Implications of the Merit-Based Incentive Payment System for Urology Practices

Abstract

Objective

To determine the implications of the merit-based incentive payment system (MIPS) for urology practices. MIPS is a Medicare payment model that determines whether a physician is financially penalized or receives bonus payment based on performance in four categories: quality, practice improvement, promotion of interoperability, and spending.

Methods

We performed a cross-sectional analysis of urologist performance in MIPS for 2017 and 2019 using Medicare data. Urologist practice organization was categorized as single-specialty (small, medium, large) or multispecialty groups. MIPS scores were estimated by practice organization. Logistic regression models were used to examine the association between urology practice characteristics, including proportion of dual eligible beneficiaries, and bonus payment adjustment as defined by Medicare methodology. Rates of consolidation (movement from smaller to larger practices) between 2017 and 2019 were compared between those who were and those who were not penalized in 2017.

Results

Urologists in small practices performed worse in MIPS and had a significantly lower adjusted odds ratio of receiving bonus payments in both 2017 and 2019 compared to larger group practices (odds ratio [OR] 0.04, 95% confidence interval [95%CI] 0.03–0.05 in 2017 and OR 0.37, 95%CI 0.30–0.47 in 2019). Increasing percent of dual eligible beneficiaries within a patient panel was associated with decreased odds of receiving bonus payment in both performance years. Urologists penalized in 2017 had higher rates of consolidation by 2019 compared to those who were not (14% vs. 5%, p<0.05).

Conclusion

Small urology practices and those caring for a higher proportion of dual eligible beneficiaries tended to perform worse in MIPS.

Introduction

In 2017, the Centers for Medicare and Medicaid Services (CMS) implemented the Merit-Based Incentive Payment System (MIPS) as part of the Medicare Access and CHIP Reauthorization Act to improve quality and reduce the cost of healthcare. MIPS determines whether a physician’s Medicare reimbursement is reduced or enhanced based on performance in four categories: quality, practice improvement, promoting interoperability, and spending.¹ With the exception of spending, which CMS calculates, physicians must collect and report measures for each category. CMS intends to raise performance thresholds and payment adjustments annually. Consequently, incentives for performing well within the program are becoming stronger.¹ Most urologists are required to participate in MIPS and will be subject to these incentives,² underscoring the importance of the program for those caring for a large number of Medicare beneficiaries.

Some worry performance in MIPS is heavily influenced by proficiency in reporting and characteristics of the patient panel rather than the quality of care delivery itself.³ This may result in disproportionate penalization of those in small practices or those predominantly caring for vulnerable populations.⁴ Small practices, for instance, may lack administrative infrastructure and financial capital to track and report quality measures in MIPS, resulting in lower scores simply due to administrative constraints.⁵ Recent data suggests those in small general surgery groups experienced higher MIPS participation costs compared to those in larger practices ($16,017 per physician vs. $4,107 per physician).⁴ This is consistent with large practices benefiting from economies of scale, allowing them to diffuse the burden of reporting across more surgeons. If small practices are subject to high reporting burden and MIPS penalization, they may be driven to join larger groups. This consolidation could have important implications such as increased healthcare costs and reduced access.⁶ Similarly, practices serving a large proportion of vulnerable patients may be excessively penalized in MIPS. Dual eligible beneficiaries, or those qualifying for both Medicare and Medicaid, exemplify a vulnerable group who have high social risk and worse health outcomes.⁷ Physicians caring for this population may not be able to meet MIPS quality measures, resulting in worse performance. Studies of hospital-based pay-for-performance programs similar to MIPS demonstrated those primarily serving high social risk patients were disproportionately penalized.^8,9 These programs are often criticized for inadequate social-risk adjustment.¹⁰ Recognizing these concerns, CMS made several allowances aimed at easing the burden of participating in MIPS for smaller practices and those serving a large proportion of dual eligible beneficiaries.¹¹ These include reduced submission requirements for small practices and awarding up to five additional bonus points based on percent of dual eligible within a practice’s patient panel. However, the effects of these remedies are unclear.

For these reasons, we performed a study to understand the association between urology practice organization and performance in the MIPS program. We assessed relationships between urologist practice organization, share of dual eligible beneficiaries cared for by the practice, and the overall MIPS score. Furthermore, we characterized the association between MIPS score and subsequent urologist consolidation. We hypothesize that small urology practices and those with higher percent of dual eligible beneficiaries would perform poorly in MIPS. Similarly, we expected penalization of small practices in MIPS would be associated with higher rates of consolidation.

Methods

Data Sources and Study Population

We conducted a cross-sectional study of urologist performance in MIPS for 2017 (first performance year) and 2019 (most recent year data was available). We used the Medicare Data on Provider Practice and Specialty (MD-PPAS) file to identify National Provider Identification (NPI) numbers and tax identification numbers for urologists submitting Medicare claims during these years (9570 and 9707 urologists in 2017 and 2019, respectively).¹² With these data, urologists were assigned to their practice based on tax identification numbers using established methods.¹³ We identified individual urologist MIPS scores in the CMS Quality Payment Program file for 2017 and 2019 using NPIs.¹⁴ We obtained additional physician-level data by linking NPIs to the publicly-available Medicare Provider & Utilization Data file.¹⁵

MIPS Performance Categories and Scoring

The composite MIPS score is based on performance in four categories: quality, improvement activities, promoting interoperability, and spending. Apart from spending, providers are required to submit a specific number of measures for each category. Providers receive points for each submitted measure based on benchmarks set by CMS. These points are summed and standardized to a 100-point scale to yield an individual component score. CMS calculates the spending component using various cost measures (e.g., episode-based measures and total per capita spending) and assigns a score based on national spending benchmarks. The composite score is a weighted sum of individual component scores. CMS adjusts weights annually, gradually increasing the weight of the spending component from 0% in 2017 to 15% in 2019 while decreasing the weight of the quality component from 60% in 2017 to 45% in 2019.¹

Using the composite score, CMS adjusts per-claim reimbursement rates which take effect two years after the performance year. Physicians may receive either a penalty, no adjustment, a positive adjustment, or an additional bonus adjustment. The specific scoring thresholds for each adjustment are shown in Supplemental Table 3.

Exposure and Outcome

Our primary exposure of interest was practice organization, in which urologists participating in MIPS were grouped into single specialty and multispecialty group practices using established methods.¹⁶ Briefly, single specialty practices (i.e., consisting of urologists) were stratified by size, with small practices containing 1–2 urologists, medium practices having 3–10 urologists, and large practices having 11 or more urologists. This size division was based on prior literature.^6,16 Multispecialty group practices were typified by a practice containing at least 1 primary care physician and were inclusive of urologists employed by a hospital system.

Our primary outcome of interest was bonus payment adjustment based on composite MIPS score at the urologist level.

Statistical Analysis

First, we performed descriptive statistics for MIPS performance and for physician-level covariates for both performance years. We used chi-squared test to assess for differences in proportions and two-tailed t-test to assess for difference in means. One-way analysis of variance was used when comparing 3 or more means.

Next, we examined differences in overall MIPS score and individual component scores between the two performance years, stratified by practice organization. To compare longitudinal changes in scores, we examined differences in scores among those that maintained the same practice organization and reported in both 2017 and 2019 (76% of urologists in 2017).

We used multivariable logistic regression models to determine the likelihood of receiving a bonus payment. We modeled MIPS performance for each performance year independently. We chose bonus payment for our dependent variable because no urologists received penalties in 2019. Within the models, we adjusted for physician-level covariates including total beneficiaries, average beneficiary age, practice rurality, average standardized payments, percent of minorities in patient panel (defined as African American, Hispanic, Native American, or other), average Hierarchical Condition Category (HCC) risk score of patient panel, percent of dual eligible beneficiaries, years in practice (based on medical school graduation year), and region of practice. We hypothesized small practices would perform worse in MIPS and we suspected small practices caring for a larger proportion of dual eligible beneficiaries would perform even worse. To further understand the impact of the proportion of dual eligible beneficiaries, we calculated predicted probabilities of receiving bonus payment for each practice organization as a function of proportion of dual eligible beneficiaries in a patient panel. We examined for heterogeneity in effect of percent of dual eligible by testing its interaction with practice organization on MIPS performance. Of note, beginning in 2018, CMS included risk adjustment by awarding up to five additional bonus points based on proportion of dual eligible beneficiaries within a patient panel and HCC risk score. To characterize the impact of this adjustment, we fit a separate model in which 2017 and 2019 data were combined and an interaction term between year and percent of dual eligible beneficiaries was included to determine if there was a differential impact across the two years. Regression models utilized all observations for which complete data was available (>95% for 2017 and 89% for 2019). Regression analysis was repeated for 2019 using imputation for missing data to ensure stability of estimates. Predicted probabilities were derived from models using margins postestimation command in STATA.

Finally, we determined the association of penalization in MIPS on subsequent urologist consolidation. We stratified the cohort into those penalized in 2017 based on CMS criteria described above. After assigning urologists to their respective practice organizations for 2017 and 2019, we determined rates of practice organization change among those penalized compared to those not penalized. We determined rates of consolidation (change of practice organization: small practice to medium/large/multispecialty groups, medium practice to large/multispecialty groups, and large practice to multispecialty groups) stratified by penalization in 2017.

Analyses were carried out using STATA 17 (College Station, Tx). All tests were two-sided with probability of type 1 error set at 0.05. This study was deemed not regulated by our Institutional Review Board.

Results

70% of urologists participated in MIPS in 2017 (6,740 out of 9570), compared to 73% in 2019 (7,092 out of 9,707). There was a decrease in number of beneficiaries per provider from 2017 to 2019 (676 vs. 627, p<0.05). The remaining physician and patient-panel characteristics were similar between the two performance years (Table 1). Urologist performance in MIPS improved from 2017 to 2019 (Table 1). The greatest score improvement was in the quality component with a mean increase of 11 points (standard deviation [SD] 28).

Table 1.

Physician-level characteristics of MIPS participating urologists. Data includes those participating in 2017 and those participating in 2019.

Variable	2017 Total, Mean (SD)	2019 Total, Mean (SD)	p-value
No.	6,740	7,092
Physician Sex, No (%)			<0.05
Male	6,195 (92)	6,192 (91)
Female	514 (8)	623 (9)
Years in Practice	24 (11)	20 (12)
Avg beneficiary age per physician patient-panel, yr	74 (2)	74 (3)	0.02
Num beneficiaries per physician patient-panel	676 (386)	627 (400)	<0.05
Submitted charges per	858,165	842,278	0.18
physician, $	(664,442)	(733,524)
Region of practice, No (%)			0.34
Northeast	1,453 (22)	1,532 (22)
Midwest	1,373 (20)	1,390 (20)
South	2,673 (40)	2,820 (40)
West	1,216 (18)	1,338 (19)
Race distribution of physician patient-panel, %(SD)			0.94
White	85 (16)	84 (16)
Non-white	16 (18)	16 (18)
Avg beneficiary HCC Risk Score of physician patient-panel	1.5 (.33)	1.4 (.41)	0.16
RUCA classification of physician practice, No (%)			0.98
Urban	6,085 (91)	6,179 (91)
Rural	616 (9)	626 (9)
Percent of dual eligibles of physician patient-panel, N(%)	16.5 (13)	17 (13)	<0.05
Practice Organization, No (%)			<0.05
Small	1,109 (17)	854 (12)
Medium	1,236 (18)	1,066 (15)
Large	1,376 (20)	1,538 (22)
MSG	3,019 (45)	3,634 (51)
MIPS Category scores			<0.05
Quality Category	76 (32)	90 (16)
Promoting Interoperability	80 (35)	73 (20)
Improvement Activity	87 (32)	98 (10)
Cost	------- *	72 (9)
Composite MIPS Score	78 (30)	86 (17)

^*The cost component was not scored in 2017.

Abbreviations: MSG: Multispecialty group; RUCA: Rural-Urban Commuting Area; HCC: Hierarchical condition category

When stratified by practice organization, small practices did worse compared to medium, large, and multispecialty practices in both performance years (Supplemental Table 1). Specifically, compared to multispecialty groups, small practices performed 45 points worse in 2017 (p<0.05) and 20 points worse in 2019 (p<0.05). In terms of individual component scores, small practices performed worse in quality, improvement activities, and promoting interoperability in both performance years relative to multispecialty groups, even after adjustment for physician-level factors. However, small practices did improve across all individual component scores in 2019. Medium, large, and multispecialty group practices had small improvements in mean quality and improvement activity component scores, with a decrease in promoting interoperability scores.

Next, we modeled the relationship between bonus payment adjustment and urologist practice organization. After adjusting for physician-level covariates, urologists in small practices had a significantly reduced odds of receiving bonus payments in both performance years, (adjusted odds ratio [aOR] 0.04, 95% confidence interval [95% CI] 0.03 to 0.05] in 2017 and (aOR 0.37 [95%CI 0.30 to 0.47]) in 2019, compared to those in large group practices (Table 3). Those who submitted more charges to Medicare had increased odds of receiving a bonus in 2017 (aOR 1.25 [95%CI 1.05 to 1.48]), but reduced odds of receiving one in 2019 (OR 0.63 95% CI [0.55 to 0.74]). In both years, increasing percent of dual eligible beneficiaries, higher average beneficiary age, and increasing years in practice were associated with decreased odds of receiving bonus payment. We examined the impact of dual eligible beneficiaries further by calculating predicted probabilities of receiving bonus payment stratified by practice organization as a function of percent dual eligible in a patient panel for each performance year (Figure 1). As percent of dual eligibles increased in the patient panel, the predicted probability of receiving bonus payment declined, across all types of practice organization in both performance years. We assessed for heterogeneity in effect of percent of dual eligibles on practice organization and found no difference in effect for both performance years. This suggests the proportion of dual eligible beneficiaries in a patient panel does not differentially impact small versus larger practices. We determined if CMS initiation of risk adjustment beginning in 2018 altered the observed effect of dual eligible beneficiaries on bonus payment by using a single model for both performance years and interacting percent of dual eligibles with year. We found that the impact was less pronounced in 2019 compared to 2017 (p<0.05)

Table 3.

Adjusted odds ratios from multivariable logistic regression models, predicting bonus payment adjustment in 2017 and 2019.

	Adjusted OR 2017 (95% CI)	Adjusted OR 2019 (95% CI)
Practice Organization
Large	REF	REF
Small	0.04 (0.03 0.05)	0.37 (0.30 0.47)
Medium	0.11 (0.08 0.16)	0.97 (0.79 1.21)
MSG	0.50 (0.36 0.69)	2.13 (1.73 2.63)
Average Beneficiary Age	0.86 (0.82 0.90)	0.91 (0.87 0.95)
Rurality of Practice
Urban	REF	REF
Rural	0.93 (0.71 1.21)	1.16 (0.87 1.53)
Total Number of Beneficiaries	0.99 (0.99 1.00)	1.00 (1.00 1.00)
Annual standardized payment *	1.25 (1.05 1.48)	0.63 (0.55 0.74)
Percent Race Minority	1.00 (0.99 1.01)	0.99 (0.99 1.00)
Avg Patient HCC Score	0.93 (0.67 1.26)	1.09 (0.80 1.49)
Percent Dual Eligible †	0.98 (0.97 0.99)	0.99 (0.98 0.99)
Region of Practice
Northeast	REF	REF
West	1.19(0.95 1.50)	0.73 (0.58 0.92)
South	0.99 (0.81 1.22)	0.75 (0.61 0.92)
Midwest	1.73 (1.34 2.22)	1.31 (1.01 1.70)
Years in practice	0.98 (0.97 0.99)	0.98 (0.97 0.99)

^*Total Medicare standardized payment is log transformed

^†Odds ratio represents change in odds per 1% change in proportion of dual eligible within a patient panel.

Note: The odds ratios for each performance year represent individual models, and therefore direct odds ratio comparisons can only be made within performance years and not between performance years.

Abbreviations: OR: Odds ratio; MSG: Multispecialty group; RUCA: Rural-Urban Commuting Area; HCC: Hierarchical condition category

Figure 1.

Adjusted predicted probabilities from logistic regression models of receiving bonus payment as a function of percent of dual eligible beneficiaries in a patient panel for A) 2017 and B) 2019.

Abbreviations: MSG: Multispecialty group

Finally, we examined the impact of performance and participation in MIPS on rates of urologist consolidation. Relative to 2017, among urologists in MIPS (Table 1), there was a 5% decrease in reporting from small practices (p<0.05) and a 6% increase in reporting from multispecialty groups (p<0.05). From 2017 to 2019, 9.2% of MIPS participating urologists changed practice organization. Those who changed practice organization tended to have larger increases in composite MIPS score compared to those who did not (Supplemental Table 2). As shown in Table 2, a larger percent of those in small and medium practices changed practice organization compared to those in the other two groups (16–20% vs 4–6%). After stratifying by penalization in 2017, a higher proportion of those who were penalized tended to change practice organization (Table 2) and join larger groups (consolidated) compared to those who were not penalized (13.7% vs. 5.3%, p<0.05, Supplemental Figure 1). Furthermore, a higher percentage of those in small practices participating in MIPS joined larger groups compared to those in small practices not participating in MIPS (15.6% vs 5.1%, p <0.05).

Table 2.

Urologists in 2017 who changed practice organization in 2019, stratified by those who did and those who did not receive penalties in 2017.

	Overall (%)	Not Penalized (%)	Penalized (%)	p-value*
Small 851	Same 718 (84)	Same 514 (86)	Same 204 (81)
	Changed 133 (16)	Changed 84 (14)	Changed 49 (19)	0.05
Medium 1,062	Same 849 (80)	Same 785 (80)	Same 64 (72)
	Changed 213 (20)	Changed 188 (20)	Changed 25 (28)	<0.05
Large 1,285	Same 1,205 (94)	Same 1,204 (94)	Same 1 (33)	<0.05
	Changed 80 (6)	Changed 78 (6)	Changed 2 (67)
MSG 2,430	Same 2,340 (96)	Same 2,287 (97)	Same 53 (88)
	Changed 90 (4)	Changed 83 (3)	Changed 7 (12)	<0.05

^*p-value: Computed using chi-square test, comparing proportion of practice organization change between penalized and non-penalized groups.

Abbreviations: MSG: Multispecialty group

Discussion

Urologists in small practices performed worse in MIPS compared to those in larger group practices. Those penalized in the first year of MIPS were more likely to change practices, with those in small and medium practices being the most likely to join larger practices. When adjusting for physician-level variables, urologists in small practices and those with a higher proportion of dual eligible beneficiaries were less likely to receive bonus payments compared to those in larger practices. Interestingly, those who submitted more charges to Medicare had increased odds of receiving a bonus in 2017, but reduced odds of receiving one in 2019, consistent with CMS implementation of the spending component in 2018.

The findings from this study and those from other specialties collectively support the notion that small practices perform worse in MIPS, even after two years into the program.^17–19 While the reasons for this have not been elucidated, several possibilities exist. First, quality may be lower in these practices, thereby resulting in lower MIPS scores. However, studies examining the relationship between practice size and quality of care, including work examining urologist practice organization and quality of prostate cancer care, have shown similar levels of quality in process measures and clinical outcomes.^13,20,21 Second, small practices may lack resources to establish an infrastructure to collect, analyze, and report quality measures. For example, smaller general surgery practices reported higher MIPS participation costs compared to larger multispecialty groups.⁴ Importantly, physicians feel that the burden of reporting in MIPS is substantial, leading to burnout, and even forcing some physicians to retire prematurely.^4,22 Larger practices can share the fixed burden of reporting across more physicians and use their financial capital to establish an infrastructure to collect and report measures. Additionally, because physicians select the measures they report, larger practices may be better positioned to know how they perform across various measures and selectively report the ones in which they already excel.³ Third, incentives to participate in MIPS vary based on practice size and the percent of revenue from Medicare claims. For example, small practices whose Medicare payments represent a small proportion of their total revenue may choose not to make the investment to report in MIPS as the penalties may be inconsequential. Conversely, larger practices caring for many Medicare patients may see substantial changes in revenue depending on their performance.²⁴ However, our analysis suggests the relationship between practice organization and MIPS performance is independent of Medicare revenue.

In addition to disproportionately penalizing smaller practices, a concern of pay-for-performance programs like MIPS is inadequate adjustment for patient social risk. Prior analyses of the first MIPS performance year have demonstrated that physicians caring for a higher number of socially disadvantaged patients performed worse in MIPS.^25,26 Following the first performance year, CMS instituted a complex patient bonus, awarding up to five bonus points based on patient panel HCC risk scores and proportion of dual eligible beneficiaries.¹ We find practices caring for higher proportion of dual eligible beneficiaries were modestly less likely to receive a bonus payments in 2019, even after implementation of this remedy. This impact is less pronounced in 2019 compared to 2017, suggesting the adjustment may be providing some benefit to urologists caring for a higher proportion of dual eligible beneficiaries. It will be important to monitor the efficacy of this adjustment as CMS increases the threshold for penalization and the associated payment adjustments to ensure practices caring for higher risk patients are not disproportionately penalized. Penalization could lead to avoidance of high-risk patients to improve MIPS performance. Avoidance of this group can lead to reduced access to care and worsening of health outcomes.

Another potential consequence of MIPS is physician consolidation. Since the early 2000s, there has been increasing movement of physicians to larger group practices.⁶ Reasons for this include the financial challenges associated with operating a practice, electronic health record requirements, and meeting quality measures, such as those imposed by programs like MIPS.²⁷ Unsurprisingly, we see fewer small practices participating in MIPS over time, with increasing number of physicians joining multispecialty groups. Although we cannot establish a causal relationship between MIPS initiation and consolidation in the current study, we explored this idea by examining rates of consolidation among urologists in small groups participating in MIPS compared to those not participating in MIPS (e.g., those exempt by CMS criteria). We see that among small practices participating in MIPS, 16% joined larger practices while 5% of small practices not participating in MIPS joined larger practices. Furthermore, among those participating in MIPS, we find that those who received penalties in 2017 had a higher rate of consolidation compared to those who did not (14% vs 5%). This suggests MIPS, and performance within the program itself, is associated with consolidation. On one hand, consolidation may allow for better infrastructure to monitor physician practices and improve care delivery.²⁸ On the other hand, consolidation may reduce competition and drive up commercial healthcare prices, which can have varying impact on quality care delivered.²⁹ As CMS makes it more difficult for physicians to perform well in MIPS, it may impose even more pressure on small practices and hasten consolidation.

This study has several limitations. First, this is a cross sectional study and therefore we cannot infer causal relationships. However, we examined two performance years to understand trends and we controlled for confounders, including various physician patient panel characteristics. Second, our analysis is dependent on CMS public data, which may contain incomplete information, inaccuracies in reporting, or omission of reporting if CMS does not deem data suitable for public release. We cannot be certain all information is captured. However, by linkage to other datasets, we find at least 70% of urologists are contained within the MIPS public dataset which is consistent with prior estimates of participation rates.² Third, this study lacked patient-level data, and therefore we cannot ascertain whether poor performance in MIPS was reflective of poor quality of care, or rather a failure to effectively report within the program.

These limitations notwithstanding, findings from this study have important implications that are relevant to other pay-for-performance programs.^8,10 First, policymakers will need to ensure small practices are not disproportionately penalized. If they continue to be, an alternative may be to score MIPS performance relative to peer equivalents to allow for comparison across similar practice organizations. Second, following risk adjustments for dual eligible beneficiaries beginning in 2018, increasing percent of dual eligibles within a patient panel remained associated with worse performance within MIPS in 2019. Therefore, social risk adjustment may still be insufficient and incentives to care for these patients need to be stronger. Third, the impact of MIPS on consolidation will need to be monitored. While there are benefits to consolidation, there are potential negative consequences including increased cost of care, variation in quality, and reduced access to care.³⁰

In conclusion, this cross-sectional analysis of two MIPS performance years shows that small urology practices perform worse compared to larger groups. Those who perform poorly tend to have a higher proportion of dual eligible beneficiaries and change practices. As CMS continues to modify MIPS, it will be critical to ensure specific practices are not disproportionately penalized.