{"subscriber":false,"subscribedOffers":{}} Little Evidence Exists To Support The Expectation That Providers Would Consolidate To Enter New Payment Models | Health Affairs

# Little Evidence Exists To Support The Expectation That Providers Would Consolidate To Enter New Payment Models

Affiliations
1. Hannah T. Neprash is a doctoral candidate in health policy at Harvard University, in Cambridge, Massachusetts.
2. Michael E. Chernew is the Leonard D. Schaeffer Professor of Health Care Policy in the Department of Health Care Policy, Harvard Medical School.
3. J. Michael McWilliams ( [email protected] ) is the Warren Alpert Associate Professor of Health Care Policy in the Department of Health Care Policy, Harvard Medical School.
PUBLISHED:Free Accesshttps://doi.org/10.1377/hlthaff.2016.0840

## Abstract

Provider consolidation has been associated with higher health care prices and spending. The prevailing wisdom is that payment reform will accelerate consolidation, especially between physicians and hospitals and among physician groups, as providers position themselves to bear financial risk for the full continuum of patient care. Drawing on data from a number of sources from 2008 onward, we examined the relationship between Medicare’s accountable care organization (ACO) programs and provider consolidation. We found that consolidation was under way in the period 2008–10, before the Affordable Care Act (ACA) established the ACO programs. While the number of hospital mergers and the size of specialty-oriented physician groups increased after the ACA was passed, we found minimal evidence that consolidation was associated with ACO penetration at the market level or with physicians’ participation in ACOs within markets. We conclude that payment reform has been associated with little acceleration in consolidation in addition to trends already under way, but there is evidence of potential defensive consolidation in response to new payment models.

The past few years have seen the rapid expansion of new payment models that hold health care providers accountable for total spending and quality of care for their patients. The Department of Health and Human Services recently announced that it achieved its goal of tying 30 percent of Medicare payments to such alternative payment models by 2016. 1 The Medicare accountable care organization (ACO) programs are the broadest alternative payment models, with over 460 participating provider organizations in 2016 that collectively covered 23.5 percent of fee-for-service Medicare beneficiaries. 2 These programs set a global budget for total spending for an ACO’s patient population, with incentives for the ACO to spend less than the budgeted amount and provide high-quality care. Commercial insurers have instituted similar payment systems, and many Medicare ACOs also participate in commercial ACO contracts. 3,4

Although designed to remedy the incentives of fee-for-service payment systems, payment models that delegate financial risk to providers for the full continuum of patients’ care have triggered concerns that providers will consolidate in response. 5,6 Consolidation may lead to higher prices (or budgets) negotiated with commercial insurers. Thus, while ACOs may reduce spending in the Medicare population, provider consolidation associated with them may increase spending for the commercially insured.

Provider consolidation generally takes one of two forms: horizontal integration (two hospitals merge) or vertical integration (a hospital system purchases a physician group). However, both may be involved in a merger or acquisition (a merger of two health systems). The association between horizontal consolidation and higher prices has been well documented in hospital markets, 7 and to a lesser extent in physician markets. 8,9 Mounting evidence suggests that financial integration between physicians and hospitals also leads to higher prices and spending. 10,11

## Reasons For Consolidation

There are many possible reasons for the horizontal and vertical consolidations that began before the establishment of the Affordable Care Act (ACA) Medicare ACO programs in 2010. Consolidation may allow providers to negotiate higher rates from commercial insurers, boost the number of referrals or admissions, amass sufficient capital to invest in lucrative services, pool their malpractice risk, compete for physician labor, reconfigure their capacity in response to technological changes that shift care settings, or lower costs (for example, of information technology) through economies of scale. Hospitals and physicians may also consolidate to take advantage of Medicare payment rules that favor providing services in hospital outpatient departments instead of physician offices. 12

## Empirical Evidence And A Widespread Concern

There has been widespread concern that a wave of consolidation would follow the launch of the Medicare ACO programs because providers might seek greater scale and scope to enter and succeed under new payment models. 13,14 However, little empirical evidence exists to support this fear. While there are conceptual reasons why larger provider organizations might be better suited to succeed in ACO contracts, research suggests that consolidation beyond a modest level may be neither necessary nor advantageous for providers operating under new payment models.

Specifically, while large physician groups exhibit greater structural capacity for care management and perform better than smaller groups on some process measures of quality, 1517 these gains may be achieved at organizational sizes far smaller than large integrated health systems and have not translated into better patient outcomes or more efficient care. 18,19 Moreover, previous studies do not support the assumption that establishing direct managerial control through ownership over the full spectrum of patient care is necessary to control spending and improve quality. Studies that compared medical groups and independent practice associations have produced mixed results, 15,20 and hospitals’ ownership of physician practices has been associated with higher spending without clear gains in quality. 21

The lack of evidence extends to payment systems that reward more efficient care. Under capitation incentives, large physician groups have exhibited lower spending levels than small practices, but no lower than the levels of medium-size practices. 18 Thus far in the Medicare Shared Savings Program, independent physician groups have generated greater savings than larger vertically integrated organizations have. 3,22 Organizations that own hospitals and specialty practices have weaker incentives than those that do not to limit use of inpatient and specialty care under ACO contracts, and evidence from Medicare and commercial ACO initiatives suggests that providers can influence the use of care in multiple settings without formal ownership arrangements that unite providers. 3,23,24

Finally, if gaining bargaining power in price negotiations with commercial insurers has been the primary motive for consolidation, one would not expect acceleration in provider consolidation to be associated with ACO contracting, because the desire to command higher prices (or budgets) and negotiate better terms exists in both fee-for-service and alternative payment models. Payment reform could even reverse some previous reasons to consolidate, such as pooling resources to invest in service lines that were profitable under fee-for-service but became cost centers under new payment models. Thus, it is not clear that providers participating or preparing to participate in new payment models will consolidate faster than other providers.

However, contrary to the standard narrative, payment reform might prompt some providers to consolidate to preserve their market position, as opposed to consolidating to enter and succeed under risk contracts. Hospitals and specialists in particular might consolidate to rebuff payer pressure to enter risk contracts or to attain sufficient market share to ensure continued referrals from ACOs that might otherwise steer patients to more efficient providers.

## Study Overview

In this study we examined the relationship between Medicare ACO program participation and multiple measures of horizontal and vertical consolidation over time, from before to after the passage of the ACA in 2010. In two complementary analyses, we compared consolidation over time between markets with more versus less ACO contracting in 2014 and within markets between physicians who had entered an ACO contract by 2014 and those who had not.

Our analyses are descriptive but nevertheless useful in gauging the extent of consolidation associated with payment reform under various scenarios. For example, under the prevailing narrative of providers’ consolidating to enter and succeed under ACO contracts, we would expect increases from the pre- to the post-ACA period to be greater in markets with higher levels of ACO program entry. Because patient populations covered by ACO contracts are defined by where patients receive outpatient care, primarily primary care, under the prevailing narrative we would expect the uptake of ACO contracts in markets to be particularly associated with acceleration in horizontal consolidation among physician practices and vertical consolidation between hospitals and physicians, especially consolidation involving primary care physicians. According to this rationale, we would also expect physician groups that entered an ACO program to exhibit greater consolidation shortly before or after that entry, compared to other physician groups in their markets. In contrast, if providers have consolidated primarily as a defensive response to payment reform, we would expect greater increases among nonparticipating physicians, with ambiguous effects on the market-level relationship between ACO program entry and consolidation over time.

## Study Data And Methods

### Data And Population

To assess provider consolidation, we used Medicare claims data for the period 2008–13, data from the American Hospital Association (AHA) Annual Survey for the same period, and data for the period 2008–15 from Irving Levin Associates’ Health Care Mergers and Acquisitions Database. We used definitions of ACOs from the Centers for Medicare and Medicaid Services (CMS) to identify physicians and practices participating in ACOs and to assess ACO contracting at the level of the Metropolitan Statistical Area. Finally, we used data from the Truven Health MarketScan Commercial Database to measure commercial health care prices at the same level, as an indirect measure of consolidation that our direct measures might not have reflected.

To calculate annual market-level measures of provider market structure, we relied predominantly on Medicare claims data for a 20 percent random sample of beneficiaries. We excluded small markets with few physicians billing Medicare (see online Appendix section I). 25 Our assessments of provider market structure based on Medicare claims data included 301,855 physician national provider identifiers used to bill under 103,745 tax identification numbers in 289 Metropolitan Statistical Areas. A tax identification number may represent a solo practitioner, a practice, or a larger provider organization. Large organizations typically bill under multiple tax identification numbers.

For within-market analyses, we included only those national provider identifiers or tax identification numbers that were present in Medicare claims data throughout the study period, so that we could assess the organizational characteristics of each provider in every study year, after determining whether or not the provider participated in an ACO contract that started in 2012, 2013, or 2014.

### Study Variables

#### Medicare ACO Participation And Penetration:

We used the ACO Provider-Level Research Identifiable File to identify tax identification numbers for provider groups participating in a Medicare Shared Savings Program ACO contract that started in 2012, 2013, or 2014, and we used CMS definitions of Pioneer ACOs to identify national provider identifiers for physicians participating in a Pioneer contract. 26 In 2012, 32 organizations entered the Pioneer program, and 114 entered the Medicare Shared Savings Program. In 2013 and 2014, an additional 106 and 115 organizations entered the Medicare Shared Savings Program, respectively. Based on that program’s rules, we attributed each beneficiary to the ACO or non-ACO tax identification number that accounted for the most allowed charges for qualifying outpatient evaluation and management services delivered to the beneficiary by a primary care physician during each year. 3 To calculate a measure of ACO penetration at the Metropolitan Statistical Area level, we divided the number of ACO-assigned Medicare beneficiaries in each area by the number of assignment-eligible Medicare beneficiaries in the area (Appendix section II). 25

For within-market comparisons, we classified each physician as participating or not participating in an ACO by 2014. Physicians were identified as participating in an ACO if their national provider identifier was included in a Pioneer contract or if they billed primarily under a tax identification number included in a Medicare Shared Savings Program contract that started in 2012, 2013, or 2014.

#### Physician-Hospital Integration:

To measure physician-hospital integration, we used place-of-service codes in Medicare claims, which distinguished between a service provided in a physician practice owned by a hospital (such as in an outpatient department) and a service provided in an office setting. 27 Specifically, for each year in the study period, we determined each physician’s share of Medicare claims for outpatient care that was billed with a hospital outpatient department setting code. We considered physicians to be practicing in a hospital-owned practice if they billed at least 90 percent of their outpatient care with a hospital outpatient department setting code (Appendix section III). 25 From this physician-level variable, we calculated the share of physicians in a Metropolitan Statistical Area who displayed billing patterns consistent with physician-hospital integration.

#### Hospital And Physician Market Concentration:

For market-level analyses, we calculated a Herfindahl-Hirschman Index—a standard measure of market concentration—for hospital and physician markets for each year in the study period. Higher values corresponded to greater concentration. Using data from the AHA Annual Survey Database, we defined each hospital’s market share as its share of total hospital admissions in a Metropolitan Statistical Area, accounting for common hospital ownership in the case of hospital systems. Using Medicare claims data for professional services and tax identification numbers to define physician groups, we defined each group’s market share as its share of total allowed charges for outpatient care in the area. We also explored alternative measures of market concentration, including the four-firm concentration ratio (the total market share of the four largest firms) (Appendix section IV). 25

#### Physician Group Size:

For between-market and within-market comparisons, we assessed physician group (tax identification number) size, defined as the number of distinct physician national provider identifiers used to bill under each tax identification number in Medicare professional claims, excluding physicians with inpatient-based specialties. For between-market comparisons, we calculated an average group size at the Metropolitan Statistical Area level, weighting each group by its share of national provider identifiers in the market. This measure can be interpreted as a physician’s average practice size in the market. For a supplementary analysis, we also assessed physician group specialty mix, defined as the percentage of national provider identifiers with a primary care specialty used to bill under each tax identification number (Appendix section V). 25

#### Mergers And Acquisitions:

To identify instances of provider consolidation directly, we used data collected by Irving Levin Associates on publicly announced mergers and acquisitions that involved physician groups or hospitals in the period 2008–15. 28 We used publicly available databases that linked practice names to tax identification numbers to identify the tax identification number or numbers for each acquired physician group. This allowed us to identify physicians who practiced in an acquired group (Appendix section VI). 25 We linked the acquired tax identification numbers and their constituent national provider identifiers to identifiers for providers that participated in a Medicare Shared Savings Program or Pioneer ACO contract. In within-market analyses, this linkage allowed us to compare changes in rates of acquisition from before to after implementation of the ACA for physicians entering ACO contracts in 2012, 2013, or 2014 versus nonparticipating physicians.

#### Commercial Prices:

With commercial claims data from the MarketScan database, we calculated a price index for inpatient and outpatient care at the Metropolitan Statistical Area level, using a group of services that covered a large share of spending. An index above 1 indicated an area in which mean services prices exceeded the national mean; an index below 1 indicated an area in which those prices were below the national mean (Appendix section VII). 25

#### Insurance Market Structure:

We used data on commercial enrollment by plan type for the period 2008–13 from HealthLeaders InterStudy 29 to create a commercial insurance market Herfindahl-Hirschman Index, using the share of covered lives as the measure of an insurer’s market share, and to assess commercial health maintenance organization (HMO) penetration in each year (calculated as the percentage of commercially insured people enrolled in an HMO). Finally, we used the Medicare Beneficiary Summary File to assess HMO penetration in Medicare (calculated as the percentage of Medicare beneficiaries in a Medicare Advantage HMO).

#### Between-Market Comparisons:

Using linear regression, we compared changes in provider consolidation from 2008–10 to 2011–13 between markets with higher versus lower ACO penetration as of 2014. We used 2008–10 as the pre period because we would not expect significant consolidation in response to the ACO programs before their enactment by the ACA in 2010. Using 2011–13 as the post period gave us 1–3 years of anticipatory consolidation and up to 2 years of consolidation following ACO entry for providers that entered ACO programs in 2012, 2013, or 2014.

We modeled each market-level measure of provider market structure (physician-hospital integration, physician group size, physician market concentration, hospital market concentration, and prices) as a function of an indicator for the post period, an interaction between ACO penetration and the post period, and Metropolitan Statistical Area indicators. The interaction estimated the differential change in provider market structure from the pre period to the post period that was associated with greater entry into Medicare ACO programs, as measured by 2014 ACO penetration. In the models, to adjust for effects of insurance market changes on provider consolidation and prices, we also included commercial insurance market concentration, commercial HMO penetration, and Medicare HMO penetration.

For each measure of provider market structure and prices, to facilitate interpretation of results, we present annual means by quartile of 2014 ACO market penetration. We also estimated overall national trends in the pre period and tested whether these trends changed in the post period. Finally, to explore potential ceiling effects (that is, to see whether ACO contracting occurred predominantly in already concentrated provider markets with less opportunity for further consolidation), we restricted our analyses to Metropolitan Statistical Areas in the lower three quartiles of the distribution for a given measure.

#### Within-Market Comparisons:

To hold market factors constant, we conducted within-market comparisons of changes in organizational characteristics from 2008–10 to 2011–13 between physicians or physician groups that entered an ACO contract by 2014 versus those that did not. The characteristics (linked to physicians via national provider identifiers or to groups via tax identification numbers) included a physician-level indicator of practicing in a hospital-owned facility, physician group size, and a physician-level indicator of practicing in a group acquired by a hospital or other group—all of which we assessed in each study year.

We modeled each characteristic as a function of Metropolitan Statistical Area indicators, a time-invariant indicator for Medicare ACO participation in 2014, an indicator for the post period, and an interaction between ACO participation and the post period. The interaction estimated the differential change from the pre to the post period in organizational structure for physicians who entered the ACO programs in 2012, 2013, or 2014, holding market factors constant. In the analysis of provider group size, we weighted each group (tax identification number) by the number of physicians at baseline in 2008–10 to facilitate our interpretation of results in terms of a physician’s average group size, consistent with our between-market analyses.

In a supplementary analysis, we similarly modeled physician group size and primary care orientation after stratifying groups based on their baseline primary care orientation to determine whether any growth in group size was primarily due to the incorporation of more primary care physicians or specialists.

### Limitations

Our study had several limitations. First, our analyses were descriptive and do not support causal conclusions about the effects of the ACO programs on provider consolidation. For example, ACO contracting could be associated with provider consolidation not because of the change in payment incentives but because providers that consolidated for other reasons were also more likely to participate in the ACO programs. Nevertheless, by assessing the relationship between ACO contracting and provider consolidation, we were able to observe whether trends were consistent with widely held expectations that new challenges from payment reform would accelerate consolidation as providers integrated to meet those challenges.

Second, changes in market-level drivers of both ACO participation and provider consolidation could have obscured or exaggerated a relationship between the two. In addition, Metropolitan Statistical Areas may not perfectly reflect the market for physician and hospital services. However, our supplemental between-physician comparisons within markets held factors at the area level constant, did not rely on market definitions to assess consolidation, and supported similar conclusions.

Third, we could assess provider consolidation only to the extent that it could be measured with claims data and publicly reported mergers and acquisitions. However, our analysis of prices should have reflected any unobservable provider consolidation, net of any independent effects of ACO contracting on price competition.

Finally, for most measures we could assess consolidation only through 2013 as related to ACO contracting through 2014, and therefore we may have missed more recent consolidation. Our post-ACA period allowed for three years of consolidation among providers planning to enter the ACO programs, however, and approximately one in five fee-for-service Medicare beneficiaries were in ACO contracts by 2014. 2

## Study Results

### Overall Trends

In the period 2008–13 all measures of provider market concentration and prices increased significantly ( $p < 0.001$ for annual changes; Appendix Exhibit A1). 25 In the study period, the average Metropolitan Statistical Area experienced a cumulative increase in physician-hospital integration of 6.3 percentage points (from 16.8 percent of physicians in a hospital-owned practice to 23.1 percent); an increase in physician concentration (Herfindahl-Hirschman Index) of 76 points; an increase in average physician group size of 22 physicians; an increase in hospital concentration (Herfindahl-Hirschman Index) of 279 points; and increases in inpatient and outpatient price indices of 28 percent and 14 percent, respectively.

For most measures of concentration and prices, trends changed minimally from the pre period (2008–10) to the post period (2011–13) (Appendix Exhibit A2). 25 However, group size grew much faster during the post period (adding an additional 1.6 physicians per group per year; $p=0.09$ ). There was also a clear surge in the number of hospital mergers in the post period, but no clear increase in mergers and acquisitions involving physician groups, apart from a spike in 2011 (Appendix Exhibit A3). 25

### Between-Market Analysis

By 2014, ACO penetration had reached an average of 21.3 percent but varied considerably across Metropolitan Statistical Areas (interquartile range: 2.7–32.6 percent) (Appendix Exhibit A4). 25 Notably, in 2008–10, markets with higher 2014 ACO penetration had significantly higher levels of physician-hospital integration but more competitive hospital and insurance markets and higher commercial HMO penetration ( Exhibit 1 ).

 2014 ACO market penetration quartile MSA-level characteristics 1 (lowest) 2 3 4 (highest) Physician-hospital integration 13.6% 15.7% 18.1% 23.8% ** Mean physician group size 96.2 77.8 75.5 118.6 Mean physician HHI 953 987 547 1,036 Mean hospital HHI 4,950 4,619 3,181 4,703 ** Mean insurer HHI 2,958 2,550 2,275 2,480 **** Medicare HMO penetration 8.0% 9.4% 11.8% 5.9% Commercial HMO penetration 16.1% 18.6% 20.9% 21.0% **** Inpatient services price index 0.80 0.80 0.78 0.77 * Outpatient services price index 0.96 0.89 0.92 0.96 Physician- or practice-level characteristics ACO nonparticipant in 2012, 2013, or 2014 ACO participant in 2012, 2013, or 2014 Physicians practicing in a hospital-owned facility 16.9% 20.2% **** Physicians practicing in a group acquired by a hospital or other group 0.1% 0.4% **** Mean physician group size 69.3 130.6 ****

SOURCE Authors’ analysis of data from the Centers for Medicare and Medicaid Services, American Hospital Association, HealthLeaders InterStudy, and Irving Levin Associates. NOTES There were seventy-two physician groups in each quartile except for quartile 1, where there were seventy-three. Mean physician group (practice) size was calculated as the average number of physicians billing for outpatient care within a tax identification number, weighting each group by its share of total physicians in the MSA. A practice refers to all providers using the same tax identification number. A price index above 1 indicates an MSA in which mean services prices exceed the national mean; a price index below 1 indicates an MSA in which mean services prices are below the national mean. Significance refers to a chi-square test for trend across quartiles. HHI is Herfindahl-Hirschman Index. HMO is health maintenance organization.

*$p < 0.1$

**$p < 0.05$

****$p < 0.001$

In comparisons of provider market consolidation by 2014 ACO penetration, we found that provider market structure differed at baseline by 2014 ACO penetration and changed over time. However, we also found that, compared to markets with lower 2014 ACO participation, those with high participation did not experience greater growth from the pre to the post period in vertical physician-hospital integration or physician group size ( Exhibit 2 ) or in physician market concentration, hospital market concentration, or commercial health care prices (Appendix Exhibits A5 and A6). 25 Our results were similar in sensitivity analyses that focused specifically on hospital integration of primary care physicians, used only Metropolitan Statistical Areas in the lower three quartiles of the distribution for each dependent variable, used an alternative measure of market concentration, or excluded insurance market structure variables.

### Within-Market Analysis

In the pre period, compared to physicians who would not be participating in an ACO by 2014, physicians in the same Metropolitan Statistical Area who would later be participating were more likely to be integrated with a hospital, to practice in a large group, and to have their practice acquired by a hospital or physician group ( Exhibit 3 and Appendix Exhibits A7 and A8). 25 Of our measures of physician organizational structure, only group size increased differentially by ACO participation. From the pre to the post period, the average group size for physicians who participated in an ACO grew by 11.4 more physicians than for physicians in the same Metropolitan Statistical Area who did not. An analysis of group size and primary care orientation after stratification by baseline specialty mix revealed that this growth in group size was driven largely by the addition of more specialists or specialty practices to organizations that were already large and composed primarily of specialists in the pre period (Appendix Exhibit A9). 25

 Not participating in ACO programs by 2014 Participating in ACO programs by 2014 Physician- or practice-level characteristics Before ACA (2008–10) After ACA (2011–13) Before ACA (2008–10) After ACA (2011–13) Difference-in-differences over time, participants vs. nonparticipants Physicians practicing in a hospital-owned facility 16.9% 19.3% 20.2% 23.7% 1.1% Physicians whose practice was acquired by a hospital, hospital system, or medical group 0.1% 0.2% 0.4% 0.3% −0.2% Mean physician group size 69.3 77.2 130.6 149.9 11.4 **

SOURCE Authors’ analysis of data from the Centers for Medicare and Medicaid Services and Irving Levin Associates. NOTES Each line results from separate within-market analyses, presenting the average pre- and post-period value of the dependent variable, by 2012, 2013, or 2014 ACO participation. Mean physician group size was calculated as the average number of physicians billing for outpatient care within a tax identification number, weighting each group by its share of total physicians in the Metropolitan Statistical Area. Percentages of practices acquired were calculated as the shares of physicians billing under a tax identification number that was identified as the target of a merger or acquisition in any given year.

**$p < 0.05$

## Discussion

Many policy analysts have predicted that providers would respond to the rapid growth of new payment models by forming larger organizations to assume financial risk and succeed under these models. However, we found little evidence to support this prediction.

Markets with greater 2014 ACO participation did not experience differential changes in physician-hospital integration, physician group size, physician market concentration, hospital market concentration or, importantly, commercial health care prices from 2008 to 2013. We found that physicians who entered a Medicare ACO program between 2012 and 2014 showed no differential increase in integration with hospitals or rates of acquisition from the pre to the post period, compared with other physicians in the same market. Physician groups that entered an ACO program did exhibit significantly greater growth in size than other practices in their market. This differential increase in group size among ACO participants was driven largely by the addition of specialists—not primary care physicians—to practices that were already specialty oriented, which suggests that they did not grow in order to become ACOs. For a specialty-oriented group to position itself to enter an ACO contract, one would expect it to reorient itself toward primary care. Similarly, we found no evidence that greater integration of primary care physicians with hospitals from the pre to the post period was related to ACO participation.

We also found an overall increase in hospital mergers after the ACA without changes in hospital market concentration related to ACO penetration, and a significant inverse relationship between hospital market concentration in the pre period and the extent of subsequent ACO contracting. These findings suggest that new payment models may have triggered some consolidation as a defensive reaction to the threat these models could pose, rather than as a way to achieve efficiencies in response to the new incentives. Hospitals and specialists in particular might consolidate both horizontally and vertically to achieve sufficient market share to resist payer pressure to enter risk contracts 18 or weaken ACOs’ ability to exploit competition in hospital and specialty markets, and compel reductions in prices and service volume. Similarly, rhetoric about the benefits of integration under new payment models may have lent credence to arguments by hospitals and specialists about the clinical efficiencies derived from mergers and acquisitions that would have faced stiffer challenges before the ACA.

## Policy Implications

In general, the overall weak relationship we found between ACO contracting and consolidation from the pre to the post period should ease the concerns that provider consolidation is an inevitable consequence of payment reform 5,6,30 —concerns that might support arguments to slow the transition away from fee-for-service payment. Similarly, our findings would not support abandoning ACO-like global budget models in favor of smaller payment bundles to avoid price increases from the types of consolidation that many researchers and policy makers have assumed are required to manage a global budget.

However, our findings do nothing to diminish the importance of the trend toward less competitive provider markets and associated price increases. In fact, we also found suggestive evidence of acceleration in specialist and hospital consolidation potentially related to payment reform but not expected to support new payment models. Our methods could not determine whether this consolidation has been defensive in nature or attributable to other factors, such as those driving consolidation before the ACA. Nonetheless, our findings question the prevailing wisdom that payment reform is driving consolidation of providers as they seek to enter and succeed under new payment models. Thus, even if there has been some defensive consolidation, the weak relationship between ACO contracting and forms of consolidation that would support ACO contracts has important implications for antitrust law enforcement. Specifically, our study supports skepticism of claims by providers that they are consolidating primarily to engage in risk contracts and achieve efficiencies.

## ACKNOWLEDGMENTS

This research was previously presented at the Sixth Biennial Conference of the American Society of Health Economists, Philadelphia, Pennsylvania, June 13, 2016, and the AcademyHealth Annual Research Meeting, Boston, Massachusetts, June 27, 2016. The research was supported by grants from the Robert Wood Johnson Foundation (Changes in Health Care Financing and Organization Grant No. 71408), the Laura and John Arnold Foundation, and the National Institute on Aging of the National Institutes of Health (Grant No. P01 AG032952). The content is solely the responsibility of the authors and does not necessarily represent the official views of the Robert Wood Johnson Foundation, the Laura and John Arnold Foundation, or the National Institutes of Health.

## NOTES

• 1 Shatto JD . Center for Medicare and Medicaid Innovation’s methodology and calculations for the 2016 estimate of fee-for-service payments to alternative payment models [Internet]. Baltimore (MD) : Centers for Medicare and Medicaid Services ; 2016 Mar 3 [cited 2016 Dec 13 ]. Available from: https://innovation.cms.gov/Files/x/ffs-apm-goalmemo.pdf Google Scholar
• 2 HHS.gov [Internet]. Washington (DC) : Department of Health and Human Services . News release, New hospitals and health care providers join successful, cutting-edge federal initiative that cuts costs and puts patients at the center of their care ; 2016 Jan 11 [cited 2016 Dec 13 ]. Available from: http://www.hhs.gov/about/news/2016/01/11/new-hospitals-and-health-care-providers-join-successful-cutting-edge-federal-initiative.html Google Scholar
• 3 McWilliams JM , Hatfield LA , Chernew ME , Landon BE , Schwartz AL . Early performance of accountable care organizations in Medicare . N Engl J Med . 2016 ; 374 ( 24 ): 2357 – 66 . Crossref, Medline
• 4 Muhlestein D . Growth and dispersion of accountable care organizations, 2015 . Health Affairs Blog [blog on the Internet]. 2015 Mar 31 [cited 2016 Dec 13 ]. Available from: http://healthaffairs.org/blog/2015/03/31/growth-and-dispersion-of-accountable-care-organizations-in-2015-2/ Google Scholar
• 5 Kocher R , Sahni NR . Hospitals’ race to employ physicians—the logic behind a money-losing proposition . N Engl J Med . 2011 ; 364 ( 19 ): 1790 – 3 . Crossref, Medline
• 6 Baicker K , Levy H . Coordination versus competition in health care reform . N Engl J Med . 2013 ; 369 ( 9 ): 789 – 91 . Crossref, Medline
• 7 Gaynor M , Town R . The impact of hospital consolidation—update [Internet]. Princeton (NJ) : Robert Wood Johnson Foundation ; 2012 Jun [cited 2016 Dec 13 ]. Available from: http://www.rwjf.org/content/dam/farm/reports/issue_briefs/2012/rwjf73261 Google Scholar
• 8 Baker LC , Bundorf MK , Royalty AB , Levin Z . Physician practice competition and prices paid by private insurers for office visits . JAMA . 2014 ; 312 ( 16 ): 1653 – 62 . Crossref, Medline
• 9 Dunn A , Shapiro AH . Do physicians possess market power? Journal of Law and Economics . 2014 ; 57 ( 1 ): 159 – 93 . Crossref
• 10 Neprash HT , Chernew ME , Hicks AL , Gibson T , McWilliams JM . Association of financial integration between physicians and hospitals with commercial health care prices . JAMA Intern Med . 2015 ; 175 ( 12 ): 1932 – 9 . Crossref, Medline
• 11 Baker LC , Bundorf MK , Kessler DP . Vertical integration: hospital ownership of physician practices is associated with higher prices and spending . Health Aff (Millwood) . 2014 ; 33 ( 5 ): 756 – 63 . Go to the article
• 12 Song Z , Wallace J , Neprash HT , McKellar MR , Chernew ME , McWilliams JM . Medicare fee cuts and cardiologist-hospital integration . JAMA Intern Med . 2015 ; 175 ( 7 ): 1229 – 31 . Crossref, Medline
• 13 Pear R . Consumer risks feared as health law spurs mergers . New York Times . 2010 Nov 20 . Google Scholar
• 14 Mathews AW . Health-care providers, insurers supersize . Wall Street Journal . 2015 Sep 21 . Google Scholar
• 15 Rittenhouse DR , Shortell SM , Gillies RR , Casalino LP , Robinson JC , McCurdy RK , et al. Improving chronic illness care: findings from a national study of care management processes in large physician practices . Med Care Res Rev . 2010 ; 67 ( 3 ): 301 – 20 . Crossref, Medline
• 16 Rittenhouse DR , Casalino LP , Shortell SM , McClellan SR , Gillies RR , Alexander JA , et al. Small and medium-size physician practices use few patient-centered medical home processes . Health Aff (Millwood) . 2011 ; 30 ( 8 ): 1575 – 84 . Go to the article
• 17 Crespin DJ , Christianson JB , McCullough JS , Finch MD . Health system consolidation and diabetes care performance at ambulatory clinics . Health Serv Res . 2016 ; 51 ( 5 ): 1772 – 95 . Crossref, Medline
• 18 McWilliams JM , Chernew ME , Zaslavsky AM , Hamed P , Landon BE . Delivery system integration and health care spending and quality for Medicare beneficiaries . JAMA Intern Med . 2013 ; 173 ( 15 ): 1447 – 56 . Crossref, Medline
• 19 Casalino LP , Pesko MF , Ryan AM , Mendelsohn JL , Copeland KR , Ramsay PP , et al. Small primary care physician practices have low rates of preventable hospital admissions . Health Aff (Millwood) . 2014 ; 33 ( 9 ): 1680 – 8 . Go to the article
• 20 Mehrotra A , Epstein AM , Rosenthal MB . Do integrated medical groups provide higher-quality medical care than individual practice associations? Ann Intern Med . 2006 ; 145 ( 11 ): 826 – 33 . Crossref, Medline
• 21 Baker LC , Bundorf MK , Kessler DP . The effect of hospital/physician integration on hospital choice [Internet]. Cambridge (MA) : National Bureau of Economic Research : (NBER Working Paper No. 21497). 2015 Aug [cited 2016 Dec 13 ]. Available for download (fee required) from: http://www.nber.org/papers/w21497 Google Scholar
• 22 McWilliams JM . Changes in Medicare Shared Savings Program savings from 2013 to 2014 . JAMA . 2016 ; 316 ( 16 ): 1711 – 3 . Crossref, Medline
• 23 Song Z , Rose S , Safran DG , Landon BE , Day MP , Chernew ME . Changes in health care spending and quality 4 years into global payment . N Engl J Med . 2014 ; 371 ( 18 ): 1704 – 14 . Crossref, Medline
• 24 Colla CH , Lewis VA , Tierney E , Muhlestein DB . Hospitals participating In ACOs tend to be large and urban, allowing access to capital and data . Health Aff (Millwood) . 2016 ; 35 ( 3 ): 431 – 9 . Go to the article
• 25 To access the Appendix, click on the Appendix link in the box to the right of the article online.
• 26 Research Data Assistance Center . Shared Savings Program accountable care organizations (ACO) provider-level RIF [Internet]. Minneapolis (MN) : ResDAC ; c 2016 [cited 2016 Dec 13 ]. Available from: http://www.resdac.org/cms-data/files/ssp-aco-provider-level-rif Google Scholar
• 27 Medicare Payment Advisory Commission . Online appendixes: Medicare payment differences across ambulatory settings [Internet]. In Report to the Congress: Medicare and the health care delivery system. Washington (DC) : MedPAC ; 2013 Jun [cited 2016 Dec 13 ]. Available from: http://www.medpac.gov/docs/default-source/reports/jun13_ch02_appendix.pdf Google Scholar
• 28 Irving Levin Associates . 2016 health care services acquisition report . Norwalk (CT) : Irving Levin Associates ; 2016 . Google Scholar
• 29 Decision Resources Group . Research and data [Internet]. Burlington (MA) : DRG ; [cited 2016 Dec 21 ]. Available from: http://hl-isy.com/hpda Google Scholar
• 30 Kocher B . How I was wrong about ObamaCare . Wall Street Journal . 2016 Jul 31 . Google Scholar
Loading Comments...