A Survey of Scholarly Literature Databases for Clinical Laboratory Science

FOCUS: INFORMATION LITERACY^

A Survey of Scholarly Literature Databases for Clinical Laboratory Science

DONNA LO’MALLEY

Ihis article reviews the use of journal literature databases including CINAHL, EMBASE, and Web of Science; sum- marizing databases including Cochrane Database of System- atic Reviews, online textbooks, and clinical decision-support tools; and the Internet search engines Coogle and Coogle Scholar, llic series closes with a practical example employ- ing a cross-section of the knowledge and skills gained from all three articles.

ABBREVIATIONS: C : D S R = Cochrane Database of Sys- tematic Reviews; CINAHL = Cumulative Index to Nursing and Allied Health Literature; CDM = gestational diabetes; SCI = Science Citation Index.

INDEX TERMS: algorithms; bibliographic databases; infor- mation storage and retrieval; Internet; medical technology; online systems.

Clin Lab Sci 2008;21( l ) :49

Donna L O’Malley MLS is library associate professor at

rhe Dana Medical Library at the University of Vermont,

Burlington VT.

Addressforcorrespondence:Donna L OMalley MLS, library

iissociate professor, Dana Medical Library, University of Ver-

mont, Medical Education Center, Burlington VT05405. (802)

656-44!5, (802) 656-0762 (fax), donna.omalley@uum.edu.

Frances Delwiche MLIS MT(ASCP) is the Foeus: Information

Literacy guest editor.

fhe Focus icctiim seeks to publish reievani and timely continuing vducation for cliuicnl laboratory practitioners. Section editors, topics, and authors art selected in advance to cover current areas of interest in each discipline. Readers Ciin ohtiiin continuing education credit (CE) through PA. C.EJ’hy completing the continuing education registration form, recording answers to the examination, and mailing a photocopy of it with the appropriate fee to the arUress designated on the form. Sugges- lions jor future Eoctis topics andatithors, and manuscripts appropriate for CE credit are encouraged. Direct all inquiries to the Clin Lab Sci Editorial Office, ICInk, 858 Saint Annes Drive, Iowa City IA 52245. {319)354-3861, (319)338-1016 (fax), ic.ink@mchsi.com

LEARNING OBJECTIVES 1. Descrihe what is meant by “primary literature” in the

health sciences. 2. Discuss the characteristics of the major primary and

summarizing databases used by health professionals. 3. Discuss the advantages and disadvantages of searching

the Internet for professional health information. 4. Illustrate how popular Internet search engines can be

used to find unique information in the health sciences.

The primary literature in the health sciences consists of re- ports of original research generally published in the form of articles in scholarly/academic journals. Ihe articles in these journals are indexed by searchable databases such as MED- LINE, CINAHL, and EMBASE, which function as an aid to finding articles on a desired topic. The primary literature has the advantage of being a direct communication from the researchers who performed the investigations. By studying the methodology used, the results of the research, and the investigators’ reasoning, readers are able to reach their own conclusions regarding the validity of the research findings. However, until they have stood the test of time, the findings of original research must be interpreted with caution.

Since the primary literature is such an enormous body of work, much of which will later be disproved or substan- tially revised, many scientists and health professionals rely on secondary and tertiary forms of literature to summarize original research. These forms of literature provide the es- sential background knowledge required to understand and interpret the primary literature and for making professional and clinical decisions. Examples of such literature include review articles, yearbooks, print and online textbooks, and clinical decision-making tools.

This final article in the FOCUS series will examine several databases of the primary literature, as well as several secondary and tertiary resources commonly used by health profession- als. It will also examine the use of popular Internet search engines for finding forms of literature not usually included in proprietary databases. Finally, the article will close with a practical example utilizing a cross-section of the concepts, techniques, and skills described in the entire series.

V0L21. NO 1 WINTER 2008 CUNICAL LABORATORY SCIENCE 49

 

 

FOCUS: INFORMATION LITERACY

PRIMARY LITERATURE DATABASES Though MEDLINE is the premier database for the ptofes- sional health sciences literature, research shows that it does not provide complete coverage of all the journals relevant to clinical laboratory science.’ To cover a topic comprehensively, other literature databases such as CINAHL, EMBASE, or Web of Science may need to be searched in addition to or instead of MEDLINE. These databases may cover a different mix of journal titles representing a slightly to markedly differ- ent subject area coverage; their records may contain unique elements of information not found in the other databases; or they may utilize alternative mechanisms for identifying related records. As a result, a search in one of these databases will frequently bring up articles not found in MEDLINE.

CINAHL The focus of the Cumulative Index to Nursing and Allied Health Literature (CINAHL)’ database is evident from its title. CINAHL includes over one million records for journal

articles, book chapters, doctoral dissertations, conference pro- ceedings, websites, and other types of publications. CINAHL covers publications from 1982 to the present. Although there is considerable overlap with MEDLINE, CINAHL indexes many publications important to health sciences professions that are not covered by MEDLINE. Access to CINAHL is available through EBSCO Information Services and Ovid Technologies. Like the MEDLINE records described in the first anicle, CINAHL records include standard bibliographic elements, such as author, title, journal name, volume, issue, page numbers, and publication date, plus abstracts and CINAHL subject headings. CINAHL records also include several unique fields, such as the Cited References field and Journal Subset (Table 1).

The search algorithm used by CINAHL depends largely on which vendor’s version is being searched. Ovid Technologies’ advanced search actively encourages the searcher to utili/c CINAHL subject headings rather than relying solely on key-

Table I. Important fields in

Field Author

Institution

Title

Source

Journal subset

CINAHL subject headings

Abstract

Publication type

Language

Citations

a CINAHL record

Contents The personal name or names of the authors of this document

Contact information for the author to whom correspondence should be addressed

The title ot the article

The information necessary to locate the document. In the case of a journal article it is the journal name, volume, issue and page numbers.

CINAHL allows searchers to limit their searches to subsets of the database. Sub- sets include subject areas, country of publication, and peer review status.

Just as Medline uses the MeSH vocabulary, CINAHL uses an approved list of subject headings. These subject headings overlap with MeSH, but include special- ized terms related to allied health and nursing.

A summary of the document

Journal article, book, audiovisual, pamphlet, software, dissertation, research instrument, etc.

Most documents referred to in CINAHL are in English.

The references in the document’s bibliography

50 VOL 21. N O l WINTER 2008 CUNICAL LABORATORV SCiENCE

 

 

FOCUS: INFORMATION LITERACY

word searching, while the EBSCO Publishing system offers ,1 unique Visual Search feature which sorts results into topics represented by colored circles. Databases that assign subject headings to their records, such as CINAHL and MEDLINE, can be searched most effectively by using both subject head- inu,.s and synonymous keywords.’** All vendors of CINAHL olkr a means of: identifying CINAHL subject headings.

Like MEDLINE, several options exist for limiting a CI- NAHL search to create a small set of highly focused results. These include the familiar limits for language, age groups, and publication types, and more specialized limits such as special interest category, geographic journal subset, and clinical queries. For example, a search can be restricted to the professional literature, thus eliminating articles written for consumers or patietits, by limiting the publication type to “Journal Article” or “Research”, or by limiting to the “Peer- leviewed Journals” subset.

EMBASE EMBASE”” is a biomediciil literature database published by Elsevier, headquartered in Amsterdam, The Netherlands. EMBASE contains records for over 1 1 million journal articles From over ^,000 biomedical journals published in 70 countries. EMBASE excels in coverage of European and non-English language publications, as well as in the subject areas of pharmaceuticals, psychiatry, toxicology, and alter- native medicine. EMBASE’s earliest articles are from 1974 and new records are entered on a weekly basis. EMBASE is available through various vendors, including Dialog/Data- star, DIMDI, Lexis/Nexis, NERAC, OVID Technologies, and S IN . Elsevier also offers its own version of EMBASE through EMBASE.com.

EMBASE records contain the usual citation details, plus abstract, subject headings, drug descriptors, medical descrip- tors, medical devices, brand names, and manufacturer names. KMBASE uses a list of approved subject headings comparable ro MEDLINE’s McSH heading.s called the EMTREE the- saurus. EMTREE allows searchers to include all narrower terms under broader terms, much like MEDLINE’s “explode” feature. Though EMTREE makes no effort to use the same subject heading.? as MeSH, the EMBASE.com version of EMBASE^ includes MeSH terms in its records in addition to its own subject headings.

lo search EMBASE, begin with a keyword search. Then, from the complete record of relevant articles, identify search- able terms that can be used to narrow, expand, or otherwise

revise the search. In addition to a search box to enter terms, a link to the EMTREE thesaurus is provided to enable brows- ing for subject headings. An EMBASE search is especially useful for identifying articles in European journals that are covered less comprehensively by MEDLINE, and for searches in subject areas in which EMBASE is particularly strong.

Web of Science/ Science Citation Index Web of Science” is produced by Thompson Scientific, an international information company with a wide range of products for academia, business, and government. Web of Science contains over 38 million records and consists of three databases: Science Citation Index, Social Sciences Citation Index, and Arts & Humanities Citation Index, each of which may be searched separately or in combination with the others. Science Citation Index (SCI) contains references to over 6,300 peer-reviewed journals in science, medicine, and engineering. SCI provides data from as far back as 1900, with weekly updates.

Although SCI has a general search function, its real strength is in cited reference searching. In addition to basic bib- liographic information (author, article title, journal name, publication date, etc.), SCI records contain the full list of references cited by the articles indexed (Table 2). The cited and citing articles are connected by the database software in such a way that a concept can be cracked backward and forward in time through the articles’ bibliographies. Suppose a searcher has located a useful but older article, and wants to identify more recent articles that have been published on that topic. By bringing up the SCI record for the original article and clicking on the “Times Cited” link, one can easily obtain a list of articles that used the original article in their bibliographies. Presumably, more recent articles that cited the original article will be on the same or a similar topic. Although other literature databases, including CINAHL, have a similar feature, because SCI indexes such a large num- ber of journals crossing many disciplines, its cited reference searching is far more useful.

Like PubMed, SCI has a “related articles” feature, but it is structured on a completely different algorithm. In PubMed, related articles are identified on the basisof their shared sub- ject headings. In contrast, related articles in SCA are ideniifed based on shared references in their bibliographies. Articles that share many of the same references with the original record will appear higher in the list than those that share just a few references.

VOL21, NO 1 WINTER2008 CLINICAL LABORATORY SCIENCE 51

 

 

FOCUS: INFORMATION LITERACY

Lacking a formal thesaurus of subject headings, SCI relies on ke)’words for general subject searching. As a search mecha- nism, this is not nearly as powerful as the subject heading

Table 2. Important Index record

Field Title

Authors

Source

DOI

Language source

Abstract

Document type

Author keywords

Keywords plus

Subject category

Bibliography

Fields in a Science Citation

Contents Title of the journal article

Journal article authors

Journal title, volume, issue and page numbers

Digital Object Identifier. Eor records that do not have page numbers and an article number, the record displays theDOi, if available.

Language in which the

article was published

A summary of the article

Journal article, letter, or review article

Keywords as listed by the author

Words or phrases that frequently appear in the titles of an article’s refer- ences, but do not necessarily appear in the title of the article or in a list of author keywords

Broad subject category of the journal

The full record also includes the references from the article’s bibliography.

searches available in MEDLINE, CINAHL, or EMBASE. Hence, general subject searches are best performed in these other databases, at least initially. However, SCTs cited refer- ence search is an excellent tool for expanding a search, espe- cially when only a small number of articles can be found.

DATABASES THAT SUMMARIZE Literature that summarizes the primary research of a given field serves a vital role in the communication of knowledge. It saves the time of the reader and highlights important points that someone less familiar with a topic might not appreciate. It is especially useful for gaining background information on topics outside one’s area of expertise. In the clinical arena, secondary literature serves as a valuable source of ready information regarding diagnostic indicators and treatment guidelines. However, since they are removed from the original documents, summarizing publications run the risk of introducing misinterpretations, biases, and omissions. With this caveat in mind, databases of secondary literature can be highly useful.

journal articles that summarize research regularly appear in MEDLINE, CINAHL, EMBASE, and Web of Science as review articles. Review articles may take the form of literature reviews, often with hundreds of references, or they may be in the form of academic tutorials, such as this article. Most literature databases have an option to limit searches to review articles only. Other databases, such as the Cochrane Database of Systematic Reviews, electronic textbooks, and clinical decision-making tools, contain only summaries.

Cochrane Database of Systematic Reviews The Cochrane Database of Systematic Reviews (CDSR)~ is a full text database of a specialized type of review article designed to summarize all relevant research on a specific healthcare intervention. High quality research studies, primarily random- ized controlled trials, are identified through an exhaustive search of the literature, coupled with a hand search of selected journals, and their findings compared following a standardized protocol. The completed systematic review anal}v,es the restilts of all the studies and derives an overall conclusion regarding the intervention. Each review concludes with implications for clinical practice and suggestions for future research. As a summary of the most rigorous research studies available, the Cochrane systematic review represents the highest level of evi- dence available for that particular intervention. Of course, not all possible interventions or treatments will have been studied, so not all topics will have a review in CDSR. Approximately 4,800 systematic reviews have been published.

52 VOL 21, NO I WINTER 2008 CLINICAL LABORATORY SCIENCE

 

 

FOCUS: INFORMATION LITERACY

The CDSR is updated quarterly, and is available from Wiley Interscienceand Ovid Technologies. Abstracts of the reviews are available tree online trom the Coehrane Collaboration .u http://wwwxochrane.org/reviews/. The CDSR is searched using either keywords or McSH headings. As a full text database, the entire records are searchable. This is a mixed blessing, as ke)’word searches often bring up articles missed by subject heading searches, but they may also retrieve more ir- iclevant articles simply because they pick up every occurrence i)f the term{s) anvwhere in the full-text of the documents.

Since CDSR is indexed by MEDLINE, the reviews can be found by conducting a topic search in MEDLINE, combined with a search on CDSR as a journal title.

Online textbooks Ihe content and structure of textbooks is familiar to any cur- rent or past student in the sciences. Online, these textbooks have tables of contents, chapters, page numbers, and indexes just like their print counterparts. Upon their initial publiaition, ihe online textbooks are identical to the print versions, but they have the advantage of being able to be updated far more fre- quently and economically. Vendors of online textbooks in the health sciences include Ovid Technologies, AccessMedicine, MDConsult and the (free) NCBI Bookshelf These companies typically attempt to include a textbook representing most of the major medical specialty areas, thus enhancing their useful- ness to busy clinicians and hospital staff. Usually only the most current edition of a textbook is available online.

The full-text of online textbooks is usually searchable. If the search is specific enough it can target a narrow topic, but quite often it will generate a long list of irrelevant hits due to incidental occurrences of the search terms. When avail- able, searching within specific chapters or using the index can help to focus a search.

Clinical decision-support databases In contrast to the seemingly anachronistic online textbooks, with their quaint tables of contents, indexes, and even page numbers, new formats for online information have emerged. Ihese are clinical decision-making tools designed to pro- vide health professionals with rapid, authoritative access to clinical information. Examples of these databases include Clin-eguide, DynaMed, eMedicine, FIRSTConstilt, InfoR- ctriever, and UpToDate. Each of these competing resources fill a unique niche, some providing lengthy, evidence-based topic reviews, and others very brief summaries that can be accessed via handheld devices.

The information in these databases is more clinically oriented than textbooks. DynaMed records, for example, include fields for ICD-9/-10 Codes, Causes and Risk Factors. Complica- tions and Associated Conditions, History, Physical, Diagno- sis, Prognosis, Treatment, Prevention, Screening, References, and Patient Education information.

Clinical decision-support databases are updated on a quar- terly, monthly, or weekly basis. These databases don’t have editions as textbooks do. When a change is made in the database, the old information is not retained.

Researchers have estimated that healthcare professionals will spend no more than two minutes researching a clinical question,** so database developers have experimented with a variety of interfaces to make these databases operate more quickly. Each database is different and can change its interface without warning. Users should observe the display carefully for instructions on how to conduct a search, knowing that programmers are continuously making improvements.

The bibliographies of records in these databases are an excel- lent source of high quality references. References may link to journal articles through PubMed or to guidelines available from professional organizations or government bodies.

THE INTERNET Some of the most heavily used literature databases were converted to electronic form on mainframe computers de- cades before the rise of the popular Internet. MEDLINE, for example, was searchable electronically through the MED- LARS system as far back as 1964.” Cancer.gov went online in 1982 as the National Cancer Institute’s Physician Data Query (PDQ*) database.’” Today, both of these databases are available free to the general public through the Internet. While some may think of the Internet as an alternative to databases of professional literature, in fact it merely serves as a conduit through which to access this and other types of information.

The difficulty with the Internet is that the high-quality, reliable, and authoritative information is so thoroughly in- termixed with everything else that it is sometimes virtually impossible to find. It requires a sophisticated searcher and a refined skill set to effectively extract high-quality scholarly and professional documents. Using the popular Internet search engine Google as an example, some of these techniques will be exatnined.

VOL21.NO1 W1t>4TER20Oe CUNICAL LABORATORY SCIENCE M

 

 

FOCUS: INFORMATION LITERACY

Google Google, also known as Google Web, is a database with un- limited scope. The authors of pages in Google range from experts in every field to charlatans deliberately promoting false information. The Google database consists of a copy of every page on the Internet that it has been able to access, now numbering in the billions. Each page becomes a record, with the words on the page sorted into fields, including page title, headings, links, and images.

(ioogle rose to prominence among Internet search engines in part due to its innovative Page Rank system. The Page Rank for a Google record is based on how often other web pages cite the page and the rank of the citing pages. It is this feature that allows Google searches to produce more relevant results than other Internet search engines. Googles ability to outperform other search engines is also due to the approaches it uses to translate and execute search queries. Google is understandably reluctant to share the specifics of its search algorithm,” but generally speaking, it displays the results of a search in order of relevance and Page Rank, with the most relevant and highest-ranked pages displayed first. This presents a distinct drawback for scientists who are often looking for the most recent information. Since recent articles have not had the benefit of years of exposure that would allow other web pages to link to them, they are often not among the most highly ranked by Google, and therefore may not appear in the top pages of a Google search.

Google is a powerful search engine and an enormous da- tabase. Research has shown that, unfortunately, the larger a database is the harder it is to locate information in it, no matter how good the search engine or how expert the searcher.” However, there are several effective strategies that will maximize success when searching Google. One of the most useful is the use of quotation marks for phrase search- ing. If a search on a phrase produces records that don’t use the words in the order expected, it may be placed in double quotation marks to ensure that Google will search exactly those words in exactly that order. Another strategy is to search on the most specific and unique terms possible.

Rather than searching on a specific topic directly, relevant results can often be found by searching for the name of the organization, society, or government body that would most likely provide the information. Access that organization’s home page, and conduct an internal search. The strategy of searching for an organization rather than the exact topic can be very helpfiil because of Google’s limitations. Google

cannot transfer a search from its search box to the search box on another website. Although Google can see all of PubMeds records, for example, it cant search like PubMed, mapping to MeSH headings and combining terms in the way PubMed does. Nor can Google generate results from a dynamic database. The Centers for Disease Control and Prevention (http^/v^-ww.cdc.gov), for example, links to many databases that will produce reports based on CDC statistics. Because these reports are generated dynamically, they only exist when someone requests them, so Google can’t include them in its database.

Some well-known publicly available databases are not avail- able through Google at all. Examples include the National Cancer Institute Clinical Trials Database, the US Patent and Trademark Database, and US Census Bureau data. The infor- mation in those databases is available through the Internet, but is not searchable by Google. Private, fee-based databases, such as CINAHL, EMBASE, Web of Science, most medical textbooks, and clinical decision support systems, are also not searchable through Google.

Google Scholar Google attempts to address the needs of scientists and profes- sionals through its Google Scholar database. Google Scholar includes journal articles, theses, books, meeting abstracts, and certain unique documents not found elsewhere. Unlike most licensed literature databases. Google has agreements with many publishers that grant it permission to search the full text of each article.

Like Google Web. Google Scholar displays the most relevant, highest ranked pages first, but Google Scholar uses cited refer- ences from bibliographies, not Page Rank, to determine rank. Publications that have been cited more frequently by other publications will be ranked higher. Again, this can impact the retrieval of recent articles. In order to address this prob- lem, Google Scholar offers a Recent Articles link at the top of the results page. When this link is clicked, a drop-down box for publication date appears, allowing searchers to limit the search to a specified year, forward.

Google Scholar has several unique features that ftirther distinguish it from its parent. When multiple versions of a given article exist in the database, it groups them together via a link attached to the original record. Eor example, there may be an official version of an article on the publishers website, a link to the preprint of the article on the authors website, and a link to the PubMed record. By clicking on

54 VOL 21, N O ! WINTER 2008 CLINICAL LABORATORY SCIENCE

 

 

FOCUS: INFORMATION LITERACY

the All Versions link, Coogle automatically displays these records as a separate set.

Secondly, articles in Google Scholar that have been cited hy other articles in Google Scholar indicate this connection ihrough a “C’ited By” link. (Clicking on this link will produce .1 list of articles chat cited the original article (Figure I). Ilius, Google Scholar provides a form of cited reference searching similar to that provided by Web of Science.

I’inally, Google Scholar employs an undisclosed mechanism lor finding related articles, using the Related Articles link ibund at the bottom of each record.

Figure 1. Google Scholar “Cited By” feature

I n w y t Vidw Naws Map.s more»

Results 1 – 4 of 4 for “Antibiotic resistance: a survival strateq

Tip: Tiy removing quotes ftom your search to gel more insults.

Antibiotic resistance a survival strategy AT Sheldon Ji – Clin Lob Za. 2005 – ncbi nini nih gov AnBbtOlIc rflBtctance: a survival strategy SKeidon AT Jr. Antibiotic md

nc. SHvv Spring. MD 20910. USA…. j,.l^i-,V^I.. “.f.aii-.li- Fimlltj^lK M- BL D(i»c(

Antibiotic resistance mechanisms among pediatrtc respiratory

enteric pathogens a current update .. – all 4 vetsions igemann – The Podialric Infectrous DisftBSB Journal. 20O6 –

ihsldon AT Jr AnHblotlc resistance a survival •gy ClmLatiSci 2005:18.170.180,…

VIdao Nawa more»

Scholar B**UI« t – ? OI J mng Sheldon Jr: AntlbtoBc reststance; a su

y resistance meciianisms among pediatric respiratory and enteric pathogens a current update .. – .iii 4 veisions -i AB Bru»ageriann – The Pediatri; Infechous Disease Journal 2006 – pldj.org Antibiotic leslstancaisaconiinuelly incieasing problem thai has. to a giaator orl«(H[ eirtant. affected virtually avaiyaraa of tha world. The {clenUfic

is abundant with papers related to antibiotics and anbbiotic …

TolC Is Required tor Pathogenicity of Xylella fastidiosa in Vrtis vintfera Grapevines – -jli “‘ .-ei^iofi? .ID ReJdy SL Roddy. DL Hopkins DW Gabriel – apsnei oig MPMI Vol 20. No. 4, 2007, pp. 403-410, doi: 10.1094/MPMI-20-t-0403.C. 2007 T I M American Phytopathological Sociaty… Joseph D. Reddy. I Stephanie L. Reddy, 1 Don L. Hopkins. 2 and Dean W. GabriaM … t Oepartmrnt of Plant > RaWedAilicles-Wetj …>,ii-li

Since Google Scholar includes articles that are not available free on the Internet, many records include a link to ohtain a copy through other means. If the article is not available for free or through an institutional subscription, the publisher may ofter the option to purchase the article outright. There may be a link through Library Search to WorldCat, a service that lists libraries that subscribe to the journal in question. Or, the BL Direct link offers the option to purchase a copy of the article through the British Library, although it is often more expensive than purchasing a copy directly from the publisher.

Google Scholar can be useful for cross-disciplinary searches. Topics in public health, healthcare administration, and education are not always well covered in databases like MEDLINE, EMBASE, and CINAHL On the other hand, owing to the enormity of Google Scholar, if a topic is well- represented in a smaller, subject-specific database, a search of that database may be more productive.

U S I N G T H E L I T E R A T U R E IN G L I N I C A L LABORATORY PRACTICE Suppose your laboratory is interested in exploring the various methods and protocols for screening for gestational diabetes (GDM), and you want to find what has been published to date on the topic. You begin by reviewing the basic facts of the disease, such as its clinical importance, incidence and prevalence, early detection, methodologies used for screen- ing, current screening recommendations, and the effective- ness of treatment.

These background questions can best be addressed by lit- erature that summarizes the current state of knowledge. From the array of clinical decision-making resources avail- able, you choose UpToDate as a starting point. Searching on “gestational diabetes”, you find the topic review entitled “Screening and diagnosis of gestational diabetes mellitus”.” llie article discusses the prevalence of [he disease, possible adverse outcomes for mother and infant, and therapeutic outcomes. It summarizes guidelines for screening and di- agnosis of gestational diabetes mellitus from the American Diabetes Association, the American College of Obstetricians and Gynecologists, the United States Preventive Services Task Force, the Canadian Task Force on Preventive Health Care and the Fourth International Workshop-Conference on Gestational Diabetes Meliitus. You note that these guidelines are not in complete agreement with one another. The review concludes with a summary and its own recommendations. The bibliography for this article includes 67 references, most with links to the MEDLINE abstract.

VOL21, NO 1 WINTER 2008 CUNICAL LABORATOtW SCIENCE 55

 

 

FOCUS; INFORMATION LITERACY

Having obtained the clinical background of the disease, you next seek laboratory-specific information from a current textbook or manual. MDConsult’s book collection contains the 21″ edition oi Hemy’s Clinical Diagnosis and Manageme7it by Laboratory Methods,^^ wherein the chapter on carbohy- drates offers a lengthy discussion on glucose measurements and CDM screening protocols, including references to the American Diabetes Association guidelines. Ovid Technologies provides access to A Manual of Laboratory & Diagnostic Tests,’ ̂ which describes the test procedure, patient preparation and aftercare, and limitations of the test. While the recommen- dations included are not referenced, they do follow W H O guidelines. Conflicting guidelines are not mentioned.

Additional information may be sought from the websites of professional societies and organizations. A quick Google search leads you to the website of the American Association for Clinical Chemistry. This site lists a number of standards, including the 2002 Guidelines and Recommendations for Labo- ratory Analysis in the Diagnosis and Management of Diabetes Mellitus.^” This guideline does acknowledge the controversy. The authors recommend following the American Diabetes Association guideline, but do not provide their rationale.

Real izing there is no consensus on issues surrounding screen- ing for CDM,’^ you begin to wonder if there is another

methodology for testing that might prove more effective than the traditional glucose tolerance test. You decide to search the journal literature indexed in MEDLINE, which covers mLich of clinical chemistry. You quickly access PubMed via your hospital library’s website. Recalling that PubMed executes a simultaneous MeSH term and textword search, you enter the word string “gestational diabetes screening” into PubMed’s search box. You limit your search to human studies, English language, and Core Clinical Journals. A quick scan of the results brings up a case-control study from 2007 describing the use of various serum markers for predicting gestational diabetes in the first and second trimesters of pregnancy.”* You are intrigued, and decide to expand your search. Checking the MeSH headings assigned to this article, yoLi re-run the search utilizing additional televant terms. Also, you follow the Related Articles link for the most relevant results to bring up additional useful articles.

In an effort to search the literature thoroughly, you repeat the search in Ovid CINAHL, which indexes some publica- tions not covered by MEDLINE. Searching on each concept separately, you generate separate sets of results for “diabetes mellitus, gestational”, “glucose tolerance test”, and “health screening”. You also generate a set of articles indexed under the subject heading “diabetes mellitus. gestational” combined with the subheading “diagnosis”. Finally you combine the

Figure 2. CINAHL search strategy

CINAHL – Cumulative Index to Nursing & Allied Health Literature

Search History

enp Diabetes Mellitus. Gestatlondl/

CKp Glucose Tolerance Test/

etp Health Screening/

Und 2 _ 1 atid 3 eip Diabetes Mehicus: Gestatlonal/dl [Diagnosis] 4 or 5 or 6 limit? to English Hnlt S to illlcd healtk joiirnsb

so: 910

738^

155

57

201

303

303

2 2

– DISPLAY

. DISPLAY

• ̂ D I J P U ^ “~^SPLA7

• •iVDISt’LAY

-; DIS PLAY

-DISPLAY

D I S ^

Enter Keyword or phrase (use “$” lor truncation): ]

HeaiJIng

Figure 3. Web of Science Cited Reference Search

Cited Reference Search Select dal4l><i»el«) and tlmespaiii

ClUtton Databases: B „ Science Citation Index Eipanded

I- V Social Sciences Citation indei (SSCl)-19a2-present r u Arts £i Humanities Citation Index (A&Ha)-1962-present

Find the citations to a person s work by entering the person s name, the work’s source, and/or publication /ear.

“” LatMt IL ^^-i ^ {updated Juhfl 29, 2007)

‘ ‘YMr [ J’T’? j j ” From [1S&; ̂ to | ;ocr- ^

(deliult I) all yeirs)

To remember these settihgt, fitst sign in or register.

r our Cited Rtftrencc Search (utonw

CITED AUTHOR: i> Enter the name ol the cltcQ author (see citad autftor Irnlai •-.). Enmple: O’BRIAH C- OW OBRIAH C |5mlrnakls KV

CITED woRKi ^ Enter the abbreviated loumai/book title In which the work appeared, a patent number, or another work [set EjtecLworljinjtei or view.^he Tti?!!l?on_IS! list of loutnal abbrtv|atlojis).

] Comput AppI Math •

CITED YEAR(5): L Enter year, or range of years, the cited work was published, EKSmples: 1943 or 1943-19*5 ^___

56 VOL21,NO1 WINTER200a CLINICAL LABORATORY SCIENCE

 

 

FOCUS: INFORMATION LITERACY

^et.s, using the Boolean AND and OR. To help eliminate duplicates from the previous PubMed search, you limit your Hnal set to Allied Health journals (Figure 2). From this set, you discover several unique hits to add to your collection.

Rounding out your search, you conduct a Cited Reference Scatch in Science Citation Index to see if any reseatch has heen published following up on the 2007 study (Figure 3). Mo references are brought up. probably because che study is so recent. Would Google Scholar have any unique hits? Since PubMed, CINAHL, and Web of Science only allow searches to match words in a limited number of fields such as the title, abstract, and subject headings, being able to search the full text of articles might be advantageous. Furthermore, Cjoogle Scholar may contain some forms of “grey” literature such as pre-prints and technical reports that are not generally included in the traditional databases. Clicking on the Recent Articles link and limiting to 2002 forward, you scan for any unique documents.

CONCLUSION Ihe research question on screening for gestational diabetes lesulted in a search of eight different databases. Fortunately, most research topics in clinical laboratory science do not require such an exhaustive (and exhausting) search, However, when a comprehensive search is needed, you’ll want to take advantage of all the resources at your disposal and draw on Ll variety of search techniques. By learning the characteristics t)i the various information resources available, cultivating the skills to search them competently, and knowing how to evaluate tlie quality of the studies you find, you’ll be well on your way to becoming truly “information literate”.

Clin Lab Sci encourages readers to respond with thoughts, questions, or comments regarding this article. Email responses lo ic.ink@mchsi.com. In the subject line, please type “CLIN LAB SCI 21(1) FO O’MALLEY”. Selected responses will ap- pear in the Dialogue and Discttssion section in a fixture issue. Responses may be edited for length and clarity. We look forward to hearing from you.

REFERENCES 1. Delwitbe FA. Mappinj; the literature of clinical laboraiory science,

J Med Libr Assoc 2003 Jiih’JI (3):303-10. 2. CINAHL. Ipswitjh. MA: EBSCO Publishing. 3. Hayncs RB, Wilczynski N. McKJbbon KA, and others. Developing

optimal search strategies for detecting clinically sound studies in MEDI.INE. J Am Med Inform Assoc t994-Nov-Dec;](6):447-58.

4. NJfilczynski NL, Haynes RB: Hedgt-s Team. Developing optimal search strategies for detecting clinically sound prognosiic studies in MEDLINE: an analytic survey. BMC Med 2004 jun 9;2;23.

5. EMBASE. Amsterdam. 7he Netherlands: Eisevier. 6. Web of Science. Pbiladelpbia: ‘Ihomson Scientific. 7. Cochrane Database ofSystematic Rt-vicws. Chichester: Wiley. 8. Ely JW, Osheroff JA, Ebell MH. and others. Analysis of questions

asked by family doctors regarding patient care. RMJ I’»y Aug 7;319(7206):358-6l.

9. Milestones in NLM History. Available from btip;//www.iilm.nib. gov/about/nlmhistory.html. Accessed 2007 May 30.

10. Nicdcrhuber JE. Celebrating 70 years of excellence in cancer research. Available from http://wvvw.cancer.gov/aboLitnci/ncia. Accessed 2007 May 30.

11. Hansell S, Google keeps tweaking its search engine. “Ihc New York Times 2007 June 3. Available from http://vwv^,nytinies. com/2007/06/03/buslness/yourmoney/03google.htnil. Accessed 2007 June 2′).

12. Motville P. Ambient Findability. Sebastopol. CA: O’Reilly Media: 2005.

] 3. Jovanovic. L. Screening and diagnosis of gestational diabetes mellitus. In: UpToDate. Rose, BD (Ed), UpToDate, Walth.ini MA 2007.

14. McPherson RA, Pincu.s MR. ediiors. Henry’s clinical diagnosis and management by laboratory methods. 21″ ed. Philadelphia: Saunders Elsevier; 2007.

15. Eischbacb FT. A manual of laboratory & diagnostic tests. 7”’ ed. PbiLidelpbia: Williams & Wilkin.s; 2004.

16. Sacks DB, Bruns DE. Coidstein DE, and others. Cuidelincs and recommendations for laboratory analysis and in the diagnosis and management of diabetes mellitus. Clin Cbem 2002;48:436-72.

17. Hollander MH, Paarlberg KM. Hiiisjcs AJ. (k-stational diabetes: a review of- the current literature and guidelines, Obsiet Cynecol Surv 2007 Feb;62(2): 125-36.

18. Smirnakis KV, Plati A, Wolf M. and others. Predicting gesrational diabetes: cboosing the optimal early serum marker. Am J Ohstet

Cynecol 2007 Apr;l96(4):410.el- 410.e7.

V O L 2 I . N O I WINTER ^ n e CUNICAL LABORATORY SCIENCE 57