A failure to translate the United States' global leadership in biomedical science into a comparable position in health care (Nature 471, 569–571; 2011) does not justify dismantling the very source of that leadership — the National Institutes of Health (NIH). The real issue is that more science, data and resources are needed by other units of the US Department of Health and Human Services (HHS) responsible for engineering the application of discoveries.

Opportunities for scientific reorganization in the NIH include improving cost-effectiveness and instrumentation of assets and weaknesses. But it is crucial to separate the engine of discovery from the engine of application. Discovery is stochastic and opportunistic; application is the stuff of engineers. That is why attempts to over-engineer discovery fail and why science should not drive its application.

There should be separate units to promote discovery, assess outcomes and engineer the healthcare system. At present, these approximate to the NIH, the Agency for Healthcare Research and Quality, and HHS units such as the Food and Drug Administration.