Identification of Surgeon Outliers to Improve Cost Efficiency: A Novel Use of the Centers for Medicare and Medicaid Quality Payment Program

Abstract

Background:

Bundled payment programs for total joint arthroplasty (TJA) have become popular among both private and public payers. Because these programs provide surgeons with financial incentives to decrease costs through reconciliation payments, there is an advantage to identifying and emulating costefficient surgeons. The objective of this study was to utilize the Centers for Medicare and Medicaid Services (CMS) Quality Payment Program (QPP) in combination with institutional data to identify cost-efficient surgeons within our region and, subsequently, identify cost-saving practice patterns.

Methods:

Data was obtained from the CMS QPP for total knee arthroplasty (TKA) and total hip arthroplasty (THA) surgeons within a large metropolitan area from January 2019 to December 2021. A simple linear regression determined the relationship between surgical volume and cost-efficiency. Internal practice financial data determined whether patients of identified surgeons differed with respect to x-ray visits, physical therapy visits, out-of-pocket payments to the practice, and whether surgery was done in hospital or surgical center settings.

Results:

There were 4 TKA and 3 THA surgeons who were cost-efficiency outliers within our area. Outliers and nonoutlier surgeons had patients who had similar body mass index, American Society of Anesthesiologists Physical Status Score, and age-adjusted Charlson Comorbidity Index scores. Patients of these surgeons had fewer x-ray visits for both TKA and THA (1.06 versus 1.11, P < .001; 0.94 versus 1.15, P < .001) and lower out-of-pocket costs ($86.10 versus $135.46, P < .001; $116.10 versus $177.40, P < .001). If all surgeons performing > 30 CMS cases annually within our practice achieved similar cost-efficiency, the savings to CMS would be $17.2 million for TKA alone ($75,802,705 versus $93,028,477).

Conclusions:

The CMS QPP can be used to identify surgeons who perform cost-efficient surgeries. Practice patterns that result in cost savings can be emulated to decrease the cost curve, resulting in reconciliation payments to surgeons and institutions and cost savings to CMS.

Total joint arthroplasty (TJA) represents one of the costliest procedures in the Centers for Medicare and Medicaid Services (CMS) budget [1,2]. In 2014, TJA procedures accounted for over 7 billion dollars in total cost to CMS [3], and in 2017, TJA represented the highest expense of any inpatient procedure [4]. The economic burden of TJA is projected to increase exponentially over the next decade [5,6]. As a result, bundled payment programs and other payment models for TJA have become increasingly common among both private and public payers [2,7-11]. Because these programs provide surgeons with financial incentives to decrease costs and improve quality by tying outcomes and cost savings to reconciliation payments, there is an advantage to identifying and emulating cost-efficient surgeons. The CMS Quality Payment Program (QPP), which is publicly available data, is one way to gain insights into the relative efficiency of orthopaedic surgeons nationally.

In 2015, Congress established the QPP as part of bipartisan legislation to address the growth in physician fees over time [12]. Among its features, the QPP established a dataset of clinician participation and performance that is searchable at the level of Tax Identification Number and National Provider Identifier (NPI), allowing scrutiny of surgical outcomes and cost data for both practice groups and individual surgeons [12]. More specifically, it allows identification of surgeons and practice groups that perform better than their local counterparts with respect to costs and complications for surgeries performed on Medicare beneficiaries. The primary objective of this study was to determine whether QPP data could be used to identify the most cost-efficient surgeons (as measured by the ratio of observed to expected costs) within one large geographic region in the Northeast. Using internal data from our large group practice, we also assessed whether the practice patterns of these cost-efficient outlier surgeons differed from those of other surgeons within our practice.

Due to concerns that improvement in cost efficiency could be associated with poor outcomes, including complications [13,14], we secondarily assessed whether there was an association between cost efficiency and adverse sequelae of surgery among regional QPP arthroplasty surgeons.

Methods

Publicly available data was obtained from the CMS QPP for total knee arthroplasty (TKA) and total hip arthroplasty (THA) surgeons who performed > 30 of either procedure within a large metropolitan area in the Northeast United States from January 2019 to December 2021. Data included case volume, National Provider Identification (NPI) numbers, expected costs based on patient and surgical factors as determined by CMS, observed costs to CMS, and the sequela rate [1]. The CMS defines a sequela as being applied to subsequent complications or conditions directly related to the surgery [1].

To identify cost-efficient outlier surgeons, we performed a simple linear regression to determine the relationship between surgical volume and cost efficiency, as measured by the log-transformed ratio of observed to expected costs. We identified outlier surgeons using a threshold of 2 standard deviations from this regression line. The NPI numbers were used to identify these surgeons. For those surgeons within our large group practice (N = 35), an internal database of utilization and payment data was then queried to determine whether patients of these low-cost surgeons differed from high-cost surgeons with respect to x-ray visits, physical therapy visits, and out-of-pocket payments to the practice using t-tests. There is no direct financial reward for surgeons in the institution to be cost-efficient. Under the current practices, in which the institution does not participate in CMS bundled payments for care improvement (BPCI), there is no direct reward for the cost-efficient surgeons. Previously to the study years, in 2018, our institution was participating in the BPCI and BPCI-Advanced. During this time, all profits from CMS bundles went directly to the institution and not the individual surgeon. The only benefit for highly cost-efficient surgeons is that they can do more cases and get paid more that way. However, in individual case, being costefficient does not directly result in financial rewards for the surgeon.

For the surgeons at our institution, TKA and THA cases were performed in a combination of hospital-based and surgical centers. The surgical centers are both hospital- and physician-owned. We calculated the potential decreased cost to CMS if those results could be duplicated across our institution’s surgeons. Simple linear regression was then used to assess whether there was an association between cost efficiency and adverse sequelae of surgery. These analyses were carried out separately for THA and TKA. Significance was defined a priori as an alpha level of < 0.05. All statistical analyses were performed using R Studio (Version 4.1.3, Vienna, Austria).

Results

Data for 77 TKA and 53 THA surgeons were available for review. Of these, 28 TKA surgeons (36%) and 20 THA surgeons (38%) were from our institution. Among surgeons performing > 30 qualifying cases within the period of study, our institution performed 51% (n = 5,418) of TKAs and 63% (n = 4,651) of THAs in our geographic area. There was a weak correlation between surgical volume and cost-efficiency for TKA (r −0.19, estimate: 0.02 per 100 cases, confidence interval (CI) [−0.03 to −0.003], P .09) and a weak-to-moderate, statistically significant correlation for THA (r = −0.39, estimate: −0.04 per 100 cases, CI [−0.06 to −0.01], P = .004).

We identified 4 TKA and 3 THA surgeons who were high-cost-efficiency outliers within our metropolitan area (Figures 1 and 2). Of these, 3 of the 4 TKA surgeons (Surgeon 1, Surgeon 3, and Surgeon 4) and all 3 THA surgeons (Surgeon 3, Surgeon 4, and Surgeon 5) were from our institution. These surgeons performed 43, 88, and 252 QPP-eligible TKAs, and 75, 224, and 91 THAs, respectively.

Fig. 1.

Identification of Highly Cost-Efficient Total Knee Arthroplasty Surgeons within the Metropolitan Area.

Fig. 2.

Identification of Highly Cost-Efficient Total Hip Arthroplasty Surgeons within the Metropolitan Area.

Financial data for the 3 total knee surgeons within our practice demonstrated that patients of these surgeons had fewer x-ray visits (mean 1.06 versus 1.11, P < .001) and lower out-of-pocket costs (mean $86.10 versus $135.46, P < .001) than their colleagues (Table 1). Outlier surgeons had a lower utilization of surgical centers than nonoutliers (mean 24.5 versus 30.3%). Among total hip surgeons, the cost-efficient outliers had fewer PT visits (mean 0.01 versus 0.28, P < .001), fewer x-ray visits (mean 0.94 versus 1.15, P < .001), and lower self-pay amounts (mean $116.10 versus $177.40, P < .001) (Table 2). These surgeons also had a lower utilization of surgical centers than their nonoutlier counterparts (mean 14.4 versus 21.9%).

Table 1.

Patient and Cost Characteristics of Outlier and Nonoutlier Surgeons Performing TKA in a Single Practice.

Characteristic	Outlier (n = 4)	Nonoutlier (n = 24)	P Value
Age	67	68	<.001
BMI	31.6	31.3	.25
ASA	2.39	2.43	.06
Age-adjusted CCI	4.09	4.13	.11
PT visits	1.73	2.05	.07
X-ray visits	1.06	1.11	<.001
Self-pay amount	$86.10	$135.46	<.001
Hospital (inpatient) (%)	75.5	69.9	-
Surgical center (outpatient) (%)	24.5	30.3	-

TKA, total knee arthroplasty; BMI, body mass index; ASA, American Society of Anesthesiology Score; Age-Adjusted CCI, age-adjusted charlson comorbidity index.

Table 2.

Patient and Cost Characteristics of Outlier and Nonoutlier Surgeons Performing THA in a Single Practice.

Characteristic	Outlier (n = 3)	Nonoutlier (n = 17)	P Value
Age	65	66	.21
BMI	29.2	29.5	.36
ASA	2.32	2.40	.08
A-CCI	4.29	4.37	.07
PT visits	0.01	0.28	<.001
X-ray visits	0.94	1.15	<.001
Self-pay amount	$116.10	$177.40	<.001
Hospital (inpatient) (%)	85.6	78.1	-
Surgical center (outpatient) (%)	14.4	21.9	-

TKA, total knee arthroplasty; BMI, body mass index; ASA, American Society of Anesthesiology Score; Age-Adjusted CCI, age-adjusted charlson comorbidity index.

Outlier surgeons had TKA patients who were younger by one year (mean 67.0 versus 68.0, P < .001). There were no statistically significant differences in the body mass index of patients in the outlier and nonoutlier groups for either THA or TKA. Neither The American Society of Anesthesiologists Physical Status Score (ASA) nor age-adjusted Charlson Comorbidity Index (CCI) differed between the TKA or THA patients of outlier and nonoutlier surgeons.

Our institution was markedly more cost-efficient than other area practices (Figures 3 and 4). Thus, when scrutinizing our internal database of utilization and payments, these outlier surgeons were compared to other surgeons within their own high-performing practice, not all surgeons participating in QPP.

Fig. 3.

Total Knee Arthroplasty Cost Performance of Author’s Practice versus Other Regional Practices.

Fig. 4.

Total Hip Arthroplasty Cost Performance of Author’s Practice versus Other Regional Practices.

For TKA, higher cost efficiency was associated with a decrease in adverse surgical sequelae (r = 0.35, estimate 20.69, CI [8.3 to 33.1], P =.001) (Figure 5). For THA, there was not a statistically significant association between higher cost efficiency and a decrease in surgical sequelae, though the direction of association was similar (r = 0.22, estimate 16.2, CI [−3.8 to 36.3], P = .11) (Figure 6). If all arthroplasty surgeons performing > 30 CMS cases annually within our practice achieved the cost efficiency of the outlier surgeons, the savings to CMS would be $16.5 million for TKA alone ($76,496,191 versus $93,028,477) and $8 million for THA alone ($52,325,005 versus $60,320,826).

Fig. 5.

TKA Cost-efficiency is Associated with Fewer Surgical Sequelae among QPP Participants. TKA, Total knee arthroplasty; QPP, quality payment program.

Fig. 6.

THA Cost-efficiency is Associated with Fewer Surgical Sequelae among QPP Participants. TKA, Total knee arthroplasty; QPP, quality payment program.

Discussion

This study highlights a novel use of the CMS QPP to identify highly cost-efficient surgeons within a geographic area. Our findings also demonstrate that patients with these cost-efficient outliers undergo fewer x-ray visits, see PT less frequently, and have lower out-of-pocket costs.

Practice patterns may not be the sole reason for the cost efficiency of outlier surgeons. An explanation for the cost efficiency of the outlier surgeons is their surgical volume, with each of the outlier surgeons at our institution performing at least 150 annual TJAs with a range of (150 to 373) annual cases. Contemporary literature has highlighted that higher-volume orthopaedic surgeons are correlated with lower operative times, blood loss and transfusion rates, and 30-day emergency department returns [15,16], and it is possible that the surgeon outliers from our institution have similar cost-savings. Another consideration for improved cost-efficiency is the inpatient or outpatient setting for TJA. Surgical centers allow for accelerated discharge and minimize the length of stay, which markedly alters costs [17]. Although surgeries included in the study took place in a combination of hospitals and surgical centers, the cost-efficient surgeon outliers had lower utilization of outpatient surgical centers than their nonoutlier counterparts for both TKA and TKA procedures (Tables 1 and 2).

Though patient factors may have been responsible, there were few differences in the patient characteristics of outlier surgeons relative to their colleagues, and these would have been captured by CMS as part of their determination of expected costs. Although the patients of outlier surgeons were 1 year younger than their counterparts, this is not a clinically meaningful difference, and there was no difference in patient body mass indices between the groups. The patients of outlier and nonoutlier TKA surgeons had similar ASA scores (2.39 versus 2.43, P = .06). Similarly, THA patients of outlier and nonoutlier surgeons had similar ASA scores (2.32 versus 2.40, P =.08).

Additionally, when age-adjusted CCI scores were compared between outliers and nonoutliers, there was no difference in scores for both TKA and THA patients. It is possible that outlier surgeons were identified by statistical chance based on a random distribution of cost-efficiency around a mean. However, 2 surgeons (Surgeon 3 and Surgeon 4) were identified as outliers for both hip and knee arthroplasties, which suggests that the outlier designation is nonrandom. It is likely that the cost-savings result from operative factors and perioperative decisions rather than patient demographics or chance alone.

Cost-optimization strategies have flourished in the era of alternative payment models, but in the arthroplasty literature, actionable changes have largely focused on decreasing the length of inpatient stays, postacute care costs, and skilled nursing facility stays [14,18-21]. To date, published studies have largely ignored the role of individual surgeons in reducing costs or the increased marketability of these surgeons when operating within a bundled payment program or other cost-sharing system. Emulating the strategies of highly cost-efficient surgeons could lead to immediate, tangible cost savings to payers and patients and increased model-dependent gainsharing payments to surgeons [2,22-24].

Arthroplasty surgeons have seen their reimbursements decrease despite an increase in other costs related to joint arthroplasties [25-27]. However, within a system with sufficient gainsharing, assuming a role as a steward of health care spending can be advantageous to the surgeon. Cost-efficient surgeons become more valuable when hospital systems benefit directly from their associated reconciliation payments. The QPP allows these surgeons to be identified by name.

Our study also found that increasing cost efficiency was not associated with increased complications and adverse events (Figures 5 and 6). The directionality of this association remains uncertain, as patients who have adverse sequelae of surgery also tend to be the costliest to treat. Nevertheless, there did not appear to be an immediate safety concern in providing lower-cost care.

There are several potential limitations to our work. This was a retrospective study, and patients were not randomized to surgeons operating within our region. Surgical decision-making is complex, and the data set arguably lacks the granularity to make direct comparisons in cost efficiency between surgeons or practice groups. However, this is precisely what the CMS program aims to do in calculating an expected procedural cost, and thus this limitation represents our reliance on the data as presented by the country’s largest payer. The CMS data is also limited in its ability to capture patients who present at outside facilities. This invited the potential for inaccuracies in patient outcomes and complication rates, thus impacting costs. Within our institutional data, we included information for submitted claims, which may not reflect a patient’s entire financial reality. For example, we include x-ray and PT visits that were completed within our institution and are available in the billing databases. Through the institutional nurse navigator program, data regarding patient attendance at outside facilities is recorded, as the program can search for and gather patient records of nearby centers. Limitations include patient presentation to centers unable to be collected by the nurse navigator program, which has the potential to alter complication, readmission, and cost data. The surgeons identified are specific to the northeast region, and the results of our study may not be generalizable to health care systems in other geographies. Within our institution, every surgeon was a fellowship-trained joint surgeon, most of whom have a high procedural volume, thus potentially limiting generalizability to other hospital networks and surgeons. Also, the study highlights cost-efficient surgeons and their patterns for reducing costs, such as fewer PT and x-ray visits, but does not elaborate on the operative techniques or specific postoperative decisions that may enable greater cost efficiency.

Conclusions

The demand for THA and TKA procedures is projected to increase exponentially, resulting in large financial implications for payers [5,7,10]. Our study has identified the most cost-efficient surgeons within a geographic region and illuminated cost-saving practice patterns associated with those surgeons, including reduced x-ray and PT visits, along with lower out-of-pocket spending for patients. Highly cost-efficient surgeons, who are valuable to hospital systems, can be identified by name using the QPP database. Providers and practices can use this novel approach to the QPP data to compare performance against others. More granular data about surgical and perioperative decision-making is necessary to identify additional cost-saving strategies.