Pitfalls of Extensive Documentation
and billing for all work legitimately performed is imperative.
3
Trainees are generally directed to document extensively.
4
Supervising physicians may also be encouraged by admin-
istrators or billing services to provide appropriate docu-
mentation to qualify for the highest appropriate level of
service.
The widespread use of electronic medical records allows
for fast and easy documentation with prepopulated tem-
plates and phrases. Templated medical records are designed
to efciently capture required histories (present illness, past
illnesses, social and family history, review of systems), objec-
tive ndings, and data review to justify coding levels. Emer-
gency medicine practice entails multitasking of clinical prob-
lems involving varying degrees of acuity and complexity.
The relative ease of computerized charting can allow for
careful review of systems pertinent to the chief complaint,
and an unsuspected issue may shed light on the present-
ing complaint. For instance, a patient presenting with leg
edema who is found to have pruritis on review of systems
would suggest a liver pathology as the cause. Further, an
incidental nding could lead to the discovery of an unre-
lated but signicant pathology. For example, a patient pre-
senting with ankle pain who is found to have unintentional
weight loss on review of systems may suggest an undiag-
nosed malignancy. Thorough use of a review of systems
may provide the physician with diagnostic acumen; however,
an extensive review of systems may also lead to notation
of system complaints that are not thoroughly investigated.
Casual reference to incidental ndings can be medicole-
gally problematic if the ndings are not fully investigated or
followed up.
What might be considered overdocumentation can also
result in overbilling. Third-party payers—such as gov-
ernment agencies, insurance companies, health mainte-
nance organizations, and employers—review emergency
medicine group practices. A group with high levels of
complex care billings could invite audits, resulting in pay-
backs and penalties. In rare cases, charges of fraud have
been leveled. For instance, Prime Healthcare Services, Inc.
agreed to pay $1.25 million to settle upcoding allegations
regarding falsied information about patient comorbidities
and complications.
5
As medical billing companies process
healthcare encounters into claims for insurance companies
and patients, they can provide protection against potential
penalties. To protect physicians and themselves, billing spe-
cialists may intentionally bill a medical record documented at
a level 5 to a lower level. In other words, careful charting by a
physician to justify a high level of service may be reimbursed
at a lower level.
3
Another consideration related to overdocumentation is
that time spent charting may be at the expense of patient
care. An observational study of ofce-based physicians
showed that physicians spent 2 hours on the computer for
every hour of patient time.
6
Although some charting is often
done after the scheduled end of shift, this study demon-
strated that charting takes more clinical time than direct
patient care.
Conceptually, given a cohort of ED patient encounters,
opportunities for reimbursement may be missed when chart-
ing is done in the context of a busy ED. Again, conceptually,
given the same cohort of emergency patients, there must be
a level at which all possible billable issues have been cap-
tured, and further documentation, from the point of view of
billing, would be overdocumentation.
While it is arguable to what degree extensively detailed
charting may be clinically valuable, a common assumption is
that adding more information to documentation will improve
reimbursement. To what degree this assumption is accurate
is the focus of this study.
METHODS
We conducted a 1-year productivity analysis of all super-
vising physicians (n=39) across 3 hospitals in an urban,
university emergency medicine training program located in
Philadelphia, PA. The health system serves an area of the
state in which 30% of the population lives below the federal
poverty level, and the majority of the payers are government-
funded programs such as Medicare and Medicaid.
7
All supervising physicians work 8-hour shifts, with a range
of 8 to 18 shifts per month. Physicians in training work during
most shifts at all 3 sites. The department at the main univer-
sity hospital is divided into 3 zones according to acuity. The
other hospitals in the system each have 1 designated main
ED and 1 minor care area. Minor care areas at all 3 sites
are staffed by advanced practice providers. No advanced
practice provider medical records were included in this
analysis.
At the study institution, a third-party company that con-
tracts with the health system handles billing. The electronic
medical record generates coding levels that the hospital’s
coding department conrms. The medical records are then
sent to the billing company for nal review.
Data for each supervising physician were assessed for
patients per hour (pt/h), RVUs per hour (RVU/h), amounts
of monies billed, and actual reimbursement from billings.
Ranges were averaged rst by physician. Because physi-
cians worked varying numbers of hours, we calculated per-
centages of reimbursement to billings to standardize the
evaluation. The relationships between RVU/h vs pt/h, RVU/h
vs revenue collected, and pt/h vs revenue collected were
determined. The R
2
value for each relationship was used
to determine variance, and a correlation coefcient was
also calculated. For trending purposes, physicians were
ranked and subsequently divided into RVU quartiles based
on billings and revenue collection percentages.
RESULTS
Among the 39 physicians, pt/h ranged from a low of 1.3
toahighof3.97.RVU/hrangedfromalowof3.97toahigh
of 10.07. A scatter plot of pt/h vs RVU/h showed a strong,
positive liner relationship (R
2
=0.7571) and a correlation coef-
cient of 0.87 (Figure 1).
Percentages of actual revenue collected vs billings ranged
from a low of 12.9% to a high of 23.1%. RVU/h vs revenue
collection yielded a moderately positive linear relationship
(R
2
=0.1752) with a correlation coefcient of 0.42, suggesting
that adding more detail to documentation to produce higher
RVUs may not translate to increased revenue (Figure 2).
The analysis of pt/h vs revenue collection yielded a weak
relationship (R
2
=0.0815), with a correlation coefcient of
0.29 (Figure 3).
A quartile analysis and interaction plot comparing RVU/h
to percent revenue collections showed an increase in RVU/h
until the third quartile (Figures 4 and 5).
300 Ochsner Journal