Biomedical Computing: Digitizing Life in the United States

2 Building Tomorrow’s Biomedicine: The National Institutes of Health’s Early Mission to Computerize Biology and Medicine

2 Building Tomorrow’s Biomedicine The National Institutes of Health’s Early Mission to Computerize Biology and Medicine T o James Shannon, the breakthroughs in Cambridge, England, during the early 1950s were an inspiration. They indicated to him that a productive, new approach to studying life was afoot. At its root, he declared, was “the fact that the research of Watson and Crick on nucleic acid structure and that of Kendrew and Perutz on the three-dimensional structure of proteins had been aided significantly by computational techniques.” After becoming director of the NIH in 1955, Shannon made it clear that he hoped the agency would foster similar “computational” triumphs in the United States. Pointing to the digital electronic computer, Shannon voiced his “expectation that this powerful and new technology would not only facilitate the solution of problems as currently formulated, but also, like the telescope, would permit the exploration of phenomena otherwise unapproachable.”1 During his thirteen-year tenure as director, Shannon would lead the golden age of NIH expansion. From almost the outset, he made computer technology a central component of his ambitious plans for establishing the NIH as the leader of a national push to rationalize and coordinate biomedical research.2 Within the NIH, however, Shannon’s embrace of computers was met with skepticism 68 Biomedical Computing and hostility. Shannon never overcame the opposition to computing within the NIH but rather circumvented it by drawing resources from beyond the agency. By 1960, the NIH was leading a massive effort to computerize the life sciences notwithstanding the unpopularity of the machines on its main campus. That effort, sponsored by the US Congress and initially led by Lee Lusted, would build on Robert Ledley’s operations research approach to using computers to try to propagate a new, mathematized way of investigating life. The computers in which the NIH initially invested were many orders of magnitude slower than today’s computers, and of course they were enormous— often taking up whole buildings and weighing in the tons. They were relatively cumbersome to use because data and instructions had to be prepared on punch cards—output, meanwhile, was printed onto tapes or sheets of paper. Nevertheless , in contrast to the computers used by Kendrew and Ledley, the transistorized “second generation” of general-purpose computers of the late 1950s and early 1960s were significantly faster, more reliable, less expensive, and much easier to program (due to the development of assembly languages and highlevel languages like FORTRAN and COBOL) than their vacuum-tube based predecessors. The computers the NIH purchased or rented starting in the late 1950s were mostly treated as ultra-fast calculators, and were therefore viewed as serving similar functions as adding machines, tabulating machines, and other devices that performed arithmetic operations. This particular approach to using computers opened many avenues of biomedical research but would also for many years restrict most biomedical computing activities to quantitative rather than more broadly logical or symbolic domains. One result, as William F. Raub, who guided NIH efforts to computerize the life sciences during the 1970s, explained during a “confession” he made to a meeting of the Association for Computing Machinery in 1971 was that “too many life scientists associated with our resources hold on to the outdated concept of the computer as a giant calculating device; not enough of them are led to appreciate that it is a powerful symbol manipulator and that arithmetic operations are only one of the many manifestations of this.”3 It is tempting to regard the NIH’s early effort to computerize the life sciences with a knowing smile and attribute to hubris and ignorance its inability, despite the many millions invested in it and despite the zeal with which it was pursued, to bring about the particular transformations its advocates sought. Indeed, by the mid-1960s it was quite evident that the life sciences were not becoming Building Tomorrow’s Biomedicine 69 more mathematical and logical just to accommodate computers. However, by introducing computers to so many areas of research and treatment, the NIH would create many opportunities its planners had not foreseen. In trying to realize its vision, the NIH would also profoundly influence the development of computer technology itself, with consequences for almost everyone who has ever used a computer. Finding a Sponsor for Biomedical Computing Before 1960, when the NIH became heavily involved in promoting computer technology on its campus and especially among...