The Impact of Nurse Practitioner Attribution in Medicare Shared Savings ACOs

ABSTRACT

Objectives: Currently, the Medicare Shared Savings Program (MSSP) requires beneficiaries to have 1 or more visits with a qualifying physician to be attributed to an accountable care organization (ACO). Allowing primary care nurse practitioners (NPs) to serve as qualifying providers for the sake of attribution could be an effective way to expand access to the program but could potentially draw in sicker beneficiaries and increase an ACO’s benchmarks.

Study Design: This observational study assesses the potential impact of changing the MSSP attribution rules to include NPs and other advanced practice clinicians.

Methods: For this analysis, we ran MSSP attribution for the entire United States with and without including NPs as attribution-eligible providers. We then compared the change in attributed population with the change in CMS hierarchical condition category (HCC) risk scores.

Results: For the majority of ACOs, allowing NPs to serve as an attribution-eligible provider had little impact on the number of attributed beneficiaries or the mean HCC score. For a small number of ACOs (n = 18), the change increased attribution by more than 10%. The mean change in HCC score was 0.

Conclusions: Allowing NPs to serve as attribution-eligible clinicians would result in more attribution to MSSP ACOs without increasing the level of patient complexity, a positive impact for those interested in growing ACOs.

The American Journal of Accountable Care. 2023;11(4):20-29. https://doi.org/10.37765/ajac.2023.89475

Population health models are dependent on defined cohorts of patients who have sustained relationships with providers over time. In the Medicare Shared Savings Program (MSSP) accountable care organization (ACO) model, these cohorts are determined using claims-based attribution methods that link a beneficiary to the provider who accounts for the plurality of their primary care.¹ However, not all provider types are eligible to link beneficiaries to an ACO under the current attribution rules. Right now, a beneficiary must receive a primary care service from an ACO physician (primary care or specialist) to be attributed to an ACO. Patients receiving the entirety of their primary care from a nurse practitioner (NP) without a visit with an ACO physician cannot be assigned to an ACO.² In part, this is due to the lack of requisite information in claims data to differentiate between primary care and specialty-focused NPs.³ When NPs function as medical specialists, NP attribution could pull in sicker patients.

Other value-based payment models established by the Center for Medicare and Medicaid Innovation, such as Realizing Equity, Access, and Community Health (REACH), allow for alignment with a physician or an NP, but REACH ACOs account for less than a quarter of all CMS-based ACOs in 2023.^4,5

The growing supply of NPs represents an important source of primary care at a time when baby boomers are aging and retiring, increasing the demand for health care while simultaneously leaving the health care workforce.⁶ In fact, in 2019, NPs made up approximately 39% of ACO providers, up from 18% in 2013.⁷

In many ways, NPs are ideal clinicians to practice in population health models of care. They are known for a patient-centered practice style⁸ and frequently work with underserved populations. For example, they are often the usual source of care for racially and ethnically diverse patients and those dually eligible for Medicare and Medicaid^.9-11 Expanded NP attribution could help ACOs increase diversity. Study findings have also shown that patients attributed to NPs receive care that is of similar quality to care provided by physicians.¹² Although there is conflicting evidence from an unpublished study in the US Department of Veterans Affairs health care system,¹³ large-scale studies demonstrate that NP-driven primary care is less expensive than physician-driven care, in part because NPs use fewer and less costly services for similar patients.^14,15 The high-value practice style of NPs aligns well with the population health goals of ACOs.¹⁶

To better understand the impact of excluding NPs from ACO attribution on the overall size and clinical severity of the attributed population, this analysis examines the results of a simulation that broadens the MSSP attribution rules to allow NPs to serve as linking clinicians. It also assesses whether ACOs that would grow under NP attribution are different from ACOs that would be unaffected by this policy change.

METHODS

For this descriptive study, we used 100% of 2019 Medicare fee-for-service claims to simulate beneficiary attribution to MSSP ACOs using the traditional attribution method and a method that allows NPs to be primary linking clinicians.

To test the impact of this policy change, we replicated the MSSP attribution rules, assigning beneficiaries to the ACO or non-ACO provider group tax identification number (TIN) that accounted for the plurality of their primary care services (eAppendix [available at ajmc.com]) in 2019. As specified in the CMS 2021 methodology, the attribution first used primary care physicians (defined by TIN–national provider identifiers [NPIs] with a specialty designation of 01, 08, 11, 37, or 38). For all unattributed beneficiaries, a second pass through the data identified those with a plurality of primary care from ACO medical specialists. This second pass allows the ACO to capture more beneficiaries, but it also tends to bring in sicker individuals. Primary care delivered by an ACO NP counts toward MSSP attribution, but only if the beneficiary also sees an ACO physician. We then reran attribution, allowing NPs to serve as primary linking clinicians.

To determine beneficiary clinical severity, we calculated CMS hierarchical condition category (HCC) scores using version 21.¹⁷ This model produces multiple risk scores for each individual, including a community score for those who live at home, an institutional score for those who live in a nursing home, and an end-stage renal disease (ESRD) score for those who qualify for Medicare by virtue of the ESRD benefit. For this analysis, we retained the community score for all beneficiaries, regardless of eligibility category or place of residence. This allowed us to see the relative clinical intensity of each beneficiary, with higher scores for those with multiple chronic conditions and greater clinical complexity.

To assess the impact of allowing NPs to be primary linking clinicians, we calculated the percentage changes in the number of attributed beneficiaries and the mean ACO/TIN HCC score between traditional and NP-enhanced attribution methodologies. Next, we assessed the relationship between ACO characteristics (rural, use of home-based primary care, risk track, hospital affiliated, percent primary care providers [PCPs]) and having a large increase in attributed beneficiaries (defined as 6% growth or more; n = 73). An ACO is defined as rural if 35% or more of its beneficiaries live in a rural county. Hospital affiliation is defined as being “high revenue” in the MSSP public use file (PUF) for 2019. This means the ACO has total Medicare Part A and B revenue for all beneficiaries (both fee-for-service and ACO) that is 35% or more of total Part A and B expenditures of the ACO-assigned beneficiaries (total Part A and B revenue / ACO Part A and B expenditures ≥ 35%). Risk track is defined as 1 sided or 2 sided based on contract type in the 2019 PUF. Finally, home-based primary care (HBPC) is identified using Current Procedural Terminology codes in Medicare claims to find beneficiaries with 1 or more visit in the year. We then calculated HBPC users per 1000 beneficiaries for each ACO. Finally, for percent PCPs, we took the number of PCPs from the MSSP PUF and divided it by total providers associated with the ACO (primary + specialty care). For the analysis, we ran logistic regression models to predict “large increase” status, controlling for ACO size. Separate models were run for each characteristic.

RESULTS

In 2019, the vast majority of beneficiaries (10 million out of 10.4 million in the program) were attributed to an MSSP ACO by way of a primary care physician. When NPs were allowed to serve as primary linking clinicians, the vast majority of ACOs saw only modest growth in assigned beneficiaries of between 0% and 4%, with an overall mean growth of 3.4% (Figure). Of the 474 ACOs in our data set, 87 saw a 5% to 10% increase in attributed beneficiaries (18% of all ACOs) and 18 ACOs (4% of all ACOs), denoted as high-growth ACOs, increased their beneficiary counts by more than 10% (Table).

After controlling for ACO size, rural ACOs were by far the most likely to benefit from NP attribution (OR, 14.9; P < .0001). ACOs with a higher concentration of PCPs also had higher odds (OR, 1.98; P < .0001) of high growth in attributed beneficiaries, whereas hospital-affiliated ACOs (OR, 0.85; P < .0001) and those in 2-sided risk contracts (OR, 0.47; P < .0001) had lower odds of high growth. HBPC visits per beneficiary appear to have minimal impact, with an OR of essentially 1.0.

Overall, the mean change in HCC scores was 0, meaning that allowing NPs to be primary linking clinicians had very little impact on overall ACO beneficiary clinical severity. A few ACOs had an increase or decrease in HCC score of greater than 1%, but they accounted for less than 10% of all ACOs in our study. The largest change in either direction for a single ACO was just 3.1%.

DISCUSSION AND CONCLUSIONS

Currently, attribution in the MSSP program excludes NPs as primary linking clinicians. In fact, this is statutory and can only be changed through legislation such as the ACO Assignment Improvement Act of 2021,¹⁸ which did not get through Congress after its initial introduction and, at the time of writing, has not been reintroduced, or the Improving Care and Access to Nurses Act, which was reintroduced in the US House of Representatives in April 2023.¹⁹ Although CMS cannot change the attribution rules independently, it recently proposed extending the attribution window to 2 years, allowing ACO physicians to see the beneficiary any time during this longer period.20 This would effectively make it easier for NPs to link beneficiaries to ACOs, except in the case of a predominantly NP or NP-only ACO.

This descriptive analysis shows that allowing NPs to serve in this capacity has little impact on the overall clinical severity of patients attributed to MSSP ACOs. However, it does increase the number of beneficiaries assigned to an ACO and could potentially increase attribution stability over time by allowing those who only see an NP in a given year to remain attributed. ACOs likely to benefit most are those in rural areas or those with a large percentage of PCPs. In both cases, NPs appear to be managing their own panel of patients, expanding access to primary care. At the same time, hospital-affiliated and risk-bearing ACOs are less likely to see a big growth in attribution from this policy change. It may be easier for these larger ACOs to ensure that beneficiaries who see an NP also see an ACO physician.

When it comes to implementation, differentiating primary care from specialty care NPs is problematic. Estimates of the proportion of all NPs in primary care range widely, from 26% to 70%,^21,22 as it is difficult to identify clinical specialty through billing data. This could be done by changing the data collected on NP specialization, which is currently limited to a single code. Alternatively, members of this team have developed a claims-based method for identifying PCPs.²³ Although “incident to” billing, through which NP care is billed under a physician’s NPI, can hide NP services, we believe this is less common in ACO-based primary care, where a lower reimbursement rate for NPs (85% of the physician fee schedule) is beneficial because it lowers the total cost of care. Either of these techniques would lend confidence to policy makers and population health providers that including NPs as linking clinicians would not distort the intent of the current attribution methodology. In fact, this policy change would result in more attribution to MSSP ACOs, a positive impact for those interested in growing accountable care in Medicare.

Acknowledgments

The authors would like to thank Meera Srinivasan for providing enhanced analysis of factors associated with increased attribution. They also want to acknowledge and thank the anonymous reviewers who provided comments and expertise that helped enhance this paper.

Author Affiliations: Brandeis University (JP), Waltham, MA; Institute for Accountable Care (JP, AP, SS), Washington, DC; Boston College, Connell School of Nursing (MO-J), Chestnut Hill, MA.

Source of Funding: Self-funded.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (JP); acquisition of data (JP, AP); analysis and interpretation of data (JP, MO-J, AP); drafting of the manuscript (JP, MO-J, AP, SS); critical revision of the manuscript for important intellectual content (JP, MO-J, SS); administrative, technical, or logistic support (SS); and supervision (JP).

Send Correspondence to: Jennifer Perloff, PhD, Brandeis University, 415 South St MS035, Waltham, MA 02453. Email: perloff@brandeis.edu.

REFERENCES

1. Attribution: Principles and Approaches Final Report. National Quality Forum. December 2016. Accessed May 15, 2023. https://www.qualityforum.org/WorkArea/linkit.aspx?LinkIdentifier=id&ItemID=84236

2. Medicare Shared Savings Program: Shared Savings and Losses and Assignment Methodology Specifications. CMS. January 2022. Accessed June 10, 2023. https://go.cms.gov/412bbw6

3. O’Reilly-Jacob M. The murky waters of nurse practitioners and Medicare claims. Med Care. 2020;58(10):851-852. doi:10.1097/MLR.0000000000001406

4. Accountable Care Organization Realizing Equity, Access and Community Health (ACO REACH) performance year 2023 participant overview. CMS. Accessed January 24, 2023. https://innovation.cms.gov/media/document/aco-reach-py23-participants

5. Accountable care organizations. CMS. January 9, 2023. Updated March 6, 2023. Accessed November 8, 2023. https://data.cms.gov/medicare-shared-savings-program/accountable-care-organizations

6. Auerbach DI, Staiger DO, Buerhaus PI. Growing ranks of advanced practice clinicians—implications for the physician workforce. N Engl J Med. 2018;378(25):2358-2360. doi:10.1056/NEJMp1801869

7. Nyweide DJ, Lee W, Colla CH. Accountable care organizations’ increase in nonphysician practitioners may signal shift for health care workforce. Health Aff (Millwood). 2020;39(6):1080-1086. doi:10.1377/hlthaff.2019.01144

8. Kippenbrock T, Emory J, Lee P, Odell E, Buron B, Morrison B. A national survey of nurse practitioners’ patient satisfaction outcomes. Nurs Outlook. 2019;67(6):707-712. doi:10.1016/
j.outlook.2019.04.010

9. Everett CM, Thorpe CT, Palta M, Carayon P, Gilchrist VJ, Smith MA. Division of primary care services between physicians, physician assistants, and nurse practitioners for older patients with diabetes. Med Care Res Rev. 2013;70(5):531-541. doi:10.1177/1077558713495453

10. Poghosyan L, Carthon JMB. The untapped potential of the nurse practitioner workforce in reducing health disparities. Policy Polit Nurs Pract. 2017;18(2):84-94. doi:10.1177/1527154417721189

11. Cody R, Gysin S, Merlo C, Gemperli A, Essig S. Complexity as a factor for task allocation among general practitioners

and nurse practitioners: a narrative review. BMC Fam Pract. 2020;21(1):38. doi:10.1186/s12875-020-1089-2

12. Buerhaus P, Perloff J, Clarke S, O’Reilly-Jacob M, Zolotusky G, DesRoches CM. Quality of primary care provided to Medicare beneficiaries by nurse practitioners and physicians. Med Care. 2018;56(6):484-490. doi:10.1097/MLR.0000000000000908

13. Chan DC Jr, Chen Y. The productivity of professions: evidence from the emergency department. National Bureau of Economic Research working paper 30608. October 2022. doi:10.3386/w30608

14. Razavi M, O’Reilly-Jacob M, Perloff J, Buerhaus P. Drivers of cost differences between nurse practitioner and physician attributed Medicare beneficiaries. Med Care. 2021;59(2):177-184. doi:10.1097/MLR.0000000000001477

15. Perloff J, DesRoches CM, Buerhaus P. Comparing the cost of care provided to Medicare beneficiaries assigned to primary care nurse practitioners and physicians. Health Serv Res. 2016;51(4):1407-1423. doi:10.1111/1475-6773.12425

16. Liu CF, Hebert PL, Douglas JH, et al. Outcomes of primary care delivery by nurse practitioners: utilization, cost, and quality of care. Health Serv Res. 2020;55(2):178-189. doi:10.1111/1475-6773.13246

17. Risk adjustment. CMS. Updated September 6, 2023. Accessed November 8, 2023. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors

18. ACO Assignment Improvement Act of 2021, HR 6308, 117th Cong (2021). Accessed July 3, 2023. https://www.congress.gov/bill/117th-congress/house-bill/6308

19. I CAN Act, HR 2713, 118th Cong (2023). Accessed July 3, 2023. https://www.congress.gov/bill/118th-congress/house-bill/2713/text

20. CMS-1784-P: revisions to payment policies under the Medicare Physician Fee Schedule Quality Payment Program and other revisions to Part B for CY 2024. CMS. July 13, 2023. Accessed July 17, 2023. https://go.cms.gov/4a4FXZq

21. NP fact sheet. American Association of Nurse Practitioners. Updated November 2022. Accessed November 8, 2023. https://www.aanp.org/about/all-about-nps/np-fact-sheet

22. Primary care workforce projections. Health Resources and Services Administration. Updated October 2023. Accessed November 8, 2023. https://bhw.hrsa.gov/data-research/projecting-health-workforce-supply-demand/primary-health

23. O’Reilly-Jacob M, Chapman J, Subbiah SV, Perloff J. Estimating the primary care workforce for Medicare beneficiaries using an activity-based approach. J Gen Intern Med. 2023;38(13):2898-2905. doi:10.1007/s11606-023-08206-3