Skip Navigation Bar

Long Range Plan 2000-2005

Table of Contents

GOALS AND OBJECTIVES
GOAL 4 - CONDUCT AND SUPPORT INFORMATICS RESEARCH 

From research laboratories to the patient bedside, biomedical knowledge is being generated at a staggering rate. This new knowledge must be captured, analyzed, and disseminated in order for it to be useful and to make a significant difference in healthcare. In addition to the traditional forms of knowledge, the use of the computer has enabled researchers, practitioners, and health care consumers to more effectively gain and use knowledge. The challenge lies in finding new approaches to deal with the increasing volume and complexity of biomedical information and thereby improve our understanding of health and disease.

OBJECTIVE 4.1 -
FURTHER MEDICAL INFORMATICS RESEARCH

FINDINGS

Research is conducted at NLM both in the in Lister Hill National Center for Biomedical Communications (LHNCBC) and in the National Center for Biotechnology Information (NCBI). Lister Hill works in the more general field of medical informatics. NCBI concentrates in work relevant to molecular biology and genomics. LHNCBC work includes language and information processing, digital library research, consumer health informatics, image processing, and advanced computing and communications for health care applications. Ongoing R&D projects include the Unified Medical Language System (UMLS), Natural Language Systems, Indexing Initiative, image and document management and delivery systems, assisted searching of NLM’s databases through Internet Grateful Med, the clinical trials database, the Visible Human project, and the Profiles in Science digital library project.

The UMLS project develops and distributes knowledge sources for improved access to biomedical information. The Metathesaurus contains information about biomedical concepts and terms from more than 40 controlled vocabularies and classifications. The Semantic Network identifies the semantic types and relationships of Metathesaurus concepts, and the SPECIALIST lexicon and associated lexical programs allow for linguistic processing of biomedical text.

There are more than 1000 licensed UMLS users who use the knowledge sources in a wide range of applications, including indexing bibliographic and clinical material, retrieving information from Web-based systems, diagnostic prompting systems, electronic medical records, and formalizing the language used in medical records and messages. NLM makes use of the UMLS in its own applications, including Internet Grateful Med, PubMed, and the clinical trials database. 

Natural language research is focused on the development of SPECIALIST, an experimental medical language processing system. Modules based on the major components of language, including the lexicon, morphology, syntax, and semantics, have been built and are used in research, particularly in the area of information retrieval effectiveness.

Document image analysis and recognition techniques play an important role in several research projects. One project (MARS) has partially automated the entry of citation data from scanned biomedical journals into the MEDLINE database. Another (DocView) aids the library patron in receiving library documents delivered through the Internet. Sample of NLM Web sites

The clinical trials database project addresses issues in consumer access to health information. When fully operational, the database will be a comprehensive resource for patients, families, and members of the public, providing easy access to clinical trials funded by the Federal government and by private industry.

DXPNet, a collaborative project among the NLM, the National Center for Health Statistics (NCHS), and the National Institute of Arthritis, Musculoskeletal, and Skin Diseases (NIAMS), brings together a collection of radiographs and related text material from the National Health and Nutrition Examination Surveys.

The Visible Human male and female data sets, consisting of MRI, CT and cryosection images, were released as national resources in 1995 and 1996 respectively. Users in over 40 countries are applying them to a wide range of educational, diagnostic, treatment planning, virtual reality, artistic, mathematical and industrial uses. Work has also begun on the next phase of development of the Visible Human, including the segmentation, classification, and three-dimensional rendering of the data sets.

The digital library research program investigates all aspects of creating and disseminating digital collections, including proposed and adopted standards, emerging technologies and formats, effects on previously \established processes, and protection of original materials. The Profiles in Science digital library site focuses on major scientific achievements of the twentieth century by presenting the archival collections of prominent biomedical scientists on the World Wide Web.

Through the Extramural Programs Division, NLM has a number of programs for the support of medical informatics research in universities, hospitals, and research institutions. These include investigator-initiated grants in medical informatics, biotechnology information, and health sciences library and information science, institutional and individual training grants, and a variety of research contracts.

A number of special programs fund, for example, a collaboration with the National Heart, Lung, and Blood Institute projects to apply medical informatics techniques to speed critical life-saving information to heart attack victims. NLM has also funded over twenty telemedicine research projects, as well as planning projects in health-care applications for the Next Generation Internet. NLM’s participation in a multi-agency Digital Libraries Initiative has resulted in several large awards for health-care related digital library research projects .

NLM has carried out in-house studies and awarded research grants over the years to determine the information needs and uses of health professionals. A landmark study of this kind employed the Critical Incident Technique to identify the impact of MEDLINE-derived information on a wide range of professional activities, including medical decision-making and patient care outcomes. Other studies surveyed NLM users to determine satisfaction with specific products and services, including beta testing of new offerings such as Grateful Med. Still others queried hospital libraries, and other organizations and individuals, with respect to their readiness to adopt Internet technology. Extramurally, NLM has awarded grants that looked at information needs, usage and value, including the value of providing information at the point of care .

PROGRAM PLANS

  • Strengthen fundamental informatics research, both through intramural research at NLM and extramural research at universities, research laboratories, and other organizations.
  • Increase support for investigator-initiated research grants.

LANGUAGE AND KNOWLEDGE PROCESSING

  • Continue to explore advanced indexing technologies for present and developing forms of information, including automated concept-based indexing techniques for the biomedical literature.
  • Continue to identify additional concepts, organizing principles, and concepts to be added to the UMLS resources.
  • Expand the UMLS Semantic Network, particularly in the areas of anatomy and the genome.
  • Make the UMLS Knowledge Sources better able to reflect the different perspectives and views of medical and health concepts exhibited by the general public, health care practitioners, health policy makers, and clinical and health services researchers.
  • Develop efficient methods for more frequent update and dissemination of the UMLS Knowledge Sources to support the need for immediate access to new vocabulary in areas such as drugs and devices.
  • Enhance the capabilities of the Internet-based UMLS Knowledge Source Server to make the UMLS components easier to use and to promote direct use by external software applications.

DIGITAL LIBRARY RESEARCH

  • Continue to participate with other agencies in the Digital Libraries Initiative - Phase 2 (DLI-2) to support innovative digital libraries research and applications.
  • Continue digital library research, addressing issues in building, maintaining, preserving and disseminating diverse multi-media digital collections, including the use of metadata for managing, displaying, and retrieving data in digital archival systems.
    • Extend the Profiles in Science digital library site.
    • Develop collaborative projects with other institutions as test-beds for digital library research.
  • Expand efforts to improve search system interfaces, search engines, and utility of intelligent agents (e.g. search assistants), both for human beings and network applications.
  • Work with the database efforts of other Federal agencies and similar international efforts to enhance interoperability among databases and ease of access to data.
  • Further develop document management technologies and applications, including scanning, optical character recognition, and document structure analysis.

VISIBLE HUMAN Photo of Visible Human image

  • Using the Visible Human data set, undertake and support standardization efforts for classifying, storing, retrieving, and displaying anatomic images.
  • Develop 3D anatomic image sets from the Visible Human data and conduct research on effective retrieval and transmission of such images over the emerging Next Generation Internet.
  • Using the Visible Human data set, conduct and support research in the development of generalizable image processing tools, such as the Visible Human Image Processing Tool Kit, a public domain, open source software toolkit which will be capable of automated segmentation and alignment of radiological and anatomical images.
  • Extend basic research in automated image indexing and retrieval, using the salient features of the Visible Human images themselves.
  • Conduct basic research in algorithms for automated image segmentation, recognition, indexing and decomposition.
  • Work towards the further integration among the UMLS, the Visible Human, and other anatomical initiatives such as the Human Brain Project.

SIMULATION

  • Conduct and support research in the development of a Visible Human Project Atlas for use in educational applications, beginning with the head and neck body regions.
  • Augment the Visible Human with plans and approaches for developing collections relating to various abnormalities and disease models.

JUST IN TIME ANSWERS

  • Encourage research, development, and innovation in “just in time” knowledge retrieval. The goal is to tailor specific answers to questions posed by clinicians at the time and place they are seeing patients, and to help patients and families to find information specific to their own immediate health concerns.
  • Conduct and support research to link patient-specific data to related knowledge-based information.

CONSUMER HEALTH

  • Conduct and support informatics research designed to produce health information systems that the public can use easily and understand.
  • Develop a consumer health terminology server to provide assistance to the increasing members of the general public who are users of NLM’s Web-based systems, including spelling correction algorithms and investigation of multi-language interfaces.
  • Use NLM databases, e.g. the clinical trials database and MedlinePlus, as test-beds for research on health information seeking behavior by the public. Photo of Dr. Lindberg and grandson at computer

EVALUATION

  • Conduct and support basic and applied research to identify health care professionals’, researchers’, and the general public’s need for, access to, evaluation of, and use of biomedical and health information.
  • Broaden end user studies beyond measures of satisfaction to include the effects of information on health outcomes.

DATA MINING AND MACHINE LEARNING

  • Promote and support research on health data mining as a method for discovering new clinical, public health, and health services information, making use of UMLS tools as appropriate.
  • Explore the use of data mining tools in databases, literature, and the NLM collection.
  • Pursue metadata approaches to the problem of data standardization at all levels, from basic electronic transfer to higher levels of information structure.