[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Effect of Attribution Length (Consistent Primary Care) on Annual Cost
Effect of Attribution Length (Consistent Primary Care) on Annual Cost

Months of continuous attribution included all months from when patients were initially attributed until the first gap in attribution. Because patients with 3 or more body systems with a chronic condition almost surely have nonzero costs, we used a standard Poisson distribution regression model rather than a zero-inflated model for estimating the effect of chronicity on annual cost. Error bars indicate 95% CI. One to 6 months of attribution indicates the reference category.

Figure 2.
Effect of Attribution Length on Annual Cost Exceeding $100 000
Effect of Attribution Length on Annual Cost Exceeding $100 000

Months of continuous attribution included all months from when the patients were initially attributed until the first gap in attribution. Error bars indicate 95% CI. One to 6 months of attribution indicates the reference category.

Table 1.  
Descriptive Statisticsa
Descriptive Statisticsa
Table 2.  
Effect of Attribution Length on Annual Use of Health Care Resoucesa
Effect of Attribution Length on Annual Use of Health Care Resoucesa
Table 3.  
Effect of Attribution Length on Annual Use of Resources by the Number of Body Systems With a Chronic Condition
Effect of Attribution Length on Annual Use of Resources by the Number of Body Systems With a Chronic Condition
1.
US Department of Health and Human Services.  About the Affordable Care Act.http://www.hhs.gov/healthcare/about-the-law/index.html. Last reviewed August 13, 2015. Accessed September 15, 2015.
2.
Centers for Medicare & Medicaid Services.  Fact sheets: Medicare ACOs provide improved care while slowing cost growth in 2014.https://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2015-Fact-sheets-items/2015-08-25.html. Published August 25, 2015. Accessed September 4, 2015.
3.
Song  Z, Safran  DG, Landon  BE,  et al.  Health care spending and quality in year 1 of the alternative quality contract.  N Engl J Med. 2011;365(10):909-918.PubMedArticle
4.
McWilliams  JM, Chernew  ME, Landon  BE, Schwartz  AL.  Performance differences in year 1 of Pioneer accountable care organizations.  N Engl J Med. 2015;372(20):1927-1936.PubMedArticle
5.
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.PubMedArticle
6.
McWilliams  JM, Landon  BE, Chernew  ME, Zaslavsky  AM.  Changes in patients’ experiences in Medicare accountable care organizations.  N Engl J Med. 2014;371(18):1715-1724.PubMedArticle
7.
Colla  CH, Wennberg  DE, Meara  E,  et al.  Spending differences associated with the Medicare Physician Group Practice Demonstration.  JAMA. 2012;308(10):1015-1023.PubMedArticle
8.
Maeng  DD, Graham  J, Graf  TR,  et al.  Reducing long-term cost by transforming primary care: evidence from Geisinger’s medical home model.  Am J Manag Care. 2012;18(3):149-155.PubMed
9.
Maeng  DD, Khan  N, Tomcavage  J, Graf  TR, Davis  DE, Steele  GD.  Reduced acute inpatient care was largest savings component of Geisinger Health System’s patient-centered medical home.  Health Aff (Millwood). 2015;34(4):636-644.PubMedArticle
10.
Christensen  EW, Dorrance  KA, Ramchandani  S,  et al.  Impact of a patient-centered medical home on access, quality, and cost.  Mil Med. 2013;178(2):135-141.PubMedArticle
11.
Hoff  T, Weller  W, DePuccio  M.  The patient-centered medical home: a review of recent research.  Med Care Res Rev. 2012;69(6):619-644.PubMedArticle
12.
Averill  RF, Goldfield  NI, Vertrees  JC, McCullough  EC, Fuller  RL, Eisenhandler  J.  Achieving cost control, care coordination, and quality improvement through incremental payment system reform.  J Ambul Care Manage. 2010;33(1):2-23.PubMedArticle
13.
Cole  ES, Campbell  C, Diana  ML, Webber  L, Culbertson  R.  Patient-centered medical homes in Louisiana had minimal impact on Medicaid population’s use of acute care and costs.  Health Aff (Millwood). 2015;34(1):87-94.PubMedArticle
14.
Allen  S.  Medicaid and pediatric accountable care organizations: a case study.  Accountable Care News.2010;1(5):1-4.
15.
Kelleher  KJ, Cooper  J, Deans  K,  et al.  Cost saving and quality of care in a pediatric accountable care organization.  Pediatrics. 2015;135(3):e582-e589.PubMedArticle
16.
Raphael  JL, Giardino  AP.  Accounting for kids in accountable care: a policy perspective.  Clin Pediatr (Phila). 2013;52(8):695-698.PubMedArticle
17.
Homer  CJ, Patel  KK.  Accountable care organizations in pediatrics: irrelevant or a game changer for children?  JAMA Pediatr. 2013;167(6):507-508.PubMedArticle
18.
Johns Hopkins Bloomberg School of Public Health. The Johns Hopkins ACG System, Technical Reference Guide. Version 10.0. http://acg.jhsph.org/public-docs/ACGv10.0TechRefGuide.pdf. Published December 2011. Accessed August 27, 2015.
19.
Agency for Healthcare Research and Quality.  Healthcare Cost and Utilization Project (HCUP): Clinical Classification Software (CCS) for ICD-9-CM.http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Last modified June 1, 2015. Accessed August 27, 2015.
20.
Agency for Healthcare Research and Quality.  Healthcare Cost and Utilization Project (HCUP): Chronic Condition Indicator (CCI) for ICD-9-CM.https://www.hcup-us.ahrq.gov/toolssoftware/chronic/chronic.jsp. Last modified June 1, 2015. Accessed August 27, 2015.
21.
Casalino  LP.  Accountable care organizations: the risk of failure and the risks of success.  N Engl J Med. 2014;371(18):1750-1751.PubMedArticle
22.
Toussaint  J, Milstein  A, Shortell  S.  How the Pioneer ACO model needs to change: lessons from its best-performing ACO.  JAMA. 2013;310(13):1341-1342.PubMedArticle
23.
Congressional Budget Office.  Lessons from Medicare’s demonstration projects on disease management, care coordination, and value-based payment. https://www.cbo.gov/sites/default/files/112th-congress-2011-2012/reports/01-18-12-MedicareDemoBrief.pdf. Accessed August 27, 2015.
24.
Kocot  SL, Dang-Vu  C, White  R, McClellan  M.  Early experiences with accountable care in Medicaid: special challenges, big opportunities.  Popul Health Manag. 2013;16(suppl 1):S4-S11.PubMedArticle
Original Investigation
February 2016

Effect of Attribution Length on the Use and Cost of Health Care for a Pediatric Medicaid Accountable Care Organization

Author Affiliations
  • 1Department of Research and Sponsored Programs, Children’s Hospitals and Clinics of Minnesota, Minneapolis
JAMA Pediatr. 2016;170(2):148-154. doi:10.1001/jamapediatrics.2015.3446
Abstract

Importance  Little is known about the effect of pediatric accountable care organizations (ACOs) on the use and costs of health care resources, especially in a Medicaid population.

Objective  To assess the association between the length of consistent primary care (length of attribution) as part of an ACO and the use and cost of health care resources in a pediatric Medicaid population.

Design, Setting, and Participants  A retrospective study of Medicaid claims data for 28 794 unique pediatric patients covering 346 277 patient-attributed months within a single children’s hospital. Data were collected for patients attributed from September 1, 2013, to May 31, 2015. The effect of the length of attribution within a single hospital system’s ACO on the use and costs of health care resources were estimated using zero-inflated Poisson distribution regression models adjusted for patient characteristics, including chronic conditions and a measure of predicted patient use of resources.

Exposures  Receiving a plurality of primary care at an ACO clinic during the preceding 12 months (attribution to the ACO).

Main Outcomes and Measures  The primary outcome measure was the length of attribution at an ACO clinic compared with subsequent inpatient hospitalization and subsequent use and cost of outpatient and ancillary health care resources.

Results  Among the 28 794 pediatric patients receiving treatment covering 346 277 patient-attributed months during the study period, continuous attribution to the ACO for more than 2 years was associated with a decrease (95% CI) of 40.6% (19.4%-61.8%) in inpatient days but an increase (95% CI) of 23.3% (2.04%-26.3%) in office visits, 5.8% (1.4%-10.2%) in emergency department visits, and 15.3% (12.5%-18.0%) in the use of pharmaceuticals. These changes in the use of health care resources combined resulted in a cost reduction of 15.7% (95% CI, 6.6%-24.8%). At the population level, the impact of consistent primary care was muted by the many patients in the ACO having shorter durations of participation.

Conclusions and Relevance  These findings suggest significant and durable reductions of inpatient use and cost of health care resources associated with longer attribution to the ACO, with attribution as a proxy for exposure to the ACO’s consistent primary care. Consistent primary care among the pediatric Medicaid population is challenging, but these findings suggest substantial benefits if consistency can be improved.

Introduction

Accountable care organizations (ACOs) grew out of the Affordable Care Act and were designed to improve the quality of health care and control rising health care costs of populations.1,2 Many ACOs have served Medicare populations, although ACO-like contracts in the private sector have also been established.3 Results so far have indicated that Medicare ACOs largely succeeded in controlling costs while providing health care of the same or better quality.46 Savings estimates for these ACOs and similar payment systems range from 1.2% to 3.8%, with some ACOs experiencing increased costs.35,7 The rationale underlying potential cost savings centers on reduced use of expensive inpatient services through more consistent and appropriate outpatient care, such as the medical home.812 The ACO contracts usually require health care professionals to assume responsibility for all health care costs for a patient population. This contract allows health care professionals flexibility to address their patients’ needs and provide case management services that are not typically reimbursed through fee-for-service (FFS) and that could decrease more expensive care, such as inpatient stays.812

Few researchers have published their ACO experience with Medicaid populations and still fewer with pediatric Medicaid populations.13,14 A notable exception has been the exclusively pediatric ACO, Partners for Kids, in Ohio. This ACO achieved lower-cost growth than the traditional pediatric Medicaid programs while generally improving the quality of health care.15 The application of an adult ACO model to a pediatric population could be problematic because adult care focuses on managing chronic conditions, whereas pediatric care generally focuses on development, and because Medicaid is generally a temporary safety net, whereas Medicare is permanent.14,16,17 The relatively few published reports of experiences with exclusively pediatric ACOs make assessment of the viability and effectiveness of pediatric Medicaid ACOs difficult. Pediatric health care organizations have little information on which to base their decision to participate in an ACO and how the ACO may be best structured to benefit all parties.

In 2013, Children’s Hospitals and Clinics of Minnesota (CHC) contracted with the State of Minnesota Department of Human Services (DHS) to develop a pediatric Medicaid ACO to provide health care for approximately 15 000 children in Minnesota. The assumption for effectiveness was that patients who chose to receive their primary care at a CHC facility would have better coordination and consistency of care, resulting in decreased use of high-cost inpatient or outpatient services. This study examined that assumption. The results should provide guidance for other health care organizations considering ACO participation.

Box Section Ref ID

At a Glance

  • The association between the length of consistent primary care in an accountable care organization (attribution) and the use and cost of health care services was assessed.

  • Continuous attribution to a primary care practice for more than 2 years was associated with a 40.6% decrease in inpatient days.

  • Attribution for more than 2 years was associated with increases of 23.3% for office visits, 5.8% for emergency department visits, and 15.3% for use of pharmaceuticals.

  • Overall, being continuously involved with a primary care practice was associated with a 15.7% reduction in cost.

Methods
Accountable Care Organization

Children’s Hospitals and Clinics of Minnesota is a nonprofit, independent entity that owns and operates 2 hospitals, an outpatient surgery center, and outpatient clinics throughout the Twin Cities metropolitan area. In 2014, approximately 14 000 inpatient, 95 000 emergency department (ED), and 400 000 other outpatient visits were noted. The CHC organization has approximately 60% of the Twin Cities market share for inpatient discharges. About 20% of Minnesota’s pediatric Medicaid patients were attributed to CHC’s ACO, and 96% of patients attributed to CHC came from the Twin Cities metropolitan area. This research was approved by the institutional review boards of CHC and the Minnesota DHS, who waived the requirement for informed consent.

ACO Contract With the Minnesota DHS

In 2013, CHC entered into an ACO contract with the DHS, termed the Integrated Health Partnership, to be accountable for the health care of about 15 000 pediatric Medicaid patients. Under this contract, attribution to the ACO was defined as (1) being in a CHC health care home or (2) receiving a plurality of primary care at a CHC clinic. Attribution as used in this study was a proxy for receiving consistent primary care at a CHC clinic. Patients were not aware that they were attributed to this ACO and could choose to receive any health care service wherever they liked. Care provided by CHC to Medicaid patients was reimbursed on an FFS basis, as was care received by patients at other facilities. However, CHC was at risk for meeting quality and risk-adjusted cost targets for attributed patients regardless of where the patients’ health care was obtained. If costs were below the target, CHC shared savings equally with the DHS; if costs exceeded the target, CHC was at risk for 50% of the losses. Individual health care professionals and clinics did not receive any of the shared savings and were not at risk for shared losses. Quality targets included appropriate numbers of well-child visits, being up-to-date with recommended immunizations, appropriate treatment for upper respiratory tract infections, the percentage of patients with well-controlled asthma, and the quality of the patients’ experience.

Study Design

This retrospective, observational study assessed the effect of a pediatric Medicaid ACO on the use and cost of health care resources. Our assumption and hypothesis for this study was that longer attribution, which was a proxy for consistent primary care, would be associated with decreased use of inpatient, outpatient, and ancillary services and decreased costs.

Data

The primary outcome measure consisted of the number of inpatient hospitalizations. Secondary outcome measures included the use of ED, other outpatient, and ancillary health care services and total costs. Patients were attributed on a monthly basis by the DHS. Each month that a patient was attributed to the ACO constituted a unit of observation for analysis (patient-attributed month). The use and cost metrics for each patient-attributed month represented a trailing 12-month measure of annual use and costs of health care services.

The DHS provided age, sex, county of residence, a predicted measure of patient use of resources, whether treatment was FFS or capitated through a third-party payer, and whether the patient had health insurance other than Medicaid at any point during their attribution to the ACO. The measure of patient use of resources was computed using the Adjusted Clinical Groups software system developed at Johns Hopkins Bloomberg School of Public Health.18 We aggregated diagnoses into groups using Clinical Classification Software for International Classification of Diseases, Ninth Revision, Clinical Modification that was produced by the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality.19 The presence of chronic disease and the number of body systems with a chronic disease were examined using Chronic Condition Indicator software from the Agency for Healthcare Research and Quality.20

Clinical Care

Patients could obtain care wherever they chose and were not restricted to a CHC facility. Thus, CHC had an incentive to provide health care that met the patients’ needs and minimized obtaining care at other facilities where the ACO did not have control. Many patients sought care at other facilities. For example, attributed patients received about 25% of ED care and 30% of inpatient care at non-CHC facilities. We assumed that even care obtained at a non-CHC facility would be coordinated through the primary care clinic. The share of patient care provided by CHC declined with age. For example, excluding the newborn nursery, CHC provided 90% of inpatient days for patients who were 1 year of age compared with 34% for those who were 18 years of age.

Measures of Use of Health Care Resources

Measures of the use of health care resources included the number of inpatient days, office visits with a physician or a nurse practitioner, ED visits, and unique substance-route pharmaceutical counts. These measures are reported as events per patient per year. As supplemental information, we also provided results for the number of inpatient stays, radiology encounters, laboratory encounters, and annual days of supply for all pharmaceuticals (eTable in the Supplement).

Cost Measures

We included 3 annualized cost measures—total costs, costs included in the ACO contract, and capped included costs. Costs were defined as the amount that the DHS paid on a claim. Total cost represented the total of all claims paid by the DHS. Included costs were the subset of paid claims included in the ACO contract. Excluded costs were those the ACO could not reasonably influence (ie, dental services, transportation, personal care services, long-term care, and residential mental health). Capped included cost was the measure used by the DHS when determining whether the ACO was eligible for shared savings; for this measure, the DHS capped each individual patient’s annual costs at $200 000. Only 0.3% of patients exceeded this cap.

Statistical Analysis

Use and cost measures were positively skewed, with many patients having zero use for some measures. We used a zero-inflated Poisson distribution regression model to estimate the effect of attribution length on the use and cost of resources. Given the longitudinal nature of the data and that the use and cost metrics for each patient-attributed month were annual values, these measures were highly correlated, particularly for patient attribution in successive months. Therefore, we used clustered bootstrap SEs (patients selected with replacement) obtained from 50 bootstrap samples. The size of these samples was 28 794 patients (approximately 249 000 patient-attributed months for the continuous attribution model or approximately 346 000 for the total attribution model). All models were adjusted for sex, age, FFS patients, presence of other health insurance, county of residence, relative use of health care resources by the patient, and the number of body systems with a chronic condition. Because many patients had attribution gaps, we conducted the analysis first on the subset of patients without an attribution gap and second, for patients with and without attribution gaps. As a sensitivity test, we examined cross-sectional data from a single month to eliminate the longitudinal correlation and determine how the results compared with the longitudinal model.

We used logistic regression to examine (1) the association between attribution length and the probability of exceeding a catastrophic cost cap, (2) the characteristics of those who stopped attribution, and (3) the characteristics of those who returned to attribution. We used STATA software (version 14.0; StataCorp) for all calculations.

Results

A total of 28 794 unique patients were attributed to CHC for at least 1 month from September 1, 2013, to May 31, 2015, for a total of 346 277 patient-attributed months. The mean (SD) patient age was 7.5 (5.2) years. The mean (SD) length of attribution to the ACO was 12.7 (7.0) months (Table 1). In any given month, approximately 5% of patients left the ACO, 3% returned after previously leaving the ACO, and 3% joined the ACO for the first time. Approximately 50% of patients were still in the ACO after 12 months and 30% after 24 months. The DHS did not provide reasons for patients leaving the ACO, but patient departure could have resulted from a change in Medicaid eligibility or attribution to a different ACO or clinic.

Use of Health Care Resources

Attribution length of 2 years or more was associated with a decrease of 40.6% (95% CI, 19.4%-61.8%) in the number of inpatient days but an increase in the number of physician and/or nurse practitioner visits, ED visits, and use of pharmaceuticals (Table 2). Longer attribution was associated with fewer radiology encounters but more laboratory encounters (eTable in the Supplement). The impact of longer attribution was nearly always larger for those with continuous attribution than for those with attribution gaps (Table 2). Patients with chronic conditions were less likely to leave the ACO and more likely to return to attribution. For example, patients with 4 or more body systems with a chronic condition had an odds ratio (95% CI) of 0.51 (0.47-0.56) for leaving and 1.93 (1.77-2.11) for returning to attribution compared with those without a chronic condition.

Attribution to the ACO seemed most effective at lowering inpatient use of resources among patients with chronic conditions. The presence of at least 3 body systems with a chronic condition was associated with a greater decrease in inpatient use of resources compared with the presence of 0 to 2 chronic conditions (Table 3). After 2 years of attribution, the number of inpatient days decreased 20.1% (95% CI%, 0.7%-40.8%) for patients with 0 to 2 body systems with a chronic condition but decreased 31.3% (95% CI, 6.3%-68.9%) for those with 3 body systems with a chronic condition. For patients with 0 to 2 body systems with a chronic condition, the number of ED visits increased by 11.7% (95% CI, 5.4%-18.1%) after 2 years of attribution compared with no significant change for patients with at least 3 body systems with a chronic condition (Table 3).

Cost

Increased attribution length was associated with decreased costs. These cost reductions were largest 13 to 18 months after attribution and diminished with longer attribution (Figure 1A). Examination of included cost by the number of body systems with a chronic condition found no significant change for patients with 0 to 2 body systems with a chronic condition (Figure 1B) but a decrease of 32.5% (95% CI, 15.1%-49.8%) for those with 3 body systems with a chronic condition.

Cost reductions were larger for included costs than for capped included costs (Figure 1A). The DHS appeared to reap a larger benefit from continued attribution in the ACO than did CHC (Figure 1A). In addition, longer attribution was associated with a 33.0% (95% CI, 16.1%-49.9%) reduction in the probability of a patient having costs in excess of $100 000. Because only 0.3% of patients exceed the annual catastrophic cap of $200 000 for this ACO, we estimated the effect of attribution length on the probability of costs exceeding $100 000 (0.7% of patients) to improve statistical power (Figure 2). The benefit of cost reduction above the cap flowed completely to the DHS until costs fell below the threshold because the DHS bore the risk above the threshold.

As sensitivity tests, we first estimated the results for inpatient days using cross-sectional data from a single month to eliminate the longitudinal correlation. The result from this model was a reduction of 30.6% (95% CI, 22.4%-38.9%) in inpatient days. This result is consistent with the reduction from the longitudinal model (40.6%; 95% CI, 19.4%-61.8%). As a second sensitivity test, we estimated the longitudinal model using only patients 15 months or older to determine whether the cost results were driven by the cost of the newborn nursery. For this sample, the reduction in included costs was 19.6% (95% CI, 3.9%-35.3%), which is consistent with the reduction of 19.0% (95% CI, 7.2%-30.9%) from the full sample.

Discussion

Our study found that longer attribution to the CHC ACO (consistent primary care) was associated with decreased use of inpatient resources and decreased annual costs (Table 2 and Figure 1A). However, we found an increase in the number of ED and outpatient clinic visits and use of pharmaceuticals. Attribution to the ACO seemed most effective at reducing the use of resources by patients with chronic conditions (Table 3 and Figure 1B). Decreases in the use of resources and increases in cost savings mostly occurred in the first 13 to 18 months of attribution to the ACO. This finding suggests that pediatric Medicaid ACOs may be substituting less expensive outpatient care for more expensive inpatient care and exerting their greatest effect in the first year of a patient’s attribution. The ACO seemed most effective at reducing costs and the use of resources incurred by patients with multiple chronic conditions (Figure 1B).

Our findings were consistent with those of other studies showing that decreases in inpatient care accounted for most of the cost reductions.5,9 The ACO, through consistent primary care, may have substituted more expensive inpatient care with less expensive outpatient care through better primary care and care management. As an alternative, the increased use of outpatient resources may be explained by patients’ having easier access to care that had previously been relatively inaccessible.

Although 2 years of attribution to the ACO were associated with a 15.7% (95% CI, 6.6%-24.8%) cost reduction, the total ACO population had a 9.8% (95% CI, 3.9%-15.8%) cost reduction because only 30% of the ACO population remained attributed for 2 years. However, a 9.8% reduction would be substantial at the population level if applied to most pediatric Medicaid patients. Policies that would increase the length of attribution to the ACO might increase cost savings but might also be challenging to implement because these patients often have social determinants that influence when and where they seek care.

When the number of ACO patients reaches a critical mass, ACO practice patterns may have beneficial spillover effects for other patient groups.5 Casalino21(p1751) argued that until this occurs, “hospitals and physicians will be in the difficult position of dealing with diametrically opposed sets of payment incentives” because they are not reimbursed for using non-ACO health care professionals to identify patients and manage their chronic conditions through means other than the traditional visit with a health care professional.22,23 To unlock the benefits of ACOs, payers must shift away from FFS reimbursement; however, a crucial balance must be struck between providing incentives to invest in the personnel and infrastructure to succeed as an ACO but not shift so quickly that participation is discouraged.24 At the same time, shifting too slowly may result in more health care professionals and organizations declining to participate in ACOs.22

Our study has some limitations. Our results are based on a single ACO in 1 geographic area. However, CHC is the only exclusively pediatric ACO in Minnesota, and lessons learned here may be applicable to other pediatric ACOs. Demographic data were limited to age and sex, as provided by the DHS. To the degree that patients who are minorities or for whom English is not their primary language were included in the ACO, social determinants were an unobserved variable in our analysis. To the extent that these factors were associated with attribution patterns, they may have influenced the results. In other words, if patients in the most difficult social or family circumstances were also the ones who were more likely to leave attribution, efforts to keep them attributed consistently would not necessarily improve their patterns of use and cost of health care resources. Finally, patients may enroll in Medicaid when they have a health crisis; hence, using 1 to 6 months of attribution as the reference period may overstate the effect of attribution length on the use and cost of health care resources and cost. If so, 7 to 12 months of attribution may be a better reference period because the health crisis that precipitated Medicaid enrollment would likely be resolved in the first few months. When we used 7 to 12 months as an alternative reference period, we observed a decrease in the use of inpatient resources and total costs associated with the length of attribution to the ACO.

Conclusions

Our findings suggest significant and durable inpatient health care resource use and cost reductions associated with longer attribution to the ACO, where attribution is a proxy for exposure to consistent primary care within the ACO. Our findings will need to be replicated in other populations.

Back to top
Article Information

Corresponding Author: Eric W. Christensen, PhD, Department of Research and Sponsored Programs, Children’s Hospitals and Clinics of Minnesota, 2525 Chicago Ave S, Minneapolis, MN 55404 (eric.christensen@childrensmn.org).

Accepted for Publication: September 15, 2015.

Published Online: December 14, 2015. doi:10.1001/jamapediatrics.2015.3446.

Author Contributions: Dr Christensen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Both authors.

Acquisition, analysis, or interpretation of data: Both authors.

Drafting of the manuscript: Both authors.

Critical revision of the manuscript for important intellectual content: Payne.

Statistical analysis: Both authors.

Administrative, technical, or material support: Payne.

Study supervision: Payne.

Conflict of Interest Disclosures: None reported.

Additional Contributions: Heather Petermann, MS, sponsored the research with the Minnesota Department of Human Services institutional review board, provided study data, and reviewed the manuscript. Tim Ryan, MPA, Minnesota Department of Human Services, provided study data and reviewed the manuscript. Eric Solnitzky, BA, Children’s Hospitals and Clinics of Minnesota, served as a source for developing the study concept and reviewed the manuscript. None received any compensation for their contributions.

References
1.
US Department of Health and Human Services.  About the Affordable Care Act.http://www.hhs.gov/healthcare/about-the-law/index.html. Last reviewed August 13, 2015. Accessed September 15, 2015.
2.
Centers for Medicare & Medicaid Services.  Fact sheets: Medicare ACOs provide improved care while slowing cost growth in 2014.https://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2015-Fact-sheets-items/2015-08-25.html. Published August 25, 2015. Accessed September 4, 2015.
3.
Song  Z, Safran  DG, Landon  BE,  et al.  Health care spending and quality in year 1 of the alternative quality contract.  N Engl J Med. 2011;365(10):909-918.PubMedArticle
4.
McWilliams  JM, Chernew  ME, Landon  BE, Schwartz  AL.  Performance differences in year 1 of Pioneer accountable care organizations.  N Engl J Med. 2015;372(20):1927-1936.PubMedArticle
5.
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.PubMedArticle
6.
McWilliams  JM, Landon  BE, Chernew  ME, Zaslavsky  AM.  Changes in patients’ experiences in Medicare accountable care organizations.  N Engl J Med. 2014;371(18):1715-1724.PubMedArticle
7.
Colla  CH, Wennberg  DE, Meara  E,  et al.  Spending differences associated with the Medicare Physician Group Practice Demonstration.  JAMA. 2012;308(10):1015-1023.PubMedArticle
8.
Maeng  DD, Graham  J, Graf  TR,  et al.  Reducing long-term cost by transforming primary care: evidence from Geisinger’s medical home model.  Am J Manag Care. 2012;18(3):149-155.PubMed
9.
Maeng  DD, Khan  N, Tomcavage  J, Graf  TR, Davis  DE, Steele  GD.  Reduced acute inpatient care was largest savings component of Geisinger Health System’s patient-centered medical home.  Health Aff (Millwood). 2015;34(4):636-644.PubMedArticle
10.
Christensen  EW, Dorrance  KA, Ramchandani  S,  et al.  Impact of a patient-centered medical home on access, quality, and cost.  Mil Med. 2013;178(2):135-141.PubMedArticle
11.
Hoff  T, Weller  W, DePuccio  M.  The patient-centered medical home: a review of recent research.  Med Care Res Rev. 2012;69(6):619-644.PubMedArticle
12.
Averill  RF, Goldfield  NI, Vertrees  JC, McCullough  EC, Fuller  RL, Eisenhandler  J.  Achieving cost control, care coordination, and quality improvement through incremental payment system reform.  J Ambul Care Manage. 2010;33(1):2-23.PubMedArticle
13.
Cole  ES, Campbell  C, Diana  ML, Webber  L, Culbertson  R.  Patient-centered medical homes in Louisiana had minimal impact on Medicaid population’s use of acute care and costs.  Health Aff (Millwood). 2015;34(1):87-94.PubMedArticle
14.
Allen  S.  Medicaid and pediatric accountable care organizations: a case study.  Accountable Care News.2010;1(5):1-4.
15.
Kelleher  KJ, Cooper  J, Deans  K,  et al.  Cost saving and quality of care in a pediatric accountable care organization.  Pediatrics. 2015;135(3):e582-e589.PubMedArticle
16.
Raphael  JL, Giardino  AP.  Accounting for kids in accountable care: a policy perspective.  Clin Pediatr (Phila). 2013;52(8):695-698.PubMedArticle
17.
Homer  CJ, Patel  KK.  Accountable care organizations in pediatrics: irrelevant or a game changer for children?  JAMA Pediatr. 2013;167(6):507-508.PubMedArticle
18.
Johns Hopkins Bloomberg School of Public Health. The Johns Hopkins ACG System, Technical Reference Guide. Version 10.0. http://acg.jhsph.org/public-docs/ACGv10.0TechRefGuide.pdf. Published December 2011. Accessed August 27, 2015.
19.
Agency for Healthcare Research and Quality.  Healthcare Cost and Utilization Project (HCUP): Clinical Classification Software (CCS) for ICD-9-CM.http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Last modified June 1, 2015. Accessed August 27, 2015.
20.
Agency for Healthcare Research and Quality.  Healthcare Cost and Utilization Project (HCUP): Chronic Condition Indicator (CCI) for ICD-9-CM.https://www.hcup-us.ahrq.gov/toolssoftware/chronic/chronic.jsp. Last modified June 1, 2015. Accessed August 27, 2015.
21.
Casalino  LP.  Accountable care organizations: the risk of failure and the risks of success.  N Engl J Med. 2014;371(18):1750-1751.PubMedArticle
22.
Toussaint  J, Milstein  A, Shortell  S.  How the Pioneer ACO model needs to change: lessons from its best-performing ACO.  JAMA. 2013;310(13):1341-1342.PubMedArticle
23.
Congressional Budget Office.  Lessons from Medicare’s demonstration projects on disease management, care coordination, and value-based payment. https://www.cbo.gov/sites/default/files/112th-congress-2011-2012/reports/01-18-12-MedicareDemoBrief.pdf. Accessed August 27, 2015.
24.
Kocot  SL, Dang-Vu  C, White  R, McClellan  M.  Early experiences with accountable care in Medicaid: special challenges, big opportunities.  Popul Health Manag. 2013;16(suppl 1):S4-S11.PubMedArticle
×