Group: NIH must hike IT spending

Biomedical researchers are starving for computing resources, and the National Institutes of Health should spend more money on computer hardware, software, training and computer science research to support them, an NIH advisory group said June 3.

A report by the Working Group on Biomedical Computing did not recommend any specific amount of funding, but it did say, "A small push by NIH could result in great changes in a short time."

"I think you can read between the lines and see that we're not talking about tens of millions, and we're not talking about billions," said David Botstein, professor of genetics at Stanford University Medical School and co-chairman of the panel.

But if NIH does not spend enough money, the study warned, programs could be delayed. "If NIH does not act, change could take another five, 10 or 20 years," according to the study.

Larry Smarr, the other co-chairman of the group and head of the National Science Foundation-funded National Center for Supercomputing Applications, said the recommendations are intended "as a minimum of a five-year program" that could be launched by giving more money to programs that NIH already has in place. "Obviously we would very much be pleased if the program were started immediately," he said.

In a prepared statement, NIH director Harold Varmus said the group had "come up with a number of thought-provoking recommendations. The institute directors and I will be studying the report closely, discussing it with others at NIH and expect to have something to say in the reasonably near future about implementing these proposals."

"As far as nuts-and-bolts details like money, nothing is out there yet," an NIH spokesman said.

According to the report, most new money would be doled out as grants to universities and other research institutions, as NIH spends most of its budget now. One of the working group's proposals is that NIH fund five to 20 National Programs of Excellence, at $1.5 million to $8 million each. The programs would enable scientists to develop better software and would increase collaboration between biologists and computer scientists.

The report also suggests that NIH provide more support to NSF's supercomputing centers, where biomedical research makes up one-fourth of computer usage.

Computers, especially midrange systems and supercomputers, have become increasingly important to researchers, the working group concluded, but most researchers do not have enough access to the systems. Without the systems, scientists will have difficulty using huge biomedical databases, such as those maintained by the National Library of Medicine's National Center for Biotechnology Information.

Brett Berlin, a consultant who advises the government on supercomputing issues, said NIH has skimped for years on investments in high-performance computing, both internally and externally, relying instead on NSF and other agencies to supply these resources. "What they're not stepping up to is the presumption that the entire future of the research establishment is at stake here," he said.

Smarr said the group's proposals "are consistent with" recommendations by the President's Information Technology Advisory Committee, which said the government should spend more money on basic research and on improving the nation's computing infrastructure. The Clinton administration wants to spend $366 million next year on such programs, while Rep. F. James Sensenbrenner Jr. (R-Wis.), chairman of the House Science Committee, is floating a plan to spend $4.8 billion over five years.

Whether NIH could get more money next year is uncertain. David Kohn, press secretary with Rep. John Porter (R-Ill.), who heads the House Appropriations Committee's subcommittee in charge of the NIH budget, said programs under the panel's jurisdiction are facing cuts under the current House budget blueprint.