For those who revere technological innovation, nostalgia is alluring and often counterproductive. Such is likely the case with the Biden administrations proposal to establish a new agency to boost medical innovation—the Advanced Research Project Agency for Health (ARPA-H). The idea is to replicate for health care what the Defense Advanced Research Projects Agency (DARPA) has done for defense technologies. Skepticism is in order.

For those who came of age in the 1960s, few things captured our imaginations more than NASA’s drive to put men on the moon via Project Apollo. Twenty days after Alan Shepards 15-minute hop on Freedom 7, President John Kennedy challenged America to commit itself to achieving the goal, before this decade is out, of landing a man on the moon and returning him safely to the Earth.” Eight years later, the challenge was met.

Thus, moonshot” become a metaphor for the government harnessing vast resources to develop new technologies. The term used to have a different meaning. Merriam-Webster cites a 1936 article where moon shot” meant a highly unlikely event—victory by a very low-rated racehorse. That dictionary entry argues Project Apollo changed the meaning of moonshot: Now, politicians who announce a moonshot … intend it in the new sense of the word, as an ambitious venture which is intended to show significant results.” The 20th centurys other great example was the Manhattan Projects six-year development of an atomic bomb.

Therein lies ARPA-H’s problem. Project Apollo and the Manhattan Project were enormous triumphs. DARPA has an impressive list of achievements, perhaps most importantly ARPANET—now known as the Internet. DARPA also played some role in Modernas development of the mRNA vaccine for COVID-19. But we remember these triumphs not because they are the rule, but rather because they are so atypical.

Success with the bomb, the moon landing, and the internet share a combination of characteristics. First, each primarily involved engineering, rather than basic science. The physics of thermonuclear explosion, rocketry, and electronic communications were already well-developed. The challenge was to turn established physics into viable machines. Second, each goal was binary and easily observable. Either we landed on the moon or we didn’t. Third, the federal government was willing to spend whatever was needed. Fourth, the geopolitical environment was so urgent as to push aside the usual bureaucratic gamesmanship and sluggishness. The Manhattan Project, Project Apollo, and ARPANET arose during times when America was threatened by foreign tyrannies.

ARPA-H is supposed to become a pollinator—devising ideas, outsourcing, coordinating research, and assembling disparate entities into teams. The White House announcement said of ARPA-H: “President Biden Reignites Cancer Moonshot to End Cancer as We Know It.” That should be a warning. The cancer moonshot, launched by then-Vice President Biden, has generated research, discussion, and so forth, but it hasn’t put us on any road to ending cancer. The environment that led to the bomb, the moon landing, and the internet were absent. Curing cancer didn’t involve simply building a machine to accomplish a task whose science was already known. Metrics for determining success and failure were mostly absent. By the standards of medical research, the cancer moonshot’s budget was tiny. And there is no compelling geopolitical reality to stifle the usual bureaucratic dysfunctions.

ARPA-H is emerging in the same sort of environment. We can see the business-as-usual environment in a recent dustup over whether to house ARPA-H in the National Institutes of Health (NIH) or in the Department of Health and Human Services (HHS). None of this means that the federal government has no role to play in spurring innovation in health care.

An alternative approach exists in an interesting but underused NIH program—the Eureka Prizes. ARPA-H, like most grant-providing agencies, would offer funding to innovators it thinks would be likely to produce innovations. The Eureka Prizes, housed in NIH, take a different tack. Here, the government states a clear, well-defined goal—say, developing a cure for one specific type of cancer by some specific date—and promises a set amount of cash for the first entity that succeeds at the goal.

Rather than ARPA-H’s approach of having the federal government pick innovators, the Eureka Prizes allow innovators to pick themselves. That’s an idea worth pursuing.