This study compares health care costs and health services utilization patterns of patients receiving care in accountable care organization (ACO), patient-centered medical home (PCMH), hybrid (ACO + PCMH), and standard (neither ACO/PCMH) facilities.
Objectives: To compare the total health expenditures and health services utilization of patients receiving care in stand-alone accountable care organization (ACO only), stand-alone patient-centered medical home (PCMH only), hybrid (ACO + PCMH), and standard (neither ACO nor PCMH) facilities.
Study Design: A retrospective cross-sectional study was conducted.
Methods: Medical Expenditure Panel Survey 2016 data were used to identify patients 18 years and older who received care at a facility designated as ACO, PCMH, hybrid, or standard. The association of health expenditures with site of care was analyzed using a generalized linear model (gamma distribution with log-link function). Unadjusted expenditures were compared after bootstrapping.
Results: We identified 3431 patients who received care in facilities designated as ACO (n = 1096), PCMH (n = 355), hybrid (n = 1219), or standard (n = 761). Unadjusted yearly total health expenditure was significantly higher in the standard care group vs hybrid ($9850 vs $8432; P < .0001), standard vs ACO ($9850 vs $8399; P < .0001), and standard vs PCMH ($9850 vs $7580; P < .0001). Additionally, unadjusted total health expenditure was significantly lower in the PCMH group compared with the ACO, hybrid, and standard care groups. After adjusting, total health expenditure was significantly lower in the ACO (β = –0.12; P = .0189) and PCMH (β = –0.25; P = .0005) groups compared with hybrid but significantly higher in standard care (β = 0.16; P = .0049) compared with hybrid.
Conclusions: Care received in ACO and PCMH facilities is associated with lower total health care costs compared with standard care. However, hybrid models are associated with slightly higher total health care costs compared with stand-alone models. Integrating innovations in health care delivery and health care reimbursement warrants further evaluation.
The American Journal of Accountable Care. 2020;8(2):14-21. https://doi.org/10.37765/ajac.2020.88438
The United States spent $3.65 trillion toward health care in 2018 and is expected to spend nearly $6 trillion per year by 2027.1 Although the United States spends excessive amounts on health care, it also grapples with issues pertaining to the quality of care, efficiency, disparities, and the patient experience in the current health care system. In short, the United States ranks first in health care spending yet lags in health outcomes among the developed nations.2
The disconnect between health spending and health outcomes is a cause of concern. Researchers have cited the fee-for-service model, fragmented care delivery models, defensive medicine, administrative costs, inflated prices, and underinvestment in addressing social determinants of health as some of the reasons that explain the disconnect.3-9 Innovation in health care delivery and reimbursement have come to the fore to address the issues mentioned above. These innovative models strive to achieve the Triple Aim of improving population health, reducing costs, and enhancing the patient experience.10 Two innovative models that have been widely adopted are patient-centered medical homes (PCMHs) and accountable care organizations (ACOs).
ACOs and PCMHs have proliferated in the health care arena. The number of ACOs and patients served by ACOs have experienced tremendous growth. The number of covered lives under ACOs has increased from 2.6 million in 2011 to 23.5 million in 2015.11 This number is expected to rise to 150 million lives by 2025.11 ACO contracts have been sponsored by CMS as well as private payers. Preliminary evidence indicates positive outcomes concerning health outcomes and health care costs compared with standard care, especially among the clinically vulnerable population.12-14 Currently, 13,000 practices housing about 67,000 clinicians deliver care to approximately 21 million patients in PCMH-certified facilities.15,16 Preliminary evidence indicates that, like patients at ACOs, those receiving care at PCMH-certified facilities have improved health outcomes and reduced health care costs compared with standard care.17-19
The goals of the ACO and PCMH models are similar; however, the means to achieve the goals differ. According to CMS, the ACO is primarily a value-based reimbursement model that incorporates “voluntary” collaboration among providers, whereas the PCMH is primarily a care delivery model involving significant collaboration as part of the certification process.16,20,21 In other words, the ACO could also be described as an incentive-based payment model that addresses the fallacies of the fee-for-service model and holds providers accountable for the outcomes while giving them a share of the earned savings. The PCMH is a systems-based approach in which structures and processes are established to ensure that holistic and coordinated care is delivered to the patient.
Additionally, ACOs are contractual agreements between groups of providers and payers, but becoming a PCMH is a certification process that facilities choose to undergo. Many facilities and providers are not only obtaining PCMH certification but also entering into ACO contracts with payers, given that these models have similar goals and a high degree of compatibility. Most of the evidence compares these models with standard care. There is limited evidence on the comparative effectiveness of these models, especially hybrid vs stand-alone (ACO only or PCMH only) models, all of which are becoming more prevalent. Exploring the effectiveness of these models can provide evidence on the interventions that add value to the health care system. Also, it may help ascertain the value of incentives, structures, and processes in our inefficient health care system. Moreover, it may provide valuable insights to assist during the decision-making process of pursuing an ACO contract and/or PCMH certification from the perspectives of payers, policy makers, health systems, and provider practices.
The primary objective of this study was to compare the total health expenditures of adult patients receiving care in stand-alone ACO (ACO only), stand-alone PCMH (PCMH only), hybrid (ACO + PCMH), and standard (neither ACO nor PCMH) facilities. The secondary objective was to compare expenditures (inpatient and outpatient) and health services utilization (emergency department [ED] and outpatient visits).
A retrospective database analysis using the 2016 Medical Expenditure Panel Survey (MEPS) data was conducted. MEPS is a nationally representative database of survey responses of noninstitutionalized Americans.22-24 Information on demographic characteristics, health conditions, health status, charges, payments, health services utilization, access to care, satisfaction with care, insurance status, income, and employment status is collected from individuals, families, medical providers, and employers.24 An overlapping panel design was used to collect data. Three data files were used for this study: full-year consolidated (FYC), medical conditions (MC), and medical organizations (MO). The FYC file was used to obtain demographic, expenditure, and health services utilization information. The MC file was used to impute the Charlson Comorbidity Index (CCI) score. The MO file was used to identify site-of-care designation. The MO file was generated using the responses from the practice site (physician or administrative staff). Additionally, respondents in the MO file were those that were identified as the usual source of office-based care.24 Each of these files included a unique patient identifier that was utilized for matching.
Adult patients (≥18 years) were placed in the 4 cohorts (ACO, PCMH, hybrid, standard) using the responses to the following questions: “Is the practice certified as a patient-centered medical home?” and “Does the practice participate in an ACO arrangement with either Medicare or private insurers?” Patients were included only if the responses to the above questions were any combination of yes and/or no.
Information was also collected on age (18-40, 41-64, ≥65 years), gender, insurance type (public, private, uninsured), education (high school or less, some college, bachelor’s degree or higher), income (high, middle, low), race (white, African American, other minorities), ethnicity (Hispanic, non-Hispanic), and employment status (employed, unemployed). Total health care expenditure in MEPS included the sum of all payments for health care irrespective of the source of payment.24 Total outpatient expenditure was imputed as the sum of all payments during a visit to an outpatient (office-based or hospital) setting.24 Total inpatient expenditure was imputed as the sum of all payments during a hospitalization event.24 Outpatient utilization was imputed as the sum of all office-based and outpatient (hospital) visits.24
Descriptive analysis was conducted using χ2 for categorical variables and analysis of variance (ANOVA) for continuous variables. Unadjusted means were calculated using bootstrapping over 1000 iterations to account for the nonparametric distribution.25 These means were compared using ANOVA followed by a Bonferroni post hoc test. Association between site of care and health expenditures was analyzed using a generalized linear model with a log-link function and gamma distribution using the following covariates: age, gender, race, ethnicity, education, employment status, income, insurance type, and imputed CCI score. The complex, multistage survey design of MEPS was incorporated using survey weights in all analyses. National estimates were calculated using person weights. All statistical analyses were conducted using SAS 9.4 (SAS Institute Inc).
A total of 3431 patients were identified who received care in facilities designated as ACO (n=1096), PCMH (n=355), hybrid (n=1219), or standard (n=761). This translates to 11.89 million, 3.92 million, 12.74 million, and 7.98 million US residents receiving care at facilities designated as ACO, PCMH, hybrid, and standard, respectively (Table 1). Comparison of baseline characteristics indicated no significant differences across all sociodemographic variables besides gender. There was a statistically significant difference in gender distribution among the groups in that women accounted for approximately 60% of the patients in the hybrid and PCMH cohorts compared with approximately 55% for the ACO and standard care cohorts.
Table 2 provides a comparison of unadjusted health expenditure estimates of the 4 cohorts. Unadjusted yearly total health expenditure was significantly higher in the standard care group vs hybrid ($9850 vs $8432; P <.0001), standard vs ACO ($9850 vs $8399; P <.0001), and standard vs PCMH ($9850 vs $7580; P <.0001). Additionally, unadjusted total health expenditure was significantly lower in PCMH compared with the ACO, hybrid, and standard care groups. Total yearly outpatient expenditure was significantly different across all groups (P <.0001). The total yearly outpatient expenditure was highest in the ACO cohort ($2688), followed by hybrid ($2629) and standard ($2597), and lowest in the PCMH cohort ($2295). Total yearly inpatient expenditure was significantly different across all groups (P <.0001). Total yearly inpatient expenditure was highest in the standard care cohort ($2828), followed by hybrid ($2112) and ACO ($1637), and lowest in the PCMH cohort ($1412).
Table 3 provides a comparison of yearly health services utilization among the 4 cohorts. Average ED visits were significantly different across all groups (P <.0001). Average ED visits were highest in the standard care cohort (0.31), followed by hybrid (0.27) and ACO (0.25), and lowest in the PCMH cohort (0.23). Average yearly outpatient visits were significantly different across all groups (P <.0001). Average yearly outpatient visits were highest in the hybrid cohort (10.82), followed by standard (10.64) and ACO (10.08), and lowest in the PCMH cohort (9.38).
Table 4 provides the association of total health expenditures with site-of-care designation after adjusting for socioeconomic factors. Total health expenditures were about 12% and 25% lower in the ACO (β= –0.1199; P = .0189) and PCMH (β= –0.2516; P =.0005) cohorts, respectively, compared with hybrid. However, total health expenditure was about 16% higher in the standard care cohort (β=0.1597; P =.0049) compared with hybrid. Inpatient expenditure was about 48% higher in the standard care cohort (β=0.4831; P =.0002) compared with hybrid. There was no difference between the other cohorts in average per capita inpatient expenditures. Outpatient expenditures were about 16% and 30% lower in the ACO (β=–0.1631; P =.0019) and PCMH (β= –0.2960; P <.0001) cohorts, respectively, compared with hybrid. There was no difference in average per capita outpatient expenditures between the hybrid and standard care cohorts.
Total health expenditures were significantly higher in individuals older than 40 years (compared with 40 years and younger), women (compared with men), unemployed individuals (compared with employed), and those with a higher CCI score. On the contrary, total health expenditures were significantly lower in uninsured individuals (compared with those with private insurance), those who completed high school or less (compared with a bachelor’s degree or higher), those with middle income (compared with high income), African Americans and other minorities (compared with whites), and Hispanic individuals (compared with non-Hispanic individuals).
Based on our review of the literature, this is the first study that compares health care costs associated with innovative models (hybrid and stand-alone). This study finds that ACO and PCMH models are associated with reduction in total health care costs compared with standard care. These findings are consistent with previous literature, which indicates that ACO and PCMH models are associated with reduction in health care costs compared with standard care.13,14,17-19 However, this study also adds that stand-alone models are associated with greater reduction in health care costs compared with hybrid models. This is contrary to the expected notion that hybrid models may generate savings that are partly cumulative of those achieved by the stand-alone models. This may be attributed to multiple reasons. Firstly, hybrid models may experience additive effects of operational and implementation challenges of the stand-alone models. ACO and PCMH are associated with significant challenges including organizational structure, culture change, employee and leadership buy-in, operational changes, and administrative resource mobilization, to name a few.26-29 These challenges may be standing in the way of hybrid models achieving their intended outcomes and performing to their full potential. Also, previous research has indicated that reduction in health expenditure is time dependent30,31—that is, reduction in health expenditure can be seen after a longer duration of evaluation. In this study, information was not collected by the survey administrators (MEPS) to determine the duration of ACO contract or PCMH certification. Next, the distribution of hybrid models may not be consistent throughout the nation, leading to diverse risk profiles. An evaluation conducted by Graf et al found significant variation in participants served under the Medicaid ACO-PCMH hybrid arrangement among different states.32 They found that approximately 80% of Vermont’s Medicaid ACO population received care at a hybrid facility compared with 72% in Minnesota and 38% in Maine.32 Future research should explore the outcomes of hybrid models compared with stand-alone models and identify opportunities to reduce operational and implementation burden. With significant overlap in the goals, similar consolidation of certification requirements (PCMH) and contractual obligations (ACO) may alleviate administrative and implementation burden on participating practices.
This study found that outpatient expenditure was lowest in the PCMH cohort. Moreover, outpatient and ED utilization was lowest in the PCMH cohort compared with the other cohorts. Evidence in the literature suggests that robust primary care can improve health outcomes and reduce health care costs.33,34 PCMH certification is contingent on the structures and processes that ensure a robust primary care center, which could explain the positive findings associated with PCMH. Additionally, some evidence indicates that PCMH certification is associated with the likelihood of ACO participation.35 Moreover, ACOs with a strong primary care base tend to perform better than ACOs with a weak primary care base.36 In the evaluation conducted by Graf et al, the authors explored the utilization of existing PCMH programs in Medicaid ACO developments.32 This results in these models working as complements to each other rather than in silos. Next, it utilizes existing frameworks of quality measurement to minimize disruption as it navigates toward a hybrid model. Similarly, stepwise rollout of PCMHs followed by ACOs may help participating health systems increase efficiency, reduce administrative burden, improve health outcomes, and reduce health care costs. Also, ACOs receive assistance in terms of capital with programs like the ACO Investment Model and Advance Payment ACO Model; similar start-up capital opportunities are not available for “medical homes” to achieve PCMH certification. In light of the evidence presented, policy makers and payers may want to explore funding for PCMH certification. Further evaluation of the value of PCMH and ACO is necessary to help payers and policy makers determine interventions that provide a better return on investment.
There are several limitations inherent to the data set. First, this study does not address any contextual factors that play an important role in the performance of ACO and hybrid models because of the nature of the data available from MEPS. Some contextual factors that could affect the results are ACO size, rural vs urban ACO, hospital-led ACO vs provider-led ACO, and underlying risks of the population served, among others.37-39 Next, the cross-sectional study design utilized may not be able to capture time-dependent performance trends of these models.30,31 Information on duration of ACO/PCMH participation is not obtained from the respondents. Also, social desirability and response bias are inherent to the MEPS data set. Additionally, the data set does not incorporate the complexities of the ACO/PCMH models.
Care received in ACO and PCMH models is associated with lower health care costs compared with standard care. However, hybrid models are associated with slightly higher health care costs compared with stand-alone models. Integrating innovations in health care delivery and health care reimbursement warrants further evaluation.
Author Affiliations: University of Toledo (PMP, VV), Toledo, OH; University of Toledo Physicians, LLC (RG), Toledo, OH.
Source of Funding: None.
Author Disclosures: Dr Gupte is employed in the division of value-based care at a multiclinic specialty practice. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (PMP, VV, RG); acquisition of data (PMP); analysis and interpretation of data (PMP, VV, RG); drafting of the manuscript (PMP, VV, RG); critical revision of the manuscript for important intellectual content (PMP, VV, RG); statistical analysis (PMP, VV); provision of study materials or patients (PMP); administrative, technical, or logistic support (RG); and supervision (VV, RG).
Send Correspondence to: Pranav M. Patel, PharmD, MS, University of Toledo, 3345 Airport Hwy, Apt 7A, Toledo, OH 43609. Email: firstname.lastname@example.org.
1. Sisko AM, Keehan SP, Poisal JA, et al. National health expenditure projections, 2018-27: economic and demographic trends drive spending and enrollment growth. Health Aff (Millwood). 2019;38(3):491-501. doi:10.1377/hlthaff.2018.05499
2. Kumar S, Ghildayal NS, Shah RN. Examining quality and efficiency of the U.S. healthcare system. Int J Health Care Qual Assur. 2011;24(5):366-388. doi:10.1108/09526861111139197
3. Schroeder SA, Frist W; National Commission on Physician Payment Reform. Phasing out fee-for-service payment. N Engl J Med. 2013;368(21):2029-2032. doi:10.1056/NEJMsb1302322
4. Berwick DM, Hackbarth AD. Eliminating waste in US health care. JAMA. 2012;307(14):1513-1516. doi:10.1001/jama.2012.362
5. Garber AM, Skinner J. Is American health care uniquely inefficient? J Econ Perspect. 2008;22(4):27-50. doi:10.1257/jep.22.4.27
6. Dieleman JL, Squires E, Bui AL, et al. Factors associated with increases in US health care spending, 1996-2013. JAMA. 2017;318(17):1668-1678. doi:10.1001/jama.2017.15927
7. Bradley EH, Elkins BR, Herrin J, Elbel B. Health and social services expenditures: associations with health outcomes. BMJ Qual Saf. 2011;20(10):826-831. doi:10.1136/bmjqs.2010.048363
8. Woolhandler S, Campbell T, Himmelstein DU. Costs of health care administration in the United States and Canada. N Engl J Med. 2003;349(8):768-775. doi:10.1056/NEJMsa022033
9. Lyu H, Xu T, Brotman D, et al. Overtreatment in the United States. PLoS One. 2017;12(9):e0181970. doi:10.1371/journal.pone.0181970
10. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health, and cost. Health Aff (Millwood). 2008;27(3):759-769. doi:10.1377/hlthaff.27.3.759
11. Muhlestein D. Growth and dispersion of accountable care organizations in 2015. Health Affairs. March 31, 2015. Accessed September 20, 2018. https://www.healthaffairs.org/do/10.1377/hblog20150331.045829/full/
12. Kaufman BG, Spivack BS, Stearns SC, Song PH, O’Brien EC. Impact of accountable care organizations on utilization, care, and outcomes: a systematic review. Med Care Res Rev. 2019;76(3):255-290. doi:10.1177/1077558717745916
13. Colla CH, Lewis VA, Kao LS, O’Malley AJ, Chang CH, Fisher ES. Association between Medicare accountable care organization implementation and spending among clinically vulnerable beneficiaries. JAMA Intern Med. 2016;176(8):1167-1175. doi:10.1001/jamainternmed.2016.2827
14. Nyweide DJ, Lee W, Cuerdon TT, et al. Association of Pioneer Accountable Care Organizations vs traditional Medicare fee for service with spending, utilization, and patient experience. JAMA. 2015;313(21):2152-2161. doi:10.1001/jama.2015.4930
15. Hing E, Kurtzman E, Lau DT, Taplin C, Bindman AB. Characteristics of primary care physicians in patient-centered medical home practices: United States, 2013. Natl Health Stat Report. 2017;(101):1-9.
16. Patient-centered medical home (PCMH). National Committee for Quality Assurance. Accessed September 27, 2018. https://www.ncqa.org/programs/health-care-providers-practices/patient-centered-medical-home-pcmh/
17. Pines JM, Keyes V, van Hasselt M, McCall N. Emergency department and inpatient hospital use by Medicare beneficiaries in patient-centered medical homes. Ann Emerg Med. 2015;65(6):652-660. doi:10.1016/
18. van Hasselt M, McCall N, Keyes V, Wensky SG, Smith KW. Total cost of care lower among Medicare fee-for-service beneficiaries receiving care from patient-centered medical homes. Health Serv Res. 2015;50(1):253-272. doi:10.1111/1475-6773.12217
19. DeVries A, Li CH, Sridhar G, Hummel JR, Breidbart S, Barron JJ. Impact of medical homes on quality, healthcare utilization, and costs. Am J Manag Care. 2012;18(9):534-544.
20. Accountable care organizations (ACOs). CMS. Accessed September 20, 2018. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/ACO/
21. Defining the PCMH. Agency for Healthcare Research and Quality. Accessed September 27, 2018. https://pcmh.ahrq.gov/page/defining-pcmh
22. Cohen JW, Cohen SB, Banthin JS. The Medical Expenditure Panel Survey: a national information resource to support healthcare cost research and inform policy and practice. Med Care. 2009;47(7 suppl 1):S44-S50. doi:10.1097/MLR.0b013e3181a23e3a
23. Cohen SB. The Medical Expenditure Panel Survey: an overview. Eff Clin Pract. 2002;5(suppl 3):E1.
24. Medical Expenditure Panel Survey (MEPS). Agency for Healthcare Research and Quality. Updated August 2018. Accessed May 27, 2019. https://www.ahrq.gov/data/meps.html
25. Haukoos JS, Lewis RJ. Advanced statistics: bootstrapping confidence intervals for statistics with “difficult” distributions. Acad Emerg Med. 2005;12(4):360-365. doi:10.1197/j.aem.2004.11.018
26. Solberg LI, Crain AL, Tillema JO, et al. Challenges of medical home transformation reported by 118 patient-centered medical home (PCMH) leaders. J Am Board Fam Med. 2014;27(4):449-457. doi:10.3122/jabfm.2014.04.130303
27. Mostashari F, Broome T. The opportunities and challenges of the MSSP ACO program: a report from the field. Am J Manag Care. 2016;22(9):564-568.
28. Mechanic R, Perloff J, Litton T, Edwards K, Muhlestein D. The 2018 annual ACO survey: examining the risk contracting landscape. Health Affairs. April 23, 2019. Accessed June 28, 2019. https://www.healthaffairs.org/do/10.1377/hblog20190422.181228/full/
29. Green EP, Wendland J, Carver MC, Hughes Rinker C, Mun SK. Lessons learned from implementing the patient-centered medical home. Int J Telemed Appl. 2012;2012:103685. doi:10.1155/2012/103685
30. Saunders RS, Muhlestein D, McClellan MB. Medicare accountable care organization results for 2016: seeing improvement, transformation takes time. Health Affairs. November 21, 2017. Accessed April 10, 2020. https://www.healthaffairs.org/do/10.1377/hblog20171120.211043/full/
31. Cotton P. Patient-centered medical home evidence increases with time. Health Affairs. September 10, 2018. Accessed April 10, 2020. https://www.healthaffairs.org/do/10.1377/hblog20180905.807827/full/
32. Graf E, Jones M, Freeman S, Kissam S, Hersey C. How state Medicaid agencies leverage existing patient-centered medical home programs in accountable care organization model development: lessons learned from the State Innovation Models Initiative evaluation. Poster presented at: AcademyHealth Annual Research Meeting; June 2-4, 2019; Washington, DC. Accessed April 10, 2020. https://academyhealth.confex.com/academyhealth/2019arm/meetingapp.cgi/Paper/30748
33. Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q. 2005;83(3):457-502. doi:10.1111/j.1468-0009.2005.00409.x
34. Shi L. The impact of primary care: a focused review. Scientifica (Cairo). 2012;2012:432892. doi:10.6064/2012/432892
35. Lin YL, Du Y, Gomez C, Ortiz J. Does patient-centered medical home recognition relate to accountable care organization participation? Popul Health Manag. 2018;21(3):188-195. doi:10.1089/pop.2017.0096
36. Christensen EW, Payne NR. Effect of attribution length on the use and cost of health care for a pediatric Medicaid accountable care organization. JAMA Pediatr. 2016;170(2):148-154. doi:10.1001/jamapediatrics.2015.3446
37. Kocher R, Sahni NR. Physicians versus hospitals as leaders of accountable care organizations. N Engl J Med. 2010;363(27):2579-2582. doi:10.1056/NEJMp1011712
38. 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):2357-2366. doi:10.1056/NEJMsa1600142
39. Heiser S, Colla C, Fisher E. Unpacking the Medicare shared savings proposed rule: geography and policy. Health Affairs. January 22, 2015. Accessed April 10, 2020. https://www.healthaffairs.org/do/10.1377/hblog20150122.044093/full/