Between 2010 and 2018, the share of primary care physicians (PCPs) integrated with hospitals more than doubled from 22 to 49%.1 The first wave of physician-hospital integration prior to the 2010 Affordable Care Act (ACA) led to higher hospital and physician prices. Using a longitudinal study of patients in the IBM MarketScan commercial insurance claims data from 2001 to 2007, when the integration rate first doubled, Baker et al. found that integration increased total spending.2 Using the same data from 2008 to 2012, Neprash et al. found that spending continued to increase with integration, with no reduction in utilization.3 Despite these early results, many industry leaders contend that this integration is indeed spurring care transformation, along with cost-savings, especially over the course of the value-based efforts of the ACA.4 In fact, recent work with MarketScan data 2010–2016 has shown a causal link between patients randomly moving to more physician-hospital integrated areas and an increase in their patient care coordination.5 Here, we revisit the earlier integration studies using MarketScan data to re-examine the relationship between integration and spending during the ACA from 2010 to 2016. We observe a reversal in this relationship, finding an association between integration and cost-savings, not for already heavily integrated markets, but for young markets that became heavily integrated during this latter ACA integration wave from 2010 to 2016.
METHODS
We follow the methods of Neprash et al. updated with 2010–2016 IBM MarketScan data.3 We track a cohort of 1.7 million non-elderly privately insured people enrolled continuously in a PPO or point of service plan, with no plan switching over the period. We measure integration as percentage of PCPs in each Metropolitan Statistical Area (MSA) owned by hospitals or integrated systems. Ownership is derived from IQVIA’s SK&A data.5 We classify 268 MSAs as either mature or young markets. Mature markets in 2010 were already heavily integrated and grew slowly in additional integration from 2010 to 2016 (below the national median in the integration growth rate from 2010 to 2016). Young markets were above-median in integration growth 2010–2016, with their integration more than doubling (see Fig. 1). For each market type, using patient fixed effects, we regress the log of total annual spending per person on integration. We also examine 6 other spending and utilization outcomes in Table 1. We control for age, sex, year, 27 comorbidities, copayment rate, MSA percent elderly, percent below poverty, unemployment rate, and the number of physicians per 100,000. The AHRQ IRB determined that this was not human subject research.
Fig. 1.
Differing trends in the relationship between annual spending and the integration of primary care physicians with hospitals, 2010–2016
Table 1.
Estimated Effects of Physician Integration on Spending and Utilization, 2010–2016
| Outcomes | Marginal effect of moving from 25th to 75th percentiles of integration | |
|---|---|---|
| Mature markets | Young markets | |
| Total spending |
1.4%** (0.01, 2.7) |
−1.4%** (−2.6, −0.2) |
| Inpatient spending (non-maternity) |
−0.7% (−1.7, 0.4) |
−1.0%** (−1.9, −0.1) |
| Drug spending |
2.2%*** (0.8, 3.6) |
−2.5%*** (−3.7, −1.3) |
| Outpatient spending |
1.3%* (−0.1, 2.7) |
−1.7%*** (−2.9, −0.5) |
| Office visit price |
1.1%*** (0.7, 1.5) |
0.2% (−0.2, 0.5) |
| Number of office visits |
4.9%*** (3.0, 6.7) |
0.1% (−1.5, 1.8) |
| Number of lab/imaging claims |
3.3% (−2.4, 9.1) |
−20.8%*** (−25.9, −15.7) |
| N | 5,460,000 | 6,350,000 |
Notes: Young markets are those with above-median growth in integration from 2010 to 2016. Mature markets are below-median growth, already highly integrated
The 25th to 75th percentiles of integration are .28 to .49 for mature markets and .17 to .33 for young markets. Integration is percentage of primary care physicians within an MSA integrated with hospitals. N=1.7 million non-elderly people continuously enrolled in their privately insured PPO from 2010 to 2016 across 268 MSAs. Spending is logged and wage and inflation adjusted to 2016 dollars. Estimates are person-fixed effect regressions controlling for 33 patient characteristics and 4 MSA characteristics. 95% robust confidence intervals are adjusted for clustering at the person. MSA, Metropolitan Statistical Area. 75th percentile is significantly different from the 25th at level: ***p<0.01, **p<0.05, *p<0.01
Sources: Authors’ calculations using MarketScan and IQVIA SK&A
RESULTS
Between 2010 and 2016, physician-hospital integration of PCPs in MSAs increased from 22 to 40% (not shown). This integration had no effect on total annual spending (not shown). However, in the figure, we see that this overall effect concealed two countervailing trends. In mature markets, total spending increased by 1.4%, and in young markets, it decreased by 1.4% as integration increased from the 25th to 75th percentiles (p<.01). In mature markets, outpatient and drug spending increased, and office prices and number of visits increased (p<.05). In young markets, inpatient, outpatient, and drug spending decreased (p<.05). Office prices and number of visits were unchanged, but lab and imaging quantities declined by 20% (p<.01).
DISCUSSION
These are the first results to report an association between physician-hospital integration and overall cost-savings. The literature shows that the first wave of integration from 2000 to 2010 aimed to secure market power, to control prices, and to control the inpatient referral network before the implementation of ACA incentives. However, we find that the latest wave of additional integration during the ACA in young markets did indeed start to focus on reducing costs in high-cost markets. Since our sample was the commercial PPO sector, it is unlikely that our results were driven by the Medicare ACO program. Our results could be driven by regression to the mean (RTM) if the young markets were randomly high-cost markets in 2010, regressing back to mean costs by 2016. However, an RTM test6 shows that, at most, only 8.7% of the cost reduction in the young markets could be due to RTM. Most of the costs-savings are indeed associated with integration. Certainly, future research should examine the exact mechanism of physician-hospital clinical integration in greater detail.
Funding
This study was funded by the Agency for Healthcare Research and Quality, AHRQ-IM16513. The views expressed in this article are those of the authors, and no official endorsement by the U.S. Department of Health and Human Services, AHRQ, or RAND is intended or should be inferred.
Declarations
Conflict of Interest
The authors declare that they do not have a conflict of interest.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Furukawa MF, Kimmey L, Jones DJ, Machta RM, Guo J, Rich EC. Consolidation Of Providers Into Health Systems Increased Substantially, 2016–18. Health Affairs. 2020;39(8):1321–1325. doi: 10.1377/hlthaff.2020.00017. [DOI] [PubMed] [Google Scholar]
- 2.Baker LC, Bundorf MK, Kessler DP. Vertical Integration: Hospital Ownership of Physician Practices is Associated with Higher Prices and Spending. Health Affairs. 2014;33(5):756–63. doi: 10.1377/hlthaff.2013.1279. [DOI] [PubMed] [Google Scholar]
- 3.Neprash HT, Chernew ME, Hicks AL, Gibson T, McWilliams JM. Association of financial integration between physicians and hospitals with commercial health care prices. J Am Med Assoc Intern Med. 2015;175(12):1932–1939. doi: 10.1001/jamainternmed.2015.4610. [DOI] [PubMed] [Google Scholar]
- 4.Noether M, May S. Hospital merger benefits: Views from hospital leaders and econometric analysis. Technical report, Charles River Associates; 2017. https://www.crai.com/sites/default/files/publications/Hospital-Merger-Full-Report-_FINAL-1.pdf (accessed July 3, 2019).
- 5.Encinosa W, Dor A, Nguyen P. Does Physician-Hospital Vertical Integration Signal Care-Coordination? Evidence from Mover-Stayer Analyses of Commercially Insured Enrollees. Cambridge, MA: National Bureau of Economic Research; 2022; working paper 29599, January, 2022.
- 6.Joshi S, Nuckols T, Escarce J, Huckfeldt P, Popescu I, Sood N. Regression to the Mean in the Medicare Hospital Readmissions Reduction Program. JAMA Intern Med. 2019;179(9):1167–1173. doi: 10.1001/jamainternmed.2019.1004. [DOI] [PMC free article] [PubMed] [Google Scholar]