ePROFILES Elizabeth Connor, Column Editor ePROFILES is a column that features interviews with practitioners, leaders,andvisionariesinthefieldsofelectronicpublishing,digital libraries, medical informatics, content licensing, and related areas. Interview with Brian S. Alper of DynaMed Elizabeth Connor SUMMARY.ThisinterviewwithBrianS.Alper,MD,MSPH,founder andeditor-in-chiefofDynaMedandthemedicaldirectorofClinicalRef- erenceProductsatEBSCOPublishing,describessystemicliteraturesur- veillance, topic creation, and peer review processes which distinguish ElizabethConnor,MLS,AHIP([email protected])isAssistantProfes- sorofLibraryScienceandScienceLiaison,DanielLibrary,TheCitadel,171Moultrie Street, Charleston, SC 29409. CommentsandsuggestionsshouldbesenttotheColumnEditor:ElizabethConnor ([email protected]). [Haworthindexingentrynote]:“InterviewwithBrianS.AlperofDynaMed.”Connor,Elizabeth.Pub- lishedinJournalofElectronicResourcesinMedicalLibraries(TheHaworthInformationPress,animprintof TheHaworthPress,Inc.)Vol.4,No.1/2,2007,pp.127-143.Singleormultiplecopiesofthisarticleareavail- ableforafeefromTheHaworthDocumentDeliveryService[1-800-HAWORTH,9:00a.m.-5:00p.m. (EST). E-mail address:[email protected]]. Available online at http://jerml.haworthpress.com ©2007 by TheHaworthPress, Inc. All rights reserved. doi:10.1300/J383v04n01_12 127 128 ELECTRONIC RESOURCES IN MEDICAL LIBRARIES theproductasaneffective,point-of-careevidence-basedmedicineresource for PCs and PDAs.doi:10.1300/J383v04n01_12[Articlecopiesavailablefor afeefromTheHaworthDocumentDeliveryService:1-800-HAWORTH.E-mail address: <[email protected]> Website: <http://www.Haworth Press.com>©2007byTheHaworthPress, Inc. All rights reserved.] KEYWORDS. Electronic resources, electronic publishers, DynaMed, EBSCOPublishing,evidence-basedmedicine,systematicreviews,clin- ical reference resources Brian S. Alper, MD, MSPH, is Founder and Editor-in-Chief of DynaMed, EBSCO Publishing. Question: Please introduce yourself and describe your current roles and responsibilities and a typical day at work atDynaMed. Alper: I am Brian S. Alper, MD, MSPH, the founder and Editor-in- Chief of DynaMed and the Medical Director of Clinical Reference ProductsatEBSCOPublishing.Myrolesincludesupervisionandtraining ofeditors,supervisionofsystematicliteraturesurveillance,supervision of topic creationand peer review, planning of technicalenhancements andcontentupgrades,andevaluationofclinicalreferenceproductsfor potential license or acquisition by EBSCO Publishing. The editors at DynaMed use a systematic literature surveillance modeltolookatover500journalseitherthroughdirectsurveillanceor viajournalreviewservices.Eacharticlegoesthroughaprocesstodeter- mineitsrelevance,andrelevantarticlesarefurtherassessedforvalidity relativetoexistingDynaMedcontent.Themostvalidarticlesaresum- marized, and those summaries are added to DynaMed. This process goes on every day, which allows us to update DynaMed daily on an as-needed basis. Question:Forthebenefitofreaderswhomaynotbefamiliarwithyour company’sproducts,brieflydescribethecompany,itsfounder,itshis- tory, and current management. Alper:WhenwesayDynaMedwascreatedbyadoctorforhealthcare professionalswemeanjustthat.Iamthatdoctor.AsamedicalstudentI realized I could not possibly memorize sufficient medical information tosupportpracticesoIbeganorganizingwhatIwaslearningelectroni- cally. Through this process I discovered that the information I had eProfiles 129 organized was more useful than textbooks or existing resources used during clinical rotations. DuringarotationinruralpracticeinEasternTennessee,Isawmyin- formation make real differences in diagnosis and treatment every day. Atthispoint,itwasnotyetinasophisticatedsystembutIrealizedthat doctorsandphysicianassistantsneededthiskindofresource,andIhave beenworkingtoservethatneedeversince.IncreatingDynaMed(Dy- namicMedicalInformationSystem),aclinicalreferencetooldesigned toprovidethemostusefulinformationtohealthcareprofessionalsatthe point of care, I created a systematic way to monitor current literature and continuously update DynaMed based on the best available evi- dence.ForyearsDynaMedwasfreeandclinician-usersvolunteeredas reviewers, authors, and editors. In2004,DynaMedreceivedagrantfromtheNationalScienceFoun- dation (NSF) to research its usefulness and ease of use for clinicians. The grant was designed to systematically prioritize user interface im- provementsbyworkingwithaspectrumofrepresentativeusers,andto determine the feasibility of using DynaMed to provide rapid access to valid clinical knowledge in primary care. These NSF-funded research results contributed to the continued improvement of the database and supportedarigorousrandomizedtrialdemonstratingthatDynaMedwas more effective than usual information sources. Inmid-2005,DynaMedbecameamemberoftheEBSCOPublishing family.IsawthisunionasthebestwaytocontinuetheDynaMedmis- sionofprovidingthemostusefulinformationtohealthcareprofession- alsatthepointofcare.AspartofEBSCOPublishing,DynaMedhas been able to grow and improve while still providing the best available evidence updated daily for its worldwide community of users. EBSCOPublishingisadivisionofEBSCOIndustries,Inc.whichhas over60yearsofexperienceinthelibrarycommunity.EBSCOPublish- ing has more than two decades of experience with libraries. EBSCO PublishingisexpandingitsmedicalproductlineandDynaMedisanim- portantpartofthatgrowth.Beingpartofagrowingcompanygivesusan opportunitytobenefitfromEBSCO’stechnologicalstrengths.EBSCO- host is the proprietarysearch system at EBSCO Publishing and one of thebestknowresearchinterfacesintheworld.DynaMedrecentlyintro- duced a new Web interface powered by EBSCOhost and we are just startingtoseehowthatwillimpactDynaMed’sabilitytogrow.Thisisa very exciting time. NowthatDynaMedispartofEBSCOPublishing,therearebenefits forlibrariesaswell.BecauseEBSCOPublishinghassuchstrongmedical 130 ELECTRONIC RESOURCES IN MEDICAL LIBRARIES databases,librariansareabletopackageproductstogether,whichsaves themmoneyandgivesthemastrongsetofresearchofferingsincluding CINAHL® Plus with Full Text, SPORTDiscus™ with Full Text, and MEDLINE®with Full Text. DynaMedstartedasagrassrootsefforttoprovidehealthcareprofes- sionals with the informationthey need to make clinicaldecisions. The companyhasgrownslowlybutwithEBSCOPublishingbehindDynaMed, we are able to get the word out to a larger audience. We believe DynaMedisagreatresourceforlibrariansandhospitalslookingforthe best content and the best design at the best price. Question:WhatdistinguishesDynaMedfromMDConsult,UpToDate, AccessMedicine,or other point-of-care resources? Alper: DynaMed (see Figures 1 and 2) is able to distinguish itself in a numberofareas.Thedatabaseistheonlypoint-of-caretoolthatistruly FIGURE 1.DynaMedStart Page Used with permission. eProfiles 131 FIGURE 2.DynaMedAlphabetical Listing Used with permission. evidence-based and updated daily. The term “evidence-based” is used todescribemedically-relatedreferenceresources.Unfortunately,theterm is often used indiscriminately and without merit. DynaMed’s editorial processes rely on a strict set of evidence-based medicine standards. Evidence-based medicine is about integrating the best current evi- dencewithclinicalexpertiseandpatientvalues.Foraclinicalreference resourcetobecalledevidence-based,conclusionsmustbebasedonthe bestavailableevidence.Conclusionscanbebasedonthebestavailable evidenceonlyiftheevidenceisconsistentlyandsystematicallyidenti- fied, evaluated, and selected.DynaMedhas seven standards. 1. Systematic identification of evidence. 2. Systematicselectionofthebestavailableevidencefromthatiden- tified. 3. Systematicevaluationoftheselectedevidence(criticalappraisal). 4. Accurate summarization of the evidence and its quality. 132 ELECTRONIC RESOURCES IN MEDICAL LIBRARIES 5. Making conclusions dependent on the evidence. 6. Synthesizing multiple bits of evidence for overall conclusions. 7. Changing the conclusions when new evidence alters the best available evidence. Key evidence is given one of three “Level of Evidence” labels: Level1(likelyreliable)Evidence–representsthemostvalidreports addressing patient-oriented outcomes. Examples include randomized trialswithatleast80%follow-up,inceptioncohortstudiesforprognos- tic information, and systematic reviews of level 1 evidence reports. Theseexamplesareonlypresentedasbriefexamples.Achievingalevel 1evidencelabelmeansthatspecificqualitycriteriaweremetbasedon the study type. Level2(mid-level)Evidence–representsreportsaddressingpa- tient-orientedoutcomes,andusingsomemethodofscientificinvestiga- tion, yet not meeting the quality criteria to achieve level 1 evidence labeling. Examples include randomized trials with less than 80% fol- low-up, non-randomized comparison studies, and diagnostic studies withoutadequatereferencestandards.Level2evidencedoesnotimply reliable evidence. Level3(lackingdirect)Evidence–representsreportsthatarenot basedonscientificanalysisofpatient-orientedoutcomes.Examplesin- cludecaseseries,casereports,expertopinion,andconclusionsextrapo- lated indirectly from scientific studies. RecommendationsarelabeledbasedontheStrengthOfRecommen- dation Taxonomy (SORT) as one of the following: • Grade A recommendation(consistent high-quality evidence) • Grade B recommendation(inconsistent or limited evidence) • Grade C recommendation(lacking direct evidence) Most point-of-care resources use variations of the traditional textbook approach in which authors who are selectedfor theircontentexpertise writechaptersorpages.Theremaybeadditionallayersofpeerreview andeditingbutthisusuallyfollowsthesameexpertauthormodel.Ex- pertauthorsarenottypicallyexpertsinfinding,selecting,orinterpret- ing the best available evidence. This content does not reach the same standards for evidence-based medicine that we impose onDynaMed. DynaMedhasbeenproventoanswermostclinicalquestionsduringa randomized trial. An article published in Annals of Family Medicine1 eProfiles 133 establishedDynaMedastheonlyevidence-basedreferencethatanswers mostclinicalquestionsinprimarycare.DynaMedwasshowntoanswer approximately70%ofunselectedclinicalquestionsinactualuseduring practice.DynaMedwascomparedtofourotherclinicalreferencetools, and DynaMed was shown to provide answers with a level of evidence that meets or exceeds the other tools combined, 87% of the time. DynaMedistheonlypoint-of-careclinicalreferencetooltobeupdated daily. For the first time, information that needs to be accessed quickly canbethebest,mostaccurateinformationavailable.TheseRecently Updatedsummaries have their own link on theDynaMedinterface. EBSCOPublishinghasstronginterfacesandusersknowwhattoex- pect. A lot of attention was paid to the new interface features so that browsingisintuitive.Thereisastructureforfindinginformationthatis consistent from summary to summary (see Figure 3). Each disease or FIGURE 3.DynaMedCarpal Tunnel Syndrome (CTS) Summary Used with permission. 134 ELECTRONIC RESOURCES IN MEDICAL LIBRARIES condition summary includes:General Information (including ICD codes),Causes&RiskFactors,Complications&AssociatedCondi- tions, History, Physical, Diagnosis, Prognosis, Treatment, Preven- tion & Screening, References including Reviews and Guidelines, Patient Information, and Acknowledgements. The summaries are easytosearchmakingtheinformationeasytofindandquestionseasyto answer. Within the standardized template, with up to 65 subheadings, thecontentiswritteninoutlineformwithconclusionsfirstsoreadersdo not need to skim the text just to find out if they need to read it. AtDynaMed,theprocessisdesignedtobetransparent.TheWebsite listswhichjournalsarereadcovertocover.Usersareinvitedtocontrib- utetothedatabase,andfeedbackisavitalpartoftheDynaMedsystem. Wesolicittheopinionsofourusers,andthoseopinionshavebecomean essential part of updates and enhancements. For instance, suggestions from users have led us to add specific conditions and to make certain improvementssuchastheICD-9CodesandPatientInformationHand- outs. As we continue to build DynaMed, those comments serve as an importantfoundationandhelpusdeterminewhereweneedtogonext. ICD-9codesareusedintheUnitedStatesformedicalbillingandare also necessary to justify lab tests and x-rays, so many clinicians need thesecodesintheirdailyworkflow,andwerespondedbyaddingthem throughoutDynaMed.Withgrowthacrosstheworld,wearenowadd- ing ICD-10 codes as these are used in many other countries. DynaMed also provides citations with links including free full-text links integrated into the content as well as links to patient handouts. NowthatweareapartofEBSCOPublishing,DynaMedwillbeable totakeadvantageoftheEBSCOsuiteoffull-textmedicaldatabasesas DynaMedbecomes fully powered by EBSCOhost. Question:DynaMedcoversmorethan500medicaljournalsfromAIDS Research and Therapy to Immunity and Ageing to World Journal of SurgicalOncology,secondarysourcessuchasreviewarticlesandjour- nal review services, practice guidelines and systematic reviews, and druginformation.Pleasedescribethedistinctionsthatyoumakeamong these different sources. Alper: These sources are part of DynaMed’s systematic literature sur- veillance. The list of content sources that are read cover-to-cover is listed on the DynaMed Web site <http://www.dynamicmedical.com>. Morethan500journals,eitherdirectlyorindirectlythroughmanyjour- nalreviewservices,areevaluatedsystematicallyby DynaMed editors. eProfiles 135 Allentries,includingletterstotheeditorandfeaturesthatcoverarticles from other publications, are evaluated first for clinical relevance. All clinicallyrelevantitemsarethenevaluatedforvalidityrelativetoexist- ing DynaMed content. The best available evidence then gets selected systematically for inclusion. (This overcomes biases that occur when expert authors selectively cite articles without systematic searches or protocol-drivenselectionprocess.)InformationincludedinDynaMedis thencriticallyappraisedandtheeasy-to-interpretlevelsofevidencela- bels are applied. As to the distinction between sources, Primary Medical Literature listsjournalswhichpublishoriginalresearchandforwhichwedirectly monitor the entire journal cover-to-cover with each issue. Secondary MedicalLiteratureisusedforbothjournalsthatserveassourcesforre- view articles(but not originalresearch) and journals or journalreview services that selectively report on important research published else- where.Weusethesesourcestowidenournetforcapturingthemostrel- evant and valid original research articles. In addition, there are many resourcesthatarenotclassifiedasprimaryorsecondarymedicallitera- ture but are valuable for keeping a clinical reference resource current. Theseincludesources of systematicreviews (such as AHRQ Evidence Reports and Cochrane Database of Systematic Reviews), the National Guideline Clearinghouse, and drug information sources. Question:Please describe your typicalDynaMedcustomer. Alper:Thereisno“typical”user.DynaMedusersincludeprimarycare physicians, physician specialists, physician assistants, nurse practitio- ners,nurses,pharmacists,physicaltherapists,librarians,educators,and veterinarians.Customersincludeindividuals,communityhealthcenters, physiciangroups,residencytrainingprograms,medicalschools,hospi- tals,universities,andstatewidenetworks.Ourcustomerbaseisworldwide. Anyonewhoneedstofindthemostaccurateinformationquicklyand easilyisgoingtobeasuccessfulDynaMeduser.Thedatabasewasini- tiallydesignedtosupportageneralphysicianwithlimitedspecialist back-up. But even more specializedinformationcan be added without interferingwithnavigation,sowehavedetailedsummariesontopics like Cyclosporine for rheumatoid arthritis and for Posner-Schlossman syndrome (a rare ophthalmologic condition). Electronic technology allowsuserstochoosethedepthsocontentofspecialinterestcanbeavail- abletothosewhoneeditwithoutbeinginthewayforthosewhodonot. 136 ELECTRONIC RESOURCES IN MEDICAL LIBRARIES Question:DoyouhaveanydramaticexamplesofDynaMedinuseina patient care setting? Alper:ApatientaskedherdoctorifsheshouldtakeAvlimilforlowli- bido.ThedoctorhadnotheardofAvlimilsohesearchedDynaMedand found the summary on hypoactive sexual desire disorder which re- ported that Avlimil was promoted on the Web and through television commercials, but the only efficacy data was based on a single unpub- lishedtrialof65womenwhowere“healthyexperimenters.”Inaddition to answering the specific question, the doctor also found evidence re- garding buproprion and testosterone so learned about other treatments for this patient. This example shows the benefit of synthesizing infor- mation from multiple sources. Inarecentcase,adoctorhadapatientwhohadabreastbiopsy(be- nign)inthepastandwhose motherhadbreastcancer.Thedoctoresti- matedthepatient’sfive-yearriskat3.1%,whichhebelievedwashigh enough to considermedication.At thetime,thedoctorhad access to a JAMA early release article posted June 5, 20062 which stated that raloxifene60mgoncedailywasaseffectiveastamoxifen20mgonce daily in reducing risk for up to five years. The JAMA study compared thetwomedicationsandshowedtheireffectivenesstobesimilarbutdid notdiscusswhethertherisks/sideeffectsoutweighedthebenefitsofthe medication in the first place. Forthebackground,thedoctorturnedtotheDynaMedsummaryfor Chemoprevention of breast cancer. DynaMed provided the numbers neededtotreatorharm(NNT/NNH)thatthedoctorusedwhendiscuss- ingwhetherthepatientshould go on eithermedication.TheDynaMed EvidenceOverviewlistedincreasedrisksforotherconditionsaswellas possible side effects for each medication. The doctor was able to cite theserisksandsideeffectsduringhisdiscussionwiththepatient,which helped her have more information on which to base her decision whether or not to go on a preventative regimen. Asanotherexample,Figures4and5showatreatmentoverviewand treatment detail for Carpal Tunnel Syndrome. Given the importance that DynaMed editors place on providing the bestinformationasquicklyaspossible,theinformationfromtheorigi- nal JAMA study was incorporated into the DynaMed summary within three days of its early online release and reviewed by up to 28 editors. Question:Pleasedescribeindividual,group,andsitelicensingoptions forDynaMedusers.