Special Feature

Access to Cancer Information, Part I: The Internet Challenge

By Susan Molloy Hubbard

The internet, a global network of interconnected computers, provides an easily accessible and inexpensive source of information, including data on health care to users worldwide. A key factor in the popularity of the internet was the development of the World Wide Web, software that provides an intuitive "point and click" interface to data on the Internet via a powerful graphic-user interface, called a browser. Browsers generate screen displays that highlight links to other data files located on the Web. When the user clicks on a link that points to the address of a data file on the internet, the browser fetches the text, image, video or sound clip, database, or software program, and any associated hyperlinks it contains from the computer "server" that houses the data. By clicking on one link after another, users can literally roam the Internet, printing and saving data at will.

Its simplicity, accessibility, and popularity are now making the Web a primary mechanism for communicating health information. However, as Web access and Web development software tools become as basic as e-mail software in standard systems, almost anyone with a computer, an internet connection, and the appropriate software can create a web site. Currently, a search for the word "cancer" on many popular search engines will retrieve more than 1 million web pages containing the word "cancer" and more than 100,000 web pages containing the word "oncology."

This plethora of information represents a mixed blessing that can make finding authoritative sources of current, accurate, and useful information a formidable task. Anecdotal, unreviewed, and undated information can be found on e-mail services, news and use-groups, and web sites. Some are advancing unproven therapies that are not medically or scientifically sound. It is difficult for a novice to determine whether the information is current, accurate, and/or complete unless the information provider clearly articulates the standards it uses for including information and links to other sites.

The International Cancer Information Center (ICIC) of the National Cancer Institute (NCI) produces and maintains computerized systems that provide access to a comprehensive source of bibliographic citations on published cancer research (CANCERLIT®) and a source of current and comprehensive information on cancer care (PDQ®). Both systems are designed to assist users in maximizing effective use of current medical knowledge and to accelerate the pace at which research advances are translated into clinical practice.

CANCERLIT is a bibliographic database containing citations and abstracts of published research in cancer biology, etiology, screening, prevention, and treatment published from 1963 to the present from more than 4,000 different sources including biomedical journals, proceedings, books, reports, and doctoral theses. CANCERLIT currently contains more than 1.2 million literature citations, and it is updated monthly with more than 7,500 new records.

PDQ (Physician Data Query) is the NCI’s comprehensive cancer information database introduced in 1984.1-3 PDQ is a dynamic peer-reviewed database that incorporates expert opinion in its synthesis of the literature. PDQ contains three main types of information: 1) literature-based, full-text statements that reflect current, state-of-the-art information on the treatment, supportive care, prevention, and screening of cancer in English and Spanish; 2) summaries of clinical trials; and 3) directories of physicians and organizations that provide cancer care and/or screening facilities. The full-text statements in PDQ are provided in two formats: one that provides detailed and technical information written to meet the information needs of health professionals and another written in nontechnical language for the public, patients, and their families. The statements for health professionals cite the key research literature on which statements in PDQ are based so that users can easily turn to the literature and make their own assessments of the strengths of the research.

The information in PDQ is peer-reviewed by four separate editorial boards comprised of more than 70 cancer specialists who meet regularly to review the current literature on cancer screening, prevention, treatment, and supportive care and update PDQ statements based on new data. The core Boards are supplemented by an advisory Board composed of more than 100 additional health professionals who also review the information.

A process has been developed to assist board members to efficiently identify literature with significant new data. Each month, more than 70 biomedical journals are screened by professionals who identify articles of potential significance. After the articles are screened, those considered the most relevant and significant are sent to appropriate board members for review. If the board members believe the data warrant a change in a state-of-the-art statement, the article is put on the agenda of an upcoming editorial board meeting. At meetings, members discuss the research and decide whether to modify statements based on the data. Changes may involve the addition of a new management option, the deletion of options no longer supported by data, the integration of new data that further substantiate existing options, and/or highlight areas of controversy. On occasions, board members review important but unpublished research accepted for publication when the research findings have significant public health implications.

PDQ’s state-of-the-art statements can be updated as frequently as each month. On average in 1996, 30% of the statements were revised each month; approximately 15% of these revisions represent substantive changes.

Board members understand the increased emphasis on physician accountability and cost-effective practice. To assist physicians apply research data to clinical practice, the Editorial boards have moved from consensus development toward evidence-based criteria for evaluating literature. The criteria used to establish levels of evidence are explained in the PDQ database and have been published.2,3 Evidence-based summaries identify significant gaps in medical knowledge that are important to consider in decision-making management decisions and to highlight additional research questions that need to be answered.

A directory of physicians contains names, addresses, and telephone numbers of more than 24,000 physicians who devote most of their practice to the treatment of cancer patients. Also included is information on their medical specialties, oncologic subspecialty board certification, and organizational affiliations. Information can be retrieved by any combination of name, specialty, city, state, country, or zip code. An organizational directory contains information on almost 12,000 health care institutions that provide care for cancer patients including approximately 10,000 FDA-approved mammography screening facilities. Information on organizations can also be retrieved by name, city, state, country, and/or zip code.

PDQ contains approximately 1,500 summaries of clinical trials that are open or approved for patient accrual. All treatment, screening, prevention, and supportive care protocols supported by the NCI are listed in PDQ. Trials can be retrieved by diagnosis, treatment modality, phase, sponsor, location, drug name, or any combination of these parameters. Approximately one third of the active trials in PDQ are not NCI-sponsored and are submitted voluntarily. An editorial board reviews the protocols voluntarily submitted by investigators before their inclusion in PDQ. Whenever possible, trial summaries are linked to meeting abstracts and peer-reviewed publications that report research results. At present, 42% of the active studies are linked to one or more published reports. An archival file of more than 9,000 closed protocols provides a rich source of data on previously completed research; 15% of the closed protocols have been linked to one or more publications.

A variety of mechanisms has been developed to increase access to PDQ and CANCERLIT by NCI staff and commercial database vendors. They fall into two general categories: online time sharing systems with dial-up and/or internet/web access; and local implementations where the database resides on a single computer, local area network, or CD-ROM.2,3 The ICIC provides internet access to these resources in several ways. CANCERLIT and PDQ can be searched interactively on MEDLARS, the National Library of Medicine’s (NLM) search and retrieval system, and through licensed database vendors that have Internet distribution systems for these databases. The NCI’s Information Associates Program (IAP), a membership service, provides telnet access to the complete PDQ database on an interactive bulletin board and on CancerNetTM, the ICIC’s web, which is located at the following internet address: http://cancernet.nci.nih.gov.

Web-based software has enabled ICIC staff to convert information from its textual databases and many of NCI’s print products into attractive, hypermedia formats that simulate interactive publications to enhance the educational value of these resources. CancerNet contains all of the PDQ state-of-the-art statements and patient-oriented summaries in English and Spanish; more than 60 topic-specific citation and abstract "digests" derived from the monthly updates to CANCERLIT; NCI news bulletins; educational booklets and bulletins produced by NCI’s Office of Cancer Communications, a variety of scientific and science policy publications, as well as information produced by NCI divisions and offices, NCI-supported cancer centers and cooperative groups, the National Institutes of Health (NIH), other federal agencies, advocacy and support organizations, and selected organizations in the public and private sector. The ease of updating a web site makes CancerNet an ideal medium for transmitting time-sensitive material, including late-breaking news from NCI and the NIH. Recent news on CancerNet featured news releases on the role of screening mammography in women aged 40-49 and the President’s Cancer Panel’s report on the impact of managed care health delivery practices on the National Cancer Program. Currently, CancerNet fulfills more than 1,200,000 "hits"/pages of information each month. A growing number of medical facilities, cancer centers, community health information networks, freenets, and health-oriented web sites link to CancerNet. In part II of this series, the author provides information on how to get access to information for your patients and what to do if you do not use the web. (Susan Malloy Hubbard is Director, International Cancer Information Center, National Cancer Institute.)


1. Hubbard, et al. N Engl J of Med 1987;316:315-318.

2. Hubbard SM, Martin NB, et al. NCI's Cancer Information Systems—Bringing Medical Knowledge to Clinicians Oncology 1995;9:302-308.

3. Hubbard SM and Thurn AL. PDQ: In Information Systems in Oncology (Chapter 68). DeVita VT, Hellman S, Rosenberg (Eds.), Principles and Practice of Oncology Philadelphia: JB Lippincott Co.; 1997:2983-2991.