Christopher M. Whaley, PhD

 To understand responses of primary care clinics to inclusion in a tiered total cost of care insurance benefit design.

 We used a qualitative design beginning with longitudinal analysis of administrative data on consumer clinic choice, clinic tier placement, and clinic actions, followed by in-depth interviews with key informants from clinics, administering health plans, and program administrators.

 We collected data via semistructured interviews with purposively sampled key informants selected from clinics that prospectively reduced prices to move to, or remain in, a tier with lower cost sharing. Data from interview transcripts were coded using qualitative coding software and analyzed for thematic responses.

 Our findings suggest that clinics respond to the incentives in the tiered cost-sharing benefit design. Two motivations cited by clinics are (1) concern over developing a reputation as a high-cost clinic and (2) concern about the possible loss of patients due to higher cost sharing. Some clinics have agreed to price reductions or risk-sharing arrangements to move to, or remain in, a tier with lower cost sharing. Clinic informants reported that price reductions alone are not scalable. They sought greater transparency in tier assignment and increased data sharing to help them reduce costly or unnecessary utilization.

 Managers of primary care clinics respond to a tiered benefit design that holds them accountable for total cost of care. They respond by offering price discounts and expressing interest in reducing costly referrals and unnecessary use of services.

Tiered total cost of care health care benefit designs may significantly reduce overall spending and offer advantages over other health care reform initiatives. In this study, we find that:

Assessing primary care clinics for total cost of care put them in the positions of gatekeeper and coordinator of care.

Consumers demonstrate an ability to choose a gatekeeper clinic and prefer clinics in tiers with lower cost sharing.

In response, clinics reduce prices by an estimated 10% to 20%.

Large employers or managed care decision makers may want to consider such consumer choice approaches along with payment reform, reference pricing, pay to shop, and other initiatives.

There is substantial variation among US health care providers in their prices and use of health care services.

 There also is a weak relationship between the cost and quality of medical services,

 suggesting that cost could be reduced without sacrificing quality of care.

 One way to encourage improved efficiency—defined as better patient health outcomes and patient satisfaction for less cost—among providers is to give patients (consumers) information and financial incentives to choose providers that offer better quality and lower cost. A simulation by Desai et al suggests that if consumers in commercial plans moved to lower-cost providers for referred services (eg, diagnostic laboratory and imaging tests, durable medical equipment), spending could be reduced by 11%.

There is growing policy interest in consumer cost sharing and price transparency,

 but many cost-sharing health insurance benefit designs such as high-deductible health plans place the burden of navigating the fragmented health care services system on consumers who lack the information or expertise to integrate services themselves.

 Studies have found that consumers have difficulty making sense of complex information in health care,

 but indications exist that consumers may choose higher-quality providers.

 Some health insurance benefit designs give consumers both the information they need to choose lower-priced providers and a financial incentive to do so. Those options include preferred provider organizations (PPOs) with in-network and out-of-network providers,

 Tiered networks and PPOs preserve consumer choices with variable cost sharing, whereas narrow networks restrict choices to only those providers within network. Tiered designs reward consumers who choose better-quality or lower-priced providers with lower out-of-pocket costs and encourage providers to become more efficient.

Although these approaches indicate that consumers will use information if paired with financial incentives, they place the burden of navigating the broader health care system of specialists and hospitals on patients and in some cases are limited to discrete procedures and services. In this study we investigate consumer choice of primary care gatekeeper clinics in a tiered total cost of care system managed by the Minnesota State Employee Group Insurance Program (SEGIP). We define tiered total cost of care systems as those in which the primary care provider or clinic is assessed for total cost, including specialist, inpatient, and pharmaceutical costs, even if they did not provide the services themselves. The SEGIP design is distinct from other forms of tiered designs that individually tier hospitals, pharmaceuticals, or individual services.

SEGIP has been in operation since 2002 and covers approximately 130,000 employees and dependents, as well as a small number of retirees. The benefit design is relatively simple from the consumers’ perspective. Plan members make only 2 choices each year: a primary care clinic and 1 of 3 administering health plans. The primary care clinic then coordinates all their patients’ care, including referrals to specialists and hospitals. The SEGIP program reviews the clinics’ historical data on the mean annual risk-adjusted total cost of care and places the clinic into 1 of 4 cost-sharing tiers for the coming year.

This analysis focuses on the response of primary care gatekeeper clinics to tiered cost sharing. The clinics’ response depends, in part, on the response of consumers to tiering, and so we provide a brief summary of the consumer response data. We then examine the steps that clinics have taken when faced with pressure to lower prices.

SEGIP is one of the longest-running tiered cost-sharing systems in the country. State employees are spread across the entire state, including densely populated urban areas and remote rural areas. SEGIP assigns all primary care clinics that participate in the SEGIP program to 1 of 4 tiers based on their per capita risk-adjusted annual total cost of care for members from the prior year (or 2 years for smaller clinic groups), divided by the mean risk-adjusted cost for all SEGIP members. Tier 1 has the lowest-cost clinics and lowest cost sharing and tier 4 has the highest. Currently, total risk-adjusted per capita health care spending for the highest-cost clinic in tier 4 is roughly 80% greater than that for the lowest-cost clinic in tier 1.

, annual deductibles in 2020 ranged from $150 to $1250 for single coverage and $300 to $2500 for family coverage. Office visit co-pays ranged from $30 for enrollees who select a tier 1 clinic to $85 per visit for enrollees who select a tier 4 clinic. Maximum out-of-pocket spending, ranging from $1200 and $2400 for single and family coverage, respectively, in tier 1 to $1600 and $3200 in tier 3, increases to $2600 and $5200 in tier 4. The maximum consumer out-of-pocket cost-sharing amount, along with all aspects of the tiered cost sharing, are set through collective bargaining.

The members’ health insurance claims are processed by 3 health plans: HealthPartners, Blue Cross and Blue Shield of Minnesota, and PreferredOne. The member’s out-of-pocket premium is the same across the 3 administering health plans, so the member’s incentive to choose a lower-cost clinic is based only on their point-of-purchase cost sharing.

Members choose a primary care clinic that acts as a “gatekeeper” and directs the member’s access to specialists and hospitals. Each family member can choose a different clinic, including pediatric and obstetrics and gynecology clinics, but all family members must choose the same health plan.

Clinics can influence their tier placement in 3 ways: (1) reducing their use of avoidable utilization and low-value care relative to other clinics; (2) prospectively reducing their reimbursement price schedule for the SEGIP program by an amount required to achieve a certain tier; or (3) participating in a risk-sharing arrangement with 1 or more of the administering health plans. In the third option, a portion of the reduction to achieve a desired tier may be placed at risk via a target set with the SEGIP administration. The price reduction option was introduced in 2004, and the risk-sharing option was introduced in 2013 and phased out in 2019 because of a lack of adoption among clinics.

We designed a qualitative study using nonprobability, purposive sampling. Our sampling strategy included first identifying, from SEGIP administrative annual data, those clinics in the prior year that had sought and been assigned to a lower-cost tier than calculated cost experience for annual total cost of care data defined by SEGIP and its Deloitte partner. With rare exceptions, every clinic that participates in the SEGIP program is included in each of the 3 administering health plans, and in 2017 there were 2786 such clinic-plan combinations. Of these, 700 (25.1%) had agreed to a specified fee reduction to move to a lower tier or retain their current tier for the coming year. We then stratified these clinics by geography and size, as measured by number of primary care physicians, to include contrasts in the sample. Finally, with help from SEGIP administrators, we selected a sample of key informants (CEOs, chief operating officers, chief medical officers, or contracting officers, hereafter referred to as “clinic managers”) from 7 clinics from across various regions, both micropolitan and metropolitan areas of the state, and various clinic sizes. We also interviewed selected key informants at each of the 3 administering health plans who are responsible for the SEGIP business. We were successful in recruiting all the informants to participate in interviews.

]) based on concepts from the literature to include core questions for all respondents and additional specific questions for each type of key informant. Questions were designed to explore clinic understanding of, experience with, and response to the SEGIP program; health plans were asked about fee reductions or risk-sharing arrangements and perceptions of experiences of clinics in their networks. We recruited individuals for participation in interviews by email in May and June 2017. Two investigators (T.M., K.M.W.) conducted telephone or in-person, 60-minute semistructured interviews with 13 key informants between July and October of the same year. A listing of the professional titles of the informants is included in 

. Interviews were audio recorded and transcribed for analysis.

Two investigators (T.M., K.M.W.) independently coded the transcripts using a codebook derived from the interview questions and inductively from the responses. Investigators met to compare coding and resolve differences through review and further discussion. Key themes were identified from the systematic, iterative review of the coded material and notes were made of emergent concepts heard during the interviews.

 shows clinic placements in tiers from 2005 to 2017. From years 2005 through 2016, between 49% (in 2006) and 68% (in 2012) of clinics were in tiers 1 and 2, with an uptick to 78% in 2017, reflecting a higher number of clinics electing to reduce their prices after initial tier placement.

 shows consumer choices of clinics from 2005 to 2017. These data suggest a strong preference by consumers for clinics in the lower 2 cost-sharing tiers. From 2005 to 2017, consumer choice of clinics in tiers 1 and 2 ranged from 78% in 2008 to 91% in 2017. During this 13-year period, the mean portion of clinics in tiers 1 and 2 was 59%, and the mean portion of consumers choosing clinics in tiers 1 and 2 was 85%.

As described earlier, clinics can change their tier placement in 3 ways. The first is to reduce avoidable utilization of health care services and low-value care, thereby establishing a lower base of total annual risk-adjusted cost of care for the next 2 years’ tiering. The second is to voluntarily reduce their prices. The third is to place a portion of their revenue at risk. After clinics receive their initial tier placement in June, SEGIP contacts the clinic and gives them the option to reduce their prices or participate in risk sharing.

 shows that over time an increasing number of clinics have voluntarily reduced their prices—up to 25% of clinics by 2017. The required price reduction for a primary care clinic to move to a lower tier is calculated in a 2-step process. First, analysts at Deloitte compute the amount that a primary care clinic’s costs would need to be reduced to improve its overall cost efficiency to the threshold for the tier that the clinic is seeking. Then, price reductions are applied only to the costs incurred by the clinic, not to referrals or prescription drugs. Deloitte computes the price reduction data for SEGIP and reports that the typical required fee reduction for clinics is between 10% and 20%. For example, if the total cost of care attributed to a primary care clinic is 5% above its goal tier and it is determined by claims analysis to control 50% of its total cost of care (the other 50% of cost being for care provided by specialists or prescription drugs), the clinic would need to reduce its own prices for the SEGIP population by 10%.

Risk sharing was introduced by program officials in 2013. Figure 3 shows that in 2013 and 2014, 8% and 10% of clinics, respectively, entered into risk-sharing arrangements. After some initial popularity, the risk-sharing option became less popular over time as clinics substituted price reductions, and it was eventually phased out by SEGIP in 2019. In our interviews, both clinics and health plan administrators reported that price reductions were administratively simpler than risk sharing.

Rationale and Decision Process for Lowering Prices

Respondents described 2 rationales for lowering prices to prevent movement to a higher-cost tier. The first was concern over developing a reputation as a high-cost clinic—a “reputation” effect. The second rationale was more directly economic, centered on potential loss of patients due to higher cost sharing.

When asked about the implications of substantial price reductions for SEGIP members, the clinics reported they did not do special analyses to determine their costs for caring for SEGIP members, and 1 clinic informant noted that their discounted prices may create an economic loss for their clinic, but considering the limited patient volume, this was tolerable in the short term. When asked what would happen should a larger portion of their total patient base be in a SEGIP-style tiered benefit design, multiple clinic respondents stated that expansion of discounted prices to a larger portion of their patient base would be unsustainable.

Clinic respondents reported that they would need more data on cost drivers to improve efficient use of resources. “We cannot just continue to discount prices without understanding where we can intervene to manage total cost of care,” one clinic stakeholder said, but the lack of information on the effects of specific types of utilization on tier placement makes improvement difficult. All clinic representatives except one, who was part of an integrated system, reported that they did not know which specialists and hospitals are higher cost, and thus base their referrals on intuition rather than data. “They are kind of a gut feel.…We don’t know how those patients are utilizing so we don’t know…where those high-cost referrals are going,” one said.

All key clinic informants noted that decisions to lower prices were primarily made by clinic managers in the contracting departments, reflecting the relatively small portion of the book of business for the 7 clinics chosen. In some clinics, senior leadership was involved with the clinic’s response to the SEGIP design.

Clinic Perceptions of Member Price Sensitivity

Clinic respondents reported that they believe that state employees are price sensitive, consistent with our findings from administrative records. One clinic informant stated, “Members in the state health plan are very competitive [price sensitive]. So you have to watch that tiering pretty close; really close.” Respondents said they hear from members if they have moved to a more expensive tier. “The only thing they [members] don’t like is when a facility [moves to] a [higher] tier. If you are a tier 2 one year and next year you’re a tier 3, they give you some feedback,” one respondent said.

The 3 health plan administrators generally were satisfied with the program and believed that they have an innovative partnership with the state, but a common critique was that the program is resource intensive to administer because the primary care gatekeeper model requires running 4 plan designs, 1 for each tier.

Opportunities exist for improvement to the SEGIP program design. Currently, the tiers are constructed using a measure based solely on total cost of care, and quality is presented to plan members through a link leading to the Minnesota Community Measurement website. Literature on consumer decision-making suggests a positive relationship between consumer use of information and the ease of accessing it.

 Clearer and more direct inclusion of quality measures at the time of tier selection or incorporating quality metrics into tier design could improve the utilization of information and incent clinics to improve care.

There are 2 sources of limitations to discuss. This study did not differentiate responses by clinic type. The SEGIP program includes a range of primary clinic types from independent small primary care clinics to large, integrated systems with their own health plans. Large care systems such as multispecialty groups with hospitals are tiered as a single system, but they also have a higher ability to control their own costs. Future research should investigate these differences, which may provide further insight into clinic actions and motivations.

The interviews reported in this study did not include care providers, SEGIP members, or labor leadership. The study would be strengthened by expanding the sample and speaking with a range of stakeholders at each clinic. Practicing physicians may have different views on tiering, pricing, and the gain or loss of patient volume depending in part on the type of clinic and degree of physician leadership and involvement in contracting decisions.

We investigated a state employee benefit program that employs a tiered total cost of care benefit design that places primary care clinics into tiers with different levels of consumer cost sharing based on total cost of care. We find that both consumers and clinics respond to participation in the tiered system. In 2017, 91% of consumers chose clinics in the 2 lowest-cost tiers and 25% of clinics chose to lower their prices for the SEGIP program by a reported typical range of 10% to 20% to avoid being moved to a tier with higher consumer cost sharing.

Clinics cited 2 motivations for their response to the possibility of being moved to a higher consumer cost-sharing tier: concern over reputation and economic concern about the possible loss of patients. Clinics reported that continued price reductions are not sustainable and they wanted better information to help them to understand how to reduce unnecessary utilization.

These findings contribute to the literature on consumer engagement and clinic incentives and to the emerging literature on tiered insurance systems. The SEGIP members’ response to a clinic’s tier is consistent with advances in measurement and information technology to improve presentation of information to consumers

 and is related to the literature on physicians’ “intrinsic” incentives.

 Easily understood, publicly available information on the total cost of care for primary care clinics compared with their competitors combines both intrinsic self-assessment and extrinsic demand-related incentives for improvement.

The authors gratefully acknowledge the assistance of the Minnesota Department of Management and Budget, especially Joshua Fangmeier, and the State Employee Group Insurance Program. They also thank Jon Christianson, Susan Ridgely, and Julie Sonier for their valuable insights and suggestions in preparation of this paper and the Robert Wood Johnson Foundation’s State Health Access Reform Evaluation program, the Donaghue Foundation, and the Horowitz Foundation for Social Policy for funding support.

 Pardee RAND Graduate School (TM), Santa Monica, CA; RAND Corporation (TM, CMW), Santa Monica, CA; University of Minnesota (KMW, TYH, BD), Minneapolis, MN.

 Robert Wood Johnson Foundation, Donaghue Foundation, and Horowitz Foundation for Social Policy.

 Dr Whaley has received grant K01AG061274 from the National Institute on Aging. Mr McDonald, Dr White, Dr Huang, and Dr Dowd report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (TM, KMW, TYH, CMW, BD); acquisition of data (TM, KMW, BD); analysis and interpretation of data (TM, KMW, TYH, CMW, BD); drafting of the manuscript (TM, KMW, TYH, CMW, BD); critical revision of the manuscript for important intellectual content (TM, KMW, TYH, CMW, BD); statistical analysis (TM, TYH, BD); obtaining funding (TM, BD); administrative, technical, or logistic support (BD); and supervision (TM, KMW, BD).

 Tim McDonald, MPP, Pardee RAND Graduate School, 1776 Main St, Santa Monica, CA 90401. Email: tmcdonal@prgs.edu.

1. Cooper Z, Craig SV, Gaynor M, Van Reenen J. The price ain’t right? hospital prices and health spending on the privately insured. 

. 2019;134(1):51-107. doi:10.1093/qje/qjy020

2. Hussey PS, Wertheimer S, Mehrotra A. The association between health care quality and cost: a systematic review. 

. 2013;158(1):27-34. doi:10.7326/0003-4819-158-1-201301010-00006

3. Whaley C. The association between provider price and complication rates for outpatient surgical services. 

.2018;33(8):1352-1358. doi:10.1007/s11606-018-4506-7

4. Berwick DM, Hackbarth AD. Eliminating waste in US health care. 

. 2012;307(14):1513-1516. doi:10.1001/jama.2012.362.

5. Desai SM, Hatfield LA, Hicks AL, Chernew ME, Mehrotra A, Sinaiko AD. What are the potential savings from steering patients to lower-priced providers? a static analysis. 

6. Mehrotra A, Schleifer D, Shefrin A, Ducas AM. Defining the goals of health care price transparency: not just shopping around. 

 Catalyst. June 26, 2018. Accessed August 5, 2021. https://catalyst.nejm.org/health-care-price-transparency-goals/

7. Executive order on improving price and quality transparency in American healthcare to put patients first. Trump White House Archives. June 24, 2019. Accessed August 5, 2021. https://trumpwhitehouse.archives.gov/presidential-actions/executive-order-improving-price-quality-transparency-american-healthcare-put-patients-first/

8. Taylor EA, Carman KG, Lopez A, Muchow AN, Roshan P, Eibner C. Consumer decisionmaking in the health care marketplace. RAND Corp. 2016. Accessed August 5, 2021. https://www.rand.org/pubs/research_reports/RR1567.html

9. Chandra A, Finkelstein A, Sacarny A, Syverson C. Health care exceptionalism? performance and allocation in the US health care sector. 

. 2016;106(8):2110-2144. doi:10.1257/aer.20151080

10. Dafny LS, Hendel I, Marone V, Ody C. Narrow networks on the health insurance marketplaces: prevalence, pricing, and the cost of network breadth. 

. 2017;36(9):1606-1614. doi:10.1377/hlthaff.2016.1669

11. Robinson JC, Brown TT, Whaley C. Reference pricing changes the ‘choice architecture’ of health care for consumers. 

. 2017;36(3):524-530. doi:10.1377/hlthaff.2016.1256

12. Whaley C, Brown T, Robinson J. Consumer responses to price transparency alone versus price transparency combined with reference pricing. 

. 2019;5(2):227-249. doi:10.1162/ajhe_a_00118

13. Prager E. Tiered hospital networks, health care demand, and prices. Duke Department of Economics. November 19, 2015. Accessed August 5, 2021. http://ipl.econ.duke.edu/seminars_dev/system/files/seminars/1386.pdf

14. Sinaiko AD, Hirth RA. Consumers, health insurance and dominated choices. 

. 2011;30(2):450-457. doi:10.1016/j.jhealeco.2010.12.008

15. Sinaiko AD, Rosenthal MB. The impact of tiered physician networks on patient choices. 

. 2014;49(4):1348-1363. doi:10.1111/1475-6773.12165

16. Sinaiko A, Landrum MB, Chernew ME. Enrollment in a health plan with a tiered provider network decreased medical spending by 5 percent. 

. 2017;36(5):870-875. doi:10.1377/hlthaff.2016.1087

17. Sinaiko AD, Rosenthal MB. Consumer experience with a tiered physician network: early evidence. 

18. Prager E. Healthcare demand under simple prices: evidence from tiered hospital networks. 

. 2020;12(4):196-223. doi:10.1257/app.20180422

30 Essential Skills for the Qualitative Researcher

The Coding Manual for Qualitative Researchers

21. Kling JR, Mullainathan S, Shafir E, Vermeulen LC, Wrobel MV. Comparison friction: experimental evidence from Medicare drug plans. 

. 2012;127(1):199-235. doi:10.1093/qje/qjr055

22. Sinaiko AD, Eastman D, Rosenthal MB. How report cards on physicians, physician groups, and hospitals can have greater impact on consumer choices. 

. 2012;31(3):602-611. doi:10.1377/hlthaff.2011.1197

23. Kolstad JT. Information and quality when motivation is intrinsic: evidence from surgeon report cards. 

. 2013;103(7):2875-2910. doi:10.1257/aer.103.7.2875

Articles

Findings of this evaluation of primary care clinic responses in a tiered total cost of care benefit design suggest that clinics respond by reducing prices.

Clinic Price Reductions in a Tiered Total Cost Benefit Design

To examine changes in hospital outpatient surgery trends and case mix for Medicare and privately insured patients needing total knee arthroplasty (TKA) following Medicare’s removal of TKA from its Inpatient Only list on January 1, 2018.

 A retrospective analysis of all hospital discharge records in Florida from 2012 through 2018.

We tracked inpatient vs outpatient performance of TKAs at the state and hospital levels. We also combined our primary data with physician practice organization information to assess variation in the policy response according to physician-hospital ownership status. Supplementary analyses examined policy-induced changes in inpatient TKA case mix.

 We observed an immediate shift of roughly 15% of Medicare TKA cases to the outpatient setting. Importantly, there was a simultaneous near doubling of the number of TKAs performed as a hospital outpatient procedure among privately insured patients younger than 60 years. Hospitals allocated a similar proportion of TKA cases to the outpatient setting across the 2 payer groups, and we found evidence of selection against the potentially riskiest Medicare TKA patients for outpatient delivery. Vertically integrated orthopedic physicians retained their Medicare and privately insured TKA cases within the inpatient (higher-cost) setting.

 Market and financial pressures are encouraging more outpatient care delivery; however, the speed of transition is dictated, in part, by regulatory constraints. Our results suggest that Medicare policy may influence surgical treatment approaches for Medicare and privately insured patients. Spillover implications need to be considered when weighing future Medicare regulatory decisions.

In the first year following Medicare’s decision to remove total knee arthroplasty (TKA) from its Inpatient Only list, Florida hospitals shifted 15% of TKA Medicare cases to the outpatient setting.

That same year, the total number of outpatient privately insured TKA cases doubled and represented 25% of all privately insured TKA cases in the state.

Medicare administrators and policy makers need to pay attention to potential for Medicare rulemaking spillovers on care delivery, and hence medical spending, for non-Medicare patients.

There is a growing shift toward outpatient care delivery throughout the US health care system, which is expected to benefit patients through lower costs and better convenience relative to inpatient treatment options.

 The hospital industry has been forced to confront these trends as outpatient revenue is now at near parity with inpatient revenue for US hospitals.

 Although market forces and financial pressures underlie some of these movements, regulatory interventions are also important. Specifically, Medicare is often a significant source of care delivery rules, which can affect the precise timing and speed of inpatient to outpatient care transitions.

A prominent and contemporary example comes from orthopedic surgery involving knee replacements (ie, total knee arthroplasty [TKA]). Each year, approximately 680,000 individuals receive a TKA, which costs more than $30,000 on average.

 In 2013, CMS first attempted to remove TKA procedures from the Medicare Inpatient Only (IPO) list, which prohibits outpatient surgery for listed procedures. The initial attempt was ultimately unsuccessful, but the policy initiative eventually gained sufficient support to culminate in TKA’s removal from the Medicare IPO list on January 1, 2018.

 The 2018 rule change granted providers the ability to perform the procedure for Medicare beneficiaries within a hospital outpatient department, although inpatient delivery remained permissible.

The Medicare TKA rule change did not take place in a vacuum, as momentum had been growing toward outpatient delivery due to clinical and procedural innovations that shortened the required length of stay for major surgery. For instance, the average length of stay for a joint replacement procedure has decreased from 9 days to 4 days in the last 20 years—with outpatient delivery almost always resulting in same-day discharge.

 Outpatient TKAs have also been estimated to be as much as 30% to 50% less expensive than inpatient TKAs,

 and some forecast up to half of all TKAs taking place within an outpatient setting over the next decade.

Thus far, there has been limited research on the 2018 TKA IPO decision. Evidence from 2 studies documents outpatient delivery take-up for Medicare cases across the United States, which averaged $3600 in lower Medicare payments and generated more than $300 million in savings,

 but neither study speaks to broader policy ramifications, particularly for non-Medicare TKA patients. Medicare accounts for about 20% of all US health care expenditures and is known to influence contracting and care delivery for the privately insured. For instance, prior research finds Medicare spillover effects onto private insurance fee schedules as well as providers’ treatment of private patients following Medicare incentive changes.

 However, Medicare’s regulatory externalities are just beginning to be understood. Geruso and Richards offer the only known study that explicitly investigates and documents spillover effects from Medicare rulemaking onto the privately insured segment of the US health care market.

 We believe such regulatory externalities are relevant in our context. In this setting, private insurers did not have uniform restrictions on the use of outpatient surgical services, but physicians may have been reluctant to perform procedures in multiple settings or bear higher perceived liability risks. Providers and private insurers could therefore have adopted the policies established by Medicare. Carey et al likewise posit that TKA removal from the Medicare IPO list could invigorate private insurer efforts to move more TKAs into the outpatient setting.

Using the universe of data from all Florida hospitals for 2012-2018, we examined trends in inpatient and outpatient TKA surgeries for 2 key payer types: traditional Medicare and the privately insured (non-Medicare). We additionally explored the degree of correspondence in postpolicy outpatient preference across payers within the same hospital as well as changes in inpatient case characteristics. Importantly, a complementary database allowed us to assess if hospital employment of surgeons moderates any change in outpatient delivery. The differences in reimbursement across settings are salient to hospitals and could have led them to discourage employed surgeons from performing TKA procedures outside the (higher-paying) inpatient setting. Taken together, our findings reveal whether an abrupt Medicare policy intervention is associated with the degree to which different health care providers substitute lower-cost treatment options.

The administrative data set was obtained from the Florida Agency for Health Care Administration (AHCA) and comprised the universe of inpatient and outpatient (ambulatory) procedure discharge records from the state of Florida over the 2012-2018 period. The data are recorded quarterly; however, we annualized the analytic data to smooth out within-year seasonality in care delivery (see the 

] for quarterly-level analyses). Each discharge record contained information on the medical procedures being performed, the providers involved (eg, surgeons and hospitals), and patient characteristics, including the primary payer (eg, Medicare or private insurance). TKA cases were identified using the corresponding 

International Classification of Diseases, Ninth Revision

) procedure codes for the inpatient discharge records and the corresponding Current Procedural Terminology (CPT) codes for the outpatient discharge records. Across both settings, we observed approximately 17,000 to 20,000 Medicare TKAs per year and 10,000 to 14,000 privately insured TKAs annually (see 

Our primary analyses longitudinally tracked the treatment setting selections for TKAs across our 2 key payer groups. We examined how often outpatient settings were used in 2018 for Medicare patients and to what extent the trend for privately insured patients corresponded to the timing of the Medicare policy change. We also examined the cross-payer concordance in 2018 outpatient use for TKAs within the same hospital.

We then conducted several supplementary analyses. The first examined the inpatient vs outpatient TKA decision for vertically integrated (ie, employed by a hospital or health system) orthopedic surgeons vs all others. We followed several recent studies and obtained information on physician vertical integration status from the SK&A physician office survey, which we linked to the discharge data via National Provider Identifier numbers.

 Next, we examined trends in site of care for all non-TKA procedures over our study period to ensure that 2018 was not remarkable in other ways. Finally, we explored changes in the inpatient case mix (ie, risk selection) for Medicare and privately insured TKAs in 2018 by comparing their patient characteristic trends against all other inpatient procedures belonging to each payer (see the eAppendix for full methodological details and results).

The study was approved by the Baylor University Institutional Review Board, and Florida AHCA was not responsible for any analyses or data interpretations.

Figure 1 displays the total number of TKA procedures performed per year among Medicare fee-for-service and private payer populations in Florida. The number of TKA procedures increased for both payers over time; however, neither payer group demonstrated meaningful TKA growth in 2018.

Figure 1 also displays annual percentage changes and hence growth in inpatient TKA delivery over our study period for each payer group of interest. From 2012 to 2016, both payers experienced similar increases in the volume of inpatient TKAs that then plateaued between 2016 to 2017. Once the Medicare prohibition on outpatient TKA delivery was lifted (2018), there was a sharp decline of inpatient TKAs. Inpatient Medicare TKA cases were 15% lower in 2018 compared with the 2017 volume, and the trends for privately insured TKA patients generally mirrored those for Medicare. Specifically, after years of positive growth, the privately insured TKA inpatient volume fell by 12% in 2018. Given that 2018 is not characterized by declines in total TKA cases, the reductions in inpatient TKAs must be offset by increases in outpatient TKAs.

 presents the number of TKA procedures among the privately insured population performed in an outpatient setting from 2012 to 2018. Starting from a low base of 264 procedures (3% of all private TKAs), more than 1000 such cases had occurred across Florida by 2015. The trend changed abruptly in 2018, however, with more than 3000 outpatient privately insured TKA cases taking place in that year (24% of private TKAs). This marked a nearly 100% increase from the previous year (2017) and was concurrent with the Medicare IPO decision for TKA. Figure 2 also presents these trends based on the employment status of the operating physician. Physicians not employed by a hospital or health system accounted for 85% of the increased use of TKAs performed in outpatient settings. Although this lack of response to the Medicare policy change by vertically integrated surgeons could be suggestive of misaligned incentives and a strategic response by hospitals, we observed a greater share of TKA patients coming from rural areas among those treated by vertically integrated orthopedic surgeons. Longer travel distances and potentially limited health care resources near their home residence could have altered the cost-benefit calculation for these specific patients (

 presents analogous trends for all non-TKA surgical procedures from 2012 to 2018 in Florida across these 2 payers. Neither the inpatient nor outpatient delivery trends indicate commensurately sharp and large changes in 2018 or any other year. The patterns in Figure 3 therefore offer further suggestive evidence that the very large change in site of care for TKAs in 2018 (ie, a doubling of outpatient delivery for the privately insured) was likely due to factors specific to TKA procedures.

, we plot the 2018 TKA outpatient shares by payer group within the same hospital. There is a central tendency to the scatter plot (ie, following the 45-degree line), implying that orthopedic surgeons and hospitals were often substituting toward the outpatient option to a similar degree for Medicare and privately insured TKA patients. There were also exceptions (ie, favoring outpatient for one payer type relative to the other), but these instances may have had more to do with the surrounding payer mix facing the surgeon and hospital than the appropriateness of outpatient TKA surgery.

We found limited changes in the case mix for inpatient TKAs following the Medicare IPO policy change (eAppendix). The single notable exception was some degree of selection against the oldest (ie, 75 years and older) Medicare beneficiaries receiving a TKA as an outpatient, which is not surprising when considering their likely elevated risks of complications and greater postoperative needs. Inpatient privately insured TKA case characteristics appeared unaffected by the sharp shift toward outpatient delivery in 2018.

An increasing share of medical care is migrating toward outpatient delivery. These trends pose strategic and financial challenges for US hospitals, but they are unlikely to be reversed. Preferences for outpatient care are not the only driving forces, however. Regulatory stances, such as those belonging to Medicare, can be influential.

We see evidence consistent with a regulatory barrier when evaluating the recent decision to move TKAs off the Medicare IPO list and thereby permit the outpatient surgery option for beneficiaries. Fifteen percent of Medicare TKA cases were reallocated from the inpatient to the outpatient setting (Figure 1), and uptake is observed across a variety of hospitals during the first deregulation year (Figure 4). The key and novel finding from our study is the alignment between the privately insured outpatient TKA trend break and the Medicare deregulation event. With an abrupt doubling of outpatient privately insured TKAs in 2018, our results are at least consistent with a Medicare policy spillover implication whereby outpatient provision was potentially restrained prior to the Medicare IPO decision. Greater use of outpatient hospital facilities following the deregulation event from Medicare may have improved care convenience for privately insured TKA patients and lowered costs for patients and their insurers. Such a positive externality as suggested by our data also comports well with Geruso and Richards’ findings from a recent study of a different Medicare surgical regulation from a decade prior.

As Medicare makes further rule changes, future research will need to explore the implications within and beyond the Medicare program. Additionally, the long-run effects of the 2018 IPO change for TKAs will need to be tracked as providers have more time to adapt.

This study used data from a single state and therefore may not generalize to other markets. This limitation is balanced, however, by the timeliness and multipayer dimensions that the Florida data offer. Florida is also second only to California in terms of Medicare population, which makes it a relevant state for Medicare policy analysis, even in isolation.

 We also relied upon an observational study design, so the associated trend changes we observe may not be causal.

Medicare treatment restrictions and accompanying payment policies ultimately shape how care is delivered to Medicare beneficiaries. We show that the CMS decision to allow outpatient TKAs was associated with a simultaneous and striking increase in outpatient TKA delivery for privately insured patients. In this way, Medicare rulemaking consequences may extend beyond Medicare patients. Potential spillover implications deserve greater attention in future policy deliberations because Medicare rules may influence the norms and the pace of change for care delivery throughout the US health care system.

The authors thank the Florida Agency for Health Care Administration (AHCA) for providing data for the study and wish to note that AHCA was not responsible for any data analyses or interpretation of results for this work. Dr Whaley would also like to thank the National Institutes on Aging (1K01AG061274) for its generous financial support.

 Hankamer School of Business, Baylor University (MRR, JAS), Waco, TX; RAND Corporation (CMW), Santa Monica, CA.

 National Institutes on Aging 1K01AG061274 (CMW).

 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.

 Concept and design (MRR, JAS, CMW); acquisition of data (MRR); analysis and interpretation of data (MRR, JAS, CMW); drafting of the manuscript (MRR, JAS, CMW); critical revision of the manuscript for important intellectual content (MRR, JAS, CMW); statistical analysis (MRR, JAS); and supervision (MRR).

 Michael R. Richards, MD, PhD, MPH, Hankamer School of Business, Baylor University, 1 Bear Pl, Waco, TX 76798. Email: Michael_Richards@baylor.edu.

1. Kacik A. Number of outpatient facilities surges as industry values more convenient, affordable care. Modern Healthcare. December 20, 2018. Accessed July 11, 2020. https://www.modernhealthcare.com/article/20181220/NEWS/181229992/number-of-outpatient-facilities-surges-as-industry-values-more-convenient-affordable-care

2. Edwards PK, Milles JL, Stambough JB, Barnes CL, Mears SC. Inpatient versus outpatient total knee arthroplasty. 

. 2019;32(8):730-735. doi:10.1055/s-0039-1683935

3. Bashinskaya B, Zimmerman RM, Walcott BP, Antoci V. Arthroplasty utilization in the United States is predicted by age-specific population groups. 

. 2012;2012:185938. doi:10.5402/2012/185938

4. Guroo. Accessed July 11, 2020. https://www.guroo.com

5. Li Y, Lu X, Wolf BR, Callaghan JJ, Cram P. Variation of Medicare payments for total knee arthroplasty. 

. 2013;28(9):1513-1520. doi:10.1016/j.arth.2013.06.001

6. Miller DC, Gust C, Dimick JB, Birkmeyer N, Skinner J, Birkmeyer JD. Large variations in Medicare payments for surgery highlight savings potential from bundled payment programs. 

. 2011;30(11):2107-2115. doi:10.1377/hlthaff.2011.0783

7. Schwartz AJ, Clarke HD, Sassoon A, Neville MR, Etzioni DA. The clinical and financial consequences of the Centers for Medicare and Medicaid Services’ two-midnight rule in total joint arthroplasty. 

. 2020;35(1):1-6.e1. doi:10.1016/j.arth.2019.08.048

8. Medicare program: hospital outpatient prospective payment and ambulatory surgical center payment systems and quality reporting programs. 

9. Jain NB, Higgins LD, Guller U, Pietrobon R, Katz JN. Trends in the epidemiology of total shoulder arthroplasty in the United States from 1990-2000. 

. 2006;55(4):591-597. doi:10.1002/art.22102

10. Memtsoudis SG, Della Valle AG, Besculides MC, Gaber L, Laskin R. Trends in demographics, comorbidity profiles, in-hospital complications and mortality associated with primary knee arthroplasty. 

. 2009;24(4):518-527. doi:10.1016/j.arth.2008.01.307

11. Berger RA, Kusuma SK, Sanders SA, Thill ES, Sporer SM. The feasibility and perioperative complications of outpatient knee arthroplasty. 

. 2009;467(6):1443-1449. doi:10.1007/s11999-009-0736-7

12. Carey K, Morgan JR, Lin MY, Kain MS, Creevy WR. Patient outcomes following total joint replacement surgery: a comparison of hospitals and ambulatory surgery centers. 

. 2020;35(1):7-11. doi:10.1016/j.arth.2019.08.041

13. Richter DL, Diduch DR. Cost comparison of outpatient versus inpatient unicompartmental knee arthroplasty. 

. 2017;5(3):2325967117694352. doi:10.1177/2325967117694352

14. Van Amerongen D, Maroulis C. Reimbursement for total joint replacement in an outpatient site of care. Johnson & Johnson Institute. 2017. Accessed July 11, 2020. https://jnjinstitute.com/sites/default/files/online_learning/files/DSUSJRC12161900%20Reimbursement%20for%20TJR%20Outpatient_WP%20FINAL.pdf

15. Barnes CL, Iorio R, Zhang X, Haas DA. An examination of the adoption of outpatient total knee arthroplasty since 2018. 

. 2020;35(6S):S24-S27. doi:10.1016/j.arth.2020.01.073

16. Haas DA, Zhang X, Davis CM III, Iorio R, Barnes CL. The financial implications of the removal of total knee arthroplasty from the Medicare Inpatient-Only list. 

. 2020;35(6S):S33-S36. doi:10.1016/j.arth.2020.01.074

17. White C. Contrary to cost-shift theory, lower Medicare hospital payment rates for inpatient care lead to lower private payment rates. 

. 2013;32(5):935-943. doi:10.1377/hlthaff.2012.0332

18. Clemens J, Gottlieb JD. In the shadow of a giant: Medicare’s influence on private physician payments. 

19. Cooper Z, Craig SV, Gaynor M, Van Reenen J. The price ain’t right? hospital prices and health spending on the privately insured. 

20. Sloan FA, Morrisey MA, Valvona J. Effects of the Medicare prospective payment system on hospital cost containment: an early appraisal. 

21. Yip WC. Physician response to Medicare fee reductions: changes in the volume of coronary artery bypass graft (CABG) surgeries in the Medicare and private sectors. 

. 1998;17(6):675-699. doi:10.1016/s0167-6296(98)00024-1

22. He D, Mellor JM. Hospital volume responses to Medicare’s Outpatient Prospective Payment System: evidence from Florida. 

. 2012;31(5):730-743. doi:10.1016/j.jhealeco.2012.06.001

23. Baicker K, Chernew ME, Robbins JA. The spillover effects of Medicare managed care: Medicare Advantage and hospital utilization. 

. 2013;32(6):1289-1300. doi:10.1016/j.jhealeco.2013.09.005

24. Weber E, Floyd E, Youngran K, White C. Peering behind the veil: trends in types of contracts between private health plans and hospitals. 

. Published online July 23, 2019. doi:10.1177/1077558719859724

25. White C. Cutting Medicare hospital prices leads to a spillover reduction in hospital discharges for the nonelderly. 

. 2014;49(5):1578-1595. doi:10.1111/1475-6773.12183

26. Geruso M, Richards MR. Trading spaces: Medicare’s regulatory spillovers on treatment setting for non-Medicare patients. SSRN. Published online March 3, 2020. doi:10.2139/ssrn.3532685

27. Iorio R, Barnes CL, Vitale MP, Huddleston J, Haas DA. Total knee replacement: the Inpatient Only (IPO) list and the two midnight rule, patient impact, length of stay, compliance solutions, audits and economic consequences. 

. 2020;35(6S):S28-S32. doi:10.1016/j.arth.2020.01.007

28. Baker LC, Bundorf MK, Kessler DP. The effect of hospital/physician integration on hospital choice. 

. 2016;50:1-8. doi:10.1016/j.jhealeco.2016.08.006

29. Richards MR, Nikpay SS, Graves JA. The growing integration of physician practices: with a Medicaid side effect. 

. 2016;54(7):714-718. doi:10.1097/MLR.0000000000000546

This article documents a 100% increase in privately insured outpatient knee replacement surgeries following Medicare’s decision to remove knee replacements from its Inpatient Only list.

Christopher M. Whaley, PhD

Articles

Clinic Price Reductions in a Tiered Total Cost Benefit Design

Removing Medicare’s Outpatient Ban and Medicare and Private Surgical Trends