Ezra Klein on ARPA-H

As usual it is best to just click and read as the column is well crafted and I can’t summarize or explain well.

The topic?

“On Monday, President Biden announced that Dr. Renee Wegrzyn, a biotech executive who previously worked at the Defense Advanced Research Projects Agency, better known as Darpa, would be the first director of the Advanced Research Projects Agency-Health, ARPA-H.”

Klein is very enthusiastic. I know something about the topic and have some thoughts.

To try to summarize. ARPA-H is an attempt to build on two (or three) great successes. FIrst DARPA is one of the key engines of economic growth responsible for, among other things, the internet. Second operation Warp Speed actually contributed to the extremely rapid development of effective SARS COV-2 vaccines. Third, the NIH’s huge budget is money well spent.

I mostly agree with Klein. I do have some doubts. I guess I will quote some more:

“The alphabet soup here obscures the ambition. Darpa is the defense research agency that was critical in creating the internet, stealth technology, GPS navigation, drones and mRNA vaccines, to name but a few. “

“Why do we need an ARPA-H when the National Institutes of Health already exists? Because the N.I.H., for all its rigor and marvels, is widely considered too cautious. ARPA-H will — in a move some lament — be housed at the Institutes, but its explicit mandate is to take the kind of gambles that Darpa takes, and the N.I.H. sometimes lets go.” 

I think Klein is too open to criticisms of the NIH. Note the two words I bolded. In fact research on the NIH campus played a very critical role in the development of mRNA (and Adenovirus based) SARS COV2 vaccines. I suppose DARPA also had a role (I admit I am ignorant and don’t know what DARPA’s contribution was). The sequence of the mRNA and Adenovirus based vaccines is not the same as the SARS COV-2 sequence. It first appeared typed on a computer by an NIH researchers (because of Covid the guy was working at home in Rockville not at the campus in Bethesda). The sequence critical to all nucleic acid based vaccines was developed over a weekend. The key is that that application was almost automatic given earlier NIH research on coronaviruses. This is absolutely not an example of how DARPA works better than the NIH.

The other bolded word is “sometimes”. I will take it literally (I like to do that). If Klein criticizes the NIH, he asserts that one should never let any gamble go. Of course he doesn’t think that. Needless to say DARPA does not invest in every project any DARPA employee proposes. There may be a case for DARPA-H separate from the NIH or, at least, adding a large new program rather than increasing and modifying the NIH budget, but it sure isn’t made by mentioning mRNA vaccines or suggesting that all gambles should be taken.

I will try to make that case. Technology is produced through research and development. DARPA stands for defense advanced research and projects administration. Most of the NIH budget goes to investigator – initiated peer reviewed grants each of which is fairly small. The current US system for development of medical treatments is a strange hybrid with publicly financed pure research and then development of products by profit seeking corporations. The hand-off from NIH to Pharma is often fumbled. I would describe ARPA-H as aiming for public development as well as public research. As Klein explains a key very costly step in the process is clinical testing. For example, an EBOLA vaccine existed before the EBOLA epidemic, but it was not approved for use in humans. There is a point where research leads to something much further from actual use. It is common to have agents which work in a test tube but not in a person. Modifying a molecule so that it is absorbed, not immediately excreted, has mild as possible side effects etc. is a long slow process mostly conducted by profit seeking pharma firms.

Now I have a concern. Biden isn’t the first Democratic President in the 21st century to attempt to increase public sector pharmaceutical development. There is an earlier effort to add a drug development effort to the NIH. My father complained constantly about dealing with them (he also had frustrating experiences with corporations). Here I have vague memories and can’t find links. I can’t write anything useful, except to suggest that someone with more knowledge and patience look into earlier efforts to add the D to NIH R& and see what happened. I suppose a third example of untypical efforts at the NIH is the NIAID vaccine development effort. This involved more hierarchical central control at NIAID (by Fauci) than was typical of the other institutes. I have certainly heard the criticism that if the boss decides everything the best researchers stay away. Note above that this program worked out rather well making a critical contribution to the development of all mRNA and adenovirus based SARS COV-2 vaccines. A third example is an NIH director who decided to be bold and redirect funding to IIRC nanotechnology or something. I do not think this went well.

I will now make an extreme small c conservative argument. It is certainly true that organizations have a conservative bias — a shift of resources takes from current employees to someone new and the insiders who lose have power. However, I think there is also a problem with the resulting efforts at bold reform. New guys at the top (say Joseph Biden) want to do something, make a mark, start something which will be associated with them. They want to be bold and can’t possibly find out about the existing effort (no one can review the huge set of NIH grants). I think the history of top down radical reform is mixed, as, of course, it must be. I support DARPA-H not because it is something new and bold, but because it addresses the handoff from research to development problem.

OK so I am getting defensive about any criticism of the NIH (my father worked there for 65 years and was a very very loyal team player). Klein mostly outsources the criticism here

“A thoughtful report from New Science backs up a complaint I’ve heard privately for years: The N.I.H. is a remarkable institution beset by a deep internal conservatism. ARPA-H is an admission of this problem, even as it is located within the N.I.H.: If the N.I.H. were making the kinds of bets ARPA-H is designed to make, there’d be no reason for ARPA-H at all. But that raises the obvious question of whether the N.I.H. should be more daring at its core.”

Klein himself pulls a bait and switch here

“But the pandemic should leave no one convinced of the infallibility of our health agencies. The N.I.H. proved unable to shift focus quickly when the pandemic hit — only 2 percent of its 2020 budget went to Covid research, one study found. The F.D.A. was excruciatingly slow to approve the same rapid tests that Europe was using long before us. The C.D.C. was, flatly, a mess. Yet none of the failures we witnessed in real time led to major reforms of these agencies. That can’t be right.”

Somehow he brings in the FDA and the CDC. Notably, setting up ARPA-H does not address any problems with the FDA and CDC. On the NIH, he looks at inputs not outputs. He does not describe something that the NIH could have done about COVID 19 which they didn’t and neglects to mention the crucial contribution of intramural NIAID research or any possible NIH funding for therapeutic mRNA (I assume the research at U Pennsylvania was NIH funded but I don’t know how to check). I think the tone of this post (mostly due to my general inclination to be combative and criticize) has a lot to do with that paragraph. It does not address the issue of DARPA-H as a new agency, a new part of the NIH or no DARPA-H at all.

I can’t resist the following parody: “no one convinced of the infallibility of our” news media. The New York times has been criticized. the National Enquirer is unreliable and corrupt. FOX News is not ideal.

Now I have my thoughts on the FDA but they are not related to founding DARPA-H.

I am going to ramble on. I don’t want people to waste time. I am going to close here noting that I agree with Klein and also don’t know what I am typing about.

—–really pointless random ramblings below —————————————————————

There is something which relates FDA delays and problems with NIH research and development. The NIH has internal controls (IRBs) including a committee to decide if a proposal is worthy of the attention of the main committee. The controls are partly budgetary (partly deciding on intramural resource allocation in a way vaguely along the line of extramural peer reviewed grants. They are partly related to ethics and the motto first do no harm, second do no harm, third do no harm. We wouldn’t want to do anything that makes people with currently incurable stage 4 cancer worse off.

It is also true that organizations, including the NIH, accumulate red tape. When something goes wrong, an administrator devises a new regulation or system of control to prevent it from ever happening again. The cost of the time of salaried employees is not calculated and does not appear in a budget. Over decades procedures get more and more complicated and the time required to make sure something bad doesn’t happen again grows and grows.

One striking aspect is anti-terrorism security. After 911 the NIH campus – previously open to the public – can be accessed only after passing security (order of bording an airplane security). Al Qaeda is not what it used to be, but security is still very tight. I do not think this is optimal.

Also woke. Considerable time is spent on required training on diversity, cyber safety etc. These are videos with brief video followed by a quiz. I am sure they appear in a budget which lists the costs of making the videos but does not consider the cost of time of people supposedly learning from the videos. Here I think there is a general pattern that the value of the time of salaried employees is not considered. I don’t think there is anything like a budget which tells someone he or she can demand no more than 1000 person hours per year of work from people who are not his or her subordinates. Notice that dealing with activists by agreeing to something training is a way for top administrators to handle criticism. Also, the activists get co-opted to train. I think this is a genuine widespread problem. An advantage of ARPA-H is that it is new and has not had decades to build up pointless time consuming procedures and regulations.

I note one argument against ARPA-H which he does not even consider — it is socialist. Some have argued that the private sector is more efficient than the public sector and that the market works, so intervention through taxing and public spending is a mistake. They will note that, like DARPA, ARPA-H will use the coercive taxing authority of the state to try to pick technological winners. I suspect that (like everything) this argument can be found on the internet (son of NSF-Net son or ARPA net). Klein does not consider this argument worthy of his attention. I absolutely agree with his choice. The evidence is clear that public sector research and development (DARPA, NIH, and NSF) has been extremely successful. As an economist, I can’t resist noting that not only is it true in reality but it can be reconciled with economic theory (as can anything). Public funding of research makes sense (as well as demonstrably working) because research produces technology which is a public good. The benefits of research can’t be captured through patents which are narrow. Also (Klein notes this) patents are an inefficient way of creating incentives for discovery — the inventor gains which is good through an artificial monopoly which is socially costly. Another reason why public intervention in research *and development* is efficient is that the returns on research projects are extremely risky with most leading nowhere and a few leading to (among other things) the internet. Venture capital works with a success rate of 10% but there are huge expected gains to exploring possibilities each with a 1% chance of success. Optimal investment in research requires extreme risk tolerance and extremely deep pockets which basically means the US Federal Government. Well, that was a long paragraph on why the topic of the paragraph should be dismissed without wasting any time on it.