WIXLIB

recommends that in cases whena database has its own indexinglanguage, or controlled vocabulary, the search be conductedwith these index terms. In thisway, the search will be the mostinclusive.Use database controlled vocabulary. For example, whenthe keyword rapid responseteam is entered into PubMed,the PubMed database matchesit to the controlled vocabularyterm “Hospital Rapid ResponseTeam.” All articles that containthe topic of hospital rapid response teams can be found bysearching with this one indexterm. Using controlled vocabulary in a search saves time andhelps prevent the chance of missing evidence that could answerthe clinical question.If the index terms matchedby the database aren’t relevantto the searcher’s keyword, thenthe keyword and its synonymsshould be used to search the database. It’s helpful, though rare,when a keyword and an indexterm match perfectly. Moreoften, the searcher will needto determine which of severaldatabase index terms is closest inmeaning to the keyword.Combine searches. Each keyword in the PICOT question issearched individually. However,keyword searches can result ina large number of articles. Forexample, a CINAHL search ofcardiac arrest resulted in morethan 2,700 articles and a searchof rapid response team resulted in100 articles. But combining thesearches using the Boolean connector “AND” (for example, car diac arrest AND rapid responseteam) yielded a more manageable12 articles that contained bothconcepts and were more likelyto answer the clinical question.(Note that databases index articles on a regular basis; therefore,ajn@wolterskluwer.comthe same search conducted atdifferent times will likely producedifferent numbers of articles.)Rebecca and Carlos want tocombine their searches becausethey’re interested in findingarticles that contain all of thekeywords (hospitalized adultsAND rapid response team ANDcardiac arrests AND ICU admis sions). After they enter each keyword into the selected databaseand search it individually, they’llcombine all the searches usingthe Boolean connector “AND.”There’s a chance, however, thatcombining the searches may re sult in few or even no articles. Forexample, the first time Rebeccasearched PubMed using its controlled vocabulary for her PICOTkeywords, and then combinedthe searches, the database cameup with only one article. She decided to refocus her search, hopingthat including only the intervention and outcomes keywords,and not the patient population,would produce articles relevantto her clinical issue.Place limits on the final combined search to further narrowthe results. This strategy caneliminate articles written in lan guages other than English orthose in which animals, and nothu mans, are the subjects. Other limits—such as age or sex ofsubjects or type of article (suchas clinical trial, editorial, orreview)—are available; however,placing too many limits on asearch may produce too few oreven no articles.Hierarchy of Evidence for Intervention Studies2Type of evidenceLevel of evidenceDescriptionSystematic review ormetaanalysisIA synthesis of evidence from all relevant random ized, controlled trials.Randomized, con trolled trialIIAn experiment in which subjects are randomizedto a treatment group or control group.Controlled trial with out randomizationIIIAn experiment in which subjects are nonrandomlyassigned to a treatment group or control group.Case-control orcohort studyIVCase-control study: a comparison of subjects witha condition (case) with those who don’t have thecondition (control) to determine characteristics thatmight predict the condition.Cohort study: an observation of a group(s) (cohort[s])to determine the development of an outcome(s)such as a disease.Systematic review ofqualitative or descrip tive studiesVA synthesis of evidence from qualitative or descrip tive studies to answer a clinical question.Qualitative or descriptive studyVIQualitative study: gathers data on human behaviorto understand why and how decisions are made.Descriptive study: provides background informationon the what, where, and when of a topic ofinterest.Opinion or con sensusVIIAuthoritative opinion of expert committee.AJN May 2010 Vol. 110, No. 543

CONDUCTING THE SEARCHRebecca begins to search thePubMed database for the evidenceto answer her PICOT question.She and Carlos will be searching the keywords rapid responseteam, the intervention of interest, and cardiac arrests and ICUadmissions, the outcomes ofinterest. To follow along, accessthe PubMed home page at www.ncbi.nlm.nih.gov/pubmed. (Notethat because new articles areadded to the database regularly,your search results may not matchthose described here.)Rebecca starts by usingPubMed’s Medical Subject Heading (MeSH) database to searchfor the intervention keyword,rapid response team. Fromthe PubMed home page, sheclicks on “MeSH Database”(see Figure 1). On the MeSHdatabase screen, she types rapidresponse team in the search fieldand clicks “Go” (see Figure 2).Rapid response team is a directmatch to the one MeSH termprovided—“Hospital RapidResponse Team” (see Figure 3).Rebecca selects this term by click ing the box next to it and thenselects “Search Box with AND”from the pull-down menu. “‘Hos pital Rapid Response Team’[Mesh]” appears in the searchbox on the next screen (see Figure 4); Rebecca clicks on “SearchPubMed.” Her search is performed and results in 19 articles(see Figure 5). She notes that mostbut not all articles appear to berelevant to the clinical question,and that they date back only to2009 because the MeSH term“Hospital Rapid Response Team”was recently introduced.Before Rebecca continues withher MeSH database searches,Lynne suggests that she use rapidresponse team in a separate searchbecause the search will be broaderthan a MeSH term search and44AJN May 2010 Vol. 110, No. 5may yield additional useful articles.From the results page, Rebeccaenters rapid response team in thesearch field and clicks “Search.”This search produces over 300articles (see Figure 6); however,many of them still don’t appearto be relevant to the clinical question. Lynne reassures Rebeccathat eventually combining hersearches will help weed out theirrelevant articles. (Because thissearch produced so many morearticles than her MeSH termsearch, which captured only themost recent articles, Lynne suggests that when Rebecca com bines her searches, she use theresults of her keyword rapidresponse team search, not her“Hospital Rapid Response Team”search.Rebecca continues to use theMeSH database to search hertwo remaining keywords. Foreach one, she starts back on thePubMed home page (click on thePubMed.gov logo on any resultspage to get to the home page).Again, she enters cardiacarrest on the MeSH databasescreen. Of the three MeSH termsprovided she selects “heartarrest,” which yields over 25,000articles. Since the keyword ICUadmissions produces no MeSHterms, Lynne advises Rebecca tosearch with the keyword inten sive care units, which matchesperfectly with the MeSH term“Intensive Care Units” andyields more than 40,000 articles.After searching her keywordand appropriate MeSH terms,Rebecca has a total of more than60,000 articles.Lynne reassures Rebecca thatshe won’t need to read all 60,000articles. She explains that the nextstep, combining the searches,will eliminate extraneous articlesand focus on the search resultsspecific to the clinical question.Combining the searches by usingthe Boolean connector “AND”will produce a list of articles thatcontain all three keywords Rebecca searched.To combine her searches,Rebecca selects the “AdvancedSearch” tab at the top of anyresults page. Each of her searchesnow appears on the AdvancedSearch page in the “SearchHistory” box. Lynne remindsRebecca to clear the search fieldat the top of the page of any keywords from past searches beforecombining the final group ofsearches.Rebecca clicks on the numberassigned to her rapid responseteam keyword search and selectsAND from the pull-down “Options” menu. Lynne shows herthat the number assigned to herkeyword search now appears inthe search field at the top of thepage. Rebecca continues to selecther individual searches and, oneby one, their corresponding numbers appear in the field above (seeFigure 7). To run the combinedsearches and view the results, Rebecca selects the “Search” tab.Her combined search produces 11 articles (see Figure 8), amuch more manageable numberto review for relevancy to theclinical question than the morethan 60,000 articles produced bythe individual keyword and controlled vocabulary searches.Rebecca asks Lynne if she canrequest the three free full-textarticles (see “Free Full Text (3)”under “Filter your results” on theupper right of the results page;Figure 8). Lynne informs her thatshe can ap ply any number of limits to her search, including “Linksto free full text.” However, themore limits applied, the narrowerthe search, and evidence to answer the clinical question may bemissed.Lynne shows Rebecca where“Limits” can be found on theajnonline.com

Figure 1. Select “MeSH Database” on the PubMed home page.Figure 2. Type rapid response team inthe search field and click “Go.”Figure 3. Select theMeSH term “HospitalRapid Response Team,”then select “Search Boxwith AND” from thepull-down menu.Figure 4. Click on “Search PubMed.”Figure 5. The “Hospital Rapid ResponseTeam” search yields 19 articles.ajn@wolterskluwer.comAJN May 2010 Vol. 110, No. 545

top of the Advanced Search page(Figure 7). She suggests that Rebecca consider limiting the agesof her population to further reduce her results. If she eliminatesthe pediatric population, forexample, the number of articlesproduced by her search shoulddecrease. But Rebecca thinks thatany articles that include childrenmay be of interest to the nurseson the pediatric unit, so she decides to limit her search to only“Humans” and “English” (Figure 9). Applying these limits toRebecca’s final combined searchreduces the re sults from 11 articles to 10.Rebecca asks Lynne if any ofthe articles retrieved in the searchare metaanalyses, which she re members is the best study designto answer her clinical question.Lynne responds that a quickway to find out is by going backto the Limits page and selecting“Meta-Analysis” (see Figure 9).Although this didn’t produceany results, limiting the search to“Randomized Controlled Trial”resulted in one article.As Rebecca’s session in searching PubMed concludes, Lynneexplains to Carlos and Rebeccathat searching is a skill that im proves with practice. Moreover,each database may have its owncontrolled vocabulary and limits.In any search, Lynne emphasizesthe importance of searching at least two databases searching one keyword at atime using the database’s controlledvocabulary when available combining the searches to yieldarticles that are manageable innumber and relate specificallyto the PICOT question applying “Humans” and “English” limits to the final searchRebecca is excited to practiceher searching skills to find theanswer to her clinical question.She and Carlos set up a timeto search the Cochrane andCINAHL databases. Carlosreminds Rebecca that althoughconsidering the level of evidencewhen making a clinical decisionis important, it’s not the only fac tor. The decision should also bebased on the quality of the evidence, the feasibility of implementing a change in the hospital,and a consideration of the patients’values and preferences.In the next article in this series,to be published in the July issueof AJN, Rebecca gathers all thearticles relevant to her PICOTquestion and meets with Carlosto learn how to critically appraisethe evidence. You’re invited toFigure 6. Type rapid response team inthe search field and click “Search”; thissearch results in more than 300 articles.Figure 7. Combine the individual searches.46AJN May 2010 Vol. 110, No. 5ajnonline.com

Figure 8. The final results.Figure 9. Using limits to narrow the search.this meeting to learn, along withRebecca, how to select “keeper”studies that, when synthesized,will help determine if a practicechange should be implemented ather hospital. Susan B. Stillwell is clinical associateprofessor and program coordinator ofthe Nurse Educator Evidence-BasedPractice Mentorship Program at Ar izonaState University in Phoenix, where EllenFineout-Overholt is clinical professorand director of the Center for the Ad vancement of Evidence-Based Practice,Bernadette Mazurek Melnyk is dean anddistinguished foundation professor ofnursing, and Kathleen M. Williamson isassociate director of the Center for theAdvancement of Evidence-Based Prac tice. Contact author: Susan B. Stillwell,sstillwell@asu.edu.Solutions to Our “Practice Creating a PICOT Question”ExerciseDid your questions come close to these?Scenario 1: A meaning question.How do family caregivers (P) with relatives receiving hospice care(I) perceive the loss of their relative (O) during end of life (T)?Scenario 2: An intervention or therapy question.In patients with dementia who are agitated (P), how does babydoll therapy (I) compared with risperidone (or antipsychotic drugtherapy) (C) affect behavior outbursts (O) within one month (T)?ajn@wolterskluwer.comREFERENCES1. Dacey MJ, et al. The effect of a rapidresponse team on major clin ical outcome measures in a community hospital. Crit Care Med 2007;35(9):2076-82.2. Melnyk BM, Fineout-Overholt E.Making the case for evidence-basedpractice. In: Melnyk BM, FineoutOverholt E, editors. Evidence-basedpractice in nursing and healthcare: aguide to best practice. 1st ed. Phila delphia: Lippincott Williams andWilkins; 2005. p. 3-24.3. DiCenso A, et al. Introduction to evidence-based nursing. In: DiCenso A,et al., editors. Evidence-based nurs ing:a guide to clinical practice. St. Louis:Elsevier Mosby; 2005. p. 3-19.4. Gibson F, Glenny A. Critical appraisalof quantitative studies: is the qualityof the study good enough for you touse the findings? In: Craig JV, SmythRL, editors. The evidence-basedpractice manual for nurses. 2nd ed.Edinburgh; New York: Churchill Livingstone Elsevier; 2007. p. 95-122.5. Fineout-Overholt E, et al. Findingrelevant evidence. In: Melnyk BM,Fineout-Overholt E, editors. Evidencebased practice in nursing and health care: a guide to best practice. 1st ed.Philadelphia: Lippincott Williamsand Wilkins; 2005. p. 39-69.AJN May 2010 Vol. 110, No. 547

42 AJN May 2010 Vol. 110, No. 5 ajnonline.com Determine the level of evi-dence. Research evidence, also called external evidence, can be viewed from a hierarchical per-spective. The best external evi-dence (that which provides the most reliable information) is at the top of the list and the least reliable is at the bottom (see Hi