I would like to see some figures on health care costs with and without bednets. There are obvious health care costs to the famiies and government which bednets prevent. Any rough idea of what these costs are?

We believe that the figures Rob cites in the comment above are overly optimistic, as is the current figure on the website. We have raised this issue with Rob and expressed our hope that the website figure will be changed.

Responses to Rob’s comment above:

• Making all of the adjustments Rob proposes to our figure – including multiplying the cost per life saved by 70% in order to account for potential lives saved among people over 5 – would result in a figure of $738.50 (this is just 70%*$1055 where $1055 is Rob’s figure after making other adjustments). $738.50 is still well outside Rob’s range of $100-$500. and additionally multiplying the cost per life saved by 1/2 in order to reflect an assumption that nets last twice as long as our estimate indicates – would result in a figure of $527.50 (this is just (1/2)*$1055 where $1055 is Rob’s figure after making other adjustments). $527.50 is still outside Rob’s range of $100-$500. We have not seen analysis supporting the $100-$500 figure.
• That said, I don’t believe any of the adjustments Rob proposes are warranted.
  • Our “2.2 years” figure is based on our best guess using the best available data. We haven’t seen analysis supporting a higher figure.
  • It is true that our “cost per life saved” includes only under-5 lives saved. We are not aware of any rigorous studies examining the impact of nets on over-5 mortality, and we do not believe it is safe to assume that lives are saved in similar proportion to the burden of deaths. In addition, our impression is that the figures for over-5 deaths are disputed; according to the most recent data we know of, over-5 deaths account for closer to 10% of all malaria deaths (see the “AFR” tab of the death tables from the Global Burden of Disease 2004; divide the sum of cells S26 and H26 by cell G26). There are debates over these figures; we find it most appropriate to use the most recent official source.
  • It is unclear to me why one should use a figure of $5 per net rather than the figures we use, which we created in conjunction with AMF as our best estimates of the average and marginal total expenses per net distributed.

$100 (or 20 nets at $5 per net) per life is based on the widely cited Lengler Cochrane-based review indicating 160 bednet years equates to one under 5 life saved (see http://www.againstmalaria.com/onechild.aspx and http://www.againstmalaria.com/onechild_netstatistics.aspx).

We accept 20 nets per life may now be an overestimate, particularly given the move to universal coverage campaigns where under 5s are not being targeted but the aim is an entire community under nets.

Some time ago we started gathering cost-effectiveness data so we can update this figure and we hope to do so in the next two months.

The impact of nets in terms of lives saved does vary from location to location according to the malaria burden so we are likely to publish a range.

Our current information, part way through data collection, indicates one life per 20 to 100 nets in place, which at $5 per net, gives a range of $100-500 per life.

We have changed the language on the AMF website to this effect.

We believe GiveWell’s estimate of $1,700 per life is too high due to input estimates being overly conservative.

For example, GiveWell use an estimated net life of 2.2 years, taking wastage into account. We believe the value is higher, and above three years. We hope study data will give further clarity on this key input in the coming several months.

GiveWell’s number is more correctly ‘$1,700 per under 5 life saved’ and does not include the deaths of those over 5 years of age. Malaria does not just kill the under 5s. It seems reasonable to us to include the over 5s in the calculation given the entire community is covered in universal coverage distributions. The percentage of under 5s that make up the mortality figures is frequently cited as 70%. If this is taken as the under 5 percentage this would seem to adjust GiveWell’s number to ‘$1,190 per (any) life saved’.

Givewell also use a total cost per net figure of $5.54. As well as the cost of the net, this includes shipping, pre-distribution, distribution and post-distribution costs. AMF uses just the cost of a net, $5 in our estimate of the cost per life. On this basis, the $1,170 would further reduce to $1,055 if to be compared to our current estimated range of $100-500.

We accept there is more work to be done to update these figures and will contribute further in several months.

In our main model of the cost-effectiveness of deworming (XLS), based on our corrections to the DCP2 model, we estimate that treatment eliminates negative health impacts of STHs and schisto in about 15% of people treated.

In addition, the severity of infections can vary widely and some infections may be completely asymptomatic, so “number of infections averted” is not necessarily a meaningful figure.

Thank you!