Plan for 2006-2016
Goal 3. Integrated Biomedical, Clinical, and Public Health Information Systems that Promote Scientific Discovery and Speed the Translation of Research into Practice
"The physician of the future will meet with a patient and his family members, access the patient’s data in real time, including importing data from the patient’s personal health record, integrate it with a host of NLM resources, access available tutorials and other educational and explanatory materials, identify relevant clinical trials, and finally export all of this just-in-time, patient-relevant information back into the personal health record – all within the space of an office visit."
NLM has been working to enable the potential research, health care, and public health payoffs of integrated access to electronic scientific data, patient records, and medical knowledge at least since the 1980s, when it launched both the Integrated Advanced Information Management Systems (IAIMS) initiative and the Unified Medical Language System (UMLS) project. NLM’s Specialized Information Services Division has worked on connecting various types of data and knowledge needed by toxicologists and the environmental health community for decades, most recently in the TOXNET system. The Lister Hill Center has conducted research and development focused on integration of image databases (the Visible Humans are the most spectacular example) with other types of information and on gateway access to disparate NLM resources.
When the rise of the World Wide Web solved many of the significant technical incompatibility and telecommunications problems of the 1980s, NLM supported new demonstration projects under its High Performance Computing and Communications initiative and actively participated in federal efforts to promote health data standardization and address public policy issues (e.g., privacy, security, telecommunications, intellectual property, public access to government information) that affect the ability to integrate data and knowledge.
Today the National Center for Biotechnology Information’s Entrez system already enables a type of traversal from underlying biological data (at multiple levels of granularity and aggregation) to published scientific papers to synthesized information in monographs for specialists to summarized information written for the lay public. A wealth of potentially useful information remains locked away in patient records. NLM should strengthen its efforts to hasten the day when electronic health records are linked to digital libraries in systems that promote scientific discovery and bring research results directly to bear on health care and public health.
"Bingo! We knew we had nailed it." This was the enthusiastic comment by a Minnesota lab director who used an NLM DNA sequence database to identify the first case of polio in the U.S. in the 21st century. The exponential growth of molecular data generated by scientists around the world will continue and present opportunities we are just beginning to imagine, for example, studying human genetic variation and its relation to disease and learning how to create personalized drugs for optimum effectiveness.
Molecular biology and understanding derived from genomics are clearly primary drivers of medical progress in the 21st century. The quantity of genomic data has grown exponentially with the mapping of the human genome and projects that have built upon it. NLM’s National Center for Biotechnology Information (NCBI) is the information hub in this biomedical revolution, developing databases and sophisticated software tools that enable researchers to make the connections that are vital to the discovery process.
Knowledge emerging from genomics has brought hope to families stricken by hereditary diseases. The Human Genome Project is transitioning from its initial emphasis on basic research on DNA sequences to activities with increasing clinical relevance. NCBI has always worked in close collaboration with the National Human Genome Research Institute. As genomics evolves from the domain of biological research into the realm of diagnosis, treatment, and prevention of disease, NCBI has increased its active partnership with other NIH components, e.g., the flu genome project with the National Institute of Allergy and Infectious Diseases, the trans-NIH PubChem development, and whole genome association studies with the National Heart Lung and Blood Institute.
NCBI should lead the development of repositories of information on human variation that will be essential tools for discovering the associations between genes and disease. These databases will need to capture in a systematic fashion a wide array of clinical and laboratory information and couple these data with genotypes and environmental factors, including pathogens and chemicals.
NCBI should continue to be the computational and database focal point in a trans-NIH program to connect clinical and genotypic data from large, long-term studies. Scores of case-control studies, such as the Framingham Heart Study, are rich sources of detailed, longitudinal phenotypic data.
For the many studies without genotype data, initiatives are underway to genotype participants and to add that information to databases housed at NLM. These resources will offer investigators an unprecedented opportunity to study the complex interactions of genes, the environment, and health in large populations. The findings will lay the foundation for the development of therapies and prevention strategies that can be targeted to a patient’s genetic makeup, ushering in an era of personalized medicine.
NLM’s publicly accessible tools for the analysis and integration of data across disciplines, and its funding of informatics research and training programs, place it at the center of the search for genomic discoveries that promise improved health for all. However, implementing this knowledge within the framework of the current health care system will require significant effort to address privacy concerns, to deliver the necessary data efficiently, to create and maintain decision support tools, and to inform the public. Families affected by hereditary diseases may wish to facilitate access to their clinical and genomic data to speed research discoveries. There should be easy mechanisms for them to do so.
NLM has an opportunity to assume a leadership role in this effort and thus contribute to the transformation of clinical, genomic, and environmental information into better evidence-based medicine and improved patient care.
To keep pace with the rapid increase in the volume of information generated in molecular biology and to respond to the new genetic, phenotypic, and environmental exposure data generated by NIH institutes, NCBI resources will need to continue to grow at a rate linked to the explosion of data, that will necessarily be greater than the rate of growth of the NIH overall. This is a direct consequence of the appearance of high throughput, high volume research technologies that produce tsunamis of data that require human intervention to covert data into information, and information into knowledge. A formula must be found for predictable and sustained resources consonant with the research support requirements of NIH initiatives.
Current electronic health records typically resemble traditional paper-based patient charts. To a variable extent, today’s electronic health records provide some or all of the record-keeping functionality of paper-based systems (e.g., recording history and physical examination notes, progress notes, consultation notes, "written" orders, laboratory and imaging study results, and so on). In addition, current electronic health records may provide digital image storage and retrieval, limited decision support tools, some cross-record analysis capabilities useful in quality assessment and research, and some patient access to health data via the Web.
Large parts of current electronic health records are still stored as free text. Even names of drugs, laboratory tests, diagnoses or therapies are often not unambiguously stored in standardized or coded form. Standardizing and structuring patient data in electronic health records are essential to enable a host of different uses, including: detection of data errors, unambiguous transfer of information between care providers and the public health system, automated connections to supporting knowledge in digital libraries, generation of alerts and reminders, automated decision support for complex cases, construction of longitudinal records for monitoring patients with chronic conditions, Web-enabled data entry by patients, and generation of databases for clinical and health services research and the assessment of health care.
Through its Unified Medical Language System (UMLS) initiative, NLM has been a leading proponent in US government efforts to promote, develop, support, and disseminate key standard clinical vocabularies, align them with electronic message standards, and map them to billing codes.
For more than a decade, NLM has commissioned or contributed to studies by such organizations as National Academy of Sciences,[51, 52] the National Committee on Vital and Health Statistics, and the Presidential Commission on Systemic Interoperability that have shaped the strategy for advancing the development and adoption of electronic health records in the US. Through research and development work in the Lister Hill Center and NCBI, NLM has also gained direct experience with developing databases of clinically significant data, including high resolution images and clinical research data.
NLM must continue and enhance these efforts in response to specific US government priorities and feedback from those attempting to implement standards in current electronic health records and personal health records, regional health information exchanges, clinical research systems, and public health applications.
In addition to facilitating incremental improvements in retrospective data, NLM should build on its long record of conducting and supporting research, development, and policy studies related to electronic health records by promoting work to define and develop the next generation of electronic health records (EHR). Areas ripe for further investigation include: sophisticated handling of images and sounds; identification of the key patient data that actually affect care and outcomes and the best way to store and use it; application of relevant developments in computer science and telecommunications to EHR problems; advanced methods for clinical and public health data acquisition from clinicians, patients, instruments, and manufacturers, with the goal of structuring data at the source; representation of genetic information in patient records; effective strategies for linking to a variety of ancillary knowledge and decision support tools; automated transmission of notifiable disease reports to the public health system; and generation of useful research datasets that protect patient privacy.
For more than four decades, NLM has conducted and supported groundbreaking research and development related to the representation, interpretation, and use of biomedical knowledge in electronic forms. NLM grants funded much of the important research on artificial intelligence in medicine, clinical reminder and alert systems, decision rules, medical logic modules, and biomedical ontologies. NLM’s intramural research and service divisions have made leading contributions to defining useful structures for controlled vocabularies (e.g., MeSH, UMLS Metathesaurus, RxNorm) , medical publications, clinical trials descriptions, electronic knowledge bases (e.g., Hepatitis Knowledge Base, Hazardous Substances DataBank, Genetic Home Reference), high resolution anatomic imaging data (the Visible Humans), semantic networks, and effective semantic links between disparate information sources (e.g., TOXNET, Entrez). The Lister Hill Center has been a major contributor to understanding of natural language and medical images.
In some respects, these remarkable developments have been unavailable or under-used in routine medical practice. Now recent changes in computer technology, public awareness of the need to eliminate gross clinical errors, and emerging public policies seem to mark a new time and a new opportunity for NLM and colleagues to bring together many of the past technological breakthroughs for the benefit of U.S. health care and public health.
The broad deployment and use of advanced electronic health records will provide expanded opportunities for access to biomedical knowledge and advanced decision support for the public, their care providers, and the public health workforce. To turn this potential into effective reality, NLM should continue to promote research and development on robust and scaleable approaches to synthesizing, representing, updating, and deploying electronic knowledge and decision algorithms for use in conjunction with electronic health records. Topics worthy of additional concentrated attention include: drug information structured to interact with personal health records, clinical trials results reporting, automated summarization of published evidence, representations of knowledge and rules that facilitate sharing and reuse, management and updating of digital libraries of clinical and public health support tools, and methods and timing for presenting knowledge and decision support that have the greatest probability of being used.
Research should continue to exploit the information content of biomedical images and to develop algorithms for indexing and retrieving biomedical images by picture content and semantic meaning. Most important will be the incorporation of handy new imaging systems within clinical patient care. Whenever possible, intramural research should partner with allied research groups in other NIH Institutes.