TheEmerging Impact of the Information
Age on OrthopaedicSurgery*. Implementation of a Computer-Based
Patient Record andan Outcomes Data-Collection System at the
Department ofOrthopaedic Surgery, University of
Iowa
BY RICHARD C.
JOHNSTON, M.D.,M.S.
American Orthopaedic Association
*Presented at the Annual
Meeting of the American Orthopaedic Association, Sun Valley, Idaho, June 7,
1999.Address for R. C. Johnston: Department of Orthopaedic Surgery,
University of Iowa Hospitals and Clinics, 200 Hawkins Drive, JPP 01016, Iowa
City, Iowa 52242. E-mail address for R. C. Johnston:
richard-johnston@uiowa.edu.
Thegoal of this
presentation is to report ourexperiences to date with
implementation of a computer-basedpatient record along with a
system for thecollection of patient health-status outcomes data as
anintegral part of routine patient care in theorthopaedic department. We are "goinglive"in our first clinic as this is written.Therefore, this is a
report of all activityleading up to this point in the project.
Background
Anexcellent argument for the collection of health-status
outcomesdata in orthopaedics was recently published(3).Health-status outcomes data are defined as data aboutthe
state of health, the primary disease, andcomorbidities before and
after an appropriate interval followingthe application of a
selected treatment strategy. Itis widely agreed that collection of
these datais important; however, they are collected rarely,
andfor practical purposes they are never collected ina
sustained, generalized manner. Most policy analyses requiringsuch
data use surrogate data, as they aremore easily, quickly, or
cheaply acquired than areactual outcomes data. For example, length
of stayis used as a surrogate for total resourceexpenditure, and satisfaction with recent hospitalization is usedas a surrogate for actual improved health status.It is
helpful to use surrogate data ifthey have been found to correlate
with theactual outcome of interest, but that is rarelythe case because that conclusion must have beenreached by the
study of the actual outcomesdata at some time in the past.
Theprocess of collecting outcomes data involves several
criticalconcepts. First, in order to provide the necessaryaccuracy and precision, a very large amount ofdata is
required. Second, the volume of requireddata is so great that it
must bemanaged in an electronic system. Third, because oftime constraints in the practice setting, the needto gather
any information over and above standardclinic processes is viewed
as an imposition. Consequently,it is extremely rare that such data,
nomatter how minimal, are collected consistently over along time-period. Fourth, data must be collected asan
integral part of standard clinic processes inthe routine care of
patients, in a mannerthat does not increase the burden of
care.Fifth, much of these data must be inthe form of
patient-administered health-status measures. Sixth, itis extremely
important that these structured data canbe stored in a relational
database that canbe used for business as well as clinicalpurposes and can be merged with a narrativeword template to
produce clinical notes. With useof such a relational database, the
saving ofresources can be great.
Collectionand use of outcomes data as a routinepart of
the clinical care of patients hasrarely been done and has not been
describedin the literature. Collection of data both withinand outside of the health-care field has beenstudied. Paper
and pencil and computer mouse andcomputer touch-screen technologies
have been compared(2).Both
types of computer technology are cheaper touse than paper and
pencil if the volumeof data is more than minimal, and usersprefer touch-screen technology to the mouse. Most computermonitors in business settings where users either arenot
familiar with computer technology or work ina rushed environment
(such as in a bankor a restaurant) utilize touch-screen
technology.
Theimplementation of such a program requires a radicalchange in the patient-encounter process in clinics. Theimplementation requires three phases. In the first phase,one
must study and document the present processwith an analysis of
patient flow and asurvey of all stakeholders regarding satisfaction
and judgmentof quality of care. This information is thenused in the second phase, the design ofthe new process, and
also as a baselinewhen the new process is evaluated, which isthe third phase.
Studyof the
Existing Process of Patient Encounters inthe Orthopaedic
Clinic
Theexisting clinic process was studied by observing patientand information flow. The patient checks in atthe main
reception desk, is sent to theadjacent registration desk if
necessary, and is thensent to the waiting room until he orshe is called into one of the clinics.The next stop is the
specific clinic receptiondesk; the patient then proceeds to the
sub-waitingroom or directly to the examining room. Thepatient waits in the examining room until heor she is seen by
the nurse, medicalstudent, resident, or faculty physician. Usually
the residentis the first to see the patient; theresident then returns to see the patient again,with the
faculty physician. The patient typically goesto the adjacent
radiology suite and returns withradiographs at some point during
the encounter. Mostcommonly, the patient goes home after seeing
thefaculty physician. Occasionally, the patient goes to theadjacent physical therapy department, the occupational therapy
department,or the prosthetist. Alternatively, these providers may
seethe patient in the examining room. The patientmay go
elsewhere in the facility for otherdiagnostic tests or
consultations. A large amount ofpaperwork must then occur. Time data:In order to understand the patient flow inmore detail, we had three student workers collecttime data in
the clinics (Table I).One was
stationed at the waiting-room door; one,at the reception
desk/sub-waiting-room area; and one, inthe hallway, to monitor the
examination rooms. Werecorded the times when the patient entered
andleft the waiting room, entered and left thesub-waiting room, and entered and left the examinationroom as
well as the times that thenurse, the medical student, the resident,
and thefaculty physician each entered and left the examinationroom. All parties may have entered and leftthe examination
room multiple times. We also recordedwhere the patient was going
upon leaving theexamination room. Data were collected with paper
andpencil, were stored, and were analyzed with acomputer-software program (Access 97; Microsoft, Redmond,
Washington).
Designof the New
Process for Patient Encounters inthe Clinic
Weenvision that the patient will use the timein the
waiting and examination rooms prior toseeing the doctor to
complete, utilizing computer touch-screentechnology, a series of
questionnaires regarding health status.Thus, an interim report will
be available tothe physician prior to seeing the patient. Thephysician can use these data as the coreof the clinical
history. The physician will examinethe patient and enter the data
into thecomputer, utilizing the same touch-screen technology. These
datathen can be merged in a narrative fashionto produce
a clinic note. This note canbe edited minimally in typical cases
and extensivelyin unusual cases, and reports can be sentto appropriate parties and places, such as thereferring
doctor and the medical record.
ClinicalContent
Theclinical content was developed in a modular fashion.At present, we have ten clinical tracks matchingthe clinical
subspecialties, based on anatomical area, problemtype, and patient
age. These tracks are spine,hip/knee reconstruction, sports,
foot/ankle, shoulder, wrist/hand, pediatric scoliosis,pediatric
other, tumor, and trauma. A series ofinitial questions was
developed to algorithmically lead thepatient to the correct
questionnaires. Each track comprisesfive types of visits: new,
short-return, definitive-return, returnby an established patient
with a new problem,and remote follow-up forms.
Eachtrack contains a generic package of forms thatare
the same for all adults. Children ortheir parents will use similar
but age-appropriate forms.This generic package consists of the
Short Form-36(SF-36) health-status questionnaire(4)as well as
questionnaires regarding expectations of treatment,comorbidities
(both systemic and musculoskeletal), family history, andsocial
history. We also have a series ofdemographic fields, which are
filled out at eachencounter from the institutional registration.
All of theseforms are either identical to or compatible withthe Musculoskeletal Outcomes Data Evaluation and Management System(MODEMS) forms of the American Academy of OrthopaedicSurgeons
(AAOS)(1).
Eachtrack contains a group of forms that arefunctionally similar from track to track but aretailored to
the specific track. These consist ofa disease/anatomical
area-specific health-status form that is alsoidentical to or
compatible with the AAOS MODEMSforms. A treatment-history form and
a series ofquestions about the patient's specific problem, such
aswhen and how it started, what makes itworse or
better, and whether it is gettingworse or better, make up the
remainder ofthe patient-administered portion of the forms. The
meantime required for 150 new patients to completethese
questionnaires was thirty minutes (range, fifteen tosixty
minutes).
Usingsimulation methods, we analyzed these times and thewaiting times mentioned above and determined that, withour
current scheduling patterns, only about 40 percentof our new
patients would have completed theforms before being seen by the
doctor. Wedetermined that we must bring new patients inat least twenty-three minutes earlier so that 95percent will
be able to complete the forms.Since it is vital that the forms
becompleted and the results be made available beforethe
patient is seen by the doctor, weare bringing in our new patients
thirty minutesearlier than we have in the past.
Itis necessary, for practical purposes, to select aminimal interval between the dates for collection ofhealth-status outcomes data. These intervals will vary greatlyaccording to the clinical problem and the rapidityof change
in the health status. Currently, wethink that six months after
baseline is soonenough to collect such data for total jointand spine problems. Carpal tunnel syndrome may warrantcollection of such data at one month. Sincepatients do not
return like clockwork, and becausewe do not want to exclude
patients withfollow-up data, these times become windows, with
nogaps in between. Thus, for a total jointor spine
problem, six months can become fourto nine months; one year, ten to
eighteenmonths; and two years, nineteen to thirty-six months.(The interval until the next follow-up would beginwith the
most recent visit.) Therefore, at intervalsof less than four
months, we are notinterested in measuring health status but simply
inrecording what has happened with the problem sincethe
patient's last visit. Simple outcomes measures, suchas the date
that the patient returned towork, are recorded. This is what we
referto as the short-return visit. It takes thepatient
about one to two minutes to completethis form.
Wehave called the visit at six months orlater the
definitive-return visit. At this visit, weseek to compare the
health status of thepatient with his or her baseline and withsome norm or norms. Therefore, we repeat thehealth-status
measures that we used for the new-patientvisit, but we do not again
collect theadditional material about general health that was
usedfor risk adjustment unless the patient indicates thatthere has been a change. If there hasbeen no change, the data
collected at baselinecan be used for adjustment at six months.It takes the patient about five to twentyminutes (mean,
twelve minutes) to complete these forms.
Ifit is impossible or impractical for the patientto
return to our clinic, these forms maybe completed from the
patient's home, either electronicallyon our Web site or with paper
andpencil. The data then can be keyed inby one of our
office personnel.
Thetime-consuming data acquisition is done, as it isfor
new patients, in waiting-room kiosks or clinicalexamination rooms,
with a touch-screen computer terminal. Aninterim report is then
produced for the physicianto use, as he or she deems
appropriate,in reviewing the clinical history with the
patient.
Considerablesavings in overhead expense might be contemplated
forthe after-visit documentation. However, additional assistance
will berequired to help the patient in the data-acquisitionstage. We plan to have an additional personin the kiosk area
and another in eachclinic. Since the extra help is needed fromthe beginning of the project and the savingsdo not come until
later, there is aninitial increase in expenses.
Theresident and/or faculty physician then examines the patientand records the data in the computer withuse of the touch
screen. The diagnosis andmanagement plan also are recorded. This
process requiresbetween one and eight minutes, depending on
thecomplexity of the examination and the specific problem.
Thesoftware associates each answer from both the patientand the clinician with certain preselected words, andthe
system produces a narrative clinic note, whichis very similar in
style to a classicdictated and transcribed note. The clinician can
editthis note as needed and can sign itelectronically;
it then can be sent to referringphysicians and others
electronically or on paper, orboth. This step can greatly reduce
the timeand expense needed for documentation of the encounter.
Eachdiagnosis in the system is associated with anICD-9
(International Classification of Diseases, Ninth Revision) code.Each question in the history and physical formsis associated
with a body system and acomponent of the "Historyand Physical."The software can count the answers and
theirlocation, and, utilizing Health Care Financing Administration
rules,it can calculate the Evaluation and Management CPT-4(Physicians Current Procedural Terminology, Fourth Revision) code
afterthe physician answers five additional questions about
complexity.From this information, an automated bill can begenerated, greatly reducing the time and expense ofproducing
billing and insurance forms. To the extentthat interpretation of
the Health Care Financing Administrationrules is correct, one can
be assured thatthe billing level and the documentation are inagreement.
Finally,and most importantly, a database with detailed
informationon each patient's health status, risk-adjustment data,
diagnosis,detailed treatment strategy, and other outcome measures
overtime is available for analysis. If we cancollect
these data on a large enough scale,we can use real rather than
surrogate datafor decision-making.
CriticalConcepts
in Collecting Outcomes Data
1.A very large amount of data is requiredto provide the
necessary accuracy and precision.
2.The volume of data is so great thatit must be managed
in an electronic system.
3.In the practice setting, because of time constraints,the need to gather any information over andabove standard
clinic processes is viewed as animposition. Consequently, it is
extremely rare that suchdata, no matter how minimal, are
consistently collectedover a long time-period.
4.If data are to be collected consistently fromall
patients over an extended period of time,they must be collected as
an integral partof standard clinic processes in the routine
careof patients in a manner that does notincrease the
burden of care.
5.Much of these data must be in theform of
patient-administered health-status measures.
6.These structured data can be stored in arelational
database, used for business as well asclinical purposes, and merged
with a narrative wordtemplate to produce clinical notes. The saving
ofresources can be great.
