Increases In Physician Professional Fees In Private Equity–Owned Gastroenterology Practices

Abstract

Consolidation of physician practices, largely driven by health systems, has motivated policy efforts to move care toward lower-price, non–health system settings. At the same time, however, private equity (PE) firms are increasingly acquiring those non–health system practices, potentially negating the prior price advantages of those practices. We used novel ownership data on gastroenterology practices linked to commercial claims for the period 2015–20 to study how PE acquisitions affect the prices and volume of care relative to both health system–affiliated practices and independent practices. We examined both professional fees and facility fees. After PE acquisition, prices increased by $92 per claim, or 28.4 percent, driven by a 78.1 percent increase in professional fees. Facility fees did not exhibit a statistically significant change. Meanwhile, utilization also increased. These findings suggest that PE firms have multiple avenues for raising prices—in this case, primarily via professional fees. For policy makers, although moving care out of higher-price health system settings remains a key strategy to lower spending, unchecked growth in professional fees in PE-acquired outpatient settings may nullify some of the intended effects.

Private equity (PE) firms have acquired an increasing number of US physician practices in recent years. Among the common targets are outpatient procedural specialty practices such as gastroenterology practices, likely because of the presence of a high number of independent practices, the profitability of ancillary services and ambulatory surgery centers, and opportunities for economies of scale and increased market power. Acquisitions of gastroenterology practices (physician offices and ambulatory surgery centers) by PE firms began in 2016 and had approached 14 percent of practicing gastroenterologists by 2021.^1,2

Existing research has examined the effects of PE consolidation relative to independent ownership. For example, one study from the period 2016–20 found that PE-acquired gastroenterology practices increased health care spending by 32 percent relative to independent practices, driven in part by increases in patient visits and service use.³ Another study estimated that PE-acquired gastroenterology practices increased prices by 18 percent relative to independent practices during 2012–21, with greater price increases in concentrated markets.⁴ Studies with comparison groups of independent practices shed light on how physician practice patterns change after independent physicians sell their practices to PE firms, which is a topic of increasing interest for physicians, professional societies, and policy makers.^3,5–7

However, to our knowledge, no prior research has compared PE ownership with the increasingly common alternative of health system ownership. Understanding potential changes in prices, if any, between physician practices acquired by PE relative to those owned by health systems is of particular importance within gastroenterology. First, the share of gastroenterologists employed by health systems rose from 23 percent to 38 percent between 2016 and 2018.⁸ Some physicians consider PE employment to be a favorable alternative to health system employment, with perceptions of improved clinical autonomy, independence, and reduced administrative burdens.^8–10 Second, the effects of PE consolidation relative to health system affiliation may be significant in the context of policy developments to shift care away from expensive hospital-based settings to office settings. Prices for the same services have been found to be significantly higher in hospital-based outpatient clinics compared with outpatient physician offices, without observed differences in quality of care.^11–14 This site-based price differential is primarily driven by facility fees that are intended to cover costs related to staffing, licensure, and technology systems,^13,14 prompting Congress to consider policy proposals to shift care away from expensive hospital-based clinics to less costly non-hospital-affiliated outpatient practices.

Given this growth of PE in gastroenterology^1,2,15 and the associated price increases, the overall effectiveness of policy efforts to promote health care affordability depends on the extent to which outpatient practices remain lower-price alternatives to health systems after PE acquisition. Specifically, price increases that result from PE acquisition^3,5,6 may erode savings associated with receiving care at independent practices relative to health systems. Despite the policy significance of this countervailing PE-related consolidation of outpatient practices, no prior research has quantified the price changes at outpatient practices after PE acquisition relative to the counterfactual of prices at health systems, or whether prices differ at all.

In this context, we linked data on PE acquisitions from the period 2015–19 to commercial claims data from the period 2015–20 and used an event study design with a difference-in-differences strategy to study the extent to which PE acquisition was associated with changes in professional and facility fees. We used two control groups. As in prior research, we compared PE-acquired practices with practices that remained affiliated with health systems for the same period. We also compared PE-acquired practices with practices that remained under independent ownership for the entire study period.

Study Data And Methods

To conduct our analyses, we built a panel data set using data from the period 2015–20 in multiple steps. Additional detail on data sources and the analytical approach is in online appendix A1.¹⁶

IDENTIFYING PE ACQUISITIONS

We combined several data sources to identify physicians affiliated with PE-acquired practices. To identify PE acquisitions in gastroenterology during 2016–19, we used proprietary data from PitchBook Inc.,^3,6 which include information on acquisition date and the names of some acquired practices. To identify stand-alone practice sites associated with each platform acquisition, we manually verified and expanded this list using a combination of press releases, industry reports, and physician practice websites.

IDENTIFYING PROVIDERS AFFILIATED WITH PRACTICES

To identify individual physicians (MDs and DOs) affiliated with PE-acquired, health system–affiliated, and independent practices, we used two databases from IQVIA, a health care data vendor: the 2016 SK&A Office Based Physicians database and the 2019 OneKey database, which use the same approaches to identify and verify affiliations.¹⁷ Both data sets are independently verified with clinician-level information (for example, age, location, specialty, clinician credentials, and National Provider Identifiers [NPIs]) and practice-level information, including ownership and corporate affiliations, on 9.7 million health professionals in the US.^18,19

LINKING ACQUISITION DEALS TO PRACTICE SITES

We used probabilistic record linkage algorithms to link exact and nonexact records of practice site names, addresses, and ownership information (for example, parent organization) in the 2019 OneKey data to reported acquisitions. For any unmatched deals, we manually matched a subset of acquisitions to the OneKey data by using publicly available information to verify practice locations and potential name changes for acquired practices. Our linking strategy produced a match rate of 71 percent across all deal years, for a total of 736 physicians in 148 PE-acquired practice sites.

IDENTIFYING HEALTH SYSTEM–AFFILIATED AND INDEPENDENT PRACTICES

To identify health system–affiliated control practices, we used OneKey to identify physician practices in gastroenterology whose corporate owner was defined as a health system or integrated delivery network. This definition included practices that had common ownership or management by a hospital or health system, and it has been used in prior studies on physician organization.^20–22 To assess the completeness and validity of IQVIA’s definition, and because there is no standard definition for practices that are owned by hospitals or health systems, we manually verified OneKey’s counts of unique physician NPIs in health system–affiliated practices against the number of employed or affiliated physicians listed on publicly available health system websites. For a random sample of health systems, this process resulted in physician counts in OneKey that were within 10 percent of physician counts listed on health system websites. We also used OneKey to identify practices whose ownership type was described as “independent,” excluding practices with other corporate ownership and hospital or health system affiliation, as has been previously applied in other research.^21–23

IDENTIFYING PHYSICIANS AFFILIATED WITH PRACTICES

Next, to facilitate linkages to claims data, we identified physicians located at PE-acquired, health system–affiliated, and independent practice sites. Because linkages between physicians and their practice sites were available only in 2016 and 2019, we limited our analytic sample to the 63 percent of physicians who remained at the same practice during this period for both PE-acquired and control practices.

LINKING PHYSICIANS TO CLAIMS DATA

Using provider NPIs, we linked our sample of physicians to FAIR Health National Private Insurance Claims patient deidentified claims data from the period January 1, 2015, through December 31, 2020. FAIR Health is an independent nonprofit that manages the largest database of commercial health insurance claims in the US. Prior studies have used FAIR Health claims data to examine imputed costs and the use of care.²⁴ Claim lines associated with each NPI were aggregated to the practice level, and data were further deidentified to mask any practice- or geographic-level identifiers, in compliance with FAIR Health data confidentiality policies. All analyses were conducted by the study authors using FAIR Health’s deidentified data.

STUDY VARIABLES

Primary outcomes at the practice-quarter level included imputed allowed amount per claim and its components, including average physician professional fees per claim and average facility fees per claim. We also examined total spending per practice-quarter.

To evaluate changes in patient risk composition, we measured the median patient Hierarchical Condition Categories (HCC) risk score.

Detailed definitions of outcomes are in appendix exhibit A2.¹⁶

STATISTICAL ANALYSIS

We used a difference-in-differences event study design to compare outcomes in PE-acquired practices with those of matched controls. Matched controls were constructed, first, from a cohort of health system–affiliated practices that did not experience any ownership change during our study period, and second, from a cohort of independent practices. PE-acquired practices were matched to controls in 2015, the year before any PE acquisition in our sample. Non-PE control practices were identified using 3:1 caliper matching without replacement that required exact match on Metropolitan Statistical Area (MSA) size, defined using the National Center for Health Statistics Urban-Rural Classification Scheme for Counties as large, medium, and small,²⁵ and within one standard deviation for preacquisition characteristics (total number of new and unique patients, median HCC score, and average allowed amount). Overall, 126 PE-acquired practices matched to 368 health system–affiliated control practices (83 percent match rate), with 120 (81.1 percent) PE practices matched to a full set of three matched controls. In addition, 126 PE-acquired practices matched to 378 independent control practices (100 percent match rate).

In the event study analyses, event time 0 denoted the quarter of acquisition. We used data from six quarters before acquisition (event time −6 through −1) through six quarters after acquisition (event time 1 through 6), with the quarter of acquisition as the reference period. The unit of analysis was the practice-quarter. A linear difference-in-differences regression compared changes in outcomes of interest in PE-acquired practices relative to matched controls, before and after acquisition. We tested for differences in preacquisition trends between acquired and control practices, using joint F-tests.

For outcomes of interest, we calculated the adjusted differential change in percentage terms by dividing the difference-in-differences estimate by the unadjusted preacquisition mean of the outcomes among PE practices. Regressions included practice and time fixed effects to account for time-invariant practice-specific attributes and secular trends. To mitigate heteroske-dasticity concerns, all regressions were weighted by the average patient volume per practice over the course of the study period. Standard errors were clustered at the level of the matched group (PE-acquired practice and its set of matched controls). Analyses were conducted using Stata, version 17.0. Oregon Health & Science University’s Institutional Review Board approved this study (IRB No. 22582).

SENSITIVITY ANALYSES

We conducted several sensitivity tests. First, we estimated regressions that included independent practices as a reference group and used regression results to calculate the difference in outcomes between PE-acquired practices and health system–affiliated practices. Using two-tailed t-tests, we tested the null hypothesis that there was no difference in outcomes between PE-acquired practices and health system–affiliated practices in the postacquisition period. Second, we used alternative regression specifications, including log-transformed outcome variables and market-by-year fixed effects. Third, we calculated predicted values from difference-in-differences regressions to compare differences in postacquisition outcomes in PE-acquired and health system–affiliated settings. Further, we tested the robustness of our results to alternative matching approaches by including a full set of health system–affiliated practices as controls. We also employed an alternative matching strategy, using MSA size only, to assess potential bias from matching on preintervention outcomes.²⁶ Additional robustness checks included using a longer preacquisition and follow-up period of eight quarters; using the quarter before acquisition, rather than the quarter of acquisition, as the reference period; and using robust estimators to assess the sensitivity of our findings to differential timing of acquisitions and heterogeneous treatment effects.²⁷ Finally, we separately estimated results for common categories of gastroenterology procedures to examine whether our results were consistent across procedures.

LIMITATIONS

We acknowledge several limitations. A chief limitation was the inability to identify practices directly acquired by health systems; we thus relied on IQVIA data to determine practices’ affiliation with health systems. As a result, we were unable to compare PE acquisition with health system acquisition, and instead we compared PE acquisition with a stable cohort of practices that were affiliated with health systems over the course of our study period. Relatedly, IQVIA’s definition of health system–affiliated practices does not differentiate between practices that are directly owned by or have strategic affiliations with health systems, and this may aggregate potentially heterogeneous entities. As a result, our estimates of physicians’ practice settings may have included measurement error, although estimates were broadly consistent with publicly reported affiliations.

Second, our analysis used claims from commercially insured patients and might not be generalizable to other populations.

Third, because of data limitations, we were unable to examine whether facility fees at PE-acquired practices were driven by certain procedures (for example, endoscopy procedures) or certain settings (for example, ambulatory surgery centers under joint ownership).

Fourth, we were unable to include ancillary providers who could have contributed to observed increases in utilization. As our study design held constant the physicians at each practice pre- and postacquisition, we also were unable to account for entering and exiting physicians during our study period.

Fifth, deidentification and aggregation of FAIR Health data precluded NPI- or service-level analyses, including evaluation of specific mechanisms (for example, particular procedures, changing service intensity, or changing site of care) that may have driven these results.

Finally, given the recency of PE acquisitions, our follow-up period included only six quarters postacquisition. Results may have underestimated the true magnitude of effects observed with greater acquisition maturity.

Study Results

CHARACTERISTICS OF PE, HEALTH SYSTEM, AND INDEPENDENT PRACTICES

The number of PE acquisitions of gastroenterology practices increased from 2016 to 2020, reaching 126 acquisitions during our study period (see appendix exhibit A3 for counts of acquisitions by year).¹⁶ In preacquisition periods, practices acquired by PE exhibited some differences compared with health system–affiliated and independent practices; however, preacquisition trends were parallel among primary outcomes (appendix exhibit A4),¹⁶ supporting the identifying assumption that group differences would have remained constant in the absence of acquisition.

CHANGES IN OUTCOMES AFTER ACQUISITION

Exhibit 1 presents descriptive statistics for the 149 PE-acquired practices and their health system–affiliated matched controls in 2015. In adjusted event study estimates (exhibit 2), average allowed amounts per claim were lower at PE-acquired practices before acquisition compared with controls, although trends appeared fairly parallel over time. After PE acquisition, acquired practices demonstrated a consistent differential increase in allowed amounts of $92 per claim, or 28.4 percent through six quarters postacquisition (exhibit 3). Total spending also increased by $7,108 per practice, or 30.16 percent through six quarters postacquisition (exhibit 3 and appendix exhibit A5).¹⁶

EXHIBIT 1.

Characteristics of private equity (PE)-acquired and health system–affiliated gastroenterology practices, before acquisition, 2015

	Before matching			After matching
Characteristics	PE-acquired (n = 149)	Health system-affiliated (n = 724)	p value	PE-acquired (n = 126)	Health system-affiliated (n = 368)	p value
Total spending per quarter ($)	42,634	53,537	0.23	38,040	39,417	0.80
Allowed amount per claim ($)	275	247	<0.01	273	263	0.35
Professional fees per claim ($)	120.28	92.97	<0.01	122.49	102.22	0.05
Facility fees per claim ($)	248.88	258.94	0.55	244.29	280.12	0.05
Charges per claim ($)	484	524	0.09	474	543	0.005
No. of new patients	63	72	0.50	57	61	0.61
No. of unique patients	78	89	0.47	72	78	0.57
Patient HCC score	4.17	4.96	0.008	4.09	4.19	0.67

SOURCE Authors’ analysis of data from PitchBook, IQVIA, and FAIR Health.

NOTES The table presents baseline characteristics of PE-acquired and health system–affiliated gastroenterology practices in 2015, before any acquisitions in our analytic sample. Health system–affiliated practices represent practices with a corporate owner identified in the IQVIA data as a health system or integrated delivery network. These practices were identified using 3:1 caliper matching without replacement, requiring exact match on Metropolitan Statistical Area size and matches within 1 standard deviation for continuous covariates (total number of new and unique patients, median patient Hierarchical Condition Categories [HCC] score, and allowed amount). “Per quarter” indicates “per practice-quarter.” Allowed amounts and fees per claim are averages.

EXHIBIT 2. Effect of private equity (PE) acquisition on allowed amount per claim for gastroenterology services, relative to health system–affiliated practices, by quarter, 2015–20.

SOURCE Authors’ analysis of data from PitchBook, IQVIA, and FAIR Health. NOTES Each point in the figure represents the coefficient obtained by estimating an event study regression. In the event study regression, we compared outcomes in PE-acquired practices with those of matched controls up to 6 quarters before acquisition (event time −6 through −1) and 6 quarters after acquisition (event time 1 through 6), with the quarter of acquisition (0) as the reference period. The unit of analysis was the practice-quarter. Event study regressions included practice and time fixed effects, and standard errors were clustered at the level of the matched cohort. Health system–affiliated practices are defined in the exhibit 1 notes.

EXHIBIT 3.

Differential changes in outcomes in private equity (PE)-acquired and health system–affiliated gastroenterology practices, before and after acquisition, 2015–20

	PE-acquired		Health system-affiliated		Difference-in-differencesa
	Pre	Post	Pre	Post	Unadjusted	Adjusted	% change	p value
Total spending per quarter ($)	23,567	33,214	22,122	24,684	7,085	7,108.42	30.16	0.01
Allowed amount per claim ($)	317	515	286	392	92	90.70	28.39	<0.01
Professional fees per claim ($)	221	432	150	192	169	172.58	78.09	<0.01
Facility fees per claim ($)	429	549	447	625	−58	−12.19	4.54	0.53
Patient HCC score	1.93	2.02	2.33	2.52	−0.1	0.03	8.80	0.74

SOURCE Authors’ analysis of data from PitchBook, IQVIA, and FAIR Health.

NOTES Unadjusted and adjusted differential changes in outcome variables were averaged at the practice level for PE-acquired practices and matched health system–affiliated controls. Adjusted regression coefficients were estimated using a linear difference-in-differences model that included practice and time fixed effects and were weighted by average patient volume per practice over the course of the study period. Standard errors were clustered at the level of the matched cohort. Adjusted percentage differential change (the column labeled “% change”) was calculated by dividing the adjusted differential change obtained from the difference-in-differences regression by the preacquisition mean for PE-acquired practices. Health system–affiliated practices are defined in the exhibit 1 notes. “Per quarter” indicates “per practice-quarter.” Allowed amounts and fees per claim are averages. HCC is Hierarchical Condition Categories.

^aDifference in differences between PE-acquired and control practices or the differential change.

Compared with health system–affiliated matched controls, PE-acquired practices demonstrated differential increases in physician professional fees per claim (exhibit 4), whereas there was no statistically significant change in facility fees per claim (appendix exhibit A6).¹⁶ Across the postacquisition period, PE-acquired practices experienced a mean increase of $173 in professional fees per claim, or a 78.1 percent increase (exhibit 3).

EXHIBIT 4. Effect of private equity (PE) acquisition of physician practices on professional fees per claim for gastroenterology services, relative to health system–affiliated practices, by quarter, 2015–20.

SOURCE Authors’ analysis of data from PitchBook, IQVIA, and FAIR Health. NOTES Each point in the figure represents the coefficient obtained by estimating an event study regression. In the event study regression, we compared outcomes in PE-acquired practices with those of matched controls up to 6 quarters before acquisition (event time −6 through −1) and 6 quarters after acquisition (event time 1 through 6), with the quarter of acquisition (0) as the reference period. The unit of analysis was the practice-quarter. Event study regressions included practice and time fixed effects, and standard errors were clustered at the level of the matched cohort. Health system–affiliated practices are defined in the exhibit 1 notes.

There was no statistically significant differential change in median patient HCC scores (appendix exhibit A7).¹⁶

Results were consistent after we compared PE-acquired practices with matched independent control practices, although they differed in magnitude (appendix exhibits A8–A10).¹⁶ As unadjusted physician professional fees in health system–affiliated settings were lower than professional fees generated from independent settings in our sample (appendix exhibit A4),¹⁶ the increase in professional fees achieved by PE firms after acquisition was greater relative to health system–affiliated practices than relative to independent practices.

SENSITIVITY ANALYSES

Joint F-tests of the hypothesis that preacquisition trends between PE-acquired and health system–affiliated practices were no different showed no statistically significant preacquisition differences between PE-acquired and health system–affiliated practices (appendix exhibit A11).¹⁶ Our findings were robust to including independent practices as a reference group (appendix exhibits A12 and A13), using log-transformed outcomes (appendix exhibit A14), including market-by-year fixed effects (appendix exhibit A15), using a full set of controls rather than matched controls (appendix exhibit A16), using alternative matching approaches that used exact matching on MSA size only (appendix exhibit A17), including longer preacquisition and follow-up periods (appendix exhibit A18), changing the reference period to the quarter before acquisition rather than the quarter of acquisition (appendix exhibit A19), and using predicted values from difference-in-differences regressions to compare postacquisition outcomes across PE-acquired and health system–affiliated practices (appendix exhibit A20).¹⁶ Our findings were robust to the 2021 estimator from Brantly Callaway and Pedro Sant’Anna for differential treatment timing and heterogeneous treatment effects (appendix exhibit A21),¹⁶ as well as to recent advances in the difference-in-differences literature that propose further adjustments to the Callaway and Sant’Anna estimator^27,28 (appendix exhibit A22).¹⁶ Procedure-specific analyses for commonly billed gastroenterology services (appendix exhibit A23) demonstrated that the allowed amount per claim increased for procedures related to the removal of polyps, tumors, and lesions (appendix exhibit A24).¹⁶ In addition, acquisitions increased volume across selected services, including evaluation and management visits and esophagogastroduodenoscopy procedures.

Discussion

Overall, our results provide empirical support for the business strategies that PE firms may pursue in physician practice markets. A novel contribution of our analysis was to examine specific mechanisms of prices increases via facility fees versus physician professional fees. Our results shed additional light relative to the existing literature that has examined PE-acquired practices compared with independent practices: We found that compared with a control group of health system–affiliated practices, PE acquisition of gastroenterology practices was associated with differentially higher prices for the same services, primarily through increases in physician professional fees. Service-specific analyses suggest that our results may have been driven by increased use of certain services such as the removal of polyps during colonoscopies. Although this was beyond the scope of our study, future research could examine specific changes in practice patterns, including for common gastroenterology procedures such as colonoscopies. Additional research also is needed to understand whether PE acquisitions result in market expansion, business stealing, increased consolidation, or staffing changes (for example, increases in nurse practitioners);²⁹ the downstream implications for patient out-of-pocket payments and quality of care; and the effects of PE exits, given emerging evidence that PE-acquired practices undergo subsequent buyouts in three years.¹⁵

To our knowledge, our study was the first analysis to compare PE-acquired practices with a matched control group of health system–affiliated practices. As Medicare physician reimbursement rates (adjusted for inflation) have continued to decline³⁰ and practice costs have risen, PE acquisition offers physician practices the potential to negotiate higher prices from commercial insurers. The ability of PE firms to command higher professional fees is a potential reason for physician practices to continue their attraction to PE investors. Whether increased revenues from higher professional fees negotiated by PE-acquired practices result in improved physician satisfaction, retention, or quality remains a key area for future research.

Our findings also raise important questions about regulatory approaches to improving the affordability and value of health care in an era of rapid consolidation by corporate entities. Prices for the same services have been found to be significantly higher in hospital-based outpatient settings than in stand-alone clinics, prompting Congress to consider a number of policy proposals to address hospital-based spending.^{11–14,24,31} Medicare has taken steps toward payment reform that would pay the same rate for the same service, regardless of where it is provided, referred to as site-neutral payment.³² As part of a broader site-neutral payment reform, a growing number of states prohibit providers from charging facility fees for certain outpatient services.³³ Although site-neutral payment reform is a promising step toward reducing incentives for hospital consolidation of physician practices and associated price increases, our results suggest that such policies will be insufficient to regulate price increases driven by PE firms. Greater antitrust enforcement, improved transparency of acquisitions, and closing of payment loopholes that are often exploited by PE firms, such as surprise billing strategies weakened by the No Surprises Act of 2021, could further mitigate the undesirable effects of PE on health care spending.³⁴

Conclusion

As the corporate transformation of physician practices accelerates, the US is likely to witness greater consolidation of physician practices by health systems and PE firms in the coming years. Our results suggest that even as policy makers work to shift care away from expensive hospital-based settings to outpatient office settings, the proliferation of PE ownership of outpatient practices, with increases in professional fees driving the price growth, has the potential to counteract policy efforts to address health care spending.