TheEmerging Impact of the InformationAge on OrthopaedicSurgery*. The Value and Promise of PatientDatabasesin Orthopaedic Surgery
BY JOHN J. HARRAST; ROBERT POSS,M.D.
American OrthopaedicAssociation
*Presented at the Annual Meeting of theAmerican Orthopaedic Association, Sun Valley, Idaho, June 7,1999.Address for J. J. Harrast and R. Poss: Department of OrthopaedicSurgery, Brigham and Women's Hospital, 75 Francis Street, Boston, Massachusetts02115-6110. E-mail address for J. J. Harrast: jjharrast@msn.com. E-mail addressfor R. Poss: rposs@partners.org.
Inthe early 1980s,orthopaedic surgeons made the transitionfrom reviewing patientcohorts on paper to storingpatient data in computerized databases.These databases haveproven to be useful tools for documentingpatientcare, performing clinical research, and fostering continuingeducation.Patient databases will play an even more importantrole in the future as orthopaedic surgeons andsoftwaredevelopers further exploit the power of theInternet. In this paper,we review our experiencewith use of a patient database at asingle institution, the strides being taken to createaregistry that can be used by manycenters, and the potential toaccelerate these effortsthrough the widespread use of theInternet. Patientdatabases store information collected from instrumentsdesigned tomeasure components of patient care. We have useda system model (Fig. 1)to considerthese individual components. In the model,patient care comprisesthe continuous and recurring cycleof evaluation, decision, andintervention, and the interactionsamong these three elements. Eachof these componentscan be measured independently. Validity,reliability, and sensitivityare measures of evaluation.Appropriateness is a measureof the decision, and outcome is ameasureof the intervention. Satisfaction is a measure ofpatient care as a whole. Todemonstrate the use of one of these measureswith apatient database, consider the example ofa study that we conductedto determine theappropriateness of total hip replacement. Wemeasured thechange in the physical scales of the ShortForm-36 (SF-36) from a point immediately before hipreplacement surgery to two years after surgery inaconsecutive group of 200 patients(1).Thepatient population was divided into three groupshigh, middle, and low on the basis of preoperative function (Fig. 2).The patientswho were in the low andmiddle-function groups preoperativelydemonstrated dramatic improvements in functionafter total hipreplacement; however, patients with relativelyhigh preoperativefunction did not improve substantially. Thefrequency of patientswho reported dissatisfaction was alsohigher in this group. Theseresults suggest that patients who are relativelyhigh-functioningrequire more intensive preoperative educationalefforts to aligntheir expectations with those of the surgeon.Thesepatients in particular must understand the limits aswell as the promise of hip replacement surgerybefore decidingwhether to undergo it. This exampledemonstrates how measurement offunction and satisfaction aspart of a patient database providesinsight thatcan lead to improved patient care.
TheTotal JointReplacement Registry at Brigham and Women'sHospital
TheTotal Joint Replacement Registry at Brigham and Women'sHospital in Boston has amassed a database withinformation onmore than 15,000 patients and morethan 20,000 joint replacementprocedures performed in thelast three decades. This databasecomprises patient demographics,histories, findings on physicalexamination, intraoperative findings, informationon complications,patient outcomes, and patient-satisfaction data thatwere collectedprospectively with informed patient consent. Theannual operatingbudget of the Registry is currently$350,000, with external grantsaccounting for $150,000 ofthe total. TheRegistry has been a substantial resource for attendingsurgeons, residents, and fellows. Since 1990, the Registryhasaccommodated more than 1500 data requests forclinical researchprojects. These projects have generated 110publications as well as140 presentations at theannual meeting of the American Academy ofOrthopaedicSurgeons. The Registry database also has been avaluable resource for collaboration with visiting scientists anda resource for collaborative studies with other departmentsatour institution, such as Rheumatology, the MultipurposeArthritisCenter, and Internal Medicine(2-6,8).Wecould not have conducted these studies withoutthis database ofprospectively collected information. In recentyears, the databasehas made it possible forthe Department of Orthopaedic Surgery to bealeader at Brigham and Women's Hospital in implementingnew patient-education programs, clinical pathways, and cost-efficiencyprograms.
LessonsLearned
Throughthe operation of the Registry, we have learnedmany lessons regarding the successful implementation of apatient database. Most importantly, the reason or reasonsforconstructing a patient database must be clearand specific;otherwise, a great deal of energyand money may be expendedcollecting data thatwill never be used. Is the reason fordeveloping a patient database to perform retrospective and/orprospective studies, to assist in influencing patient care,orto document patient outcomes? The more reasonsthat there are, themore complex, time-consuming, expensive,and, in the end, difficultthe task. Thereare practical limits to the quantity and frequencyof data collection. Our experience suggests that patientencounters for obtaining outcomes information should be limitedto ten minutes. The number of encounters duringwhich thepatient is asked to provide informationalso must be limited. Wehave found thatpatient cooperation declines when the assessment ofoutcomesfor the same condition occurs more than onceayear. The number of patients for whomsubstantial data are collectedmust be targeted tothe research question being asked. It isimpracticalto collect large amounts of data from allpatients visiting a busy orthopaedic clinic.
AMulticenterDatabase: The Hip Society Registry
Prospectivelycollected data from multiple centers offer thepromiseof providing sufficiently large cohorts to serve asan early-warning system for implant failures and toformulateresearch questions far earlier than would bepossible at anyindividual institution. The Hip Societyhas initiated a pilotproject, at ten participatingcenters, to begin to construct amulticenter clinicaland radiographic database. The initial projectwould requireonly a limited data set. Thehypothesis of the first study to be conductedis thatthe impact of dislocation following totalhip arthroplasty on thepatient's perception of outcomeis greater than surgeons haveappreciated. With informedpatient consent, participating centerswill collect data preoperatively,at surgery, and at follow-up, andthey willsend digitized radiographs to the Hip SocietyRegistryto be electronically integrated with the clinicaldata.The study will enroll all patients undergoing primarytotal hip arthroplasty during a one-year period atthe tenparticipating centers. We have projected thatthis cohort will besufficient to answer theresearch questions posed. Simultaneously,the Hip Society Registry will perform a pilotaudit that records a limited data set todocument the date ofimplantation and the typesof components that are being removedduring revisiontotal hip replacement(8)at thesefew centers over a period ofone year.
FutureApplications
TheInternet affords the opportunity to outsource themanagementof patient databases by using the World WideWeb and ubiquitous Web-browser software. Off-site databases aregrowing rapidly in all fields. Now, a data-managementcompanycan maintain the hardware, software, and dataat its externallocation, while a personal computerand browser are all that theorthopaedic surgeonneeds in order to enter and access hisor her data. In1999, we decided to transfer the management oftheBrigham Total Joint Replacement Registry to anoff-sitedata-management company. Reciprocal data transfers will beconductedthrough the World Wide Web. We realizedimmediate cost-savings byrelinquishing expensive rental space andreducing the number offull-time employees from 7.5to three. We budgeted the total costforthe year 2000 to be the same asthat for 1998, but weanticipate that, insucceeding years, the cost of maintaining theRegistrywill be less because of outsourcing. TheWeb provides a unique opportunity for multicenter registrydatabases to flourish. The Hip Society Registry willbeavailable to participants online so that real-timeinteractions canoccur; these interactions would have beenvirtually impossiblewithout the Web. The Web alsoallows orthopaedic departments andpractices to participate moreeasily in multicenter "bestpractice"groups to study the effectiveness ofprograms. Inaddition to single-center and multicenter registries andstudies,the Web is transforming the training and certificationprocesses in medicine through its influence on patientdatabases. For example, one requirement of the AmericanBoardof Orthopaedic Surgery's examination process is thatcandidatessubmit to the Board a six-month caselist of operative procedures.In previous years, thistask has proven arduous for candidates usingpenand paper or a disk-based system. This year,candidates for the part-II certification examination were ableto electronically record their cases in an ongoingmanner andthen to submit them to theBoard through the Web. The resultingpatient database,consisting of more than 92,000 cases submittedbymore than 700 candidates, provides a snapshot oforthopaedic surgery practice in the United States duringthissix-month window of time. Finally,patient databases and the World Wide Web nowoffer residency programs the possibility of automatically generatingthe information needed for the Residency Review Committee'sProgram Information Form. Residents tracking their cases withuse of online databases generate the statistics requiredforthe report, and program directors are ableto follow, in real time,both the experiencethat their programs are providing to residentsandthe performance of residents in their programs. TheWorld Wide Web is enabling patient databases toreachtheir full potential to provide timely andaccurate information toorthopaedic surgeons. Its impact willcontinue to grow for years tocome. NOTE:We are indebted to Maurice E.Müller,M.D., for his counsel and collaboration in thedevelopment and implementation of many of the conceptsexpressed in this paper. We also gratefully acknowledgetheMaurice E. MüllerFoundations of Switzerland and North Americafor theirsupport of the Brigham Total Joint ReplacementRegistry;the contributions of Sonya van der Meer, JohnKwon, and Tad Bergman to the success ofthe Brigham TotalJoint Replacement Registry; and thecontributions of Beth Imrem ofData Harbor, Incorporated,to the Hip Society Registry. Finally, weacknowledgeand thank the Council on Research of theAmerican Academy of Orthopaedic Surgeons, the MusculoskeletalEvaluationand Research Institute, and the Orthopaedic ResearchandEducation Foundation for their support of the HipSociety Registry.
