Cochrane review of the evidence regarding ITN distribution discussed in our ITN distribution intervention report. (This Cochrane review appears to us to be the mostwidely cited evidence of the efficacy of ITNs.) We have published our notesfrom this conversation.
        • Speaking with two scholars that Professor Lengeler referred us to on this question, Dave Smith and Thomas Smith. We have published notes from each of these conversations.
        • Reviewing literature that these scholars referred us to.

        In our main model of the cost-effectiveness of distributing ITNs (XLS), we assumed that half of the benefits of ITN distribution for saving children come from directly covering individual kids, while the other half comes from interrupting malaria transmission in the community. Roughly speaking, the greater proportion of the benefits of LLIN distribution that come from community effects, the more effective universal distribution will be relative to distribution targeted at children.

        What we did

        To continue our investigation of this question, we asked Dario Amodei to run a simulation comparing universal and targeted ITN distribution campaigns. Using software produced by the Swiss Tropical and Public Health Institute, Dario was able to compare the expected outcomes of a universal and a targeted ITN distribution campaign. The software runs probabilistic simulations, which means it might not get the same answer every time except with very large sample sizes. The models appear to sensitive to a number of assumptions in ways that we haven't fully worked out.

        Once we had the results of Dario's analysis, we asked Rob Mather at the Against Malaria Foundation about them. He consulted with scholars from AMF's Malaria Advisory Group, who went though some back and forth, and then sent us the record of their exchange.

        Thanks to Dario and Alex Dingle, who helped him set up the modeling software. 

        What we found

        Universal bednet distribution was generally as good or better than targeted distribution with respect to mortality in the modeling scenarios that Dario ran for us, though we're not sure to what extent this conclusion is sensitive to assumptions. We don't fully understand what drives the outcomes in these scenarios and have done any assessment of whether the modeling is reasonable.

        One of Dario's other tentative findings was that higher transmission of malaria does not necessarily translate into greater effectiveness for ITNs. We contacted Rob Mather to ask about this, and he sent the question to the Malaria Advisory Group. Thomas Smith, a scholar originally referred to us by Christian Lengeler who also serves on AMF’s Malaria Advisory Group, wrote in his reply:

        We would expect a higher coverage of ITNs to have a substantial effect on transmission measures whatever the starting level of transmission (>0), but the effect on measures of morbidity and mortality would also depend on the age groups considered and the time horizon. While it is clearly the case that there is no benefit in using ITNs if there is no malaria at all, the effects on morbidity and mortality are greatest at intermediate levels of transmission and the relationship between the public health benefits of given levels of scaling up and the initial EIR is complicated. It is certainly not the case that ITNs have least utility in low endemic areas, since it is precisely there that there is a chance they will interrupt transmission.

        We've posted additional files related to this analysis (e.g. charts from Dario's analysis, source code, and emails from the AMF Malaria Advisory Group) here (ZIP file; requires this software to run the malaria models).


      Your message has been successfully submitted and would be delivered to recipients shortly.