RVU Ready? Preparing Emergency Medicine Resident Physicians in Documentation for an Incentive-based Work Environment
Address for correspondence and reprints: Luan Lawson, MD; e-mail: email@example.com.
Objectives: The emergency medicine (EM) job market is increasingly focused on incentive-based reimbursement, which is largely based on relative value units (RVUs) and is directly related to documentation of patient care. Previous studies have shown a need to improve resident education in documentation. The authors created a focused educational intervention on billing and documentation practices to meet this identified need. The hypothesis of this study was that this educational intervention would result in an increase in RVUs generated by EM resident physicians and the average amount billed per patient.
Methods: The authors used a quasi-experimental study design. An educational intervention included a 1-hour lecture on documentation and billing, biweekly newsletters, and case-specific feedback from the billing department for EM resident physicians. RVUs and charges generated per patient were recorded for all second- and third-year resident physicians for a 3-month period prior to the educational intervention and for a 3-month period following the intervention. Pre- and postintervention data were compared using Student’s t-test and repeated-measures analysis of variance, as appropriate.
Results: The evaluation and management (E/M) chart levels billed during each phase of the study were significantly different (p < 0.0001). The total number of RVUs generated per hour increased from 3.17 in the first phase to 3.71 in the second phase (p = 0.0001). During the initial 3-month phase, the average amount billed per patient seen by a second- or third-year resident was $282.82, which increased to $301.94 in the second phase (p = 0.0004).
Conclusions: The educational intervention positively affected resident documentation resulting in greater RVUs/hour and greater billing performance in the study emergency department (ED).
Since its creation by the Health Care Financing Administration (HCFA, now the Centers for Medicare and Medicaid Services [CMS]) in 1992, the relative value unit (RVU) has become an important measurement tool for the work performed by physicians. Each emergency physician (EP) should have a thorough understanding of this unit to appreciate its use in making comparisons between physicians in regards to work output, billing for physician services, and determining and justifying staffing needs. However, the RVU has most recently gained attention for its relation to physician reimbursement. A Schumacher Group survey of emergency department (ED) administrators found that many EPs are not well prepared for incentive-based salaries.1 To be compensated well in this setting, resident physicians must understand RVUs and proper documentation.
Relative value units were developed by HCFA to construct a fee schedule for reimbursement of physician services. The total RVU has three components: the work RVUs for physician time and effort, practice RVUs to cover equipment and supplies, and malpractice RVUs for malpractice premiums. RVUs are given for both the documented procedures performed and the evaluation and management (E/M) level of the chart. The E/M level of a chart is determined by set criteria, key among them being history, physical examination, and medical decision-making. The more detailed the history and physical examination, and the more complex the decision-making, the higher the E/M level of chart, but only if it is documented accordingly. Downcoding results when a patient encounter warrants a higher level than is billed due to missing elements of history, physical examination, or medical decision-making in the physician’s documentation.
Although the Accreditation Council on Graduate Medical Education (ACGME) recommends that educational activities involving billing and coding be included as a part of the core competency of professionalism,2 studies have shown that resident physicians do not feel comfortable with the amount of education they are receiving in the areas of billing and coding.3,4 An abstract presented at the 2006 American College of Emergency Physicians Research Forum found that resident charts are more frequently downcoded than attending physician charts.5
Educational interventions have previously been tried for emergency medicine (EM) resident physicians involving a financial incentive for complete documentation,6 as well as for internal medicine resident physicians by implementing a template history and physical form,7 both of which resulted in improvements in the documented level of care by resident physicians. The authors previously completed an unpublished study at our academic facility and identified a need to improve EM resident physician education on billing and documentation. In that study, a survey given to both EM residents and attending physicians regarding billing and documentation practices found that 91% of resident physicians and 95% of attending physicians feel that this knowledge will be an important part of their future jobs. Eighty-two percent of resident physicians reported wanting more education in this area. An educational intervention would both comply with the ACGME core competency requirement and better prepare our resident physician graduates for incentive-based reimbursement.
We created a focused educational intervention on billing and documentation practices to meet this identified need. The intervention included focused education in complete documentation of charts and the appropriate use of E/M codes. It also provided case-specific feedback and involved identifying and creating templates to document billable EP procedures. The hypothesis of this study was that this educational intervention would result in an increase in RVUs generated by EM resident physicians and the average amount billed per patient.
We used a quasi-experimental design. All protocols were reviewed and approved by the University Medical Center Institutional Review Board.
Study Setting and Population
The study setting is an academic ED with a 3-year EM residency program consisting of 12 first-year resident physicians, 13 second-year resident physicians, and 11 third-year resident physicians; we did not include a control group. The ED sees an annual volume of about 70,000 patients.
In our ED, documentation is entirely electronic. The commercial documentation program has been in use since June 2007; a different commercial documentation program had been in place prior to that time. The documentation system provides historical prompting based on a template that correlates with the chief complaint.
For our study, a 1-hour lecture on documentation and coding was presented to 18 of the 24 second- and third-year EM resident physicians from the East Carolina University EM Residency Program at the beginning of the study period. The focus of the lecture was to address the requirements for each E/M documentation level and to highlight billable procedures in an effort to raise resident awareness regarding common EM procedures. Proper documentation of these procedures was included in the lecture. A pocket card on documentation, including procedures and E/M documentation levels, was distributed to all 36 EM resident physicians. Handouts from the lecture were available to all residents following the lecture, including those who were not present.
For a 3-month time period after the lecture, the EM resident physicians were given weekly case-specific feedback on the level of charts they generated and any missing elements that resulted in downcoding. They were not given the opportunity to edit any charts that had been downcoded. A biweekly newsletter written by the study authors was distributed to all EM resident physicians with an emphasis on aspects of documentation identified by our billing department as areas needing improvement. Examples of these areas included the importance of the Review of Systems section, what constitutes E/M codes, and how procedures generate RVUs.
Methods of Measurement
The true unit of analysis in this study is resident physician performance in documentation, which can be reflected in chart levels and billing. Billing and documentation analysis was provided by the billing department. Our billing department consisted of two personnel providing the coding and one person reviewing the encounters involved in our study after they were identified by the coders. The study investigators were blinded to the individual resident physician, as each physician was identified by a number indicating only resident training level. Chart E/M documentation levels, RVUs generated, and charges billed were assessed for each second- and third-year resident for a control period of 3 months prior to the lecture (September–November 2007) and again for the study period of 3 months after the lecture (January–March 2008). From these data, RVUs per hour were calculated as the RVUs generated for each resident divided by the number of hours worked by that resident, and RVUs per patient were calculated as the RVUs generated for each resident divided by the number of patients seen by that resident. The RVUs generated for each resident were further broken down into RVUs generated based solely on chart level and those based solely on procedures performed by each resident. Charges generated were calculated in a similar fashion. No charts were returned to physicians for editing of incomplete documentation. The billing schedule was based on 2007 charges for both study phases and was the same regardless of insurance type or self-pay. All second- and third-year categorical EM resident physicians were included in the study. Each of these resident physicians rotated through the ED during both the first and second 3-month period.
First-year resident physicians were excluded from the study for several reasons. Previous studies have demonstrated a significant increase in productivity from first year to second year that may be associated with the rapid growth of the first-year residents as they become familiar with their new role as resident physicians and are able to see greater numbers of patients.8 Furthermore, first-year residents in our department did not all rotate through the ED during both the control period and the study period.
Continuous variables (RVUs, number of patients, and charges) are reported as means and were compared pre- versus postintervention using a Student’s t-test. Categorical variables (chart level) are reported as percentages and were compared pre- versus postintervention using repeated-measures analysis of variance, using billing level as the dependent variable and study phase as the independent variable. In both cases, p < 0.05 indicated significance.
A total of 13 second-year and 11 third-year EM resident physicians rotated in the ED during the initial 3-month period, and all 24 also rotated in the second 3-month time period. Seventy-five percent of these residents attended the initial lecture. Five resident physicians worked in the ED 6 of these months, 8 worked 5 months, 4 worked 3 months, and 7 worked 2 months. Owing to vacation requests and educational conferences, the first study period included 7,556 resident hours and the second period included 7,335 resident hours worked. The average number of patients seen per hour by the resident physicians was 1.08 in the first study period and 1.17 in the second study period (p = 0.0269).
The difference in patient population between the two phases was not statistically significant in number or acuity, as determined by patients seen and admission rate. The average number of patients seen per resident in the initial 3-month period was 339 and in the second 3-month period was 364 (p = 0.412). The overall ED admission rate during the first phase was 33.5% and during the second phase was 32.6%.
E/M Chart Level
The E/M chart levels billed during the phases of the study were significantly different (p < 0.0001), and a post hoc analysis shows that the difference lies in a decrease in Level 3 charts from 42% to 36% (p = 0.0005) and an increase in Level 5 charts from 19% to 27% (p < 0.0001). The Level 4 chart percentages stayed nearly the same, with a change from 30% to 29% in Phase 2 (p = 0.3793) (see Table 1).
Table 1. E/M Chart Levels for Each Study Period
RVUs and Procedures
The total number of RVUs generated per hour (which includes both documentation chart level and billable procedures) increased from 3.17 in the first phase to 3.71 in the second phase (p = 0.0001), with all but one resident showing an absolute increase in RVUs/hour (see Table 2). The RVUs generated based solely on the chart level increased from 2.88 to 3.41 (p < 0.0001), while the RVUs generated on billable procedures did not show a significant increase, from 0.28 to 0.30 (p = 0.6135). The RVUs/patient increased from 2.94 to 3.17 (p = 0.0005; see Table 3). The total number of billable procedures among second- and third-year resident physicians increased from 418 (0.0514 procedures/patient) to 561 (0.0642 procedures/patient). The procedures with the greatest change in frequency were arthrocentesis, splint application, venipuncture, arterial stick, arterial line placement, and intraosseous line placement (see Table 4).
Table 2. RVUs per Individual Resident for Each Study Period
Table 3. RVUs Generated for Each Study Period
Table 4. Most Frequently Billed Procedures for Each Study Period in Order of Frequency
|Laceration repair (0.0236/patient)||Laceration repair (0.0238/patient)|
|Drainage of skin abscess (0.0067/patient)||Endotracheal intubation (0.0069/patient)|
|Central line placement (0.0058/patient)||Drainage of skin abscess (0.0590/patient)|
|Endotracheal intubation (0.0053/patient)||Central line placement (0.0058/patient)|
|Procedural sedation (0.0044/patient)||Treat fracture or dislocation (0.0031/patient)|
|Lumbar puncture (0.0031/patient)||Procedural sedation (0.0026/patient)|
|Treat fracture or dislocation (0.0018/patient)||Lumbar puncture (0.0024/patient)|
|Arthrocentesis (0.0007/patient)||Arthrocentesis (.0022/patient)|
|Apply splint (0.0007/patient)||Apply splint (0.0019/patient)|
|Anoscopy (0.0006/patient)||Anoscopy (0.0017/patient)|
|Venipuncture, arterial line, arterial stick, or IO (0.0006/patient)||Venipuncture, arterial line, arterial stick, or IO (0.0014/patient)|
|Total billed procedures: 482||Total billed procedures: 561|
During the initial 3-month period, the average amount billed per patient seen by a second- or third-year resident was $282.82, which increased to $301.94 in the second period (p = 0.0004). The average billed per hour increased from $303.79 to $354.08 (p = 0.0002). The increase of $46.77 in billing per hour based on the chart was significant (p = 0.0001); however, the increase of $3.44 in billing per hour based on procedures alone was not significant (p = 0.1929; see Table 5). Overall, these changes in billing resulted in an increase of $365,348 billed in the second 3-month period by second- and third-year categorical EM resident physicians.
Table 5. Billing Fees Generated by Resident Documentation and Procedures for Each Study Period
The seemingly small, yet significant increase in billing per second- and third-year resident physician work hour and per patient seen resulted in a relatively large increase in billing performance for our ED over a 3-month period. This increase projected over a full year would result in nearly $1.5 million in additional increased billing generated by this educational intervention on documentation and billing practices. These interventions not only improve the financial stability of the academic ED, but also improve the medical record and help to appropriately document the patient’s ED visit. Most importantly, this intervention may help to better prepare EM resident physicians for an incentive-based salary after graduation.
E/M Chart Level
According to the University Health Consortium and the Association of American Medical Colleges, the national averages for ED chart levels are 1% of charts are Level 1, 8% are Level 2, 33% are Level 3, 35% are Level 4, and 23% are Level 5.9 This results in roughly 42% of charts billed as Level 3 and below. In our first 3-month study period, just over 50% of our charts were billed at a Level 3 or below. The second study period was more consistent with the national average, with 44% of our charts billed at a Level 3 or below. The change was most notable as a decrease in Level 3 charts with an associated increase in the Level 5 charts. This does correlate with our educational intervention, specifically identifying which elements of the chart are required for each level of chart, as well as the importance of documenting medical decision-making, also an important factor in billing and coding. This shift to a higher level of billing is more consistent with our patient populations’ acuity levels, where roughly 34% of patients evaluated are subsequently admitted. The nearly unchanged admission rates during the respective study periods suggest similar patient acuity and that the improvements noted in phase two are a result of the educational intervention.
RVUs and Procedures
The RVUs generated per patient and per hour increased in the second phase of the study. This increase may be attributed to an increase in the E/M levels charted and billed. Although the educational intervention focused on both charting and billable procedures, the procedures performed did not significantly increase the RVUs generated per resident. However, there were 79 more documented procedures in Phase 2 of our study, with an associated increase in the number of previously undocumented minor procedures such as venipunctures, arterial sticks, and splint placements.
As with the E/M chart level and RVUs, there was an increase in billing from the first to the second study period. Again, this was attributed to an increase in the E/M chart level rather than to an increase in the number of procedures documented.
The lecture was given during a mandatory lecture session; however, only 75% of the residents included in the study were present for this lecture. Handouts from the lecture and the biweekly newsletters were provided to all residents.
Using a quasi-experimental design limits the study due to the lack of a control group. Introducing a control group would have resulted in smaller study numbers and presented the challenge of preventing those residents from viewing the newsletters or hearing about the lecture. Had we included a control group, we would have been able to control for the natural progression of resident physicians over the 6-month study period as they familiarized themselves with procedures and documentation. However, we attempted to limit this by excluding interns from the study, as it was felt that interns experience a more rapid improvement in many areas of ED practice, from documentation to patients seen per hour.
The increase in patients seen per hour by the second- and third-year resident physicians was statistically significant between the two study periods. However, the difference between 1.08 patients/hour and 1.17 patients/hour is just 1/10 of a patient/hour and may not be clinically significant. Also, the acuity levels of patients, and thus those qualifying for higher E/M codes and those requiring procedures, may have been different between the two time periods despite the unchanged admission level.
Although we did not openly disclose that resident documentation and RVU generation was being studied during either phase, it is possible that residents may have realized this was occurring and therefore put forth more effort in their documentation during Phase 2, resulting in an increase in chart level and RVUs generated.
We did not keep track of hours spent finishing charts after a scheduled shift for either study period. The improved documentation may have come at the cost of increased hours in the department, effectively “off the clock.” This study is also limited in that it was confined to one academic ED, and it is likely that each ED will have different areas of strength and weakness with regards to billing and documentation practices.
The study was performed using a single commercial documentation program and therefore might not be applicable to settings using a different program, dictation, or written documentation. Future areas of research could include evaluating resident physician documentation and billing after graduation for those resident physicians receiving the educational intervention and for those resident physicians who graduated prior to the study. Future studies could also evaluate whether ongoing education and feedback is required to maintain this change or if a one-time intervention is sufficient to affect long-term change.
The educational intervention did result in a statistically significant change in our ED documentation practices and consequently our billing. Although the intervention was focused on both charting and billable procedures, the impact was noted to be significant only with regard to the resident physicians’ chart levels. Further efforts to increase documentation of billable procedures are currently under way.
The authors thank Dr. Tim Reeder, Kim Carter, and the billing department for their help and support.