DOI: 10.1542/pir.32-8-3502011;32;350Pediatrics in Review 

Christine Briccetti and Peter RoweClinical Literature

Research and Statistics: Searching for Answers: Strategies for Searching the

http://pedsinreview.aappublications.org/content/32/8/350located on the World Wide Web at:

The online version of this article, along with updated information and services, is

Pediatrics. All rights reserved. Print ISSN: 0191-9601. Boulevard, Elk Grove Village, Illinois, 60007. Copyright © 2011 by the American Academy of published, and trademarked by the American Academy of Pediatrics, 141 Northwest Pointpublication, it has been published continuously since 1979. Pediatrics in Review is owned, Pediatrics in Review is the official journal of the American Academy of Pediatrics. A monthly

at Queens Univ on August 24, 2014http://pedsinreview.aappublications.org/Downloaded from at Queens Univ on August 24, 2014http://pedsinreview.aappublications.org/Downloaded from

a

Author Disclosure

Drs Briccetti and Rowe have

disclosed no financial relationships

relevant to this article. This

commentary does not contain a

discussion of an unapproved/

investigative use of a commercial

product/device.

Searching for Answers: Strategies forSearching the Clinical LiteratureChristine Briccetti, MD,* Peter Rowe, MD*

Case StudyYou are evaluating a 5-day-old in-fant whose newborn screen was re-ported as positive for primary congen-ital hypothyroidism. While you awaitthe results of his thyroid stimulatinghormone (TSH) and thyroid assess-ment, the family asks you about theaccuracy of the newborn screening.They want to know how likely it is thattheir baby has primary congenital hy-pothyroidism. How would you get theinformation needed to answer theirquestions?

Searching the LiteratureMEDLINE is the National Libraryof Medicine’s database of biomedicalarticles. PubMed (www.PubMed.org) is the free resource that is usedmost commonly to search the MED-LINE database. PubMed ClinicalQueries (found under “PubMedTools”) is a feature that can be ac-cessed from the PubMed home page.This tool filters out information thatis not likely to be useful clinically.Although not suited for extensive lit-erature searches, this procedure is anexcellent method to search for usefulclinical information in a few minutes.

The first step is to decide whatquestion or information to enter intothe search engine to obtain the mostuseful results. One approach is tophrase the query in the “PICO” for-mat. (1) PICO is an acronym thatwas developed as a strategy to helpbreak down questions into their mostimportant subcomponents and trans-late them into terms that enable ef-fective searches for related evidence-based information. The “P” in PICO

stands for patient, population, orproblem. (2) What are the character-istics of an individual patient that areimportant? Is the patient’s sex, age,condition, or some other factor ofparticular importance? The “I”stands for intervention, which couldinclude exposure, diagnostic test,prognostic factor, or treatment. (2)“C” represents comparison or the al-ternative to the intervention. (2) Forexample, when searching for infor-mation about a new therapy, thecomparison might be the standardtherapy or no therapy. A diagnostictest would be compared to the diag-nostic test that is considered the goldstandard. Finally, the “O” is the out-come. (2) Depending on the ques-tion, outcome could represent symp-toms, adverse effects, mortality, oraccuracy of diagnosis. Splitting aclinical question into these parts canaid in identifying the primary con-cepts to query with a search engine(Table 1). A clear, structured, andsearchable question developed withthis process for the search in the casestudy is: In newborns, how accurateis a positive newborn screen for con-genital hypothyroidism comparedwith serum TSH in diagnosing con-genital hypothyroidism?

Study DesignThe next step is to determine whattype of study best answers the clinicalquestion (Table 2). The best type ofstudy design to evaluate a diagnosticquestion is a cohort study, in whichall patients receive both the test be-ing evaluated and the diagnostic goldstandard. (2) Evidence obtainedfrom systematic reviews or meta-analyses of randomized, controlled*Johns Hopkins University, Baltimore, MD.

research and statistics

350 Pediatrics in Review Vol.32 No.8 August 2011

at Queens Univ on August 24, 2014http://pedsinreview.aappublications.org/Downloaded from

trials is considered the highest levelof evidence.

Search ResultsThe next step is to find search termsbased on the key concepts as identi-fied in the PICO question. Depend-ing on the variety and the usefulnessof the articles obtained, the searchersmay include some or all of the keyconcepts. Alternative spellings orterms may be needed to include alarger number of articles. Some ofthe search terms that may be appro-priate for the question on the accu-racy of the newborn screen for pri-mary congenital hypothyroidism arelisted in Table 3.

Combining search terms requiresthe use of Boolean operators. Allsimilar terms are combined with theBoolean operator “OR” to indicatethat they are synonymous. Key con-cepts are separated by the Booleanoperator “AND” so that all conceptsare included in the results. For exam-ple, a search of newborn screeningAND congenital hypothyroidism un-

der the category of diagnostic studiesyields 20 studies and 16 review arti-cles, including “Screening for con-genital hypothyroidism: UnitedStates Preventive Services reaffirma-tion recommendation.” (3) Accord-ing to this article, 1 in every 25 new-borns who has a positive screeningresult for hypothyroidism ultimatelyreceives the diagnosis of congenitalhypothyroidism.

Case ProgressionYou share the information you ob-tained from your literature searchwith the family. At this point, confir-matory test results have returned foryour patient and you diagnose congen-ital hypothyroidism. You prescribelevothyroxine and schedule the familyfor an appointment with a pediatricendocrinologist. You explain to theparents that compliance with levothy-roxine is crucial to prevent develop-mental delay. They ask you what thechances are that their baby will haveintellectual disability, given optimummedical treatment.

You design another PICO questionto guide another literature search.Your “P” is children who have congen-ital hypothyroidism, “I” is treatmentwith levothyroxine and alternative isno treatment, and “O” is future cog-nitive ability. Because you are investi-gating the long-term outcomes of a pa-tient who has a certain condition, youinvestigate cohort studies. In general,cohort studies follow patients who havea certain exposure (in this case, treatedhypothyroidism) to a particular out-come (presence of intellectual impair-ment). Cohort designs are powerful formaking an association between causeand effect. At the PubMed ClinicalQueries page, you use the same keywords of newborn screening ANDcongenital hypothyroidism but searchunder “prognosis.” This time, in ad-dition to getting the same review ar-ticles, you find 71 articles specificallyrelated to prognosis. After limitingyour results to papers published in thepast 5 years in English, your totalnumber of articles becomes 19. Youfind a clinical report from the Amer-ican Academy of Pediatrics thatstates that if treated within 2 weeksof birth, children who have primarycongenital hypothyroidism have sim-ilar intelligence, school performance,and neuropsychological testing re-sults as their peers. (4)

ConclusionAlthough most busy practitioners donot have the time for exhaustive lit-erature searches for each clinicalquestion they encounter, they canuse the few steps outlined in thisarticle to find answers to questionsthat matter most to them and theirpatients. Other articles in this serieswill give clinicians the skills to evalu-ate the quality and results of the arti-cles they find.

Table 1. Example of the PICO Process to Identifya Research Question

P Patient or Population NewbornsI Intervention or Indicator Newborn screen for congenital hypothyroidismC Comparison Serum TSH measurementO Outcome Diagnosis of congenital hypothyroidism or the

“accuracy of the test”

Table 2. Choosing the Type of Research Study (2)

Type of Question Best Type of StudyDiagnosis Cohort study in which patients receive both the test being

evaluated and the diagnostic gold standardTherapy Double-blind, randomized, controlled trialPrognosis Longitudinal cohortCause Cohort studies (start with patients with and without the

risk factor)Case-controlled studies (start with patients with and

without the disease)

research and statistics

Pediatrics in Review Vol.32 No.8 August 2011 351 at Queens Univ on August 24, 2014http://pedsinreview.aappublications.org/Downloaded from

References1. Richardson WS, Wilson MC, NishikawaJ, Hayward RS. The well-built clinical ques-tion: a key to evidence-based decisions.ACP J Club. 1995;123:A12–A132. Evidence-Based Answers to Clinical Ques-

tions for Busy Clinicians. Melbourne, Aus-tralia: The Centre for Clinical Effectiveness,Southern Health; 20093. United States Preventive Services TaskForce. Screening for congenital hypothy-roidism: United States Preventive Services

reaffirmation recommendation. Ann FamMed 2008;6:1664. American Academy of Pediatrics. Up-date of newborn screening and therapyfor congenital hypothyroidism. Pediatrics.2006;117:2290–2303

Table 3. Sample Search TermsKey Concepts Search Term

P Newborns Newborn or infantI Newborn screen for congenital hypothyroidism Newborn screen for congenital hypothyroidism or

congenital hypothyroidism diagnosisC Compared with TSHO Accuracy of test Accuracy or diagnosis

research and statistics

352 Pediatrics in Review Vol.32 No.8 August 2011

at Queens Univ on August 24, 2014http://pedsinreview.aappublications.org/Downloaded from

DOI: 10.1542/pir.32-8-3502011;32;350Pediatrics in Review 

Christine Briccetti and Peter RoweClinical Literature

Research and Statistics: Searching for Answers: Strategies for Searching the

ServicesUpdated Information &

http://pedsinreview.aappublications.org/content/32/8/350including high resolution figures, can be found at:

Referenceshttp://pedsinreview.aappublications.org/content/32/8/350#BIBLThis article cites 3 articles, 2 of which you can access for free at:

Subspecialty Collections

ods_-_statistics_subhttp://pedsinreview.aappublications.org/cgi/collection/research_methResearch Methods & Statisticstion_subhttp://pedsinreview.aappublications.org/cgi/collection/medical_educaMedical Educationfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://pedsinreview.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://pedsinreview.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

at Queens Univ on August 24, 2014http://pedsinreview.aappublications.org/Downloaded from