Strategies and Codes for Finding Cancer Information Online

Article excerpt

For cancer information, CANCERLIT is the bibliographic database of choice. Currently 70,000 citations per year are added to CANCERLIT, with about 60,000 being journal citations extracted from MEDLINE. The remaining citations are indexed by International Cancer Data Bank personnel from selected journals, monographs, book chapters, conference proceedings, government reports, theses, and dissertations. While CANCERLIT is fairly comprehensive, other databases, either because of their size and broad coverage of the scientific literature or because of specialized subject coverage or document type provide unique citations about the field of cancer. Of course, some of these databases also overlap with CANCERLIT, but duplicates can now be removed on DIALOG and can be MERGEd offline on BRS. Cancer 'subsets" are a good starting point for retrieval of cancer information from various online databases but may not be comprehensive. Subsets utilize free-text terms and controlled indexing terms. One way to create a cancer subset is to use subject codes which are assigned during the indexing process. Use of codes is an easy way to retrieve large numbers of records with relatively few keystrokes. In some databases the equivalent words for the code are also searchable. For databases without relevant codes, the hedge of free-text cancer terms listed in Table I can be used to search descriptors, identifiers, titles, and other specific fields. Pick the words you want depending on the focus of the search. Combine them with other words such as benign with cancer or oncogenic with viruses. This hedge contains only two histologic types of cancer, carcinoma and leukemia. Add other histologic types and other terms of your choice. This article will focus on some databases clustered by subject or by database producer. For other databases, create your own cancer subsets by looking at search guides and thesauri using the Cancer Hedge as a starting point.

THE NLM CLUSTER

While MEDLINE citations comprise a large portion of CANCERLIT, other NLM databases contain unique information. "LINE (1975-) covers computer software and audiovisual materials. The audiovisuals are targeted at the health professional for educational purposes. BIOETHICSLINE (1973-) covers journals, books, legal documents, and other publications concerning ethical and related public policy and legal aspects of biomedical research and health care. DIRLINE is a directory database that contains information about many organizations involved in health care, biomedical research, poisoning, self-help groups, etc. HEALTH PLANNING AND ADMINISTRATION (1975-) covers management, organization, planning, financing, manpower, and other business aspects of health facilities and health care. journal articles, monographs, and technical reports are included. All of these databases are indexed with Medical Subject Headings (MeSH), which can be exploded using MESH Tree Numbers or treeword names. BIOETHICSLINE is also indexed with controlled keywords from the Bioethics Thesaurus. HISTLINE covers the history of medicine and related sciences and includes historical information about diseases, drugs, persons, etc., and does include cancer. The controlled keywords used to index HISTLINE are based on MESH, but are not exactly the same. Use the Cancer Hedge in Table I to search HISTLINE. To create a cancer subset in the other NLM databases, use the following MeSH Tree numbers:

  C4         Neoplasms
   D26.191    Carcinogens
   D22.204    Antineoplastic Agents.

Look through the Trees and select codes that apply to your search or explode the whole Tree. In addition, use the following untruncated free-text words in the descriptor field: cancer, tumor, neoplasm, neoplastic, carcinogens, oncologic, and the truncated word oncogen$ to retrieve terms with these stems. (The $ symbol is used for truncation.) In DIRLINE, consider searching for cancer terms in the Name Field as well. …