More than 1,000 accountable care organizations (ACOs) are responsible for the health and healthcare costs of 32 million Americans (Muhlestein et al., 2018). The Medicare Shared Savings Program (MSSP) offers incentives to risk-bearing entities to improve the cost-effectiveness and value of care (Muhlestein & McClellan, 2016). By definition, people with serious illness (SI) have a condition with a high risk of mortality, poor quality of life with high symptom burden, and limited function—with likely caregiver stress. Seriously ill beneficiaries are responsible for a high proportion of overall Medicare spending (Kelley, 2013, 2014). People living with SI have complex needs and often experience fragmented, poor quality healthcare (Aldridge & Bradley, 2017; Institute of Medicine, 2015). Most ACOs have identified seriously ill patients. However, few have implemented programs to manage care for complex, high-risk patients (Bleser, Saunders, Winfield et al., 2019; Meier et al., 2017); enhanced understanding of the potential for savings in that subgroup may motivate health systems to implement new models for SI care.
In the MSSP, incentives for targeted SI care models are indirect and vary depending on ACO patient population and contract characteristics (Kaufman et al., 2017; Larson et al., 2012). Broad interventions likely generated the majority of shared savings in early Medicare ACO populations and may have only incidentally improved care for seriously ill beneficiaries (McWilliams et al., 2017). Comprehensive care management for management-amenable chronic conditions, expanding access and scope of primary care, and reducing unnecessary imaging have been common steps that MSSP ACOs have taken toward shared savings. Implementation of models targeting SI care often requires substantial additional investment in risk stratification, partnerships, and workforce. These investments may be particularly problematic for MSSP ACO entities with little experience managing population health (Bleser, Saunders, Winfield et al., 2019; Colla et al., 2016).
Spending on seriously ill and medically complex patients may shift ACO per capita spending and potential for shared savings, but little is known about ACO spending for this high-cost, high-need patient population. Our study evaluated factors associated with SI spending in MSSP ACOs and their relationship with shared savings success. Specifically, we looked at differences between ACOs in the highest and lowest quartiles for per capita spending on seriously ill beneficiaries and how this difference could potentially change ACO shared savings performance.
This cross-sectional study, funded by the Gordon and Betty Moore Foundation to the Duke-Margolis Center for Health Policy at Duke University, includes SI beneficiaries attributed to MSSP ACOs (2014–2016) using 100% Medicare beneficiary summary files and the final assignment indicator in the MSSP beneficiary files. We included fee-for-service Medicare beneficiaries who met an established SI definition including three criteria defined in the performance year (PY): any inpatient claim, any claim for home health or skilled nursing or durable medical equipment (DME), and either a serious chronic illness or multimorbidity (Kelley, 2014; Kelley et al., 2017). Comorbidities were defined using chronic conditions warehouse (CCW) algorithms for 27 common conditions and other chronic and disabling conditions (Centers for Medicare & Medicaid Services [CMS], n.d.) Serious chronic illnesses included advanced liver disease or cirrhosis; lung, colorectal, or endometrial cancer; chronic obstructive pulmonary disease; congestive heart failure; Alzheimer’s disease or related dementias; diabetes with complications including ischemic heart disease or peripheral vascular disease; hip fracture; or renal failure as indicated by any use of dialysis.
For each ACO-year observation, we defined binary indicators for the lowest quartile and highest quartile for adjusted per capita spending on SI beneficiaries. Per beneficiary per year (PBPY) Medicare expenditures were calculated by ACO contract year using the Medicare Master Beneficiary Summary File. Beneficiary-level Part A and B Medicare expenditures were annualized to mirror MSSP calculations and standardized to remove variation from inflation and geography using the Consumer Price Index (2016 = 100) and CMS geographic variation public use file. We applied the MSSP stop-loss rule by truncating costs at 0 and the average 99% percentile in national fee-for-service population by year within each Medicare eligibility group.
Demographics and comorbidities were determined using the master beneficiary summary file and CCW. Patient covariates from claims data included patient demographics (gender, age, race), Medicare eligibility category (age, disabled, end-stage renal disease, or both disabled and end-stage renal disease), Medicaid and Medicare dual eligibility, 27 CCW comorbidities, and binary indicators for five or more comorbidities, as well as death during the PY. County demographics, mean hierarchical condition category score, Medicaid enrollment, and Medicare Advantage market share were linked using the CMS geographic variation public use file. ACO variables included binary indicator variables for governance (hospital leadership vs. not), MSSP track (2 or 3 vs. 1), any commercial contract, and advanced payment participation. Continuous ACO variables included MSSP ACO composite quality performance score, size of the attributed population to the ACO, ACO participating clinicians per 1,000 beneficiaries, updated benchmark, and percentage of rural beneficiaries in the ACO.
We evaluated patient-level Medicare cost and use data to identify high- and low-spending ACOs. Then, we evaluated the associations of ACO characteristics and ACO contract characteristics with low spending on seriously ill beneficiaries for each ACO contract year. The cross-sectional design is common in similar evaluations of cost variation in ACOs (Comfort et al., 2018; Kyle et al., 2020; Schulz et al., 2018; Zhu et al., 2019). A Hausman test of systematic differences in coefficients supported the use of fixed instead of random effects. Standard errors were clustered by ACO year to account for the correlation of patient outcomes within an ACO each year. Spending was standardized to account for geographic variation using the CMS geographic variation public use file. Detailed estimates from the regression analysis and other supporting data are provided in an appendix to this article, published as Supplemental Digital Content at http://links.lww.com/JHM/A55.
Defining High- and Low-Spending ACOs
Annual Medicare expenditures for patients with SI (N = 2,109,573) were predicted using a generalized gamma model with a log link and a unique intercept for each of the 1,157 ACO contract years (2014–2016) (Barber & Thompson, 2004; Manning et al., 2002). First, the difference in spending attributed to the ACO, independent of patient and county factors, was generated for each ACO year by subtracting the predicted mean for each ACO population from the grand mean as is recommended for analyses of variation (Moran & Solomon, 2014; Mustillo et al., 2012). This difference is the marginal effect. Then, the ACO marginal effects (n = 1,157) were stratified into quartiles. The results present comparisons between the ACO contract years in the top and bottom quartile for marginal effect on SI PBPY spending.
Evaluating Associations of ACO Characteristics With Outcomes
We aggregated patient and county variables to create an ACO-year data file to evaluate ACO outcomes using logistic regression (N = 1,157). ACO characteristics associated with low spending (binary) and the achievement of shared savings (binary) were evaluated using logistic regression models. ACO shared savings success is defined by the indicator provided by CMS for achieved shared savings in the MSSP PY files. For ease of interpretation, continuous ACO measures were included using a binary indicator for the top quartile. ACO annual benchmark and average hierarchical condition category composite score were excluded because of collinearity with the percent seriously ill. ACO observations with fewer than 100 attributed SI beneficiaries (n = 4) were excluded. Standard errors were clustered by year to account for the correlation of outcomes by year. Analyses were performed using Statacorp Stata (Release 13); continuous variables were reported as mean (SD), except where otherwise indicated, and statistical significance was ascribed at P ≤ .05. This study was approved by the institutional review board where the work was conducted.
The MSSP SI sample included 2,109,573 beneficiary-year observations for 333 ACOs in 2014, 392 ACOs in 2015, and 432 ACOs in 2016. ACOs had a wide range of attributed beneficiaries who were seriously ill, less than 1%–30%. Mean ACO PBPY for SI beneficiaries was $51,684. High-spending ACOs (highest SI PBPY quartile) spent $14,865 more per SI beneficiary than low-spending ACOs (lowest SI PBPY quartile), with unadjusted means of $60,386 and $45,520 (Table 1). Adjusted for patient and county factors, predicted SI PBPY was $53,552 versus $51,223 among high- and low-spending ACOs. Compared to high-spending ACOs, low-spending ACOs had lower rates of SI beneficiaries who were dual-eligible for Medicaid and fewer SI beneficiaries identified as Black, Asian, or Hispanic. Low-spending ACOs had lower rates of Alzheimer’s disease or related dementias, higher rates of chronic obstructive pulmonary disease, and lower-mortality rates than high-spending ACOs. Other diagnoses varied by less than 5 percentage points between high- and low-spending ACOs. The average risk-adjusted benchmark for high-spending ACOs was 30% higher than low-spending ACOs, indicating a sicker attributed population.
TABLE 1 -
ACO and Beneficiary Characteristics for ACOs in the Lowest and Highest Quartiles of Spending on Seriously Ill Beneficiaries
||Low Spending Quartile N = 289
||High Spending Quartile N = 290
|Per beneficiary per year (PBPY)a, USD
| Number of beneficiaries
| Updated benchmark, USD
| Readmissions/1,000 admissions
| ACO quality score
| Participating provider/1,000 beneficiaries
| Rural beneficiaries, %
| ACO contract type, %
| Hospital-led or co-led
| Commercial contract
| Advance payment model
| MSSP Track 2 or 3
Seriously ill beneficiary characteristics
| Demographics, %
| Dual-eligible for Medicaid
| Race/ethnicity, %
| Native American
| Health status, %
| Died in performance year
| Five or more comorbid conditions
| Alzheimer’s and related dementias
| Chronic obstructive pulmonary disorder
| Rheumatoid arthritis
| Transient ischemic attack
Note. ACO = accountable care organization; MSSP = Medicare Shared Savings Program; USD = 2016 United States dollars.
aBeneficiary-level Parts A and B Medicare expenditures were annualized to mirror MSSP calculations and standardized to remove variation resulting from inflation and geography using the Consumer Price Index and CMS Geographic Variation Public Use File. We applied the MSSP stop-loss rule by truncating costs at 0 and the average 99% percentile in national fee-for-service population by year within each Medicare eligibility group.
Compared to high-spending ACOs, low-spending ACOs had, on average, 2,150 fewer beneficiaries (14,178.8 vs. 16,328.9), a higher-percentage of rural beneficiaries (30% vs.13%), and a lower-readmissions rate (160.7 vs. 183.3). Low-spending ACOs were less likely to have a commercial contract (18.3% vs. 34.3%), more likely to participate in the advance payment model (10.0% vs. 6.9%), and less likely to be hospital-led (20.3% vs. 33.5%) than high-spending ACOs. Relative to high-spending ACOs, low-spending ACOs also had a higher-MSSP quality performance score (91.6 vs. 89.6).
Utilization in High- and Low-Spending ACOs
Low-spending ACOs (vs. high-spending ACOs) had fewer SI beneficiaries with any occurrence of readmission, skilled nursing, home health, and Medicare Part B use. All SI patients utilized inpatient care following the inclusion criteria (Figure 1). Compared to high-spending ACOs, low-spending ACOs had higher rates of use for DME and ambulatory surgery. Similarly, low-spending ACOs had lower per-patient spending than high-spending ACOs for all categories except ambulatory surgery and DME (Figure 2).
ACO Factors Associated With Low Serious Illness Spending
Large ACOs with hospital governance and commercial contracts were less likely to be in the lowest quartile for SI spending (Figure 3). ACOs with high MSSP quality scores and high percentages of rural beneficiaries were more likely to be in the lowest quartile for SI spending. ACOs in the top quartile for the percentage of the population that is seriously ill were less likely to be in the lowest quartile for SI spending. The probability of being in the lowest spending quartile was 10 percentage points higher for ACO contracts in 2016 compared to 2014.
After controlling for ACO factors, the probability of an ACO achieving MSSP shared savings was 16.4 (confidence interval [CI] = 9.3, 23.6) percentage points higher for ACOs in the lowest versus highest spending quartile (Figure 4). The magnitude of the association was second only to participation in MSSP Track 2 or 3, which was associated with an increase of 29.8 percentage points in the probability of achieving shared savings.
Lower spending for seriously ill Medicare beneficiaries in ACOs was associated with achieving ACO shared savings in MSSP. For most ACOs, the seriously ill contribute approximately half of the spending yet constitute 8%–13% of the attributed population. Patient and geographic (county) factors explained $2,329 of the observed difference in PBPY spending on seriously ill beneficiaries between high- and low-spending ACOs. The remaining $12,536 may indicate variation as a result of potentially modifiable factors. Consequently, if 10% of attributed beneficiaries were seriously ill, then an ACO that moved from the worst to the best quartile of per capita SI spending could realize a reduction of $1,200 PBPY for the ACO population overall. Though the prevalence and case mix of seriously ill populations vary across ACOs, this association suggests that care provided for seriously ill patients is an important consideration for ACOs to achieve MSSP shared savings.
ACOs with lower percentages of attributed SI beneficiaries were more likely to have low SI spending. Although evidence for ACO selection (e.g., cherry-picking healthy patients) is mixed, the potential for ACOs to achieve spending targets by avoiding attribution of high-risk patients could explain this correlation (Markovitz et al., 2019; McWilliams et al., 2020). Consistent with prior evidence that physician-led versus hospital-led ACOs produce greater savings in the general attributed population, we found that physician-led ACOs were more likely to have low spending for the seriously ill population (Kaufman et al., 2017; McWilliams et al., 2018, 2020). For ACOs starting in 2012 or 2013, most per capita savings were generated among low-risk populations through the first 2 years of the MSSP, as well as through reductions in postacute care (Colla et al., 2019; McWilliams et al., 2017; Schulz et al., 2018). The correlations between ACO shared-savings success and low spending in the SI may follow broad interventions to improve care coordination and reduce unnecessary utilization, which reduced costs for the entire attributed population and included seriously ill by extension.
ACO Track 2 or 3 participation was the strongest driver of shared savings achievement in this descriptive study. This finding contributes to research that suggests entities with both upside and downside risk are more likely to achieve savings (Ouayogodé et al., 2017). The Advance Payment ACO Model, which provides participants with monthly upfront payments, was also associated with low spending. There is little evidence evaluating the 35 ACOs participating in the advance payment model, though early research suggests participants successfully used advance payments to develop their care coordination infrastructure to achieve savings (Trombley et al., 2019). The probability of low spending among ACOs was higher for contracts in 2015 and 2016 relative to 2014, which may reflect the growth in experience of ACOs. In later years, ACOs were more likely to have had more time to implement care management strategies and more experience managing population health (Bleser, Saunders, Muhlestein et al., 2019). However, the general decline in inpatient stays nationally could also explain reduced per-patient spending (McDermott et al., 2017). As SI care models gain traction with CMS demonstration models such as Primary Care First, the downward trend in SI spending is likely to continue.
As more ACOs assume downside risk through the 2018 Pathways to Success Model, the next steps in population health management may include better methods to stratify patients by risk and identify high-investment, high-margin care models as new sources of savings (Gupta et al., 2019). ACOs with low spending in SI were smaller, physician-led, and performed better in the MSSP quality measurement for the attributed population. Unlike physician-led ACOs, hospital-led ACOs must balance the lure of immediate fee-for-service revenue for inpatient stays with the long-term potential for shared savings. The reduced ability to spread risk may increase the need for small ACOs to be proactive in managing complex patients, particularly in rural communities where residents tend to be older and sicker (Nattinger et al., 2016). Larger ACOs with hospitals may have more capital to invest in complex illness care and have more organizational integration, which is associated with savings in clinically complex populations (Colla et al., 2020).
While seriously ill patients constitute a meaningful proportion of ACOs, the high churn rates make it difficult to determine which subgroups of patients have care needs that are actionable. Each year, a large proportion of the seriously ill experience either death or improvement in health status. Early adopters of care models for high-cost, high-need populations provide examples of ACO strategies for improving SI care, such as providing palliative care through home- or community-based programs (Bleser, Saunders, Winfield et al., 2019). Of ACOs participating in a national survey, 10% selected palliative care/hospice as a top priority, but most hospital- and community-based palliative care was reported as being only partially implemented, suggesting that ACOs are still in the early stages of establishing such programs (Roiland et al., 2020). Interviews suggest some ACOs are using care plans to manage the care of complex patients, and most care plans were developed and maintained by care management rather than clinical staff (Fraze et al., 2020). Promising pilots for models of care, including home-based palliative and primary care for frail and functionally limited persons, have produced savings and reduced unnecessary or harmful care (Ruiz et al., 2017, 2018; Szanton et al., 2018; Yosick et al., 2019). As the data on ACOs continue to accumulate, rigorous studies evaluating interventions should augment this descriptive study of the SI population.
These findings were limited to the subgroup of the MSSP population that meets the criteria for SI during the ACO PY. The subgroup was conditioned on inpatient and skilled nursing facility or home health or DME use. An ACO that is successful in preventing hospitalization may have a smaller and systemically different SI population captured using this definition compared to ACOs that do not minimize hospitalizations. We were unable to determine how much of the variation in costs is related to appropriate care; thus, we cannot draw any conclusions about whether the spending for high-spending ACOs is modifiable. We were unable to calculate patient-level benchmarks because of a lack of claims-level data for attributed patients to evaluate savings relative to the benchmark for the SI cohort. Smaller ACOs have greater random variation in outcomes measures, including cost, that may not reflect actual performance (Barr et al., 2018). We evaluated whether this random variation explained the association between rural ACO and low-spending ACOs using another model predicting the high spending, which found that rural ACOs were less likely to be high spending.
In our study, lower spending for seriously ill Medicare beneficiaries and risk-bearing contracts in ACOs was found to be associated with the successful achievement of MSSP shared savings. The substantial variation between ACOs in spending for the seriously ill, particularly on acute and postacute care, suggests the potential return on investment from interventions targeting SI care may be substantial. The largely untapped potential of ACOs to improve SI care underscores the need for a better understanding of organizations that do this well under risk-bearing payment models. Opportunities may exist for Medicare ACOs to produce savings by improving care for frail and seriously ill patients, and further research is needed to understand which ACO strategies are effective for improving SI care.
Aldridge M. D., Bradley E. H. (2017). Epidemiology and patterns of care at the end of life: Rising complexity, shifts in care patterns and sites of death. Health Affairs (Millwood)
, 36(7), 1175–1183. https://doi.org/10.1377/hlthaff.2017.0182
Barber J., Thompson S. (2004). Multiple regression of cost data: Use of generalised linear models. Journal of Health Services Research and Policy
, 9(4), 197–204. https://doi.org/10.1258/1355819042250249
Barr L., Loengard A., Hastings L., Gronniger T. (2018, May 11). Payment reform in transition—scaling ACOs for success. Health Affairs (Millwood)
Bleser W. K., Saunders R. S., Muhlestein D. B., McClellan M. (2019). Why do accountable care organizations leave the Medicare shared savings program? Health Affairs (Millwood)
, 38(5), 794–803. https://doi.org/10.1377/hlthaff.2018.05097
Bleser W. K., Saunders R. S., Winfield L., Japinga M., Smith N., Kaufman B. G., Crook H. L., Muhlestein D. B., McClellan M. (2019). ACO serious illness care: Survey and case studies depict current challenges and future opportunities. Health Affairs (Millwood)
, 38(6), 1011–1020. https://doi.org/10.1377/hlthaff.2019.00013
Centers for Medicare & Medicaid Services. (n.d.). Chronic conditions data warehouse condition algorithms
Colla C. H., Lewis V. A., Stachowski C., Usadi B., Gottlieb D. J., Bynum J. P.W. (2019, April 18). Changes in use of postacute care associated with accountable care organizations in hip fracture, stroke, and pneumonia hospitalized cohorts. Medical Care
Colla C. H., Lewis V. A., Tierney E., Muhlestein D. B. (2016). Hospitals participating in acos tend to be large and urban, allowing access to capital and data. Health Affairs (Millwood)
, 35(3), 431–439. https://doi.org/10.1377/hlthaff.2015.0919
Colla C., Yang W., Mainor A. J., Meara E., Ouayogode M. H., Lewis V. A., Shortell S., Fisher E. (2020, October 26). Organizational integration, practice capabilities, and outcomes in clinically complex medicare beneficiaries. Health Services Research
, 55(Suppl 3), 1085–1097. https://doi.org/10.1111/1475-6773.13580
Comfort L. N., Shortell S. M., Rodriguez H. P., Colla C. H. (2018). Medicare accountable care organizations of diverse structures achieve comparable quality and cost performance. Health Services Research
, 53(4), 2303–2323. https://doi.org/10.1111/1475-6773.12829
Fraze T. K., Beidler L. B., Briggs A. D.M., Colla C. H. (2020). Translating evidence into practice: ACOs’ use of care plans for patients with complex health needs. Journal of General Internal Medicine
, 36(1), 147–153. https://doi.org/10.1007/s11606-020-06122-4
Gupta R., Roh L., Lee C., Reuben D., Naeim A., Wilson J., Skootsky S. A. (2019, April 9). The population health value framework: Creating value by reducing costs of care for patient subpopulations with chronic conditions. Journal of the Association of American Medical Colleges
, 36(1), 147–153. https://doi.org/10.1097/acm.0000000000002739
Institute of Medicine (2015). Dying in America: Improving quality and honoring individual preferences near the end of life
. National Academies Press.
Kaufman B. G., Spivack B. S., Stearns S. C., Song P. H., O’Brien E. C. (2017, December 1). Impact of accountable care organizations on utilization, care, and outcomes: A systematic review. Medical Care Research Review
, 36(1), 147–153. https://doi.org/10.1177/1077558717745916
Kelley A. S. (2013). Epidemiology of care for patients with serious illness. Journal of Palliatiave Medicine
, 16(7), 730–733. https://doi.org/10.1089/jpm.2013.9498
Kelley A. S. (2014). Defining “serious illness”. Journal of Palliatiave Medicine
, 17(9), 985. https://doi.org/10.1089/jpm.2014.0164
Kelley A. S., Covinsky K. E., Gorges R. J., McKendrick K., Bollens-Lund E., Morrison R. S., Ritchie C. S. (2017). Identifying older adults with serious illness: A critical step toward improving the value of health care. Health Services Research
, 52(1), 113–131. https://doi.org/10.1111/1475-6773.12479
Kyle M. A., McWilliams J. M., Landrum M. B., Landon B. E., Trompke P., Nyweide D. J., Chernew M. E. (2020). Spending variation among ACOs in the Medicare Shared Savings Program. American Journal of Managed Care
, 26(4), 170–175. https://doi.org/10.37765/ajmc.2020.42834
Larson B. K., Van Citters A. D., Kreindler S. A., Carluzzo K. L., Gbemudu J. N., Wu F. M., Nelson E. C., Shortell S. M., Fisher E. S. (2012). Insights from transformations under way at four Brookings-Dartmouth accountable care organization pilot sites. Health Affairs (Millwood)
, 31(11), 2395–2406. https://doi.org/10.1377/hlthaff.2011.1219
Manning W., Basu A., Mullahy J. (2002). Modeling costs with generalized Gamma Regression. [National Bureau of Economic Research Technical Working Paper No 293]
Markovitz A. A., Hollingsworth J. M., Ayanian J. Z., Norton E. C., Moloci N. M., Yan P. L., Ryan A. M. (2019). Risk adjustment in Medicare ACO program deters coding increases but may lead ACOs to drop high-risk beneficiaries. Health Affairs(Millwood)
, 38(2.), 253–261. https://doi.org/10.1377/hlthaff.2018.05407
McDermott K., Elixhauser A., Sun R. (2017, June 27). Patient trends in hospital inpatient stays in the United States, 2005-2014 [HCUP Statistical Brief No 225]
McWilliams J. M., Chernew M. E., Landon B. E. (2017). Medicare ACO program savings not tied to preventable hospitalizations or concentrated among high-risk patients. Health Affairs (Millwood)
, 36(12), 2085–2093. https://doi.org/10.1377/hlthaff.2017.0814
McWilliams J. M., Hatfield L. A., Landon B. E., Chernew M. E. (2020). Savings or selection? Initial spending reductions in the Medicare Shared Savings Program and considerations for reform. Milbank Quarterly
, 98(3), 847–907. https://doi.org/10.1111/1468-0009.12468
McWilliams J. M., Hatfield L. A., Landon B. E., Hamed P., Chernew M. E. (2018). Medicare Spending after 3 years of the Medicare Shared Savings Program. New England Journal of Medicine
, 379(12), 1139–1149. https://doi.org/10.1056/NEJMsa1803388
Meier D. E., Back A. L., Berman A., Block S. D., Corrigan J. M., Morrison R. S. (2017). A national strategy for palliative care. Health Affairs (Millwood)
, 36(7), 1265–1273. https://doi.org/10.1377/hlthaff.2017.0164
Moran J. L., Solomon P. J. (2014). Fixed effects modelling for provider mortality outcomes: Analysis of the Australia and New Zealand Intensive Care Society (ANZICS) Adult Patient Data-base. PLoS One
, 9(7), e102297. https://doi.org/10.1371/journal.pone.0102297
Muhlestein D., McClellan M. (2016, April 21). Accountable care organizations in 2016: Private and public-sector growth and dispersion. Health Affairs (Millwood)
. http://healthaffairs.org/blog/2016/04/21/ accountable-care-organizations-in-2016-private- and-public-sector-growth-and-dispersion/
Muhlestein D., Saunders R. S., Richards R., McClellan M. (2018). Recent progress in the value journey: growth of ACOs and value-based payment models in 2018. Health Affairs (Millwood)
Mustillo S., Landerman L. R., Land K. C. (2012). Modeling longitudinal count data: Testing for group differences in growth trajectories using average marginal effects. Sociological Methods & Research
, 41(3), 467–487. https://doi.org/10.1177_0049124112452397
Nattinger M. C., Mueller K., Ullrich F., Zhu X. (2016). Financial performance of rural Medicare ACOs. Journal of Rural Health
, 34(1), 98–102. https://doi.org/10.1111/jrh.12205
Ouayogodé M. H., Colla C. H., Lewis V. A. (2017). Determinants of success in shared savings programs: An analysis of ACO and market characteristics. Healthcare
, 5(1–2), 53–61. https://doi.org/10.1016/j.hjdsi.2016.08.002
Roiland R., Bleser W., Muhlestein D., Saunders R. (2020, January 7). How are ACOs prioritizing palliative care and other serious illness strategies? Duke-Margolis Center for Health Policy
. https://healthpolicy.duke.edu/publications/ how-are-acos-prioritizing-palliative-care-and- other-serious-illness-strategies
Ruiz S., Snyder L. P., Giuriceo K., Lynn J., Ewald E., Branand B., Parashuram S., Loganathan S., Bysshe T. (2018). Innovative models for high-risk patients use care coordination and palliative supports to reduce end-of-life utilization and spending. Innovation in Aging
, 1(2.). https://doi.org/10.1093/geroni/igx021
Ruiz S., Snyder L. P., Rotondo C., Cross-Barnet C., Colligan E. M., Giuriceo K. (2017). Innovative home visit models associated with reductions in costs, hospitalizations, and emergency department use. Health Affairs (Millwood)
, 36(3), 425–432. https://doi.org/10.1377/hlthaff.2016.1305
Schulz J., DeCamp M., Berkowitz A. S.A. (2018). Spending patterns among Medicare ACOs that have reduced costs. Journal of Healthcare Management
, 63(6), 374–381. https://doi.org/10.1097/jhm-d-17-00178
Szanton S. L., Alfonso Y. N., Leff B., Guralnik J., Wolff J. L., Stockwell I., Gitlin L. N., Bishai D. (2018). Medicaid cost savings of a preventive home visit program for disabled older adults. Journal of the American Geriatrics Society
, 66(3), 614–620. https://doi.org/10.1111/jgs.15143
Trombley M. J., Fout B., Brodsky S., McWilliams J. M., Nyweide D. J., Morefield B. (2019). Early effects of an accountable care organization model for underserved areas. New England Journal of Medicine
, 381:543–551. https://doi.org/10.1056/NEJMsa1816660
Yosick L., Crook R. E., Gatto M., Maxwell T. L., Duncan I., Ahmed T., Mackenzie A. (2019). Effects of a population health community-based palliative care program on cost and utilization. Journal of Palliative Medicine
, 22(9), 1075–1081. https://doi.org/10.1089/jpm.2018.0489
Zhu X., Mueller K., Huang H., Ullrich F., Vaughn T., MacKinney A. C. (2019). Organizational attributes associated with Medicare ACO quality performance. Journal of Rural Health
, 35(1), 68–77. https://doi.org/10.1111/jrh.12304