Discussion Forum

By Brent D. Fulton

Health Care Market Concentration Trends In The United States: Evidence And Policy Responses

ABSTRACT Policy makers and analysts have been voicing concerns about the increasing concentration of health care providers and health insurers in markets nationwide, including the potential adverse effect on the cost and quality of health care. The Council of Economic Advisers recently expressed its concern about the lack of estimates of market concentration in many sectors of the US economy. To address this gap in health care, this study analyzed market concentration trends in the United States from 2010 to 2016 for hospitals, physician organizations, and health insurers. Hospital and physician organization markets became increasingly concentrated over this time period. Concentration among primary care physicians increased the most, partially because hospitals and health care systems acquired primary care physician organizations. In 2016, 90 percent of Metropolitan Statistical Areas (MSAs) were highly concentrated for hospitals, 65 percent for specialist physicians, 39 percent for primary care physicians, and 57 percent for insurers. Ninety-one percent of the 346 MSAs analyzed may have warranted concern and scrutiny because of their concentration levels in 2016 and changes in their concentrations since 2010. Public policies that enhance competition are needed, such as stricter enforcement of antitrust laws, reducing barriers to entry, and restricting anticompetitive behaviors.

T he US health care system relies on competition in the provider and health insurer markets to lower costs and improve quality. Howev- er, the market concentration of

hospitals and insurers has been a matter of con- cern for several decades.1–6 More recently, Mar- tin Gaynor and colleagues reviewed studies of the competitive landscape of hospitals, health insurers, and physician services and found that hospital and health insurer markets have be- come more concentrated since the 1990s.7

To measure market concentration, the Anti- trust Division of the Department of Justice (DOJ) and Federal Trade Commission (FTC) of- ten use the Herfindahl-Hirschman Index (HHI),

which is calculated by squaring the market shares of each firm competing in a market and summing those values across all firms, resulting in a range from 0 to 10,000.8 Gaynor and col- leagues reported that 65 percent of Metropolitan Statistical Areas (MSAs) had highly concentrat- ed hospital markets (those with HHIs greater than 2,500) in 1990, and that share had in- creased to 77 percent by 2006.7 By way of exam- ple, an HHI of 2,500 could result from each of four firms in a given market having a 25 percent market share. There is less historical informa- tion on concentration in physician markets, but Gaynor and colleagues reported that those mar- kets were generally unconcentrated (with HHIs less than 1,500), particularly for primary care

doi: 10.1377/hlthaff.2017.0556 HEALTH AFFAIRS 36, NO. 9 (2017): 1530–1538 ©2017 Project HOPE— The People-to-People Health Foundation, Inc.

Brent D. Fulton (fultonb@ berkeley.edu) is an assistant adjunct professor in the School of Public Health at the University of California, Berkeley.

1530 Health Affairs September 2017 36:9

Market Concentration

o n S

e p te

m b e r 2

5 , 2

0 1 7

b y H

W T

e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

physicians. Although provider concentration could pro-

duce efficiencies that benefit purchasers of health care services, the evidence does not point in that direction. For example, reviews of studies of hospital markets have found that concentrat- ed markets are associated with higher hospital prices, with price increases often exceeding 20 percent when mergers occur in such mar- kets.7,9 Of even greater concern, the reviews found that these price increases did not appear to improve quality: In some cases, higher hospi- tal concentration was associated with higher mortality rates. In comparison to the number of hospital

market studies, a relatively small number of studies has examined the impact of physician organization concentration. Overall, these stud- ies found that higher concentration was associ- ated with higher physician prices across a range of services, including three types of commonly billed office visits,10 office visits across ten prom- inent specialties,11 orthopedics,12 cardiology and orthopedics,13 and common outpatient pro- cedures.14

A significant share of health care services in the United States is purchased by health insurers via employer-sponsored insurance and the indi- vidual market, including the Affordable Care Act Marketplaces. Increased health insurer con- centration could result in lower premiums to employers and consumers along two pathways: having efficiencies from economies of scale and insurers’ negotiating lower prices with hospitals and physician organizations that are attempting to charge prices above the competitive level. To some extent, research shows that this is happen- ing, with higher health insurer concentration being associated with lower hospital15–17 and phy- sician prices.10,17 However, the evidence shows that these price reductions are not passed on to consumers. A number of studies have found that higher health insurer concentration leads to higher premiums,7,18 including for employers19–21

and for individuals purchasing Marketplace plans.22 (This effect was moderated in Covered California, the Marketplace in an active-purchas- er state that selectively contracts and negotiates premiums with insurers.)23

Objectives The Council of Economic Advisers recently ex- pressed concern about the lack of estimates of market concentration across many sectors of the US economy.24 This study fills that gap for the most recent period for a significant portion of the health care sector. First, it shows how hospi- tal, physician organization, and health insurer

market concentration changed at the MSA level from 2010 to 2016. Second, it reports the num- ber of MSAs that may have warranted concern and scrutiny, based on the market concentration level in 2016 and the change in concentration since 2010. Third, because of the large increase in primary care physician market concentration during this period, it examines whether the oth- er health care subsectors’ concentration levels and changes were associated with this increase.

Study Data And Methods Market Concentration Measure And Data Sources Market concentration was measured by the HHI, because of its widespread use. The sources used to estimate HHIs included the fol- lowing annual data for the period 2010–16 (as of January 1 each year): for hospitals, the American Hospital Association (AHA) Annual Survey data- base; for physicians, the SK&A Office Based Physicians Database provided by IMS Health (now Quintiles); and for insurers, the Managed Market Surveyor File from HealthLeaders Inter- Study (now Decision Resources Group). The market shares of hospitals, physician organiza- tions, and insurers were based on the numbers of inpatient admissions, physicians, and enrollees, respectively. Market concentration was measured using the

product and geographic market definitions that are discussed next. Product Market Definitions The hospital

product was defined as a cluster of all inpatient services from short-term general hospitals. Hospital systems were treated as a single firm because they typically negotiate with payers as a system. The physician organization products included

services from the following five types of pro- viders: primary care physicians, cardiologists, oncologists/hematologists, radiologists, and or- thopedists. These types were selected because in each case at least 10,000 physicians were includ- ed in the data source, and all of the types except for primary care physicians are among the most highly compensated types. Physician organiza- tions owned by a group medical practice, hospi- tal, or health care system (which always included at least one hospital) were treated as a single firm. The four specialist physician organization products were combined into one HHI by calcu- lating the mean HHI across the specialties, weighted by the number of physicians in each specialty. The health insurer product was defined as

point-of-service plans and plans with preferred provider organizations, exclusive provider or- ganizations, and health maintenance organiza-

September 2017 36:9 Health Affairs 1531

o n S

e p te

m b e

r 2 5

, 2 0 1

7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

tions for the employer-sponsored market—in- cluding fully and self-insured plans—and the individual market outside of the Marketplaces (because these two markets could not be sepa- rated in the data). This product was constructed because an insurer’s market share of enrollees across these lines of business is correlated with the insurer’s market share of purchased hospital and physician services, thus indicating the con- centration level of the primary purchasers of provider services. Marketplace, Medicare Advantage, and Med-

icaid managed care enrollees were excluded, be- cause provider reimbursement rates for these enrollees are low, constrained by Medicare fee- for-service prices and Medicaid managed care premiums, or both. Additional information on the product definitions is in the online Ap- pendix.25

Geographic Market Definitions The geo- graphic market for hospitals, specialist physi- cian organizations, and insurers in this study was an MSA, which has been commonly used in other studies on market concentration of hospitals,15,26 specialist physicians,27 and insur- ers.15,26,28 The Office of Management and Budget delineated 382 MSAs in the United States,29 in which 278 million people (86 percent of the US population) resided in 2016, according to the Census Bureau.30 The geographic market for primary care physician organizations was a pri- mary care service area (PCSA), the smallest geographic area that can be a discrete service area for primary care physicians.31 The Health Resources and Services Administration defined 7,144 PCSAs in the United States,32 and PCSA- level HHIs were aggregated to the MSA level by weighting them using the PCSA population. Summary Measure To Evaluate Changes In

Market Concentration When the DOJ or FTC evaluates a proposed merger or acquisition, the agency considers how the proposed action will change the market concentration and what the resulting concentration will be. The agencies’ Horizontal Merger Guidelines8 specify the follow- ing postmerger HHI thresholds and the changes in the HHI that warrant different levels of con- cern and scrutiny, as the basis for further inves- tigation: The level is high if the HHI is more than 2,500 and the change in HHI is more than 200, because the merger or acquisition is “presumed to be likely to enhance market power”; moderate either if the HHI is more than 2,500 and the change in HHI is at least 100 and not more than 200, or if the HHI is at least 1,500 and not more than 2,500 and the change in HHI is at least 100, because the merger or acquisition “potentially raise[s] significant competitive concerns and of- ten warrant[s] scrutiny”; and low if either the

HHI is less than 1,500 or the change in HHI is less than 100, because in this case the merger or acquisition is “unlikely to have adverse competi- tive effects and ordinarily require[s] no further analysis.” Thesecriteria wereappliedtoevaluate changes

in market concentration at the MSA level in the period 2010–16 and to identify the number of MSAs that might warrant either a high or mod- erate level of concern and scrutiny. The change in HHI in each MSA for each product or product grouping was based on regressing the HHI on theyear, toestimate thechangeusingup toseven data points (for more details, see the Appen- dix).25 Because the criteria were based on both changes in concentration and ending levels of concentration, an MSA that was highly concen- trated (for example, one with an HHI of 5,000) in a particular subsector in 2016 would not even warrant moderate concern and scrutiny unless the HHI increased by at least 100 points. There- fore, these criteria were relevant for analyzing concentration trends and ending levels, not merely concentration levels at a point in time. Regression Models From 2010 to 2016, the

concentration of primary care physician organ- izations increased more than that of any other subsectors I examined. Therefore, I explored fac- tors associated with the 2016 levels and this change using four regression models. In models 1 and 2, the dependent variable was the primary care physician organization HHI in 2016, and the key independent variables were the HHIs of hospitals, specialist physician organizations, and insurers in the same year. These models tested whether primary care physician concen- tration in 2016 was associated with concentra- tion levels in other subsectors. Model 2 added health care and demographic control variables described in the Appendix.25

In models 3 and 4, the dependent variable was

Although provider concentration could produce efficiencies that benefit purchasers, the evidence does not point in that direction.

Market Concentration

1532 Health Affairs September 2017 36:9

o n S

e p te

m b e

r 2 5

, 2 0 1

7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

the annual change in the primary care physician organization HHI from 2010 to 2016. The key independent variables were the 2010 HHIs and the annual changes in the same period in the HHIs of hospitals, specialist physician organiza- tions, and insurers. These long-differenced mod- els (described in the Appendix)25 tested whether the increase in the concentration of primary carephysician organizations wasassociated with levels and changes in concentration in other subsectors. Model 4 added health care and demographic control variables (described in the Appendix).25

Limitations The study had several limitations. First, the AHA defines a system as “an entity with two or more hospitals owned, leased, sponsored, or contract managed by a central organiza- tion,”33 but it has been known to inaccurately record which hospitals are part of hospital sys- tems.34 Therefore, I used Irving Levin Associates’ annual Health Care Services Acquisition Report35 to check whether the AHA captured those mergers during the years of my study period, and I made a few minorcorrectionsto theAHA data to account for them. My reported hospital market concen- trations may still be understated, because some affiliations among hospitals that did not meet the AHA definition of a system may operate as a system when negotiating with payers. Second, the geographic definitions of MSAs

and PCSAs have limitations, because some of the areas are either too small or too large for some products. Because my study primarily ex- amined trends, if MSAs or PCSAs are sometimes too small or too large, they are likely to remain so over time. It was not feasible to define markets using formal hypothetical monopolist tests and structural models.8,36 However, when I compared my mean concentration estimates to those of a study that estimated market concentration boundaries using fixed travel times to cardiolo- gists and orthopedists, my estimates were con- sistent with that study.13

Third, to identify MSAs that may warrant ei- ther a high or moderate level of concern and scrutiny, I examined the HHI level in 2016 and the change in HHI since 2010, based on the thresholds in the Horizontal Merger Guidelines.8

However, the DOJ and FTC use many nuanced criteria to evaluate markets and do not apply these thresholds strictly.37 Thus, using only the thresholds in this manner should be considered as simply an initial screen. Moreover, these thresholds are used along with other criteria to evaluate proposed mergers. Thus, the thresholds that these agencies might use to evaluate changes in market concentration in general may differ.

Study Results From 2010 to 2016, the mean HHIs for hospitals and specialist physician organizations each in- creased by about 5 percent, while the HHI for health insurers declined by less than 1 percent (Exhibit 1). The HHI for primary care physician organizations increased the most among the subsectors, by nearly 29 percent. As a result of these changes, 90 percent of MSAs had an HHI for hospitals of more than 2,500 in 2016 (Exhib- it 2). Nearly 39 percent of MSAs had an HHI of that level for primary care physician organiza- tions in 2016—an increase of 85 percent from 2010. Ninety-one percent of the 382 MSAs analyzed

(in which 202 million people resided in 2016) may have warranted either a high or moderate level of concern and scrutiny for a least one of the four products or product groupings based on the Horizontal Merger Guidelines8 (Exhibit 3). This concern may have been warranted for hospitals in 184 MSAs, specialist physician organizations in 196 MSAs, primary care physician organiza- tions in 205 MSAs, and insurers in 170 MSAs. However, there was some overlap across sub-

sectors. For example, concern may have been warranted for hospitals and primary care physi-

Exhibit 1

Mean Metropolitan Statistical Area Herfindahl-Hirschman Index (HHI) for hospitals, physician organizations, and health insurers, 2010–16

SOURCES Author’s analysis of data from the American Hospital Association Annual Survey, the SK&A Office Based Physicians Database from IMS Health, and the Managed Market Surveyor File from HealthLeaders InterStudy. NOTES The HHI calculations are explained in the text. Specialist physi- cians include providers in the fields of cardiology, oncology or hematology, radiology, and orthope- dics. Insurers include preferred provider organization, exclusive provider organization, point-of- service plan, and health maintenance organization products in both the group and non-Marketplace individual markets, as explained in the text. HHIs for hospitals and specialist physician organizations increased 5.2 percent; for insurers, they declined 0.9 percent; and for primary care physician organ- izations, they increased 28.8 percent.

September 2017 36:9 Health Affairs 1533

o n S

e p te

m b e r 2

5 , 2

0 1 7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

cian organizations in 110 MSAs. When all four products or product groupings were combined, concern may have been warranted in only 36 MSAs. This is because the pairwise correlations of the annual change in HHI from 2010 to 2016 were low among the four products or product groupings (for the pairwise correlations, see Appendix Exhibit A4).25

From 2010 to 2016, the share of primary care physicians working in organizations owned by a hospital or health care system increased from 28 percent to 44 percent—a dramatic increase of 57 percent—while the shares in independent solo practice or organizations owned by a medi- cal group decreased (Exhibit 4). In the regression analysis, a one-point

increase in the HHI for specialist physician or- ganizations in 2016 was associated with a 0.24 increase in the primary care physician organiza- tion HHI in 2016 (p < 0:001) (Appendix Exhib- it A6, model 2).25 In contrast, a one-point in- crease in the HHI for insurers in 2016 was associated with a 0.13 decrease in the primary care physician organization HHI in 2016 (p < 0:05). The hospital HHI in 2016 was not associated with the primary care physician orga- nization HHI in 2016. Model 4 shows that a 1-point higher annual change in the HHI for specialist physician organizations from 2010

to 2016 was weakly associated with 0.10-point higher annual change in the HHI for primary care physician organizations (p ¼ 0:12). No such association was found for changes in the HHI for hospitals or insurers.

Discussion The markets for hospitals, specialist physician organizations, and primary care physician or- ganizations at the Metropolitan Statistical Area level became more concentrated across the Unit- ed States between 2010 and 2016. In 2016, 90 percent of MSAs were highly concentrated for hospitals, 65 percent for specialist physi- cians, 39 percent for primary care physicians, and 57 percent for insurers. The hospital concen- tration trends from the 1990s and early 2000s have continued.7 Furthermore, my results are consistent with the fact that more physicians are joining larger physician organizations38 and that more physicians work either directly for hospitals or in organizations that are owned by hospitals.39

Based on the criteria for HHI levels and changes in them in the Horizontal Merger Guide- lines,8 in 2016, 202 million people resided in 346 MSAs where either a high or a moderate level of concern and scrutiny may have been warranted for at least one of the four products or product groupings described above. However, relatively few MSAs were found to warrant scrutiny across all of the products or product groupings, which suggests that systematic market-level factors do not drive concentration increases across all subsectors. This scrutiny typically includes examining other competitive factors that may influence the potentially harmful impact of increased concentration, such as the likelihood of coordinated or unilateral exercise of market power—including ease of entry, significant con- solidation-specific efficiencies, and the presence of powerful buyers. The increased concentration in primary care

physician organizations was associated with more physician organizations being owned by hospitals or health care systems. However, no relationship was found between this increased concentration and either hospital market con- centration trends or levels. On the one hand, larger physician organiza-

tions, particularly those owned by hospitals or health care systems, have been linked to using more health information technology, electronic medical records, and care management process- es, which enables them to more easily adapt to value- and risk-based reimbursement.40,41 On the other hand, if a physician organization had mar- ket power before being acquired, the acquisition

Exhibit 2

Percentages of Metropolitan Statistical Areas (MSAs) whose Herfindahl-Hirschman Index (HHI) was above 2,500 for hospitals, physician organizations, and health insurers, 2010–16

SOURCES Author’s analysis of data from the American Hospital Association Annual Survey, the SK&A Office Based Physicians Database from IMS Health, and the Managed Market Surveyor File from HealthLeaders InterStudy. NOTES The HHI calculations are explained in the text. Specialist physi- cians and insurers are explained in the Notes to Exhibit 1. The percentage of MSAs with HHIs above 2,500 increased 4.2 percent for hospitals, 8.7 percent for specialist physician organizations, 2.3 per- cent for insurers, and 85.2 percent for primary care physician organizations.

Market Concentration

1534 Health Affairs September 2017 36:9

o n S

e p te

m b e r 2

5 , 2

0 1 7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e

d fro

m

could increase hospital market power if it fore- closed rival hospitals from access to physician services,42 a concern recently discussed by the FTC.43 In fact, hospitals that have stronger affili- ations with physicians have been found to have higherprices.44,45 Similarly,if a hospital orhealth care system had market power before it acquired a physician organization, the acquisition could increase physician market power if it foreclosed rival physician organizations from access to hos- pital services. The combined effect of higher hospital and physician prices results in health insurance premiums becoming even more un- affordable to many employers and consumers. A recent article provides a framework that can be used to assess vertical mergers and advocates for the DOJ’s 1984 Non-Horizontal Merger Guidelines to be updated.46

Policy Responses Given the evidence that increased concentration leads to higher health care prices and health insurance premiums, several important recent articles have suggested policy responses to en-

Exhibit 3

Percentages of Metropolitan Statistical Areas (MSAs) possibly warranting a high or moderate level of concern and scrutiny based on market concentration in 2016, and changes in that concentration, 2010–16

SOURCES Author’s analysis of data from the American Hospital Association Annual Survey, the SK&A Office Based Physicians Data- base from IMS Health, and the Managed Market Surveyor File from HealthLeaders InterStudy. NOTES The HHI calculations are ex- plained in the text. Specialist physicians and insurers are explained in the Notes to Exhibit 1. “Physician organization” includes both specialist and primary care physician organizations, with both types separately analyzed. In the ratios in parentheses, the numerator is the number of MSAs that warrant a high or moderate level of concern and scrutiny (hereafter, “concern”), and the denominator is the number of MSAs analyzed (which could be less than 382 MSAs because some MSAs were without a given type of product or product grouping). The population numbers are the 2016 populations (in millions) of the MSAs that may warrant a high or moderate level of scrutiny. Levels of concern are based on the Horizontal Merger Guidelines of the Department of Justice and Federal Trade Commission (see Note 8 in text). aThe level of concern and scrutiny in an MSA was based on the product or product grouping with the highest level. bThe level of concern and scrutiny in an MSA was high if all products or product groupings were at the high level. It was moderate if all products or product groupings were at the moderate or high level, but all were not at the high level.

Exhibit 4

Percentages of primary care physicians working in organizations, by ownership type, 2010–16

SOURCE Author’s analysis of data from the SK&A Office Based Physicians Database from IMS Health. NOTE The percentage of physicians working in an organization owned by a hospital or a health care system increased 57.0 percent, while the percentages in independent solo practice and in a medical group declined 15.1 percent and 30.9 percent, respectively.

September 2017 36:9 Health Affairs 1535

o n S

e p te

m b e r 2

5 , 2

0 1 7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h

ttp ://co

n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d

e d

fro m

hance market competition,47–50 and I discuss the following five below: enforce antitrust laws, re- duce barriers to entry, restrict anticompetitive behaviors, revise anticompetitive reimburse- ment incentives, and set provider and health insurer rates. First, proposed mergers and acquisitions

should be scrutinized by the federal and state governments to evaluate whether the net result is pro- or anticompetitive, as the DOJ and attor- neys general from multiple states did in the previously proposed Anthem-Cigna and Aetna- Humana insurer mergers.51 This scrutiny in- cludes evaluating whether the procompetitive effects could be achieved without the merger, as was ruled possible in the St. Luke’s case— which involved a hospital’s acquisition of a phy- sician group.52 This scrutiny could be enhanced at the state level. For example, the Massachusetts Health Policy Commission provides analyses of proposed health care mergers for the state’s at- torney general and the public.53

Second, policies that restrict market entry— such as hospital certificate-of-need laws, any-willing-provider regulations, and restrictive provider licensing and scope-of-practice regulations—should be evaluated to determine whether they enhance consumer welfare. Third, to enhance competition, particularly in

markets where providers are already highly con- centrated, anticompetitive behaviors should be restricted. Examples of these behaviors include antitiering clauses that force insurers to include a provider in the top tier and tying agreements that force insurers to contract with all hospitals in a system. Fourth, reimbursement policies that reduce

competition should be revised. For example, my study found an increase in the share of pri- mary care physicians working in organizations owned by a hospital or health care system. One reason for this increase is because of Medicare’s facility fee, which is paid to hospitals for physi- cian services that are provided outside the hos- pital at a site where the overhead is lower.47 The facility fee should be adjusted to reflect the site’s actual overhead rate. Otherwise, it provides an incentive for these acquisitions, which have the

potential to reduce competition among hospitals and among physician organizations.42

Fifth, seven states that began to regulate hos- pital rates in the 1970s generally had lower hos- pital spending growth.54 However, all but two of the states discontinued this practice because of private insurers’ shift to managed care and Med- icare’s shift to reimbursements based on diagno- sis-related groups. At the insurer level, health insurance rate review and Marketplace active- purchaser states that use selective contracting have both been found to be associated with lower growth rates inpremiums.23,55 Althoughprovider and insurer rate setting may be promising, it is challenging for regulators to set rates that account for changes in technology and input costs, and rating setting is subject to regulatory capture—whereby regulators become overly influenced by the regulated industry.

Conclusion Future studies should evaluate the impact of these policy options on competition and con- sumer welfare. Given health care market concen- tration levels and trends, it will be important to closely monitor market competition and to im- plement policies that help enhance competition, to ensure that competition plays a key role in lowering health care costs and improving health care quality. ▪

An earlier version of this article was presented at the “Impact of Healthcare Market Concentration on Healthcare Prices and Premiums: What Can and Should Be Done” conference at the Robert F. Wagner Graduate School of Public Service, New York University,

New York City, on April 14, 2017. The meeting was organized and funded by the Nicholas C. Petris Center on Health Care Markets and Consumer Welfare, University of California, Berkeley. The author thanks participants at the conference for their comments. Funding

for this study was provided by the Commonwealth Fund (Grant Nos. 20160413 and 20170922). The author thanks Daniel Arnold, Jason Tilipman, and Grayson Dimick, of the University of California, Berkeley, for their research assistance.

Policies that restrict market entry should be evaluated to determine whether they enhance consumer welfare.

Market Concentration

1536 Health Affairs September 2017 36:9

o n S

e p te

m b e

r 2 5

, 2 0 1

7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

NOTES

1 Gaynor M, Haas-Wilson D. Change, consolidation, and competition in health care markets. J Econ Perspect. 1999;13(1):141–64.

2 Haas-Wilson D. Managed care and monopoly power: the antitrust challenge. Cambridge (MA): Har- vard University Press; 2003.

3 Frech HE III. Competition and mo- nopoly in medical care. Washington (DC): American Enterprise Institute Press; 1996.

4 Robinson JC. Consolidation and the transformation of competition in health insurance. Health Aff (Mill- wood). 2004;23(6):11–24.

5 American Medical Association. Competition in health insurance. Chicago (IL): AMA; 2002–2016.

6 Department of Justice, Federal Trade Commission. Improving health care: a dose of competition [Internet]. Washington (DC): FTC; 2004 Jul [cited 2017 Jul 31]. Available from: https://www.ftc.gov/sites/default/ files/documents/reports/ improving-health-care-dose- competition-report-federal-trade- commission-and-department- justice/040723healthcarerpt.pdf

7 Gaynor M, Ho K, Town RJ. The in- dustrial organization of health-care markets. J Econ Lit. 2015;53(2): 235–84.

8 Department of Justice, Federal Trade Commission. Horizontal merger guidelines [Internet]. Washington (DC): FTC; 2010 Aug 19 [cited 2017 Jul 21]. Available from: https:// www.ftc.gov/sites/default/files/ attachments/merger-review/ 100819hmg.pdf

9 Gaynor M, Town R. The impact of hospital consolidation—update [In- ternet]. Princeton (NJ): Synthesis Project, Robert Wood Johnson Foundation; 2012 Jun [cited 2017 Aug 1]. (Policy Brief No. 9). Available from: http://www.rwjf.org/content/ dam/farm/reports/issue_briefs/ 2012/rwjf73261

10 Roberts ET, Chernew ME, McWilliams JM. Market share mat- ters: evidence of insurer and pro- vider bargaining over prices. Health Aff (Millwood). 2017;36(1):141–8.

11 Baker LC, Bundorf MK, Royalty AB, Levin Z. Physician practice compe- tition and prices paid by private in- surers for office visits. JAMA. 2014; 312(16):1653–62.

12 Sun E, Baker LC. Concentration in orthopedic markets was associated with a 7 percent increase in physi- cian fees for total knee replace- ments. Health Aff (Millwood). 2015; 34(6):916–21.

13 Dunn A, Shapiro AH. Do physicians possess market power? Journal of Law and Economics. 2014;57(1): 159–93.

14 Schneider JE, Li P, Klepser DG, Peterson NA, Brown TT, Scheffler

RM. The effect of physician and health plan market concentration on prices in commercial health insur- ance markets. Int J Health Care Fi- nance Econ. 2008;8(1):13–26.

15 Melnick GA, Shen Y-C, Wu VY. The increased concentration of health plan markets can benefit consumers through lower hospital prices. Health Aff (Millwood). 2011;30(9): 1728–33.

16 Moriya AS, Vogt WB, Gaynor M. Hospital prices and market structure in the hospital and insurance in- dustries. Health Econ Policy Law. 2010;5(4):459–79.

17 Scheffler RM, Arnold DR. Insurer market power lowers prices in con- centrated provider markets: who benefits? Health Aff (Millwood). 2017;36(9):1539–46.

18 Dafny LS. Health insurance industry consolidation: what do we know from the past, is it relevant in light of the ACA, and what should we ask? [Internet].Washington (DC): Senate Committee on the Judiciary, Sub- committee on Antitrust, Competi- tion Policy, and Consumer Rights; 2015 Sep 22 [cited 2017 Jul 25]. Available from: http://www .hbs.edu/faculty/Profile%20Files/ Testimony%20to%20Senate%20in %20re%20Insurance%20Industry %20Mergers%20-%209.2015_ 050cdb4e-db12-4a9d-9d50- 48d917d39e2a.pdf

19 Dafny LS. Are health insurance markets competitive? Am Econ Rev. 2010;100(4):1399–431.

20 Dafny L, Duggan M, Ramanarayanan S. Paying a premium on your pre- mium? Consolidation in the US health insurance industry. Am Econ Rev. 2012;102(2):1161–85.

21 Trish EE, Herring BJ. How do health insurer market concentration and bargaining power with hospitals af- fect health insurance premiums? J Health Econ. 2015;42:104–14.

22 Dafny L, Gruber J, Ody C. More in- surers lower premiums: evidence from initial pricing in the health insurance Marketplaces. Am J Health Econ. 2015;1(1):53–81.

23 Scheffler RM, Arnold DR, Fulton BD, Glied SA. Differing impacts of mar- ket concentration on Affordable Care Act Marketplace premiums. Health Aff (Millwood). 2016;35(5):880–8.

24 Council of Economic Advisers. Ben- efits of competition and indicators of market power [Internet]. Washing- ton (DC): CEA; 2016 Apr [cited 2017 Aug 1]. (Issue Brief). Available from: https://obamawhitehouse.archives .gov/sites/default/files/page/files/ 20160414_cea_competition_issue_ brief.pdf

25 To access the Appendix, click on the Appendix link in the box to the right of the article online.

26 Gaynor M.Health care industry

consolidation [Internet]. Washing- ton (DC): House Committee on Ways and Means; 2011 Sep 9 [cited 2017 Aug 1]. Available from: https:// waysandmeans.house.gov/ UploadedFiles/Gaynor_ Testimony_9-9-11_Final.pdf

27 Kleiner SA, White WD, Lyons S. Market power and provider consol- idation in physician markets. Int J Health Econ Manag. 2015;15(1): 99–126.

28 Emmons DW, Guardado JR, Kane CK. Competition in health insur- ance: a comprehensive study of U.S. markets. Chicago (IL): American Medical Association; 2016.

29 Office of Management and Budget. Revised delineations of Metropoli- tan Statistical Areas, Micropolitan Statistical Areas, and Combined Statistical Areas, and guidance on uses of the delineations of these areas [Internet]. Washington (DC): OMB; 2015 Jul 15 [cited 2017 Aug 1]. (OMB Bulletin No. 15-01). Available from: https://obamawhitehouse .archives.gov/sites/default/files/ omb/bulletins/2015/15-01.pdf

30 Census Bureau. Metropolitan and Micropolitan Statistical Area popu- lation totals tables: 2010–2016 [In- ternet]. Washington (DC): Census Bureau; [cited 2017 Aug 21]. Avail- able from: https://www.census.gov/ data/1tables/2016/demo/popest/ total-metro-and-micro-statistical- areas.html

31 Goodman DC, Mick SS, Bott D, Stukel T, Chang CH, Marth N, et al. Primary care service areas: a new tool for the evaluation of primary care services. Health Serv Res. 2003; 38(1 Pt 1):287–309.

32 Health Resources and Services Ad- ministration. Primary care service area version 3.1 methods [Internet]. Rockville (MD): HRSA; 2013 Sep 17 [cited 2017 Aug 1]. Available from: https://datawarehouse.hrsa.gov/ DataDownload/PCSA/2010/PCSA_ Version3.1_methods_9_17_2013 .pdf

33 American Hospital Association. AHA annual survey database: fiscal year 2015. Chicago (IL): AHA; 2016, p. 7.

34 Madison K. Hospital-physician af- filiations and patient treatments, expenditures, and outcomes. Health Serv Res. 2004;39(2):257–78.

35 Irving Levin Associates. The health care services acquisition report. Norwalk (CT): Irving Levin Associ- ates; 2011–16.

36 Gaynor MS, Kleiner SA, Vogt WB. A structural approach to market defi- nition with an application to the hospital industry. J Ind Econ. 2013; 61(2):243–89.

37 Shapiro C. The 2010 horizontal merger guidelines: from hedgehog to fox in forty years. Antitrust Law Journal. 2010;77(1):49–107.

September 2017 36:9 Health Affairs 1537

o n S

e p te

m b e

r 2 5

, 2 0 1

7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

38 Muhlestein DB, Smith NJ. Physician consolidation: rapid movement from small to large group practices, 2013– 15. Health Aff (Millwood). 2016; 35(9):1638–42.

39 Kane CK. Updated data on physician practice arrangements: inching to- ward hospital ownership [Internet]. Chicago (IL): American Medical As- sociation; 2015 [cited 2017 Aug 1]. Available from: https://www.m3 globalresearch.com/img/resources/ AMA_PRP_Physician_Practice_ Arrangements.pdf

40 Rittenhouse DR, Ramsay PP, Casalino LP, McClellan S, Kandel ZK, Shortell SM. Increased health information technology adoption and use among small primary care physician practices over time: a na- tional cohort study. Ann Fam Med. 2017;15(1):56–62.

41 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 prac- tices. Med Care Res Rev. 2010;67(3): 301–20.

42 Gaynor M. Is vertical integration anticompetitive? Definitely maybe (but that’s not final). J Health Econ. 2006;25(1):175–80.

43 Brill J. A common goal: the U.S. Federal Trade Commission’s health- care enforcement program and its implications for ACOs [Internet]. Washington (DC): Federal Trade Commission; 2015 Jun 17 [cited 2017 Aug 1]. Available from: https://www .ftc.gov/system/files/documents/ public_statements/673881/ 150617aco-summit.pdf

44 Baker LC, Bundorf MK, Kessler DP. Vertical integration: hospital own- ership of physician practices is as- sociated with higher prices and spending. Health Aff (Millwood). 2014;33(5):756–63.

45 Neprash HT, Chernew ME, Hicks AL, Gibson T, McWilliams JM. As- sociation of financial integration between physicians and hospitals with commercial health care prices. JAMA Intern Med. 2015;175(12): 1932–9.

46 Salop SC, Culley DP. Revising the U.S. vertical merger guidelines: pol- icy issues and an interim guide for practitioners. Journal of Antitrust Enforcement. 2016;4(1):1–41.

47 Gaynor M, Mostashari F, Ginsburg PB. Making health care markets work: competition policy for health care. JAMA. 2017;317(13):1313–4.

48 Fuse Brown EC, King JS. The double- edged sword of health care integra- tion: consolidation and cost control. Indiana Law J. 2016;92(55):55–112.

49 National Academy of Social Insur- ance. Addressing pricing power in health care markets: principles and policy options to strengthen and shape markets: the final report of the academy’s panel on pricing power in health care markets [Internet]. Washington (DC): The Academy; 2015 Apr [cited 2017 Aug 1]. Avail- able from: http://www.urban.org/ sites/default/files/publication/ 50116/2000212-Addressing-Pricing- Power-in-Health-Care-Markets.pdf

50 Gaynor M, Mostashari F, Ginsburg PB. Making health care markets work: competition policy for health care: actionable policy proposals for the executive branch, Congress, and

the states [Internet]. Pittsburgh (PA): Carnegie Mellon University Heinz College; 2017 Apr [cited 2017 Aug 1]. Available from: https:// www.brookings.edu/wp-content/ uploads/2017/04/gaynor-et-al-final- report-v11.pdf

51 Dafny LS. Good riddance to big in- surance mergers. N Engl J Med. 2017;376(19):1804–6.

52 Singer T. New health care sympo- sium: unpacking the issues of verti- cal and horizontal consolidation— the St. Luke’s case. Health Affairs Blog [blog on the Internet]. 2016 Mar 3 [cited 2017 Aug 1]. Available from: http://healthaffairs.org/blog/ 2016/03/03/new-health-care- symposium-unpacking-the-issues- of-vertical-and-horizontal- consolidation-the-st-lukes-case/

53 Mass.gov. Health Policy Commission [Internet]. Boston (MA): Common- wealth of Massachusetts; c 2017 [cited 2017 Aug 1]. Available from: http://www.mass.gov/anf/budget- taxes-and-procurement/oversight- agencies/health-policy-commission/

54 Atkinson G. State hospital rate-set- ting revisited [Internet]. New York (NY): Commonwealth Fund; 2009 Oct [cited 2017 Aug 1]. (Issue Brief). Available from: http://www .commonwealthfund.org/~/media/ files/publications/issue-brief/2009/ oct/1332_atkinson_state_hospital_ ratesetting_revisited_1015.pdf

55 Karaca-Mandic P, Fulton BD, Hollingshead A, Scheffler RM. States with stronger health insurance rate review authority experienced lower premiums in the individual market in 2010–13. Health Aff (Millwood). 2015;34(8):1358–67.

Market Concentration

1538 Health Affairs September 2017 36:9

o n S

e p te

m b e

r 2 5

, 2 0 1

7 b

y H W

T e a m

H e a lth

A ffa

irs b

y h ttp

://co n te

n t.h

e a lth

a ffa

irs.o rg

/ D

o w

n lo

a d e d fro

m

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.