a couple of billion dollars we invest in global health areas. But if you take the aggregate industry investment in R&D—which has stabilized now at about $130 billion a year—what we have is a drop in the ocean. It’s not as if we can just be very profligate in our choices.
X: To clarify, isn’t it about $1 billion from the Gates Foundation per year?
TM: In human healthcare R&D, it’s about $600 million a year.
X: So it’s $600 million from the Gates Foundation a year, out of a total pie of $130 billion in aggregate pharmaceutical industry R&D? A drop in the ocean…
TM: Right. We’ve got about $2 billion a year, when you include healthcare delivery, and agriculture, and all the sanitation, hygiene work. Across the broader portfolio, it’s close to $3 billion a year. But strictly in R&D, it’s about $600 million.
So you have to make choices.
For example, we’ve got a proposal in which people have shown there’s tremendous value if you give families a third bednet in Zambia for malaria. The program with bednets for malaria in Zambia has been very successful, one of our most successful countries. And there’s a proposal if you give families a third bednet, you could really make a lot more progress. And that’s versus developing a tuberculosis vaccine. How do you value those two things?
Then we have male surgical circumcision for HIV transmission versus a new drug for malaria. Then there’s a new diagnostic, second generation to the Cepheid nucleic acid diagnostic test for TB.
So you have these radically different modalities. Some things are not even therapies. It could be a new algorithm for triaging sick kids with fever, which would take into account their respiratory rate, their oxygen saturation, and would be a simple algorithm for a healthcare worker in the field to say, ‘This kid’s probably got a cold, and this kid has probably got serious pneumonia and needs to be shipped as fast as we can to some facility.’ Developing an algorithm like that is another kind of intervention. How do you value those radically different things?
X: There are a lot of people with a lot of ideas.
TM: Yes, there are a lot of ideas, and they are radically different. And they have radically different segments of the population that they can have impact on.
The first exercise that I thought was necessary was figuring out how to put them in some kind of format so that we have some kind of idea of their value. This is so we don’t make bad mistakes—putting a huge amount of resources into something that actually, although it seems important, has a really low return in terms of disability and death averted. Or it could be incredibly costly for every death averted.
Now we’re working at the public health level. We’re thinking about the 2012 number, of the 6.9 million kids [under age 5] that are dying. We developed something, which I took from industry approaches, which is a portfolio matrix. It’s a great graphic. All it does is replace net present value estimates of the cost-effectiveness or measure of return—we replace that with death/disability averted or disability-adjusted life years. It’s a summary measurement. And we want the cost of that, so we look at dollar cost per disability life year averted.
X: Is that really how they do it in industry, or are you taking a page out of the payer’s playbook?
TM: You need both approaches, actually. You want to address the major burdens of disease. We depict that on our bubble chart. The size of the bubble is the burden of disease, so you can see graphically, instantly, that HIV is huge, TB is huge, malaria is huge. It’s contrasted to sort of what you could do about that disease with existing measures. Let’s say we effectively deploy what already exists with TB drugs. Then you have a relative burden of disease that can be addressed by your new intervention.
X: What’s an example of a past program that might not have passed this test? Would the (GlaxoSmithKline) RTS,S vaccine for malaria have passed? I know that was a big investment, and we have some data now.
TM: How do we generate this portfolio matrix? We develop a target product profile. What would the criteria be, for example, for a vaccine we’d want for malaria? We’d want something very effective. Better than 60 percent effective in reduction of severe disease, and effective in young infants. A vaccine that meets those parameters is very cost-effective and addresses a very big burden of disease.
We are in the debate now whether RTS,S itself meets that kind of product profile and can have that kind of impact. It clearly is not in the optimal situation with its product profile. But
