Background
Medicare payment for many surgical procedures covers not only the procedure itself but also most post-operative care provided over a fixed period of time (the “global period”).1 When the Centers for Medicare & Medicaid Services (CMS) sets payment rates, it assumes that a certain number and type of post-operative visits specific to each procedure typically occur. In other RAND Corporation research (Kranz et al., 2021; Crespin et al., 2021a; Crespin et al., 2021b), we found that the number of visits actually performed was lower than CMS's assumptions when setting payment rates.
This study describes how new claims-based data on the number of post-operative visits could be used to adjust valuation for procedures with 10- and 90-day global periods. The idiosyncrasies of the resource-based relative value scale (RBRVS) system used to determine payment for Medicare services result in some ambiguity about how procedures should be revalued to reflect reductions in post-operative visits. We intend for the results presented in the study to be a starting point for further policy development for revaluation.
Current Approach to Collect Information on Post-Operative Visits
Currently, the number of post-operative visits that CMS assumes typically occur during global periods is informed by practitioner surveys administered by the American Medical Association/Specialty Society Relative Value Scale Update Committee (the RUC) and its individual specialty society members. The primary purpose of the surveys is to collect information on the practitioner work and time associated with individual procedures and other health care services (based on Healthcare Common Procedure Coding System [HCPCS] codes), including an estimate of the total work involved in furnishing the service and related post-operative care.2 When a procedure has a 10- or 90-day global period, the surveys also ask practitioners to report the number and type of post-operative visits that typically occur during the global period. Respondents use evaluation and management (E&M) visit codes, including codes for office/outpatient and inpatient visits of different levels, discharge visits, and critical care visits, to describe the number and level of these post-operative visits. CMS, when determining the valuation for the procedure, may adjust the counts of visits recommended by the specialty societies. The number of post-operative visits assumed to typically occur during the global period is published by E&M HCPCS code in the Physician Time File (hereafter, the Time File), which is posted annually with the Medicare Physician Fee Schedule. The Time File also includes an estimate of the physician time spent on post-operative visits.3
Summary of Prior RAND Studies and Implications for Revaluation
The Medicare Access and CHIP Reauthorization Act of 2015 required CMS to collect information on the number and level of post-operative visits actually provided and to potentially revalue misvalued procedures using these newly collected data and other information. In response, CMS collected information on the number of post-operative visits by requiring select practitioners4 to report post-operative visits following certain high-volume or high-cost procedures5 using the no-pay HCPCS code 99024. CMS also collected information on the level of visits for three chosen procedures using a provider survey focusing on the time, activities, staff, and work associated with post-operative visits following the three procedures (cataract surgery, hip arthroplasty, and complex wound repair).
RAND researchers analyzed data collected through both of these channels. In the most recent report, RAND researchers’ analysis of the number of visits reported using HCPCS code 99024 found that only 4 percent of procedures with 10-day global periods had any post-operative visits reported (Crespin et al., 2021b).6 Although 70 percent of procedures with 90-day global periods had at least one associated post-operative visit, only 38 percent of the total number of expected post-operative visits for these procedures were reported.
These findings imply that procedures with 10- and 90-day global periods are overvalued; that is, they are valued as having too many relative value units (RVUs). Overvaluation of procedures with 10- and 90-day global periods leads to overpayment for these procedures. Because changes to Medicare valuations must be budget neutral, overpayment for services with 10- and 90-day global periods reduces payment for other services paid under Medicare's Physician Fee Schedule.7 Over- and underpayment for services can distort provider incentives to provide services and affect beneficiary cost-sharing.
Our analysis of information on the level of post-operative visits also has implications for revaluation. As noted earlier, our survey effort collected information on the level of post-operative visits following three types of procedures (cataract surgery, hip arthroplasty, and complex wound repair). When post-operative visits were provided following these procedures, we found that the level of work differed between post-operative visits and the E&M codes used by Medicare as approximations for post-operative visits during the valuation process: slightly less in the case of cataract surgery and hip arthroplasty and slightly more for complex wound repair (Gidengil et al., 2019).
The goal of this study is to describe an approach in which these newly collected data—particularly the claims-based data on the number of post-operative visits—could be used to revalue global surgery procedures and determine the impact of this approach.
Valuation Background and Revaluation Approach
Each procedure's overall valuation is in terms of RVUs, with separate work, practice expense (PE), and malpractice RVU components. There are several links between the number of post-operative visits and Physician Fee Schedule valuation. In the current valuation system, the link between these visits and work RVUs is indirect; reducing the number of bundled post-operative visits does not automatically result in a reduction in work RVUs because physician work RVUs are estimated using magnitude estimation, where an entire surgical global package is valued holistically in comparison with similar services, rather than using a building-block approach, where the valuation of individual components sums to a total valuation. (We describe these approaches further next.) Although respondents to RUC surveys reported the number and level of bundled post-operative visits, it is not clear whether the respondents fully incorporate the post-operative visits in their estimates of total work. Furthermore, CMS's final decisions regarding valuation likely are based on multiple factors, including factors other than the number and level of post-operative visits. In contrast, there is a direct link between post-operative visits and direct PE inputs and physician time. Physician work, physician time, and direct PE inputs have important impacts in the allocation of indirect PE and malpractice RVUs. Changes in physician work, physician time, and direct PE inputs for an individual procedure will, in turn, affect the allocation of PE and malpractice RVUs to other Physician Fee Schedule services.
The ambiguity associated with changes to work RVUs stems from an intrinsic tension in the RBRVS related to the alignment between information on the discrete “building blocks” that contribute to physician work (such as the number of post-operative visits) and estimates of the total work for the global service. As noted earlier, the RUC/specialty society surveys collect—and CMS publishes—information about most, but not all, of the building blocks required to calculate total physician work, such as the time involved in different components of a procedure and the number of post-operative visits. Each of these building blocks contributes work RVUs to the total work for the procedure, and changes to an individual component that contributes to a global service can be applied through a “reverse building-block” method of adding or subtracting a specific number of RVUs.
However, total work is estimated via surveys using magnitude estimation, in which respondents select an already-valued service that is most similar to the service being valued and then compare them in terms of total work, including post-operative visits that are assumed to be delivered in global periods. It is conceptually possible that a procedure's consensus total work estimate from magnitude estimation would not change even if the number of assumed post-operative visits decreases. Even in such a case, however, the direct PEs and physician time associated with that code would be clearly incorrect, which would have implications for PE RVUs.
To provide estimates to frame the discussion of improving payment for global services, we revalued procedures using the reverse building-block approach by adjusting work RVUs, physician time, and direct PE inputs based on the difference between the number of post-operative visits observed via claims-based reporting and the expected number of post-operative visits used during valuation. These changes led to a different allocation of indirect PE RVUs and malpractice RVUs to codes with and without global periods. As a last step, we applied an updated conversion factor, which is a dollar-per-RVU amount used by Medicare to convert valuations into dollar terms, to determine Medicare payments for different specialties.
Data and Methods
We combined Medicare claims data and the Time File posted with the 2019 Medicare Physician Fee Schedule to calculate the share of post-operative visits that were reported for each procedure for which reporting was required. The data and methods related to our analysis of post-operative visits reported via claims are discussed in prior reports (Crespin et al., 2021a; Crespin et al., 2021b; Kranz et al., 2021). We used regression models to impute the share of reported, relative to assumed, post-operative visits for procedures with 10- and 90-day global periods for which reporting was not required.
For revaluation, our starting point was work, PE, and malpractice RVUs for procedures with 10- and 90-day global periods, as listed in the calendar year (CY) 2019 Medicare Physician Fee Schedule. The baseline CY 2019 valuations were associated with the assumed number of post-operative visits included in the global period, as listed in the Time File. We subtracted a share of work RVUs, direct PE inputs, and physician time based on the percentage of post-operative visits currently included in valuation but not typically reported. For changes in work, we explore how our results change when we use three additional observed visit metrics: the mean, modal, and 75th percentile of reported post-operative visits rather than the median.8 As a final step, we estimated the impacts of reductions in post-operative visits on work, PE, and malpractice RVUs together, including the allocative implications on indirect PE and malpractice RVUs using updated work RVUs, physician time, and direct PE inputs based on the median of observed post-operative visits.
We report the impacts of revaluation, first on work alone, next for PE alone, and finally adjusting all components together, by applying the status quo and updated valuations to the CY 2019 fee-for-service Medicare volume of procedures with 10- and 90-day global periods. We report results for each of the 291 procedure codes for which reporting was required and results by specialty, reflecting the relative volume of services across all services billed by the specialty. We also report results in terms of aggregated payments across services using an updated conversion factor to offset the change in total RVUs.
Results
Figure 1 reports updated work RVUs after removing work RVUs associated with post-operative visits that were assumed but not provided.9 Depending on which observed visit metric was used as an input in revaluation, the updated work RVUs were between 18 percent and 32 percent lower for procedures with 90-day global periods and between 39 percent and 40 percent lower for procedures with 10-day global periods compared with current work valuations.
Across all Medicare Physician Fee Schedule services, our revaluation steps reduced total work RVUs by 3 percent (result not illustrated).
Figure 1. Share of Work RVUs Remaining After Revaluation Using Different Observed Visit Metrics for the 291 Procedures for Which Reporting Was Required
SOURCE: RAND analysis of 2019 claims data for reported post-operative visits and the Medicare CY 2019 Physician Fee Schedule and Time File.
NOTES: Results reflect the 2019 Medicare volume mix across the 291 procedures for which reporting of post-operative visits was required. Pctl. = percentile; PFS = Physician Fee Schedule.
In a separate analysis, we estimated the impact of reducing only direct PE inputs under the rationale that direct PE inputs for visits that are not occurring should not contribute to procedure valuations. We found that this change reduced PE RVUs and total RVUs for procedures with 10- and 90-day global periods by 14 percent and 6 percent, respectively.10
Our adjustments to work RVUs, physician time, and direct PE inputs (rather than just work or direct PE inputs individually, as presented in the previous sections) resulted in a 28.5-percent reduction in total RVUs for procedures with 10- and 90-day global periods and a slight increase (0.4 percent) for all other Physician Fee Schedule services. The net reduction in RVUs was 2.6 percent across all Physician Fee Schedule services.11 The impact on procedure-focused specialties was larger; the largest impact was a 20.3-percent reduction in total RVUs for plastic and reconstructive surgery. We found small increases in RVUs for primary care, neurology, cardiology, and diagnostic radiology, which were caused by increases in allocated PE and malpractice RVUs for services without 10- and 90-day global periods. The net impact for specialties that bill primarily for services without 10- and 90-day global periods (e.g., cardiology) was positive.
As a final step, we estimated the change in Medicare payments under the Physician Fee Schedule by calculating an updated conversion factor to preserve budget neutrality.12 Because the overall number of RVUs decreased, the conversion factor (which is defined as the funds available to pay for Physician Fee Schedule services divided by the sum of RVUs) increased. As a result, the reductions in total RVUs for surgical specialties, such as cardiac surgery, surgical oncology, and thoracic surgery, yielded slightly smaller reductions in payments (Figure 2).13 For some specialties (e.g., interventional radiology), a small reduction in total RVUs was offset by a higher conversion factor to yield a small increase in payments. Modest increases in total RVUs for other specialties (e.g., cardiology, neurology, and the specialties that report collectively as primary care) yielded a larger (but still modest) increase in payments.
Figure 2. Percentage Change in Physician Fee Schedule Payments After Revaluation, by Specialty
SOURCE: RAND analysis of 2019 claims data for reported post-operative visits and the Medicare CY 2019 Physician Fee Schedule and Time File.
NOTES: “Change in Physician Fee Schedule Payments” is the percentage change from status quo total RVU valuations to updated total RVU valuations. Primary care includes family practice, general practice, and internal medicine. Asst. = assistant.
Discussion and Conclusion
This article describes how the reverse building-block approach could be used to adjust valuation of procedures with 10- and 90-day global periods using claims-based data on the number of post-operative visits performed. Total RVUs are driven by several components, including work RVUs and direct PE RVUs. Depending on which statistic describing the number of observed visits we used (e.g., mean, median), updated work RVUs were between 18 percent and 30 percent lower for procedures with 90-day global periods and between 38 percent and 40 percent lower for procedures with 10-day global periods compared with current work RVU levels. Adjusting direct PE inputs for the number of post-operative visits, without adjustment to work RVUs, resulted in relatively modest reductions in PE. In terms of total RVUs, changes ranged from reductions of between 5.1 percent (vascular surgery) and 20.3 percent (plastic and reconstructive surgery) among proceduralist specialties to small increases among some other specialties (e.g., cardiology, neurology, and primary care specialties). Because the reduction in total RVUs results in a higher conversion factor, reductions in actual payments to surgical specialties were lower than the reduction in total RVUs. Payments to primary care practitioners increased by roughly 3 percent.
In our revaluation approach, we make three key assumptions. First, we assume that the bundled post-operative visits that were not observed did not occur. Our earlier reports address this assumption in depth and conclude that it is unlikely that underreporting by practitioners explains why we observe fewer post-operative visits than expected (Crespin et al., 2021a; Crespin et al., 2021b; Kranz et al., 2021; Mulcahy, Mehrotra, et al., 2021). Second, we assume that the amount of physician work involved in post-operative visits is the same as the amount of work involved in the corresponding E&M visits indicated in the Time File. This assumption is consistent with our earlier survey-based findings (Gidengil et al., 2019). Third, and most importantly, our approach removes all of the work RVUs that are associated with visits that did not occur. As noted earlier, the reverse building-block approach that we used assumes that total work is the sum of work associated with discrete components of the procedure and global package (including post-operative visits). It is impossible to know how the number of post-operative visits from the surveys affected the estimates of total work made by RUC/specialty society respondents and CMS.
There are several potential paths forward for revaluation. Consistent with the approach shown in this report, CMS could simply revalue all procedures with 10- and 90-day global periods using the reverse building-block approach to reflect the actual number of post-operative visits provided during global periods. For specific codes for which there are concerns about the resulting valuation, CMS and the RUC could revalue using the usual survey-based approach. These changes could be phased in over time. A second approach would be to first revalue all procedures using this reverse building-block approach and then convert some or all global procedures to 0-day global procedures. For example, CMS might convert all procedures that currently have 10-day global periods to procedures with 0-day global periods. For these new 0-day global procedures, practitioners would bill separately for post-operative visits than they would for other services. Because bundled payments can incentivize efficient provision of care, one disadvantage of this approach is that it might discourage more-innovative means of delivering post-operative care, such as telemedicine.
In the longer term, CMS may pivot to a valuation system that is consistent with the building-block approach. Such a system would allow for more-direct adjustments to valuation based on changes in the number of empirically observed post-operative visits (or other inputs, such as physician time).
