An excellent argument for the collection of health-status outcomes
data in orthopaedics was recently published3.
Health-status outcomes data are defined as data about the state
of health, the primary disease, and comorbidities before and after
an appropriate interval following the application of a selected
treatment strategy. It is widely agreed that collection of these
data is important; however, they are collected rarely, and for practical
purposes they are never collected in a sustained, generalized manner.
Most policy analyses requiring such data use surrogate data, as
they are more easily, quickly, or cheaply acquired than are actual
outcomes data. For example, length of stay is used as a surrogate
for total resource expenditure, and satisfaction with recent hospitalization
is used as a surrogate for actual improved health status. It is
helpful to use surrogate data if they have been found to correlate
with the actual outcome of interest, but that is rarely the case because
that conclusion must have been reached by the study of the actual
outcomes data at some time in the past.
The process of collecting outcomes data involves several critical
concepts. First, in order to provide the necessary accuracy and
precision, a very large amount of data is required. Second, the
volume of required data is so great that it must be managed in an
electronic system. Third, because of time constraints in the practice
setting, the need to gather any information over and above standard
clinic processes is viewed as an imposition. Consequently, it is
extremely rare that such data, no matter how minimal, are collected
consistently over a long time-period. Fourth, data must be collected
as an integral part of standard clinic processes in the routine
care of patients, in a manner that does not increase the burden
of care. Fifth, much of these data must be in the form of patient-administered
health-status measures. Sixth, it is extremely important that these structured
data can be stored in a relational database that can be used for
business as well as clinical purposes and can be merged with a narrative
word template to produce clinical notes. With use of such a relational
database, the saving of resources can be great.
Collection and use of outcomes data as a routine part of the
clinical care of patients has rarely been done and has not been
described in the literature. Collection of data both within and
outside of the health-care field has been studied. Paper and pencil
and computer mouse and computer touch-screen technologies have been
compared2. Both types of computer
technology are cheaper to use than paper and pencil if the volume
of data is more than minimal, and users prefer touch-screen technology
to the mouse. Most computer monitors in business settings where
users either are not familiar with computer technology or work in
a rushed environment (such as in a bank or a restaurant) utilize touch-screen
technology.
The implementation of such a program requires a radical change
in the patient-encounter process in clinics. The implementation requires
three phases. In the first phase, one must study and document the
present process with an analysis of patient flow and a survey of all
stakeholders regarding satisfaction and judgment of quality of care.
This information is then used in the second phase, the design of the
new process, and also as a baseline when the new process is evaluated,
which is the third phase.
The existing clinic process was studied by observing patient
and information flow. The patient checks in at the main reception
desk, is sent to the adjacent registration desk if necessary, and
is then sent to the waiting room until he or she is called into
one of the clinics. The next stop is the specific clinic reception
desk; the patient then proceeds to the sub-waiting room or directly
to the examining room. The patient waits in the examining room until
he or she is seen by the nurse, medical student, resident, or faculty
physician. Usually the resident is the first to see the patient;
the resident then returns to see the patient again, with the faculty
physician. The patient typically goes to the adjacent radiology
suite and returns with radiographs at some point during the encounter. Most
commonly, the patient goes home after seeing the faculty physician.
Occasionally, the patient goes to the adjacent physical therapy department,
the occupational therapy department, or the prosthetist. Alternatively,
these providers may see the patient in the examining room. The patient
may go elsewhere in the facility for other diagnostic tests or consultations.
A large amount of paperwork must then occur.
Time data: In order to understand the patient flow in more detail,
we had three student workers collect time 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, in the hallway, to monitor the examination rooms. We recorded
the times when the patient entered and left the waiting room, entered
and left the sub-waiting room, and entered and left the examination
room as well as the times that the nurse, the medical student, the
resident, and the faculty physician each entered and left the examination
room. All parties may have entered and left the examination room
multiple times. We also recorded where the patient was going upon leaving
the examination room. Data were collected with paper and pencil,
were stored, and were analyzed with a computer-software program
(Access 97; Microsoft, Redmond, Washington).
The clinical content was developed in a modular fashion. At present,
we have ten clinical tracks matching the clinical subspecialties, based
on anatomical area, problem type, 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
of initial questions was developed to algorithmically lead the patient
to the correct questionnaires. Each track comprises five types of
visits: new, short-return, definitive-return, return by an established
patient with a new problem, and remote follow-up forms.
Each track contains a generic package of forms that are the same
for all adults. Children or their parents will use similar but age-appropriate
forms. This generic package consists of the Short Form-36 (SF-36)
health-status questionnaire4 as
well as questionnaires regarding expectations of treatment, comorbidities
(both systemic and musculoskeletal), family history, and social
history. We also have a series of demographic fields, which are
filled out at each encounter from the institutional registration. All
of these forms are either identical to or compatible with the Musculoskeletal
Outcomes Data Evaluation and Management System (MODEMS) forms of
the American Academy of Orthopaedic Surgeons (AAOS)1.
Each track contains a group of forms that are functionally similar
from track to track but are tailored to the specific track. These
consist of a disease/anatomical area-specific health-status form
that is also identical to or compatible with the AAOS MODEMS forms.
A treatment-history form and a series of questions about the patient's
specific problem, such as when and how it started, what makes it
worse or better, and whether it is getting worse or better, make
up the remainder of the patient-administered portion of the forms.
The mean time required for 150 new patients to complete these questionnaires
was thirty minutes (range, fifteen to sixty minutes).
Using simulation methods, we analyzed these times and the waiting
times mentioned above and determined that, with our current scheduling
patterns, only about 40 percent of our new patients would have completed
the forms before being seen by the doctor. We determined that we
must bring new patients in at least twenty-three minutes earlier
so that 95 percent will be able to complete the forms. Since it
is vital that the forms be completed and the results be made available
before the patient is seen by the doctor, we are bringing in our
new patients thirty minutes earlier than we have in the past.
It is necessary, for practical purposes, to select a minimal
interval between the dates for collection of health-status outcomes
data. These intervals will vary greatly according to the clinical
problem and the rapidity of change in the health status. Currently,
we think that six months after baseline is soon enough to collect such
data for total joint and spine problems. Carpal tunnel syndrome
may warrant collection of such data at one month. Since patients do
not return like clockwork, and because we do not want to exclude
patients with follow-up data, these times become windows, with no gaps
in between. Thus, for a total joint or spine problem, six months
can become four to nine months; one year, ten to eighteen months;
and two years, nineteen to thirty-six months. (The interval until
the next follow-up would begin with the most recent visit.) Therefore,
at intervals of less than four months, we are not interested in
measuring health status but simply in recording what has happened
with the problem since the patient's last visit. Simple outcomes
measures, such as the date that the patient returned to work, are
recorded. This is what we refer to as the short-return visit. It takes
the patient about one to two minutes to complete this form.
We have called the visit at six months or later the definitive-return
visit. At this visit, we seek to compare the health status of the
patient with his or her baseline and with some norm or norms. Therefore,
we repeat the health-status measures that we used for the new-patient
visit, but we do not again collect the additional material about
general health that was used for risk adjustment unless the patient
indicates that there has been a change. If there has been no change,
the data collected at baseline can be used for adjustment at six
months. It takes the patient about five to twenty minutes (mean,
twelve minutes) to complete these forms.
If it is impossible or impractical for the patient to return
to our clinic, these forms may be completed from the patient's home,
either electronically on our Web site or with paper and pencil.
The data then can be keyed in by one of our office personnel.
The time-consuming data acquisition is done, as it is for new
patients, in waiting-room kiosks or clinical examination rooms,
with a touch-screen computer terminal. An interim report is then
produced for the physician to use, as he or she deems appropriate,
in reviewing the clinical history with the patient.
Considerable savings in overhead expense might be contemplated
for the after-visit documentation. However, additional assistance will
be required to help the patient in the data-acquisition stage. We
plan to have an additional person in the kiosk area and another
in each clinic. Since the extra help is needed from the beginning
of the project and the savings do not come until later, there is
an initial increase in expenses.
The resident and/or faculty physician then examines the patient
and records the data in the computer with use of the touch screen.
The diagnosis and management plan also are recorded. This process
requires between one and eight minutes, depending on the complexity
of the examination and the specific problem.
The software associates each answer from both the patient and
the clinician with certain preselected words, and the system produces
a narrative clinic note, which is very similar in style to a classic
dictated and transcribed note. The clinician can edit this note
as needed and can sign it electronically; it then can be sent to
referring physicians and others electronically or on paper, or both.
This step can greatly reduce the time and expense needed for documentation of
the encounter.
Each diagnosis in the system is associated with an ICD-9 (International
Classification of Diseases, Ninth Revision) code. Each question
in the history and physical forms is associated with a body system
and a component of the "History and Physical." The software can count
the answers and their location, and, utilizing Health Care Financing
Administration rules, it can calculate the Evaluation and Management
CPT-4 (Physicians Current Procedural Terminology, Fourth Revision)
code after the physician answers five additional questions about
complexity. From this information, an automated bill can be generated,
greatly reducing the time and expense of producing billing and insurance
forms. To the extent that interpretation of the Health Care Financing Administration
rules is correct, one can be assured that the billing level and
the documentation are in agreement.
Finally, and most importantly, a database with detailed information
on each patient's health status, risk-adjustment data, diagnosis,
detailed treatment strategy, and other outcome measures over time
is available for analysis. If we can collect these data on a large
enough scale, we can use real rather than surrogate data for decision-making.
1. A very large amount of data is required to provide the necessary
accuracy and precision.
2. The volume of data is so great that it must be managed in
an electronic system.
3. In the practice setting, because of time constraints, the
need to gather any information over and above standard clinic processes
is viewed as an imposition. Consequently, it is extremely rare that
such data, no matter how minimal, are consistently collected over
a long time-period.
4. If data are to be collected consistently from all patients
over an extended period of time, they must be collected as an integral
part of standard clinic processes in the routine care of patients
in a manner that does not increase the burden of care.
5. Much of these data must be in the form of patient-administered
health-status measures.
6. These structured data can be stored in a relational database,
used for business as well as clinical purposes, and merged with
a narrative word template to produce clinical notes. The saving
of resources can be great.
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.