A critique of the Harvard Resource-Based Relative Value Scale

Rebuilding the Relative Value Unit-Based Physician Payment System

This Viewpoint explores the use of relative value units assigned by the Resource-Based Relative Value Scale in US physician payment systems and the need to rebuild this scale to reflect changes in modern clinical practice.

Temporal Trends in Revision Total Hip and Knee Arthroplasty from 2008 to 2018: Gaps and Opportunities

BACKGROUND

An overall assessment of how patient demographic characteristics and comorbidities are improving or worsening can allow better understanding of the value of revision total joint arthroplasty (TJA). Therefore, the purpose of this study was to identify patient demographic characteristics and comorbidities trends and episode-of-care outcome trends from 2008 to 2018 in patients undergoing revision TJA.

METHODS

The National Surgical Quality Improvement Program database was queried to identify patient demographic characteristics, comorbidities, and episode-of-care outcomes in patients undergoing revision TJA from 2008 to 2018 (n = 45,706). Pairwise t tests and pairwise chi-square tests were performed on consecutive years with Bonferroni correction. Trends were assessed using the 2-tailed Mann-Kendall test of the temporal trend.

RESULTS

Among patients undergoing revision TJA, there was no clinically important difference, from 2008 to 2018, in age, body mass index (BMI), percentages with >40 kg/m2 BMI, diabetes (18.8% to 19%), chronic obstructive pulmonary disease (4.1% to 5.4%), congestive heart failure within 30 days (0% to 1%), or acute renal failure (0% to 0.2%). However, modifiable comorbidities including smoking status (14.7% to 12.0%; p = 0.01), hypertension (66% to 26.0%; p = 0.02), anemia (34.5% to 26.3%; p < 0.001), malnutrition (10.4% to 9.3%; p = 0.004), and overall morbidity or mortality probability have improved, with a decrease in the hospital length of stay and 30-day readmission and a significant increase in home discharge (p < 0.001 for all).

CONCLUSIONS

Time-difference analysis demonstrated that the overall health status of patients undergoing revision TJA improved from 2008 to 2018. However, formal time-trend analysis demonstrated improvements to a lesser degree. The multidisciplinary effort to improve value-based metrics including patient comorbidity optimization and episode-of-care outcomes for primary TJA has been shown to potentially have an impact on revision TJA.

LEVEL OF EVIDENCE

Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Designed to Fail? the Future of Primary Care

Primary care is widely viewed as being in crisis despite its purported central role in addressing population issues related to healthcare cost, quality, access, and equity. Despite this pivotal role, the nature of the clinical practice today has largely emerged by default. We review the evolution of clinical practice in primary care from its genesis in small practices with paper charts and telephonic patient communication to managed care, pay-for-performance, and today's era of the electronic medical record, value-based payment, and consumerism. We suggest a necessary "reset" of expectations that focuses on today's practice structure and the historic face-to-face patient care expectations. Only by doing so can we successfully meet the demands of patients, society, and practicing internists.

Are We Appropriately Compensated by Relative Value Units for Primary vs Revision Total Hip Arthroplasty?

BACKGROUND

Relative value units (RVUs) are used to evaluate the effort required for providing a service to patients in order to determine compensation. Thus, more complicated cases, like revision arthroplasty cases, should yield a greater compensation. However, there are limited data comparing RVUs to the time required to complete the service. Therefore, the purpose of this study is to compare the (1) mean RVUs, (2) mean operative times, and (3) mean RVU/minute between primary and revision total hip arthroplasty (THA) and (4) perform an individualized idealized surgeon annual cost difference analysis.

METHODS

A total of 103,702 patients who underwent primary (current procedural terminology code 27130) and 7273 patients who underwent revision THA (current procedural terminology code 27134) were identified using the National Surgical Quality Improvement Program database. Mean RVUs, operative times (minutes), and RVU/minute were calculated and compared using Student t-test. Dollar amount per minute, per case, per day, and year was calculated to find an individualized idealized surgeon annual cost difference.

RESULTS

The mean RVU was 21.24 ± 0.53 (range, 20.72-21.79) for primary and 30.27 ± 0.03 (range, 30.13-30.28) for revision THA (P < .001). The mean operative time for primary THA was 94 ± 38 minutes (range, 30-480 minutes) and 152 ± 75 minutes (range, 30-475 minutes) for revision THA (P < .001). The mean RVU/minute was 0.260 ± 0.10 (range, 0.04-0.73) for primary and 0.249 ± 0.12 (range, 0.06-1.0) for revision cases (P < .001). The dollar amounts calculated for primary vs revision THA were as follows: per minute ($9.33 vs $8.93), per case ($877.12 vs $1358.32), per day ($6139.84 vs $5433.26), and a projected $113,052.28 annual cost difference for an individual surgeon.

CONCLUSION

Maximizing the RVU/minute provides the greatest "hourly rate." The RVU/minute for primary (0.260) being significantly greater than revision THA (0.249) and an annualized $113,052.28 cost difference reveal that although revision THAs are more complex cases requiring longer operative time, greater technical skill, and aftercare, compensation per time is not greater.

Measuring the differences in work ratios between pediatric and adult ophthalmologic examinations

PURPOSE

To assess the differences in work needed for pediatric and adult ophthalmologic examinations.

METHODS

Seven ophthalmology residents conducted slit-lamp and refraction examinations on children 3-7 years of age and adults 20-69 years of age. The examiners reported the magnitude estimate (ME) of their work in relation to two references (cross-reference ME): average adult examination and average pediatric examination. The examination time was also measured.

RESULTS

For the slit-lamp examination, 50 children and 58 adults were recruited. The ME was 1.45 (95% CI, 1.30-1.62) times higher for the pediatric examinations than for the adult examinations when the reference was an average adult case. With respect to time, the pediatric examinations took 1.22 (95% CI, 1.06-1.41) times longer than the adult examinations. For the refraction examinations, 58 children and 96 adults were recruited. The ME was 1.35 (95% CI, 1.21-1.52) times higher for the pediatric examinations. The pediatric examination took 1.32 (95% CI, 1.16-1.50) times longer than the adult examination. The cross-reference ME ratios measuring the pediatric over adult examinations against both the pediatric and adult reference cases were equivalent in both the slit-lamp and the refraction examinations; however, the ME and time ratios of the pediatric over the adult examinations were not equivalent for the slit-lamp or for the refraction examinations.

CONCLUSIONS

The cross-reference ME showed that pediatric ophthalmologic examinations require more work than the adult examination with validity and reliability. The time estimate was insufficient as a single indicator for work estimation.

A multidisciplinary approach to vascular surgery procedure coding improves coding accuracy, work relative value unit assignment, and reimbursement

BACKGROUND

Vascular surgery procedural reimbursement depends on accurate procedural coding and documentation. Despite the critical importance of correct coding, there has been a paucity of research focused on the effect of direct physician involvement. We hypothesize that direct physician involvement in procedural coding will lead to improved coding accuracy, increased work relative value unit (wRVU) assignment, and increased physician reimbursement.

METHODS

This prospective observational cohort study evaluated procedural coding accuracy of fistulograms at an academic medical institution (January-June 2014). All fistulograms were coded by institutional coders (traditional coding) and by a single vascular surgeon whose codes were verified by two institution coders (multidisciplinary coding). The coding methods were compared, and differences were translated into revenue and wRVUs using the Medicare Physician Fee Schedule. Comparison between traditional and multidisciplinary coding was performed for three discrete study periods: baseline (period 1), after a coding education session for physicians and coders (period 2), and after a coding education session with implementation of an operative dictation template (period 3). The accuracy of surgeon operative dictations during each study period was also assessed. An external validation at a second academic institution was performed during period 1 to assess and compare coding accuracy.

RESULTS

During period 1, traditional coding resulted in a 4.4% (P = .004) loss in reimbursement and a 5.4% (P = .01) loss in wRVUs compared with multidisciplinary coding. During period 2, no significant difference was found between traditional and multidisciplinary coding in reimbursement (1.3% loss; P = .24) or wRVUs (1.8% loss; P = .20). During period 3, traditional coding yielded a higher overall reimbursement (1.3% gain; P = .26) than multidisciplinary coding. This increase, however, was due to errors by institution coders, with six inappropriately used codes resulting in a higher overall reimbursement that was subsequently corrected. Assessment of physician documentation showed improvement, with decreased documentation errors at each period (11% vs 3.1% vs 0.6%; P = .02). Overall, between period 1 and period 3, multidisciplinary coding resulted in a significant increase in additional reimbursement ($17.63 per procedure; P = .004) and wRVUs (0.50 per procedure; P = .01). External validation at a second academic institution was performed to assess coding accuracy during period 1. Similar to institution 1, traditional coding revealed an 11% loss in reimbursement ($13,178 vs $14,630; P = .007) and a 12% loss in wRVU (293 vs 329; P = .01) compared with multidisciplinary coding.

CONCLUSIONS

Physician involvement in the coding of endovascular procedures leads to improved procedural coding accuracy, increased wRVU assignments, and increased physician reimbursement.

Variation in specialist fees: evidence from New Zealand insurance claims

OBJECTIVES

To examine the parity between specialties in reimbursements for surgical procedures in a private, fee-for-service setting and to ascertain whether differences exist after accounting for factors suggested by a Resource-Based Relative Value Scale (RBRVS).

METHODS

A routinely updated database covering several private insurers (n = 8294 procedures from 1997 to 2002) was used to examine differences in overall and hourly reimbursement. Multiple regression analysis was used to control for factors that might be responsible for differences in payment (location, year, sex of patient and the associated anaesthetist fee). The resulting regression residuals were compared between specialties.

RESULTS

Large differences between specialties in reimbursements were found in the overall amount paid. For most specialties, these differences were explained by factors such as time for procedure, location, complexity of procedure and sex of patients. However, hourly reimbursements for ophthalmologists were substantially higher (more than 50% above general surgery overall and 72% higher on an hourly basis). Some other smaller differences in overall and hourly reimbursement were also found.

CONCLUSIONS

These results indicate that specialist fees vary significantly but many of the differences are explainable by factors incorporated into the RBRVS. However, significant variation remains for some specialties, most notably ophthalmology. Explanations for the results are discussed, including the possibility that political factors may influence the setting of specialist fees. This raises questions concerning the fairness of reimbursements and resulting solidarity within the medical profession.

Comparison of work and time estimates by chiropractic physicians with those of medical and osteopathic providers

BACKGROUND

Resource-based relative value scales (RBRVS) have become a standard method for identifying costs and determining reimbursement for physician services. Development of RBRVS systems and methods are reviewed, and the RBRVS concept of physician "work" is defined.

OBJECTIVE

Results of work and time inputs from chiropractic physicians are compared with those reported by osteopathic and medical specialties. Last, implications for reimbursement of chiropractic fee services are discussed.

METHODS

Total work, intraservice work, and time inputs for clinical vignettes reported by chiropractic, osteopathic, and medical physicians are compared. Data for chiropractic work and time reports were drawn from a national random sample of chiropractors conducted as part of a 1997 workers' compensation chiropractic fee schedule development project. Medical and osteopathic inputs were drawn from RBRVS research conducted at Harvard University under a federal contract reported in 1990. Both data sets used the same or similar clinical vignettes and similar methods. Comparisons of work and time inputs are made for clinical vignettes to assess whether work reported by chiropractors is of similar magnitude and variability as work reported by other specialties.

RESULTS

Chiropractic inputs for vignettes related to evaluation and management services are similar to those reported by medical specialists and osteopathic physicians. The range of variation between chiropractic work input and other specialties is of similar magnitude to that within other specialties. Chiropractors report greater work input for radiologic interpretation and lower work input for manipulation services.

CONCLUSIONS

Chiropractors seem to perform similar total "work" for evaluation and management services as other specialties. No basis exists for excluding chiropractors from using evaluation and management codes for reimbursement purposes on grounds of dissimilar physician time or work estimates. Greater work input by chiropractors in radiology interpretation may be related to a greater importance placed on findings in care planning. Consistently higher reports for osteopathic work input on manipulation are likely attributable to differences in reference vignettes used in the respective populations. Research with a common reference vignette used for manipulation providers is recommended, as is development of a single generic approach to coding for manipulation services.

Comparison of 1997 Medicare relative value units for gender-specific procedures: is Adam still worth more than Eve?

BACKGROUND

On January 1, 1992, Congress implemented a Medicare payment system based on relative value units (RVUs). The total RVU (which is made up of work, practice, and malpractice RVUs) is multiplied by a dollar conversion factor to set the reimbursement for all procedures covered by Medicare. In a previous study, we found that significant gender bias exists in Medicare reimbursement for female-specific services. Recently, HCFA approved increases (beginning January 1997) in the work RVU for many gynecologic procedures. This study was undertaken to compare work and total RVUs for gender-specific procedures effective January 1, 1997.

METHODS

Using the May 1996 Federal Register, we compared work and total RVUs for 24 pairs of gender-specific procedures. The groups were matched so that the amount of work and level of difficulty would be similar, if not identical. We validated our selection of procedures for comparison by also evaluating the average time required to perform these procedures.

RESULTS

Comparison of work RVUs for the 24 paired procedures revealed that in 19 cases (80%), male-specific procedures had a higher RVU; in 3 cases (12%), female-specific procedures were higher; and in 2 cases, there was no difference. On average, work RVUs were 49% higher for urologic procedures than for gynecologic procedures. Comparison of total RVUs revealed that in 20 cases (83%), urologic procedures had a higher total RVU and in 3 cases (12%), gynecologic procedures were higher. On average, male-specific surgeries are reimbursed at an amount which is 37% higher than that for female-specific surgeries.

CONCLUSION

Recent increases in work RVUs for many gynecologic procedures have resulted in improved reimbursement. However, even with these improvements, significant gender bias still exists in the Medicare reimbursement of female-specific procedures. This gender bias is further magnified as more private insurance carriers use the system to set reimbursement.

Is Adam worth more than Eve? The financial impact of gender bias in the federal reimbursement of gynecological procedures

PURPOSE

On January 1, 1992, Congress implemented a Medicare payment system based on relative value units (RVU). The RVU multiplied by a dollar conversion factor sets the reimbursement for all procedures covered by Medicare and many other private insurers. This study was undertaken to evaluate discrepancies in federal reimbursement for gender-specific procedures.

METHODS

Using the December 1995 Federal Register and the regional Medicare conversion factor ($40.08/RVU), we compared the work RVU and total reimbursement of 24 groups of gender-specific surgical procedures. The groups were matched as carefully as possible so that the amount of work and level of difficulty would be similar, if not identical. Some examples of comparisons are as follows: biopsy of male vs female genitals, hysterectomy vs prostatectomy, staging for ovarian vs testicular cancer, and exenteration for cervical vs prostate cancer.

RESULTS

In the 24 matched procedures, the male-specific procedures were reimbursed at a higher amount in 19 (79%) cases. The female-specific procedures were reimbursed at a higher amount in 3 (12%) cases (P = 0.004). There was no difference in reimbursement for two of the comparisons. Overall, we found that male-specific procedures are reimbursed at an amount which is 44% higher than female-specific procedures. Comparison of work RVU revealed that male-specific procedures were assigned higher values in 19 cases and, overall, male gender-related surgeries had work RVU that were 50% higher than female gender-related surgeries.

CONCLUSION

There is significant gender bias against the Medicare reimbursement of female-specific services. This results in a lower net reimbursement for gynecologic procedures. In addition, since many private sector insurance carriers now use the resource-based relative value scale system, this gender bias is further potentiated.

Provider, payor, and patient outcome expectations in back pain rehabilitation

Return to work is used routinely to define successful back pain treatment. This study examined how the patient and three professional groups defined success for each of five categories of patients: not sponsored and working, sponsored and working, sponsored on modified duty, sponsored and off work for less than 10 weeks, and sponsored and off work for more than 10 weeks. The groups sampled were treating staff (n = 98), referring physicians (n = 98), third-party sponsors (n = 133), and patients (n = 648) representing all five patient categories. Each group provided a priority ranking for six objectives of treatment: Return to Work, Pain Control, Functional Improvement, Increased Strength and Range of Movement, Positive Attitude Shift, and Acquired Knowledge. The results indicated that the work status of the patient had a significant effect on the ranking of objectives by the treating staff and third-party sponsors. Physicians and patients considered pain control most important regardless of the patient category. Success in back pain rehabilitation is defined by different criteria. The determination of successful outcome must consider the patient's circumstances and acknowledge the perspective of the individual who is defining success.

Reflections on the cost consequences of the new gene technology for health policy

This article presents a preliminary and necessarily tentative and subjective assessment of the impact of new gene technology on health care costs. In the short term, diagnosis and treatment of genetic disease are likely to increase costs. Treatment with nongene therapy will continue to be far less expensive than gene therapy where it is available. Research developments to monitor as indicators of forthcoming cost reductions in genetic therapy are set forth. Some forms of genetic screening may soon reduce health care costs, and an example is provided. Genetically engineered pharmaceuticals are described and their impact on costs reviewed. Conditions under which they are likely to reduce health care costs are indicated.

Designing fee schedules by formulae, politics, and negotiations

Fee-for-service cannot be used successfully by organized health insurance without a fee schedule. America first tried to pay doctors under Medicare by an involved formula system without a fee schedule, but the effort has failed. The United States has now commissioned a research project to design a unique fee schedule that will precisely reflect physicians' effort and practice costs and that will represent the prices produced by a perfectly competitively market. The primary goal is the same as that pursued recently by reformers in all countries: viz., narrow the spread in fees and income between surgical and cognitive fields. There are serious technical limitations on this effort, despite the talent of the research team. An additional difficulty lies in the nature of the subject: paying the doctor involves conflicts of interest between payers and all doctors as well as among the medical specialties, and the conflicts cannot be resolved by any formulae calculated by any single research team. Methodological and political compromises will be necessary, in order to adopt a reform. The new method may be just as politically driven, complicated, and disputed as the old one, despite America's pretenses that it prefers free markets and opposes excessive government.

Refinement and expansion of the Harvard Resource-Based Relative Value Scale: the second phase

The Harvard resource-based relative value scale (RBRVS) for physician services has assumed a critical role in physician payment reform. We have demonstrated that the relative resource costs of providing physician services can be defined and measured in a rational and systematic way and that the results are reliable and valid. Consequently, the RBRVS is a viable basis for national payment policy and could be used for establishing a national fee schedule for physician services or to identify "mispriced" physician procedures. Since the release of the final report of the first phase of the Harvard RBRVS study in September of 1988, there has been extensive review, discussion, and criticism of the RBRVS. Dr. Laurence F. McMahon, Jr., in the accompanying article, provides a further critique of our research. In this paper, we review the RBRVS study and results and respond to the major criticisms that have been raised by Dr. McMahon and others. We then describe the tasks we are currently undertaking to expand and validate our research and address the important criticisms and limitations.