Teaching residents coding and documentation:
Effectiveness of a problem-oriented approach
Sawsan As-Sanie, MD, MPH,a,bDenniz Zolnoun, MD, MPH,aMary Ellen Wechter, MD,a
Georgine Lamvu, MD, MPH,aFrank Tu, MD, MPH,a,cJohn Steege, MDa
Department of Obstetrics and Gynecology, Division of Advanced Laparoscopy and Gynecologic Surgery,
University of North Carolina, Chapel Hill, NCa; Department of Obstetrics and Gynecology, The University of
Michigan, Ann Arbor, MIb; Department of Obstetrics and Gynecology, Northwestern University, Chicago, ILc
Received for publication March 10, 2005; revised May 29, 2005; accepted August 1, 2005
Objective: We sought to assess the effectiveness of a problem-oriented approach to teaching
residents accurate coding and documentation of ambulatory gynecology visits.
Study design: This was a pilot before-and-after study. Nine resident volunteers underwent 4
individual instructional sessions on coding and documentation with a trained faculty member
over 6 weeks. Outcomes were assessed by comparing the appropriateness of procedure and
diagnostic codes billed in participant continuity clinic prior to and in the 6 to 9 months following
Results: Following the intervention, participants demonstrated an increase in the accuracy of
coding the correct category of the evaluation and management service, an increase in the
appropriate use of modifiers, and a decline in undercoding errors.
Conclusion: Problem-oriented interactive learning appears to be an effective method of teaching
residents proper coding and documentation.
? 2005 Mosby, Inc. All rights reserved.
Accurate and consistent coding and documentation
of outpatient office visits is the foundation of operating
a compliant and productive office practice. Unfortu-
nately, many US residents in obstetrics and gynecology,
as well as those in other subspecialties, graduate with a
limited understanding of proper coding and documen-
To address this ubiquitous problem, the Accredita-
tion Council for Graduate Medical Education includes
system-based education as 1 of 6 general competencies
required for resident development.3System-based edu-
cation expects that residents demonstrate ‘‘an awareness
of and responsiveness to the larger context and system
of health care.’’ As such, residents should be knowledge-
able about coding, reimbursement, and the management
of a medical practice. Identification of these general com-
petencies was the first step in the Accreditation Council
for Graduate Medical Education’s long-term effort to
emphasize outcome assessment in resident education
and the accreditation process.
Despite the obvious need for an effective educa-
tion paradigm, there are no published guidelines on
teaching or assessing the competency of coding and
Presented at the Annual Meeting of Council on Resident Educa-
tion in Obstetrics and Gynecology and the Association of Professors of
Gynecology and Obstetrics, March 2-5, 2005, Salt Lake City, UT.
Reprints not available from the authors.
0002-9378/$ - see front matter ? 2005 Mosby, Inc. All rights reserved.
American Journal of Obstetrics and Gynecology (2005) 193, 1790–3
documentation in an ambulatory gynecology setting.
Thus, the objectives of this study were: (1) to develop a
problem-oriented approach to teaching proper coding
and documentation of ambulatory gynecology visits;
and (2) to assess the feasibility and effectiveness of such
an approach by evaluating changes in documentation
competencies and billing patterns.
Material and methods
This was an institutional review board approved pilot
before-and-after study conducted at the University of
North Carolina (UNC) Department of Obstetrics and
Gynecology between January 2003 and February 2004.
Nine obstetrics and gynecology residents, postgraduate
years (PGY) 2 to 4, were asked and consented to
participate in this study. Each resident attended 4
individual instructional sessions with 1 of 4 trained
faculty members over a 6-week period.
Each session was an individual 1-hour learner-based
examples of successes and errors from that individual
resident’s dictations in the prior 3 months. The curricu-
lum was designed to address the most common coding
and documentation errors identified in a sample of
resident dictations audited prior to the educational inter-
vention. Prior to meeting with the residents, each par-
ticipating faculty member underwent the hospital’s
mandatory coding and documentation course and an
additional 1-hour training course, which reviewed the
Session 1 provided an overview of Medicare stan-
dards for coding and documentation. This focused on
the usage of Current Procedural Terminology (CPT)
Evaluation and Management (E/M) codes and their
linkage to the International Classification of Diseases
(9th revision) diagnostic codes. The medical history,
physical exam requirements, and complexity of medical
decision making for each level of service provided were
reviewed. The appropriate use of modifiers-24, -25, and -
51, which are used to assure proper reimbursement of a
payable service provided, was highlighted in this session.
For example, modifier-24 is used when providing an
unrelated E/M service in the postoperative period of a
major surgical procedure. Modifier-25 is used when the
same physician performs a significant and separately
identifiable E/M service on the same day as a minor
procedure or a preventive health maintenance examina-
tion. Modifier-51 is used when multiple but separate
procedures are performed on the same visit.
Session 2 focused on the differences between the 1995
and 1997 Medicare guidelines for the documentation of
the physical examination. During this encounter, the
faculty member reviewed the 1995 requirements of a
multiorgan approach as well as the 1997 allowance for a
focused genitourinary examination. This session also
provided an overview of 2 methods of choosing the level
of service: billing according to evaluation and complex-
ity versus billing for time-based counseling.
Session 3 reviewed the recommended components of
an annual health maintenance examination, based on the
American College of Obstetricians and Gynecologists
guidelines on preventive care.4The emphasis was on
women ages 18 to 39 years, the most prevalent age group
Carolina preventable causes of morbidity and mortality
served as a reminder for residents to look beyond the
breast examination and Papanicolaou smear when pro-
then reviewed on the appropriateness of their coding and
documentation for an annual health maintenance visit.
Finally, session 4 was used as an opportunity to
review 5 audited resident dictations for errors defined in
the first 3 sessions. Residents were asked to provide
feedback on their dictations and coding and received
case-specific guidance from the instructor. At the end of
the session, a posttest was administered, followed by
immediate feedback on the answers.
Coding and documentation competencies were the
primary outcome evaluated. Four general categories
of competencies were developed based on the UNC
Teaching Physicians Oversight Committee guidelines for
computing a compliance error rate: class 1 (overcoding)
errors, class 2 (undercoding) errors, appropriate use of
modifier-25, and correct coding of a preventive medicine
Class 1 errors are overcoding errors that cannot be
billed as reported by the physician. The most common
class 1 errors are category errors in which the wrong
category of visit is chosen and overcoding by 1 or more
levels of service. These errors are considered the most
egregious because they bill for a higher fee than the
documentation supports. For example, billing CPT code
99385 (preventive medicine, new patient 18-39 years old)
when the appropriate code is CPT 99395 (preventive
medicine, established patient 18-39 years old) would be
considered a class 1 error.
Class 2 errors are undercoding errors that may be
billed as coded but are also serious because they gener-
ally undercode, causing lost revenues. For example,
billing CPT code 99213 (expanded problem-focused
visit, established patient) when the documentation sup-
ports the level CPT 99214 (detailed problem-focused
visit, established patient) is a class 2 error.
Coding competencies were assessed according to stan-
dard Medicare guidelines. A UNC billing compliance
auditor randomly selected 10 outpatient gynecology
visits provided by study participants: 5 dictations in the
3 months prior to the intervention and 5 dictations
between 6 and 9 months following the completion of the
intervention. To assure adequate diversity in the selected
As-Sanie et al1791
dictations, 2 of the 5 dictations in each time period were
selected from preventive medicine visits. The remaining
3 were problem-oriented visits. Within these 2 categories,
dictations were randomly selected from the appropriate
time period. Each dictation was audited to identify errors
in the coding competencies of interest. Also, a billing and
coding database was queried for all outpatient gynecol-
ogy visits provided by study residents in the same period.
Diversity in levels of service coded as well as appropriate
coding of annual health maintenance visits were exam-
ined using this database. Descriptive and bivariate sta-
tistics, including the Wilcoxon signed-rank test, were
performed using STATA 8.0 (STATA Corp, College
Eleven residents volunteered to participate in this study;
2 residents were excluded because they completed the
educational sessions within 1 month of graduation.
Among the remaining 9 residents, 5 were PGY2 and 4
were PGY3. The average postintervention examination
score was 94% correct.
After the educational intervention, residents showed
a decline in overcoding errors, primarily because of an
improvement in assigning the correct category of service
rendered (Table). Nearly a third of audited dictations
prior to the intervention were coded in the wrong
category of service; this declined to 13% following the
intervention (P = .05). There was also a trend toward a
decrease in coding errors that result in underbilling,
including meaningful improvements in billing for pro-
cedures provided and increases in the appropriate use
of modifier-25. For example, 2 of 4 procedures (50%)
performed in the preintervention period were not billed;
this declined, although not statistically significant, to a
failure to bill for 3 of 8 procedures (38%) in the post-
Based on the preintervention audit of resident dicta-
tions, the most common error when coding an annual
health maintenance exam was linking the International
Classification of Diseases (9th revision) code V72.3
(annual visit with Papanicolaou smear) to a problem-
oriented CPT E/M code (99201-99205 or 99212-99214)
rather than to a preventive medicine code (99385-99387,
99395-99397). This coding error was billed in 54% of the
preventive medicine visits prior to the intervention and
declined to 37% of preventive medicine visits in the 6 to
9 months after the intervention (P = .32).
To assure adequate diversity in the types of visits
reviewed, 40% of audited dictations were preventive
medicine visits; the remainder was problem-oriented
visits. Among the problem-oriented visits, the most
common diagnoses were pelvic pain (30%), abnormal
uterine bleeding (22%), uterine fibroids (13%), contra-
ception (9%), and first-trimester abortion (9%). There
was no relationship between the primary diagnosis and
frequency of coding error.
The diversity of E/M levels of service coded appeared
to improve following the intervention. For example,
among all outpatient problem-oriented gynecology visits
provided by participating residents in the study periods,
the use of upper-level E/M codes (99204, 99214, 99244)
increased from 11.7% to 20.6% of visits (P = .05). High-
level E/M codes (99205, 99215, 99245) also increased,
from 0% to 4% (P = .03), whereas use of low-level
E/M codes (99201, 99202, 99211, 99212, 99241, 99242)
declined from 27.9% to 15.3% (P = .21). Reflecting
this change in the diversity of E/M levels of service bil-
led, the average charge billed per encounter rose from
$109.48 to $120.29.
Despite the importance of mastering accurate coding and
prior and 6 to 8 months after an educational intervention
Frequency of coding errors among randomly sampled encounters by UNC obstetrics and gynecology residents (n = 9) 3 months
Coding ErrorNo. errors*% No. errors*%P valuey
Class 1 errors (unbillable errors)
Overcoded due to wrong category billed
Overcoded by 2 or more levels of service
Overcoded by 1 level of service
Class 2 errors (billable errors)
Consult opportunity missed
Undercoded by 1 level of service
Undercoded by 2 or more levels of service
Procedure not billed
* Total number of errors among all residents per opportunity to make error.
yWilcoxon signed rank test.
1792 As-Sanie et al
for teaching these skills in residency programs. In fact, Download full-text
many residents in obstetrics and gynecology, as well as
those in other subspecialties, graduate with a limited
understanding of proper coding and documentation. In
1 survey of obstetrics and gynecology residents attending
a regional research conference, only 29% of residents
felt confident in coding problem-oriented visits and 43%
felt confident in coding preventive care visits.2
Understanding how to correctly use modifiers was
equally lacking in this group of residents: only 29% felt
confident using modifier-25, 14% understood the use
of modifier-24, and no residents were familiar with
modifier-51. This trend is not limited to residents in
obstetrics and gynecology. In a large survey of emer-
gency medicine residents, the majority of residents rated
their confidence in their ability to accurately code as
‘‘minimal’’ (26%) or ‘‘not at all’’ (42%).1
To our knowledge, this is the first study that attempts
to address this deficit in resident education in obstetrics
and gynecology. We found that after individual instruc-
tion, university-based residents demonstrated meaning-
ful improvement in the knowledge and application of
coding principles. Residents tended to show declines in
the instances of noncompliant billing, both in assign-
ment of E/M category and level of service provided.
There was a trend toward a decline in undercoding
errors as well as an improvement in billing for proce-
dures performed. The appropriate use of modifiers to
accurately bill for services provided seemed to increase,
as did the appropriate coding of annual Papanicolaou
smears during preventive medicine visits.
Residents in this pilot study also showed improve-
ment in increasing the diversity of the E/M levels of
service coded, which may be another marker for im-
proved facility in coding for different levels of service.
As a result, the average charges billed per patient seen
also increased. Assuming correct application and docu-
mentation of the corresponding code for each visit, this
suggests that trained residents demonstrate improved
ability to use their coding and documentation skills to
reflect the complexity of services rendered. Although
this rise in average charge per patient seen was approx-
imately $10.81, this modest amount could translate into
substantial increases in billing for the same number of
services provided. For example, in a residency program
with 24 residents, providing 10 outpatient gynecology
visits per week for 47 weeks of the year, this increased
diversity in coding could result in the capture of lost
revenues conservatively estimated at $120,000 per year.
Indeed, this value underestimates the net benefit of
coding education because it does not account for other
improvements in coding knowledge, such as increased
use of modifiers to appropriately avoid denials of claims.
It is important to note that this training program
involved a small number of participating residents, and
the number of audited dictations in this study was
few. Thus, we had limited ability to detect statistically
significant improvements. Also, the follow-up was rela-
tively short and the 5 dictations may have been unchar-
acteristic of the resident or patient population as a
whole. In such a small study, results cannot necessarily
be generalized to other resident or patient populations.
The participating residents were volunteers and may
have represented a particularly motivated sample of phy-
sicians committed to improving their coding knowledge.
The long-term effectiveness of this training program
should be evaluated in future studies.
Although there was substantial time investment in
face-to-face teaching sessions with faculty instructors on
4 separate occasions, preliminary results of our previous
work teaching a 1-hour problem-oriented module sug-
gest that this approach may be effective in improving
knowledge of coding and documentation.2Thus, this
study, in addition to our previous work, is valuable in
that it demonstrates the feasibility of improving resident
coding skills. It also illustrates the potential valuable
financial gains that improved coding could have for the
department and for the individual physician.
Today in a climate of escalating costs, falling reim-
bursement, and legal scrutiny, skills for proper coding
and documentation are vital to the financial endurance
of health care providers and institutions. Indeed, coding
skills are becoming as essential to the success of the
graduating resident as any other skill acquired in resi-
dency. With cooperation and communication among
teaching programs, successful training in coding skills
should be rapidly developed and implemented in ob-
stetrics and gynecology training programs.
The authors acknowledgeDrRobertCefalo,MsHeather
Scott, and Ms Christy Davis for their contribution to
the design and execution of this study.
1. Howell J, Chisholm C, Clark A, Spillane L. Emergency medicine
resident documentation: results of the 1999 American Board of
Emergency Medicine in-training examination survey. Acad Emerg
2. Zolnoun D, Tu F, Tresolini C, As-Sanie S, Lamvu G, Steege J.
Teaching residents coding and documentation: a problem-oriented
approach. Obstet Gynecol 2004;103(Suppl 4):15S.
1.3. Available at: http://www.acgme.org/outcome/comp/compFull.
asp. Accessed February 18, 2005. Chicago (IL): Accreditation
Council for Graduate Medical Education; 1999.
4. ACOG Committee Opinion. Primary and preventive care: periodic
assessments. Obstet Gynecol 2003;102:1117-24.
As-Sanie et al1793