108th Congress

NIH: Moving Research from the Bench to the Bedside

Dr. Donald Lindberg
Director of the National Library of Medicine
July 10, 2003

Statement to U.S. House of Representatives Committee on Energy and Commerce Subcommittee on Health

Thank you, Mr. Chairman, for this opportunity to brief you and the Subcommittee about the National Library of Medicine, which is part of the National Institutes of Health within the Department of Health and Human Services. The role of the Library is central to the Nation's health, and this is due in large part to the strong support we have received in the Congress.

Progress in health care is a cyclical process, much as the title of your hearing implies. It starts with a problem recognized by a medical practitioner. This leads to experiments or experimental observations. Scientists describe the results in what we now call the peer-reviewed scientific literature. This informs the next cycle of experiments, which in turn are read by clinicians to use in patient care and by patients to inform their participation in the treatments and cures.

The National Library of Medicine - NLM - collects about 27,000 scientific periodicals from across the world and includes about 5,000 of the very best in the printed Index Medicus and the on-line Medline file. The print version started in 1879, the computer version effectively in 1965. NLM is the biggest medical library in the world. The Library is a major scientific and medical resource in the U.S. and abroad.

Let me give a measure of the scope of information that is available to today's practitioner. Medline holds the descriptions of over 14 million scientific reports. Each year we add 500,000 new ones. Clearly no doctor or scientist can possibly know all the discoveries that are described in this library. Consider the case of a conscientious medical practitioner. Let us imagine the doctor faithfully reads every night two articles from the specialty journals he or she buys. May one imagine then that the doctor has by this method kept up with progress? Really, no. By the end of such a year, this good doctor will have fallen 648 years behind on reading the new publications. So in reality what good doctors do is search the NLM files - without charge and available on Internet night and day - and read the best one or two articles for the particular patient problem of the moment. We can tell you countless examples of getting a tough diagnosis made through this system, of selecting the best new drug for treatment, and even for coming to understand new terms and ideas through reading the right paper at the right time.

There is also a version of this knowledge that is aimed at patients, families, and the public. We call this MedlinePlus. This is organized into about 600 Health Topics.

An additional important computer resource for linking laboratory discoveries to clinical practice is ClinicalTrials.gov. Here one can find out about over 7700 clinical trials in over 75,000 American communities, including the purpose of the trial, the enrollment requirements, and the telephone number of the investigator who can take new patients. The system began in 1998. It was created by NLM with initially the participation and support of all NIH Institutes, and subsequently inclusion too of trials supported by the major pharmaceutical manufacturers. The stimulus for creation of this system was the 1997 FDA Modernization Act, which authorized FDA, NIH, and NLM to make some such system for all serious or life threatening disorders.

Mr. Chairman and members of the Subcommittee, so far I have described three major NLM computer-based information systems that provide the fundamental infrastructure that connects doctors, scientists, and patients with worthwhile writings and publications on human health. This has worked well for more than 100 years, but now new science challenges us: the Human Genome Project. This and similar genomic studies on literally thousands of animals, plants, and micro-organisms make our traditional books to some extent inadequate. The human genome alone contains billions of nucleotide bases, tens of thousands of genes, hundreds of thousands of biological proteins to do the work of the genes. I am sure my colleague Francis Collins from NIH's National Human Genome Research Institute discussed this with you during the May 22, 2003, hearing before this Subcommittee - and doubtless more skillfully than I. The simple point I want to make now is that the genomic information simply is not readable from printed books. It is accessible only through a computer system that can answer questions and present the right portions of the data along with the desired relationships. This is comparable to the child looking at a drop of pond water. The life of the teeming protozoa and bacteria is visible to today's schoolchild just as it was to Leuwenhoek centuries ago only through the lens of a microscope. At NLM, that microscope to modern medicine is the National Center for Biotechnology Information (NCBI), which was authorized by Congress in 1989. It has the responsibility to collect, annotate, and provide creative access to all the human genome data from the U.S. and abroad - as well as much else. The spectacular new anti-cancer drug Gleevec, for example, came directly from clever use of these data by scientists in academia and at Novartis Labs.

Taking together all of the NLM computer knowledge sources I have mentioned, these are used on-line more than one million times each day, 500 million uses per year!

I apologize for describing only the outline of these systems, in order to stay within my time. I am submitting for the record a more detailed description of these services. If you wish, I would happily go into more detail now or do my best to answer any questions.

Additional Material for the Record to U.S. House of Representatives Committee on Energy and Commerce Subcommittee on Health

The National Library of Medicine has a number of databases and services that are involved in biomedical research, health care delivery, and information for the public. Three of the most important are PubMed/MEDLINE, MEDLINEplus, and ClinicalTrials.gov.

PubMed/MEDLINE

The "literature" is the touchstone of progress in medical research and practice. In the health sciences, the standard reference source since 1879 has been NLM's published bibliography, Index Medicus. For the past 30 years it has been supplemented by MEDLINE, an online database derived from the Index Medicus. MEDLINE (and its backfiles) is a constantly growing online resource that at last count contained more than 14 million references and abstracts to articles from about 5,000 medical journals. When it appeared in 1971, it was truly a pioneering effort in information technology, and it is today the most authoritative entry point into an ever-expanding biomedical literature. The MEDLINE files extend from the nineteen fifties to the present, and the Library is now adding data from even earlier years. PubMed/MEDLINE is by far the most widely used medical information database in the world. Each day the 14 million records are queried more than 1.3 million times by 220,000 unique users. This is roughly a half billion searches per year.

The sophisticated yet easy-to-use access system for searching MEDLINE on the Web is called PubMed. Since the launch of PubMed in 1997, continual improvements have been introduced, and today it offers a high degree of flexibility. For example, there are now Web links to almost 4,000 of the journals represented in MEDLINE, allowing users to have access to the full text of articles referenced in the database. In addition, NLM has introduced CAM on PubMed, which provides the public with access to citations from the MEDLINE database regarding complementary and alternative medicine.

An increasingly popular service on the Web for the scientist and health professional is an extension of PubMed known as PubMedCentral. This is a digital archive of life sciences journal literature, created by NLM's National Center for Biotechnology Information (NCBI). Publishers electronically send peer-reviewed research articles, essays, and editorials to be included in PubMedCentral. A journal may deposit material as soon as it is published, or it may delay release for a specified period of time. NLM undertakes to guarantee free access to the material; copyright remains with the publisher or the author. There are at present more than 50 journals in PubMedCentral, with more soon to come online.

MEDLINEplus

The National Library of Medicine, in 1998, introduced an information service directed at the general public-MEDLINEplus. MEDLINEplus is a source of authoritative, full-text health information from the NIH institutes and a variety of non-Federal sources. The main features of MEDLINEplus: more than 600 "health topics," from Abdominal Pain to Yeast Infections, detailed and consumer-friendly information about 9,000 brand name and generic and over-the counter drugs, an illustrated medical encyclopedia and medical dictionaries, directories of hospitals and health professionals, a daily health news feed from the major print media, and 150 interactive and simply presented tutorials (with audio and video) about diseases and medical procedures. With one click in MEDLINEplus, one can even do a search using PubMed/MEDLINE to retrieve references and abstracts (and in some cases, full text) of biomedical journal articles. The most recent usage figures for MEDLINEplus attest to its growing popularity among the public and health professionals. In June 2003 there were more than two million unique visitors who viewed almost twenty million MEDLINEplus pages.

The Library has learned that many health professionals are finding MEDLINEplus to be an excellent source of information. They use it to keep current on medical subjects outside of their specialty. Others are referring their patients to MEDLINEplus for up-to-date and authoritative information about their health conditions. One reason physicians feel comfortable in doing this is that they trust the imprimatur of the National Institutes of Health and the National Library of Medicine. They know that highly trained NLM information specialists follow strict guidelines in selecting Web pages that are appropriate to the audience level, well-organized, easy to use, educational in nature, and not selling a product or service. NLM receives a constant stream of testimonials from both the public and health professionals about how useful-clear and comprehensive-the system is.

Like MEDLINE, MEDLINEplus is a constantly evolving system. Links are checked daily and new health topics added weekly. In the days following September 11, entries on anthrax, smallpox, and other bioterrorism-related subjects were quickly compiled and for a while were even more heavily accessed than cancer information. The latest improvement is MEDLINEplus en Espaņol, introduced in September 2002. It provides hundreds of links to health information in Spanish and is being constantly expanded. The next major improvement in MEDLINEplus will be to link users to local resources-city, county, state, and regional agencies and support groups. In this regard, a successful prototype of a statewide system has been developed with NLM support and introduced in North Carolina.

ClinicalTrials.gov

The MEDLINEplus health topics have links to a database of ongoing and planned scientific studies-ClinicalTrials.gov. Trials are conducted when there is no proven treatment for a specific disease, or to test which treatment works best for a particular disease of condition. ClinicalTrials.gov is a registry of some 7,700 protocol records sponsored by NIH and other Federal agencies, the pharmaceutical industry, and nonprofit organizations in over 75,000 locations, mostly in the United States and Canada, but also in some 70 other countries. The stimulus that brought the FDA, NIH, and NLM together to create ClinicalTrials.gov was the 1997 FDA Modernization Act. NLM designed the system and coordinates all input from the National Institutes of Health and, through the FDA, from industry.

ClinicalTrials.gov includes a statement of purpose for each study, together with the recruiting status, the criteria for patient participation in the trial, the location of the trial, and specific contact information. The site is used extensively by patients and health professionals, and hosts over 8,000 visitors daily. NLM has worked with the Food and Drug Administration in crafting FDA's "Guidance for Industry: Information Program on Clinical Trials for Serious or Life-Threatening Diseases and Conditions." Within six months following its release, ClinicalTrials.gov received over 400 protocols from pharmaceutical industry sponsors.

Human Genome Information

The National Center for Biotechnology Information, a component of the NLM authorized by the Congress in 1989, designs and develops databases to store genomic sequence information and creates automated systems for managing and analyzing knowledge about molecular biology and genetics. With the release of the "working draft" of the human genome, the global research focus is turning from analysis of specific genes or gene regions to whole genomes, which refers to all of the genes found in cells and tissues. To accommodate this shift in research focus, NCBI has developed a suite of resources to support the comprehensive analysis of the human genome and is thus a key component of the NIH Human Genome Project.

One of the principal resources is the GenBank database, a publicly available, annotated, collection of all known DNA sequences. The NCBI is responsible for all phases of GenBank production, support, and distribution, including timely and accurate processing of sequence records and biological review of both new sequence entries and updates to existing entries. GenBank is growing rapidly with contributions received from scientists around the world and now contains more than 15 million sequences and more than 14 billion base pairs from over 100,000 species; it is accessed on the web 200,000 times each day by some 50,000 researchers.

Scientists use not only the sequence data stored in GenBank, but avail themselves of the sophisticated computational tools developed by NCBI intramural investigators, such as the BLAST suite of programs for conducting comparative sequence analysis. Entrez is NCBI's integrated database search and retrieval system. It allows users to search enormous amounts of sequence and literature information with techniques that are fast and easy to use. Using this system, one can access NCBI's nucleotide, protein, mapping, taxonomy, genome, structure, and population studies databases, as well as PubMed, the retrieval system for biomedical literature.

Continued progress in our understanding of the relation between genes and disease requires that our information-handling capabilities keep pace with the voluminous data being generated by scientists. The assembled and annotated human genome sequence is allowing researchers to identify diseases genes, decipher biological mechanisms underlying disease, and design and develop therapeutic strategies for treating and preventing disease.