CMS has recently proposed to remove total knee arthroplasty (TKA) surgical procedures from the Medicare "inpatient-only" list, which would allow those procedures to be performed on an outpatient basis. Observers have noted that this change would have a significant effect on the financial results in the BPCI and CJR programs of the remaining inpatient episodes, since the patients who continue to receive inpatient surgery are expected to be more clinically intense than those who received outpatient surgery and who are therefore no longer included in the major joint replacement (MJR) episodes. As lower-intensity patients left the bundled payment program, the historically-based targets (which were created including the clinically-simpler patients) would no longer reflect the population during the performance period. Consequently, hospitals that performed a significant number of TKA procedures on an outpatient basis would experience a significantly higher-complexity mix of the patients remaining in the BPCI and CJR programs, and are therefore expected to experience greater financial losses.

Effect of the policy change on episode costs

We performed some preliminary analysis on the effect of this change in episode composition, and the results are shown in the graphic below. This analysis makes several gross approximations about the TKA patients who would receive surgery on an outpatient basis, assuming that those patients constituted the lowest-cost 20% of all TKA patients. In this analysis we included only TKA patients in DRG 470, since patients in DRG 469 are unlikely to receive outpatient surgery. We used a 20% estimate because some of our hospitals are performing about 20% of their commercial MJR cases on an outpatient basis. As shown below, the outpatient-eligible cases have an average episode cost of almost $10,000 below the average of all episodes, while the 80% of episodes that would remain in an inpatient setting have an average cost about $4000 above the average of all cases. Assuming the target was equal to the average cost of all episodes, moving the bottom 20% of episodes into an outpatient setting would create an average loss of approximately $4000 per episode on all remaining episodes.

Approaches to mitigation

Several approaches to mitigating this issue have been proposed by various parties, two of which are described below. In one proposed approach targets are adjusted to accommodate the removal from BPCI and CJR of patients receiving outpatient surgery. In the other approach, outpatient surgery is included as an episode-initiating event for those programs, and a new target is developed for episodes initiated by outpatient surgery. The latter approach has several other advantages related to retaining in the episodes patients upon whom outpatient surgery is performed.

Revised targets to reflect only inpatient-remaining episodes

In one approach, CMS would estimate the effect of removing the outpatient cases from the historical baseline data and would recast the targets based only on the episodes that were expected to remain in an inpatient setting. We see several problems with this approach:

• Identifying outpatient-eligible cases – To perform this analysis, CMS would need to identify cases occurrint during the baseline that would have been performed on an outpatient basis. This analysis would need to be done using clinical criteria available from claims data. However, several hospitals have told us that the determination of surgical setting for these patients is not strictly based on clinical issues, but rather on such local issues as physician preference and operating room availability. These factors would be impossible to simulate retrospectively.
• Validating the outpatient identification methodology – A methodology could be developed using clinical criteria that would attempt to identify outpatient-eligible patients during the baseline; however because outpatient surgery was not performed during that period there is no way to validate that these criteria would actually identify outpatient-eligible patients.
• Hospital-specific variations – The use of outpatient surgery by individual hospitals could vary widely depending on facility availability, surgical preference, and other local factors. Therefore, these adjustments the targets would need to be specific to an individual hospital, which would be difficult to simulate.
• Opportunity for "gaming" – The target adjustment process would have to assume a certain percentage of cases would be performed on an outpatient basis, since the target is established to reflect the remaining inpatient cases. However, hospital could "game" this process by performing fewer outpatient surgeries than were estimated in the target-setting process. This would apply the higher targets to the simpler cases, and be financially advantageous to the hospital. However, creating this financial incentive would be counterproductive to the implementation of outpatient surgery for TKA. (This issue would be mitigated by the “two-midnight rule” discussed below, which may classify certain procedures as outpatient regardless of the setting in which they were performed.)
• Loss of episodes - TKA patients who received outpatient surgery would no longer be counted as episodes in BPCI and CJR. This would create lower episode volumes, with all of the related disadvantages. It would also remove these cases from the BPCI and CJR data and also from the gainsharing rules that apply to these programs.

Adding outpatient surgery cases to the MJR episodes

Another proposed approach is to include the TKA cases performed on an outpatient basis in the bundled payment program, and assign different targets to these cases adjusting for the difference between the OPPS payment and the DRG payment. Adding outpatient surgery episodes to the CJR program was suggested in the comments to the proposed rule, and at that time CMS opined that such a move was unnecessary because these surgeries were on the inpatient-only list.  However, that issue would become moot with the proposed remove of those surgeries from that list, which means that CMS might now be willing to consider such a change.

This approach assumes that the postacute cost of outpatient MJR patients would be the same as the postacute cost for those same patients if surgery was performed in an inpatient setting. In conversations with several hospitals, we have heard that these low-complexity patients who would be eligible for outpatient surgery have very little postacute care, and that the protocol for designing their postacute services would be identical regardless of the surgical setting. If this is true, the adjustment to the Medicare payment for the "index surgery" would be the only adjustment necessary to the target. This adjustment could be done for each hospital, since the DRG and OPPS payments are published for all hospitals.

The table below shows an example of this calculation. In this example a target rate of $22,000 (shown as “22”) has been established for all episodes under the current rules. Of that amount, $11,000 is the index admission payment, leaving the remaining $11,000 attributable to postacute cost. In the “current” example, the net loss is $23,000 for those eight patients.

In the “proposed” example, the three lowest-cost patients are assumed to have received outpatient surgery. The OPPS payment for the corresponding CPT code is $5000; therefore the outpatient target is established at ($22k - $11k DRG payment + $5k OPPS payment) $16,000. The patients receiving outpatient surgery have the same postacute costs as they would have in an inpatient setting, and therefore the overall financial results are identical.

This approach has the following advantages:

• It does not require making assumptions about outpatient-eligible patients during the baseline period,
• It doesn't create financial disincentives to perform surgery on an outpatient basis,
• It maintains the current number of MJR episodes in the bundled payment system, and
• It is simple, transparent, and straightforward to implement.

Other regulatory issues

Removal of TKA procedures from the inpatient-only list bring several other Medicare regulations into play. Currently, hospitals are bound by rules that reduce the DRG payment for patients having a one day hospital stay followed by certain post-acute services. In addition, if TKA procedures become outpatient-eligible, the "two-midnight" rule may apply.

Reduced DRG payments

Currently DRG payments for the index admission may be reduced if the inpatient stay is only one day and the patient receives home health services immediately after discharge. The effect of this issue on the methodology proposed above would be that the DRG payment to be removed from the episode target would not be constant for a specific hospital; instead it would be lower for small number of patients. In our experience, approximately 13% of TKA patients experience a one day length of stay, so an appropriate factor could be developed to adjust the DRG payment.

Two-midnight rule

The "two midnight rule" applies to inpatient stays that do not include two midnights. Absent other considerations, these stays may be reclassified by CMS as outpatient services and be paid at the OPPS rate rather than at a DRG rate. Currently, TKA procedures are protected from this rule because they are on the inpatient-only list, which creates an exception. However, under the proposed change many TKA inpatient procedures may be reclassified as outpatient services, although CMS has postponed this chance for two years for TKA surgeries. This would remove these procedures from the bundled payment programs, with the resulting elimination of data, gainsharing rule exceptions, and other positive aspects of these programs. This issue supports the inclusion of outpatient TKA procedures in the BPCI and CJR MJR episodes.

Physician incentive issues

An additional issue related to the reclassification of TKA patients from inpatient to outpatient, and therefore no longer being part of a bundled payment episode, is that orthopedists could no longer receive gainsharing payments for these patients. In some cases these payments are significant and can add an additional 50% to the surgeon’s Medicare payment. Since the selection of surgical setting is almost entirely made by the surgeon, it’s unlikely that surgeons participating in gainsharing arrangements would choose to perform any outpatient procedures on TKA patients. This would obviously be counterproductive.

Hopefully this analysis will be helpful in discussing ways of dealing with this proposed change to Medicare rules.