[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Table 1.  
Processes That Contribute to Index of Care Management and Coordination Activities by Index Tertile: NSACO Wave 4 Respondents Linking to Older Adult Medicare Beneficiaries With Complex Needs
Processes That Contribute to Index of Care Management and Coordination Activities by Index Tertile: NSACO Wave 4 Respondents Linking to Older Adult Medicare Beneficiaries With Complex Needs
Table 2.  
Descriptive Characteristics of Fee-for-Service Medicare Beneficiaries With Complex Needs Attributed to MSSP ACOs, 2016a
Descriptive Characteristics of Fee-for-Service Medicare Beneficiaries With Complex Needs Attributed to MSSP ACOs, 2016a
Table 3.  
Association Between ACO Intensity of Care Management and Coordination Index and Outcomes for 1 402 582 Fee-for-Service Medicare Beneficiaries With Complex Needs, 2016a
Association Between ACO Intensity of Care Management and Coordination Index and Outcomes for 1 402 582 Fee-for-Service Medicare Beneficiaries With Complex Needs, 2016a
2.
Lewis  VA, Schoenherr  K, Fraze  T, Cunningham  A.  Clinical coordination in accountable care organizations: a qualitative study.  Health Care Manage Rev. 2019;44(2):127-136. doi:10.1097/HMR.0000000000000141PubMedGoogle ScholarCrossref
3.
Hsu  J, Price  M, Vogeli  C,  et al.  Bending the spending curve by altering care delivery patterns: the role of care management within a pioneer ACO.  Health Aff (Millwood). 2017;36(5):876-884. doi:10.1377/hlthaff.2016.0922PubMedGoogle ScholarCrossref
4.
McWilliams  JM, Chernew  ME, Landon  BE.  Medicare ACO program savings not tied to preventable hospitalizations or concentrated among high-risk patients.  Health Aff (Millwood). 2017;36(12):2085-2093. doi:10.1377/hlthaff.2017.0814PubMedGoogle ScholarCrossref
5.
McDonald  KM, Sundaram  V, Bravata  DM,  et al.  Closing the Quality Gap: A Critical Analysis of Quality Improvement Strategies. Rockville, MD: Agency for Healthcare Research and Quality; 2007. Report 04(07)-0051-7.
6.
Donelan  K, Barreto  EA, Michael  CU, Nordby  P, Smith  M, Metlay  JP.  Variability in care management programs in Medicare ACOs: a survey of medical directors.  J Gen Intern Med. 2018;33(12):2043-2045. doi:10.1007/s11606-018-4609-1PubMedGoogle ScholarCrossref
7.
O’Malley  AS, Rich  EC, Sarwar  R,  et al. How accountable care organizations use population segmentation to care for high-need, high-cost patients.  Issue Brief (Commonw Fund). 2019;2019:1-17. https://www.commonwealthfund.org/publications/issue-briefs/2019/jan/how-acos-use-segmentation-high-need-high-cost. Accessed Jube 17, 2019.
8.
Centers for Medicare & Medicaid Services.  Medicare Shared Savings Program Shared Savings and Losses and Assignment Methodology Specifications: Applicable Beginning Performance Year 2016. Woodlawn, MD: Centers for Medicare & Medicaid Services; 2016.
9.
Equator Network. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. 2015. http://www.equator-network.org/reporting-guidelines/strobe/. Accessed April 30, 2019.
10.
Peck  K, Usadi  B, Mainor  A, Newton  H, Meara  E.  How ACOs Are Caring for People with Complex Needs. New York, NY: Commonwealth Fund; 2018.
11.
US Census Bureau. How the Census Bureau measures poverty. 2016. https://www.census.gov/topics/income-poverty/poverty/about.html. Accessed April 7, 2016.
12.
Kim  DH, Schneeweiss  S.  Measuring frailty using claims data for pharmacoepidemiologic studies of mortality in older adults: evidence and recommendations.  Pharmacoepidemiol Drug Saf. 2014;23(9):891-901. doi:10.1002/pds.3674PubMedGoogle ScholarCrossref
13.
Joynt  KE, Figueroa  JF, Beaulieu  N, Wild  RC, Orav  EJ, Jha  AK.  Segmenting high-cost Medicare patients into potentially actionable cohorts.  Healthc (Amst). 2017;5(1-2):62-67. doi:10.1016/j.hjdsi.2016.11.002PubMedGoogle ScholarCrossref
14.
Bynum  JPW, Chang  CH, Austin  A, Carmichael  D, Meara  E.  Outcomes in older adults with multimorbidity associated with predominant provider of care specialty.  J Am Geriatr Soc. 2017;65(9):1916-1923. doi:10.1111/jgs.14882PubMedGoogle ScholarCrossref
15.
Centers for Medicare & Medicaid Services. Risk adjustment. 2018. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors.html.Accessed September 28, 2019.
16.
Centers for Medicare & Medicaid Services.  Nursing Facility Services (Codes 99304-99318). Woodlawn, MD: Centers for Medicare & Medicaid Services; 2006.
17.
Agency for Healthcare Research and Quality.  AHRQ Quality Indicators: Guide to Prevention Quality Indicators: Hospital Admission for Ambulatory Care Sensitive Conditions. Rockville, MD: Agency for Healthcare Research and Quality; 2002.
18.
Bynum  JP, Meara  E, Chang  C-H, Rhoads  JM, Bronner  KK.  Our Parents, Ourselves: Health Care for an Aging Population: A Report of the Dartmouth Atlas Project. Lebanon, NH: Dartmouth Institute for Health Policy and Clinical Practice; 2016.
19.
Nyweide  DJ, Anthony  DL, Bynum  JP,  et al.  Continuity of care and the risk of preventable hospitalization in older adults.  JAMA Intern Med. 2013;173(20):1879-1885. doi:10.1001/jamainternmed.2013.10059PubMedGoogle ScholarCrossref
20.
McWilliams  JM.  Cost containment and the tale of care coordination.  N Engl J Med. 2016;375(23):2218-2220. doi:10.1056/NEJMp1610821PubMedGoogle ScholarCrossref
21.
Colla  CH, Lewis  VA, Kao  L-S, O’Malley  AJ, Chang  C-H, 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.2016.2827PubMedGoogle ScholarCrossref
22.
Coleman  EA, Smith  JD, Frank  JC, Min  SJ, Parry  C, Kramer  AM.  Preparing patients and caregivers to participate in care delivered across settings: the Care Transitions Intervention.  J Am Geriatr Soc. 2004;52(11):1817-1825. doi:10.1111/j.1532-5415.2004.52504.xPubMedGoogle ScholarCrossref
23.
Coleman  EA, Parry  C, Chalmers  S, Min  SJ.  The care transitions intervention: results of a randomized controlled trial.  Arch Intern Med. 2006;166(17):1822-1828. doi:10.1001/archinte.166.17.1822PubMedGoogle ScholarCrossref
24.
Parrish  MM, O’Malley  K, Adams  RI, Adams  SR, Coleman  EA.  Implementation of the care transitions intervention: sustainability and lessons learned.  Prof Case Manag. 2009;14(6):282-293. doi:10.1097/NCM.0b013e3181c3d380PubMedGoogle ScholarCrossref
25.
Voss  R, Gardner  R, Baier  R, Butterfield  K, Lehrman  S, Gravenstein  S.  The care transitions intervention: translating from efficacy to effectiveness.  Arch Intern Med. 2011;171(14):1232-1237. doi:10.1001/archinternmed.2011.278PubMedGoogle ScholarCrossref
26.
Farrell  T, Tomoaia-Cotisel  A, Scammon  D, Day  J, Day  R, Magill  M.  Care Management: Implications for Medical Practice, Health Policy, and Health Services Research. Rockville, MD: Agency for Healthcare Research and Quality; 2015.
27.
Vogeli  C, Spirt  J, Brand  R,  et al.  Implementing a hybrid approach to select patients for care management: variations across practices.  Am J Manag Care. 2016;22(5):358-365.PubMedGoogle Scholar
28.
Blumenthal  D, Chernof  B, Fulmer  T, Lumpkin  J, Selberg  J.  Caring for high-need, high-cost patients: an urgent priority.  N Engl J Med. 2016;375(10):909-911. doi:10.1056/NEJMp1608511PubMedGoogle ScholarCrossref
29.
Hibbard  JH, Greene  J.  What the evidence shows about patient activation: better health outcomes and care experiences; fewer data on costs.  Health Aff (Millwood). 2013;32(2):207-214. doi:10.1377/hlthaff.2012.1061PubMedGoogle ScholarCrossref
30.
Greene  J, Hibbard  JH, Sacks  R, Overton  V, Parrotta  CD.  When patient activation levels change, health outcomes and costs change, too.  Health Aff (Millwood). 2015;34(3):431-437. doi:10.1377/hlthaff.2014.0452PubMedGoogle ScholarCrossref
31.
Ivey  SL, Shortell  SM, Rodriguez  HP, Wang  YE.  Patient engagement in ACO practices and patient-reported outcomes among adults with co-occurring chronic disease and mental health conditions.  Med Care. 2018;56(7):551-556. doi:10.1097/MLR.0000000000000927PubMedGoogle ScholarCrossref
32.
Lee  KK, Yang  J, Hernandez  AF, Steimle  AE, Go  AS.  Post-discharge follow-up characteristics associated with 30-day readmission after heart failure hospitalization.  Med Care. 2016;54(4):365-372. doi:10.1097/MLR.0000000000000492PubMedGoogle ScholarCrossref
33.
Figueroa  JF, Feyman  Y, Zhou  X, Joynt Maddox  K.  Hospital-level care coordination strategies associated with better patient experience.  BMJ Qual Saf. 2018;27(10):844-851. doi:10.1136/bmjqs-2017-007597PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 1,221
    Original Investigation
    Health Policy
    July 12, 2019

    Association Between Care Management and Outcomes Among Patients With Complex Needs in Medicare Accountable Care Organizations

    Author Affiliations
    • 1The Dartmouth Institute for Health Policy & Clinical Practice, Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire
    • 2The National Bureau of Economic Research, Cambridge, Massachusetts
    • 3Department of Internal Medicine, University of Michigan, Ann Arbor
    JAMA Netw Open. 2019;2(7):e196939. doi:10.1001/jamanetworkopen.2019.6939
    Key Points español 中文 (chinese)

    Question  What is the association between care management and outcomes among patients with complex needs in Medicare accountable care organizations?

    Findings  In this cross-sectional analysis of 1 402 582 US Medicare beneficiaries who had frailty or multiple chronic conditions, beneficiaries attributed to accountable care organizations in the top vs bottom tertile of reported care management and coordination activities did not have statistically different quality, utilization, spending, or health care system interaction measures.

    Meaning  Accountable care organization–reported care management and coordination activities were not associated with improved outcomes among patients with complex needs.

    Abstract

    Importance  People with complex needs account for a disproportionate amount of Medicare spending, partially because of fragmented care delivered across multiple practitioners and settings. Accountable care organization (ACO) contracts give practitioners incentives to improve care coordination to the extent that coordination initiatives reduce total spending or improve quality.

    Objective  To assess the association between ACO-reported care management and coordination activities and quality, utilization, spending, and health care system interactions in older adults with complex needs.

    Design, Setting, and Participants  In this cross-sectional study, survey information on care management and coordination processes from 244 Medicare Shared Savings Program ACOs in the 2017-2018 National Survey of ACOs (of 351 Medicare ACO respondents; response rate, 69%) conducted from July 20, 2017, to February 15, 2018, was linked to 2016 Medicare administrative claims data. Medicare beneficiaries 66 years or older who were defined as having complex needs because of frailty or 2 or more chronic conditions associated with high costs and clinical need were included.

    Exposures  Beneficiary attribution to ACO reporting comprehensive (top tertile) care management and coordination activities.

    Main Outcomes and Measures  All-cause prevention quality indicator admissions, 30-day all-cause readmissions, acute care and critical access hospital admissions, evaluation and management visits in ambulatory settings, inpatient days, emergency department visits, total spending, post–acute care spending, health care contact days, and continuity of care (from Medicare claims).

    Results  Among 1 402 582 Medicare beneficiaries with complex conditions, the mean (SD) age was 78 (8.0) years and 55.1% were female. Compared with beneficiaries assigned to ACOs in the bottom tertile of care management and coordination activities, those assigned to ACOs in the top tertile had identical median prevention quality indicator admissions and 30-day all-cause readmissions (0 per beneficiary across all tertiles), hospitalization and emergency department visits (1.0 per beneficiary in bottom and top tertiles), evaluation and management visits in ambulatory settings (14.0 per beneficiary [interquartile range (IQR), 8.0-21.0] in both tertiles), longer median inpatient days (11.0 [IQR, 4.0-33.0] vs 10.0 [IQR, 4.0-32.0]), higher median annual spending ($14 350 [IQR, $4876-$36 119] vs $14 229 [IQR, $4805-$36 268]), lower median health care contact days (28.0 [IQR, 17.0-44.0] vs 29.0 [IQR, 18.0-45.0]), and lower continuity-of-care index (0.12 [IQR, 0.08-0.20] vs 0.13 [IQR, 0.08-0.21]). Accounting for within-patient correlation, quality, utilization, and spending outcomes among patients with complex needs attributed to ACOs were not statistically different comparing the top vs bottom tertile of care management and coordination activities.

    Conclusions and Relevance  The ACO self-reports of care management and coordination capacity were not associated with differences in spending or measured outcomes for patients with complex needs. Future efforts to care for patients with complex needs should assess whether strategies found to be effective in other settings are being used, and if so, why they fail to meet expectations.

    Introduction

    For older adults who are frail or have multiple chronic conditions (ie, patients with complex needs), fragmentation of care across settings and practitioners can lead to difficulty navigating the health care system, unnecessary spending, and gaps in care quality. Consistent with this challenge, a small proportion of patients (5%) account for more than half of medical spending.1 Accountable care organizations (ACOs) are redesigning care for these patients by building robust care management and care coordination strategies based on the idea that focusing on these patients is likely to yield the greatest cost savings.2 Moreover, care management and coordination activities may represent important strategies for achieving savings and quality standards under ACO programs.3

    Qualitative research has found that many ACO care management and coordination programs are focused on transforming primary care through increased team-based care, strengthening practice-based care management, and developing new roles and activities that span care settings.2 However, the evidence for the effectiveness of care management and coordination use in ACOs to date is mixed. One study3 of a Pioneer ACO evaluated its care management program and found consistent reductions in spending and utilization over time, with a strong association between length of time in the program and results. Another article4 studied unplanned hospitalizations for select chronic conditions and found no association between ACO participation and reduction in hospital use, concluding that care coordination efforts have minimal associations with hospitalization of patients with chronic diseases. However, this work aggregated all ACOs regardless of how comprehensive their care management and coordination activities were, potentially obscuring significant savings in ACOs where care management and coordination was prioritized.

    One of the challenges for the research in this area is defining and measuring what is meant by care management and care coordination activities. The notion of care management and coordination is fraught with disagreement over the importance of individual components and interventions.5 Previous research2 has distinguished care management from care coordination activities in ACOs. Both care management, which refers to programs and systems that aim to help manage patients’ health and medical conditions, and care coordination, which refers to activities to integrate care across practitioners, require organizational investments because the associated services typically cross boundaries of professions and settings. Despite ongoing investment in care management and coordination programs by ACOs,6,7 there is little empirical research examining the heterogeneity across ACOs in these activities or the association of those activities with outcomes.

    By using detailed, survey-based information collected from ACOs participating in the Medicare Shared Savings Program (MSSP), we described and quantified the care management and coordination efforts implemented by ACOs and assessed whether more comprehensive efforts were associated with better outcomes (quality, utilization, and spending) for older adults with complex needs in Medicare.

    Methods

    For this cross-sectional study, we used the 2017-2018 National Survey of ACOs (NSACO) to evaluate care management and coordination strategies reported as being used by MSSP ACOs to provide care to patients with complex needs. We used the Centers for Medicare & Medicaid Services MSSP attribution method to assign beneficiaries to the provider organization with the greatest allowed charges for eligible primary care services.8 This study was reviewed and approved by the institutional review board of Dartmouth College. Written informed consent was obtained from survey respondents, and data were deidentified. The study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.9

    NSACO Wave 4 Survey Data

    We gathered data on MSSPs’ reported care management and coordination activities from the 2017-2018 NSACO. This survey was conducted from July 20, 2017, to February 15, 2018. We identified Medicare ACOs for the survey from Centers for Medicare & Medicaid Services data. The response rate for Medicare ACOs was 69%. The ACO-affiliated leadership personnel (eg, executive directors, vice presidents, chief executive officers, medical directors, chief operating officers, director, and account managers) responded to the survey either online or via a paper survey. The survey included questions related to ACO organizational characteristics and strategies used by ACOs to meet contract goals, including several questions on care management and coordination programs and activities.

    The NSACO contained questions on care management and care coordination strategies relevant to patients with complex needs, including processes and programs for predictive risk stratification and segmentation, chronic care management, active patient involvement in decisions that involved their care and self-management of their condition, and care transition processes and services.10 These topics, drawn from 6 multipart survey questions, resulted in 12 variables for inclusion (eTable 1 in the Supplement). We computed an index score—referred to as care management and coordination index—that measured reported care management and coordination activities for each ACO and their ability to simultaneously implement care management and coordination processes. We assigned each response discrete point values that ranged from 0 for scores of 1 to 3, 1 for scores of 4 to 6, and 2 for scores of 7 to 9 on a 9-point Likert scale (4 questions, with 1 indicating no processes or programs in place and 9 indicating comprehensive processes or programs in place), and we used values of 1 or 0 for binary yes-or-no response options, resulting in a total of 0 to 16 points maximum. The care management and coordination index was similar to a factor analysis–based index and easier to interpret; the correlation between the first factor score and our computed point-based index was 0.94 (eTable 1 in the Supplement). The ACOs were ranked into tertiles based on their level of activity captured by the care management and coordination index. The main exposure variable was whether a beneficiary with complex needs (with frailty or multiple chronic conditions) was attributed to an ACO with a high reported care management and coordination index.

    Patient Population

    We used Parts A and B 2016 Medicare fee-for-service claims to capture care patterns, diagnoses, use, and spending and the beneficiary summary file to capture patient demographic characteristics, enrollment status, and date of death. We used beneficiaries’ zip codes to identify whether they lived in high-poverty census tracts (≥20% of residents below the federal poverty level)11 and to determine hospital referral regions of residence.

    To identify patients with complex needs in Medicare claims, we identified eligible Medicare beneficiaries who were 66 years or older, had full Parts A and B fee-for-service coverage, and were living in the 50 states or the District of Columbia, and we limited our sample to 2 groups: frail, older patients and patients with multimorbidities. Following the claims-based frailty indicators defined by Kim and Schneeweiss12 and Joynt et al,13 we created a cohort of frail, older patients (eTable 2 in the Supplement). We used International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes to identify beneficiaries in the multimorbidity cohort, which included those with at least 2 of a list of 18 chronic conditions14 selected for their chronic nature and association with mortality and costs (eTable 3 in the Supplement).15

    Variable Descriptions

    We reported beneficiary demographic characteristics (age in years, sex, race/ethnicity, and socioeconomic status) and dual eligibility for Medicaid. We identified nursing home residents using Medicare Parts A and B claims if they had eligible place of service and Common Procedural Terminology codes defining use of a nursing facility (eTable 4 in the Supplement).16 For this covariate, we did not distinguish between short-term and long-term stay residents. For each beneficiary, we presented the number of reported hierarchical condition categories in 2016.15

    The primary outcomes of interest included quality of care, health care utilization, spending, and interactions with the health care system for patients with complex needs. We measured quality with annual measures for the number of all-cause prevention quality indicator admissions17 and 30-day all-cause readmissions. We measured health care utilization by examining acute care or critical access hospital admissions, evaluation and management (E&M) visits in ambulatory settings, inpatient days, and emergency department (ED) visits not leading to admission. We measured spending with annual total spending and computed beneficiary total post–acute care spending. Interactions with the health care system were measured with health care contact days, defined as the number of days per year in an inpatient setting or with a practitioner visit, procedure, imaging study, or laboratory test in an ambulatory setting18 and, for patients with at least 4 ambulatory visits, a continuity-of-care index calculated as a Herfindahl-Hirschman index for patients. The continuity-of-care index is higher for patients getting more of their care from fewer practitioners, with a value of 1 if all care is received by a single practitioner.19 We also reported 1-year mortality status among beneficiaries.

    Statistical Analysis

    We compared characteristics among Medicare enrollees with complex needs attributed to ACOs across the care management and coordination index using multivariate regression models. We modeled beneficiary outcomes as functions of the beneficiary-attributed ACO–reported care management and coordination index by tertile, controlling for demographic characteristics, dual eligibility for Medicaid, nursing home residency, frailty, and multiple chronic condition measures. Least squares regression models were used for all outcome measures, and heteroscedasticity-robust SEs were clustered at the ACO level. In sensitivity analyses, we restricted the sample to include beneficiaries with frailty and those with at least 3 of the 18 selected chronic conditions. We also reestimated our initial multivariate regression models and controlled for geography measured by census regions. Statistical tests conducted in this analysis were 2-sided, with P < .05 considered to be statistically significant.

    Results

    The sample consisted of 244 MSSP ACOs with NSACO survey data on care management and coordination processes. A total of 1 402 582 Medicare beneficiaries with complex conditions (mean [SD] age, 78 [8.0] years; 55.1% female) were studied. Characteristics of beneficiaries attributed to the survey respondents (n = 1 402 582) were similar overall to those attributed to nonrespondents (n = 937 557) except for differences in the proportions of females (55.1% among respondents vs 55.5% among nonrespondents) (eTable 5 in the Supplement), individuals with nonwhite race/ethnicity (13.9% vs 20.6%), beneficiaries living in high-poverty census tracts (17.0% vs 18.3%), nursing home residents (29.6% vs 28.5%), and dually eligible beneficiaries (18.1% vs 21.6%).

    The ACOs were grouped into tertiles of the care management and coordination index score (84 in the bottom tertile, 97 in the middle tertile, and 63 in the top tertile) (Table 1), which was skewed to the right and ranged from 2 (lowest intensity) to 16 (highest intensity). The ACOs reported considerable variation in their care management and transition activities, even among those most commonly provided. More than 48 ACOs (76.2%) in the highest care management and coordination index tertile reported that they segmented high-risk patients, dividing them into groups based on common care needs, compared with only 42 (50.0%) in the lowest tertile. Other care management activities were largely similar among ACOs within tertiles, with chronic care management processes being the most comprehensive (median, 8.0 [interquartile range (IQR), 7.0-9.0] for the highest tertile and 6.0 [IQR, 5.0-7.0]) and shared decision-making process being the least (median, 7.0 [IQR, 6.0-8.0] for the highest tertile and 5.0 [IQR, 3.0-6.0] for the lowest tertile on a 9-point scale). Among potential strategies to reduce the risk of readmission for hospitalized patients undergoing a care transition to home or post–acute care, more than 90% of ACOs in the highest tertile of the care management and coordination index reported providing medication reconciliation (58 [92.1%]), transmitting discharge summaries to practitioners accepting care of the patients (58 [92.1%]), having standardized processes in place to ensure timely follow-up with primary and specialty care (61 [96.8%]), and following up via telephone within 72 hours of discharge (60 [95.2%]). In contrast, ACOs reported that the strategy adopted least was in-home follow-up within 72 hours of discharge across each tertile (24 [38.1%] in the highest tertile vs 2 [2.4%] in the lowest tertile). Overall use of patient navigators and care managers for most or all patients varied widely across tertiles (highest tertile: 57 [90.5%]; lowest tertile, 17 [20.2%]), as did the use of care managers or health coaches after discharge (highest tertile, 59 [93.7%]; lowest tertile, 13 [15.5%]).

    Patients with complex needs (1 402 582 overall, with 445 829 in the bottom tertile, 566 633 in the middle tertile, and 390 120 in the top tertile of the care management and coordination index) represented 29.5% of all MSSP-attributed Medicare beneficiaries 66 years or older. There was overlap among categories; more than 14% of the sample with complex needs met criteria for both frailty and 2 or more chronic conditions. The ACO-attributed beneficiaries with complex needs were similar across tertiles of the care management and coordination index with a few exceptions (Table 2). Patients with complex needs had a median number of hierarchical condition categories of 3.0 (IQR, 2.0-5.0) (all tertiles). Comparing patients attributed to ACOs in the top vs bottom tertiles, patients in the top tertile were less likely to have nonwhite race/ethnicity (13.1% vs 15.2%), live in a high-poverty census tract (16.5% vs 16.8%), and have dual eligibility for Medicaid (17.6% vs 18.1%) but were more likely to reside in a nursing home (30.7% vs 29.5%).

    Unadjusted outcomes in patients with complex needs varied slightly when examined across care management and coordination index tertiles. Compared with beneficiaries assigned to ACOs in the bottom tertile of care management and coordination activities, those assigned to ACOs in the top tertile had identical median prevention quality indicator admissions and 30-day all-cause readmissions (0 per beneficiary across all tertiles), hospitalization and emergency department visits (1.0 per beneficiary in bottom and top tertiles), evaluation and management visits in ambulatory settings (14.0 per beneficiary [interquartile range (IQR), 8.0-21.0] in both tertiles), longer median inpatient days (11.0 [IQR, 4.0-33.0] vs 10.0 [IQR, 4.0-32.0]), higher median annual spending ($14 350 [IQR, $4876-$36 119] vs $14 229 [IQR, $4805-$36 268]), lower median health care contact days (28.0 [IQR, 17.0-44.0] vs 29.0 [IQR, 18.0-45.0]), and lower continuity-of-care index (0.12 [IQR, 0.08-0.20] vs 0.13 [IQR, 0.08-0.21]) (Table 2). The MSSPs with the highest reported care management and coordination activities had a higher mean mortality rate than those in the bottom tertile (12.3% vs 11.9%).

    In multivariate regression analyses, quality of care and health care utilization outcomes were measured per 100 beneficiaries for ease of presentation. We did not find statistically different prevention quality indicator admissions (−0.03; 95% CI, −0.91 to 0.85; P = .95) and 30-day readmissions (−0.02; 95% CI, −0.52 to 0.48; P = .94) between patients with complex needs attributed to MSSPs in the top care management and coordination index tertile and those attributed to MSSPs in the bottom tertile (Table 3). In patients in organizations with the highest reported care management and coordination activities, changes were not significant for E&M visits (−13.5; 95% CI, −116.3 to 89.4; P = .80), acute care or critical access hospital admissions (−0.3; 95% CI, −3.1 to 2.4; P = .81), inpatient days (−0.1; 95% CI, −1.0 to 0.9; P = .89), and number of ED visits (−0.2; 95% CI, −6.20 to 5.84; P = .95) compared with patients in MSSPs reporting less care management and coordination activities. We did not find statistically significant differences in spending (−$505; 95% CI, −$1756 to $746; P = .43 for change in total spending; −$192; 95% CI, −$731 to $347; P = .48 for change in post–acute care spending) for patients with complex needs in MSSPs in the top care management and coordination tertile compared with those in the bottom tertile. We estimated no difference in continuity-of-care scores (−0.004; 95% CI, −0.01 to 0.004; P = .33) or health care system contact days (−0.7; 95% CI, −2.4 to 1.0; P = .42) for patients attributed to MSSPs with the highest reported care management and coordination activities compared with those in MSSPs reporting less activity. Results were unchanged after restricting the sample to include beneficiaries with 3 or more chronic conditions and after adjusting for geography (eTable 6 in the Supplement).

    Discussion

    To measure whether ACO efforts to manage and coordinate care are associated with health care utilization and spending outcomes in groups likely to benefit from such services, we used data from the NSACO to form a care management and coordination index that plausibly targets older patients with complex health needs. We found wide variation in reported care management and coordination activities across MSSP participants. For example, when index scores were ranked by the level of care management activity reported, we observed large differences in the use of evidence-based approaches to care transitions for hospitalized patients. In multivariate analyses of patient-level measures for 244 ACO respondents, we found that high levels of care management and coordination intensity were not associated with significant differences in hospitalizations, inpatient days, E&M visits in ambulatory settings, spending, and health care system contact days despite large samples of beneficiaries in each group of ACOs.

    Our results are in line with recent ACO literature4,20 noting the limitations of care management and coordination. This earlier work4,20 on ACOs used indirect assessment of care management and coordination activities. Our study is unique because we linked health care organization information on management and coordination activities with claims-based performance data, permitting study of whether and how measures of care use and quality are associated with an ACO’s care management and coordination efforts. Prior studies4,20 that combined many ACOs implementing limited management and coordination efforts with those intensively pursuing evidence-based strategies reported a lack of measured benefit even if the strategies used by some ACOs were effective. In articles4,21 that examined ACO performance compared with non-ACO performance in clinically vulnerable populations, some savings per beneficiary were found in the clinically vulnerable groups; however, how the ACOs achieved those savings was undefined.

    The evidence base around care transition interventions, such as home visits after discharge, use of care navigators, and other features included in the care management and coordination index, is substantial, with randomized clinical trials22-25 supporting their efficacy at reducing hospitalizations and spending. Our work can be viewed as augmenting the evidence on how such strategies may play out in the community. When focusing on whether care management and coordination efforts by ACOs were associated with outcomes in older adults with complex needs, we did not find any statistically significant associations between care management and coordination and measures of care quality, utilization, spending, or interactions with the health care system. Findings in the care transition interventions literature suggest that a reduction in cost and utilization outcomes measured in our study may be desirable except for E&M visits and continuity of care, for which a favorable direction of the association was less clear.

    We believe that the first step to better understanding the promise or limitations of care management and coordination is a better understanding of what health care organizations are implementing to address care needs for patients with complex needs. Given ample evidence to support the care management and coordination strategies that we queried13,26-32 and given external factors, such as the hospital readmission reduction program, which might add incentives to provider organizations to implement strategies for patients most at risk of hospitalization, we found a low reported use of evidence-based strategies designed to reduce readmission among ACOs in the bottom tertile of care management and coordination activities.

    This study contributes to an important research gap—assessing the association between the self-reported care coordination activity of ACOs with performance for vulnerable Medicare beneficiaries. Mixed-methods work appears to be needed for care coordination activities to help better explain why evidence-based strategies do not generate the expected changes in spending, utilization, and quality.

    Although our study is not definitive and includes several limitations, our estimates on various outcomes are indistinguishable from zero, indicating a call to invest in these areas with caution and to track local results in careful evaluations. Our results suggest that organizations should consider the effectiveness of investing heavily in care coordination activities that are difficult to implement. These results have the potential to inform practitioners and policy makers about ACO activities under the umbrella of care coordination to identify new areas of focus or improvement. As ACOs focus on expanding care coordination and management efforts, it is important to continually evaluate the effectiveness of those programs for different patient populations.

    Limitations

    Our results should be interpreted considering several limitations. Measures of care management and coordination were self-reported by ACO leaders who may have incomplete information on care management activity at clinical front lines regarding the quantity and quality of services provided. We were unable to measure and control for variability in the case loads and for care management and coordination program delivery and duration at the ACO level.

    Furthermore, there may be substantial differences in execution of reported activities that may result in important heterogeneity in the effectiveness of these activities. This is a challenge for any innovation being diffused to a heterogeneous set of practices with heterogeneous capabilities. In this article, we aimed to measure whether these efforts had an association. Although we did not find significant differences in cost or utilization outcomes and were unable to measure patient experience and physician satisfaction, care management and coordination programs might improve these outcomes.33 Although our study focused on the largest Medicare ACO program, the MSSP, our results may not be generalizable to other models.

    The most important limitation of our study is its cross-sectional nature, which does not allow for causal interpretation. Because we used claims-based methods, residual, unmeasured differences between the comparison groups were possible. If ACOs with high levels of care management had sicker patients not captured by our claims-based measures, illness would have not been adequately adjusted for and an underlying association between care management and coordination and outcomes would not have been observed. Analyses using longitudinal data would allow the evaluation of activity duration and intensity and the assessment of the association of length of exposure to care management and coordination with outcomes. Although we used the most recent data available and closer to the time of our survey, we recognize that the claims predate our survey data. The lack of statistical differences in beneficiary outcomes across ACOs could be justified if ACOs had statistically indistinguishable levels of care management and coordination activities in 2016. However, it seems unlikely that ACOs with comprehensive care management and coordination activities began those efforts only after 2016. Another limitation is that we categorized MSSPs using an index of distinct care management and coordination activities that we hypothesized could influence care quality, outcomes, and costs for patients with complex needs, but the measure does not distinguish the individual components of care management activity. We did, however, present the underlying distribution of components contributing to the index.

    Although our survey questions asked about the basic structural components that facilitate care management, they were not exhaustive. Additional components of care management and coordination occurring in conjunction with or outside the health care system (social and economic support, transportation, and food and housing)6 are important in caring for the whole patient across settings. Also, ACOs with high reported care management and coordination activities had fewer participants with nonwhite race/ethnicity compared with those reporting less activity and non–survey respondents, suggesting that the survey may tell us little about organizations that serve nonwhite populations. Although our study is not definitive, our estimates on various outcomes suggest the potential need for caution in future investments in care management and coordination.

    Conclusions

    In this study, the ACO self-reports of care management and coordination capacity were not associated with differences in spending or measured outcomes for patients with complex needs. Future efforts to care for patients with complex needs should assess whether strategies found to be effective in other settings are being used, and if so, why they fail to meet expectations.

    Back to top
    Article Information

    Accepted for Publication: May 21, 2019.

    Published: July 12, 2019. doi:10.1001/jamanetworkopen.2019.6939

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Ouayogodé MH et al. JAMA Network Open.

    Corresponding Author: Mariétou H. Ouayogodé, PhD, The Dartmouth Institute for Health Policy & Clinical Practice, Geisel School of Medicine at Dartmouth, One Medical Center Drive, Williamson Translational Research Bldg, Level 5, Lebanon, NH 03756 (marietou.h.l.ouayogode@dartmouth.edu).

    Author Contributions: Dr Ouayogodé 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.

    Concept and design: Ouayogodé, Meara, Bynum, Colla.

    Acquisition, analysis, or interpretation of data: Ouayogodé, Mainor, Meara, Bynum.

    Drafting of the manuscript: All authors.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Ouayogodé, Meara, Bynum, Colla.

    Obtained funding: Meara, Colla.

    Administrative, technical, or material support: Ouayogodé, Mainor, Meara, Colla.

    Supervision: All authors.

    Conflict of Interest Disclosures: Dr Ouayogodé reported receiving grants from the Six Foundation Collaborative (including The Commonwealth Fund, The John A. Hartford Foundation, the Peterson Center on Healthcare, the Robert Wood Johnson Foundation, The SCAN Foundation, and the Milbank Memorial Fund) during the conduct of the study and having work supported by grants to Dartmouth College from the Agency for Healthcare Research and Quality. Mr Mainor reported having work supported by grants to Dartmouth College from the Six Foundation Collaborative during the conduct of the study. Dr Meara reported receiving grants from the National Institutes of Health, The Commonwealth Fund, the Six Foundation Collaborative, and the Agency for Healthcare Research and Quality during the conduct of the study and having work supported by grants to Dartmouth College from the National Institutes of Health and the Agency for Healthcare Research and Quality. Dr Bynum reported having work supported by grants to Dartmouth College from the National Institutes of Health. Dr Colla reported receiving grants from the Six Foundation Collaborative, the California Healthcare Foundation, the National Institutes of Health, and the Agency for Healthcare Research and Quality during the conduct of the study.

    Funding/Support: Wave 4 of the National Survey of Accountable Care Organizations was supported by grant 20160616 from The Commonwealth Fund, grant R01MH109531 from the National Institute of Mental Health of the National Institutes of Health, and grant 20249 from the California Health Care Foundation. This work was supported by the Six Foundation Collaborative: grant 20171072 from The Commonwealth Fund, grant 18011 from the Peterson Center on Healthcare, grant 74883 from the Robert Wood Johnson Foundation, grant 17-013 from The SCAN Foundation, The John A. Hartford Foundation, the Millbank Memorial Fund, and grant 1U19HS024075 from the Agency for Healthcare Research and Quality's Comparative Health System Performance Initiative.

    Role of the Funder/Sponsor: The funding sources reviewed the manuscript but had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation or approval of the manuscript; and the decision to submit the manuscript for publication.

    Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

    References
    2.
    Lewis  VA, Schoenherr  K, Fraze  T, Cunningham  A.  Clinical coordination in accountable care organizations: a qualitative study.  Health Care Manage Rev. 2019;44(2):127-136. doi:10.1097/HMR.0000000000000141PubMedGoogle ScholarCrossref
    3.
    Hsu  J, Price  M, Vogeli  C,  et al.  Bending the spending curve by altering care delivery patterns: the role of care management within a pioneer ACO.  Health Aff (Millwood). 2017;36(5):876-884. doi:10.1377/hlthaff.2016.0922PubMedGoogle ScholarCrossref
    4.
    McWilliams  JM, Chernew  ME, Landon  BE.  Medicare ACO program savings not tied to preventable hospitalizations or concentrated among high-risk patients.  Health Aff (Millwood). 2017;36(12):2085-2093. doi:10.1377/hlthaff.2017.0814PubMedGoogle ScholarCrossref
    5.
    McDonald  KM, Sundaram  V, Bravata  DM,  et al.  Closing the Quality Gap: A Critical Analysis of Quality Improvement Strategies. Rockville, MD: Agency for Healthcare Research and Quality; 2007. Report 04(07)-0051-7.
    6.
    Donelan  K, Barreto  EA, Michael  CU, Nordby  P, Smith  M, Metlay  JP.  Variability in care management programs in Medicare ACOs: a survey of medical directors.  J Gen Intern Med. 2018;33(12):2043-2045. doi:10.1007/s11606-018-4609-1PubMedGoogle ScholarCrossref
    7.
    O’Malley  AS, Rich  EC, Sarwar  R,  et al. How accountable care organizations use population segmentation to care for high-need, high-cost patients.  Issue Brief (Commonw Fund). 2019;2019:1-17. https://www.commonwealthfund.org/publications/issue-briefs/2019/jan/how-acos-use-segmentation-high-need-high-cost. Accessed Jube 17, 2019.
    8.
    Centers for Medicare & Medicaid Services.  Medicare Shared Savings Program Shared Savings and Losses and Assignment Methodology Specifications: Applicable Beginning Performance Year 2016. Woodlawn, MD: Centers for Medicare & Medicaid Services; 2016.
    9.
    Equator Network. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. 2015. http://www.equator-network.org/reporting-guidelines/strobe/. Accessed April 30, 2019.
    10.
    Peck  K, Usadi  B, Mainor  A, Newton  H, Meara  E.  How ACOs Are Caring for People with Complex Needs. New York, NY: Commonwealth Fund; 2018.
    11.
    US Census Bureau. How the Census Bureau measures poverty. 2016. https://www.census.gov/topics/income-poverty/poverty/about.html. Accessed April 7, 2016.
    12.
    Kim  DH, Schneeweiss  S.  Measuring frailty using claims data for pharmacoepidemiologic studies of mortality in older adults: evidence and recommendations.  Pharmacoepidemiol Drug Saf. 2014;23(9):891-901. doi:10.1002/pds.3674PubMedGoogle ScholarCrossref
    13.
    Joynt  KE, Figueroa  JF, Beaulieu  N, Wild  RC, Orav  EJ, Jha  AK.  Segmenting high-cost Medicare patients into potentially actionable cohorts.  Healthc (Amst). 2017;5(1-2):62-67. doi:10.1016/j.hjdsi.2016.11.002PubMedGoogle ScholarCrossref
    14.
    Bynum  JPW, Chang  CH, Austin  A, Carmichael  D, Meara  E.  Outcomes in older adults with multimorbidity associated with predominant provider of care specialty.  J Am Geriatr Soc. 2017;65(9):1916-1923. doi:10.1111/jgs.14882PubMedGoogle ScholarCrossref
    15.
    Centers for Medicare & Medicaid Services. Risk adjustment. 2018. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors.html.Accessed September 28, 2019.
    16.
    Centers for Medicare & Medicaid Services.  Nursing Facility Services (Codes 99304-99318). Woodlawn, MD: Centers for Medicare & Medicaid Services; 2006.
    17.
    Agency for Healthcare Research and Quality.  AHRQ Quality Indicators: Guide to Prevention Quality Indicators: Hospital Admission for Ambulatory Care Sensitive Conditions. Rockville, MD: Agency for Healthcare Research and Quality; 2002.
    18.
    Bynum  JP, Meara  E, Chang  C-H, Rhoads  JM, Bronner  KK.  Our Parents, Ourselves: Health Care for an Aging Population: A Report of the Dartmouth Atlas Project. Lebanon, NH: Dartmouth Institute for Health Policy and Clinical Practice; 2016.
    19.
    Nyweide  DJ, Anthony  DL, Bynum  JP,  et al.  Continuity of care and the risk of preventable hospitalization in older adults.  JAMA Intern Med. 2013;173(20):1879-1885. doi:10.1001/jamainternmed.2013.10059PubMedGoogle ScholarCrossref
    20.
    McWilliams  JM.  Cost containment and the tale of care coordination.  N Engl J Med. 2016;375(23):2218-2220. doi:10.1056/NEJMp1610821PubMedGoogle ScholarCrossref
    21.
    Colla  CH, Lewis  VA, Kao  L-S, O’Malley  AJ, Chang  C-H, 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.2016.2827PubMedGoogle ScholarCrossref
    22.
    Coleman  EA, Smith  JD, Frank  JC, Min  SJ, Parry  C, Kramer  AM.  Preparing patients and caregivers to participate in care delivered across settings: the Care Transitions Intervention.  J Am Geriatr Soc. 2004;52(11):1817-1825. doi:10.1111/j.1532-5415.2004.52504.xPubMedGoogle ScholarCrossref
    23.
    Coleman  EA, Parry  C, Chalmers  S, Min  SJ.  The care transitions intervention: results of a randomized controlled trial.  Arch Intern Med. 2006;166(17):1822-1828. doi:10.1001/archinte.166.17.1822PubMedGoogle ScholarCrossref
    24.
    Parrish  MM, O’Malley  K, Adams  RI, Adams  SR, Coleman  EA.  Implementation of the care transitions intervention: sustainability and lessons learned.  Prof Case Manag. 2009;14(6):282-293. doi:10.1097/NCM.0b013e3181c3d380PubMedGoogle ScholarCrossref
    25.
    Voss  R, Gardner  R, Baier  R, Butterfield  K, Lehrman  S, Gravenstein  S.  The care transitions intervention: translating from efficacy to effectiveness.  Arch Intern Med. 2011;171(14):1232-1237. doi:10.1001/archinternmed.2011.278PubMedGoogle ScholarCrossref
    26.
    Farrell  T, Tomoaia-Cotisel  A, Scammon  D, Day  J, Day  R, Magill  M.  Care Management: Implications for Medical Practice, Health Policy, and Health Services Research. Rockville, MD: Agency for Healthcare Research and Quality; 2015.
    27.
    Vogeli  C, Spirt  J, Brand  R,  et al.  Implementing a hybrid approach to select patients for care management: variations across practices.  Am J Manag Care. 2016;22(5):358-365.PubMedGoogle Scholar
    28.
    Blumenthal  D, Chernof  B, Fulmer  T, Lumpkin  J, Selberg  J.  Caring for high-need, high-cost patients: an urgent priority.  N Engl J Med. 2016;375(10):909-911. doi:10.1056/NEJMp1608511PubMedGoogle ScholarCrossref
    29.
    Hibbard  JH, Greene  J.  What the evidence shows about patient activation: better health outcomes and care experiences; fewer data on costs.  Health Aff (Millwood). 2013;32(2):207-214. doi:10.1377/hlthaff.2012.1061PubMedGoogle ScholarCrossref
    30.
    Greene  J, Hibbard  JH, Sacks  R, Overton  V, Parrotta  CD.  When patient activation levels change, health outcomes and costs change, too.  Health Aff (Millwood). 2015;34(3):431-437. doi:10.1377/hlthaff.2014.0452PubMedGoogle ScholarCrossref
    31.
    Ivey  SL, Shortell  SM, Rodriguez  HP, Wang  YE.  Patient engagement in ACO practices and patient-reported outcomes among adults with co-occurring chronic disease and mental health conditions.  Med Care. 2018;56(7):551-556. doi:10.1097/MLR.0000000000000927PubMedGoogle ScholarCrossref
    32.
    Lee  KK, Yang  J, Hernandez  AF, Steimle  AE, Go  AS.  Post-discharge follow-up characteristics associated with 30-day readmission after heart failure hospitalization.  Med Care. 2016;54(4):365-372. doi:10.1097/MLR.0000000000000492PubMedGoogle ScholarCrossref
    33.
    Figueroa  JF, Feyman  Y, Zhou  X, Joynt Maddox  K.  Hospital-level care coordination strategies associated with better patient experience.  BMJ Qual Saf. 2018;27(10):844-851. doi:10.1136/bmjqs-2017-007597PubMedGoogle ScholarCrossref
    ×