Association Between Patient Cognitive and Functional Status and Medicare Total Annual Cost of Care: Implications for Value-Based Payment | Health Disparities | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Table 1.  Mean Annual Costs and Prevalence of Patient Risk Factors of Fee-for-Service Medicare Beneficiaries
Mean Annual Costs and Prevalence of Patient Risk Factors of Fee-for-Service Medicare Beneficiaries
Table 2.  The Association of Patient Cognitive and Functional Status With TACC in the Medicare Fee-for-Service Patient Population
The Association of Patient Cognitive and Functional Status With TACC in the Medicare Fee-for-Service Patient Population
Table 3.  The Association of Patient Local Area Supply and Economic Characteristics With TACC in the Medicare Fee-for-Service Patient Population
The Association of Patient Local Area Supply and Economic Characteristics With TACC in the Medicare Fee-for-Service Patient Population
Table 4.  Previously Unobserved Medicare Patient Characteristics Added to Standard the CMS-HCC Cost Modela
Previously Unobserved Medicare Patient Characteristics Added to Standard the CMS-HCC Cost Modela
1.
Buntin  MB, Ayanian  JZ.  Social risk factors and equity in Medicare payment.  N Engl J Med. 2017;376(6):507-510. doi:10.1056/NEJMp1700081PubMedGoogle ScholarCrossref
2.
Joynt  KE, De Lew  N, Sheingold  SH, Conway  PH, Goodrich  K, Epstein  AM.  Should Medicare value-based purchasing take social risk into account?  N Engl J Med. 2017;376(6):510-513. doi:10.1056/NEJMp1616278PubMedGoogle ScholarCrossref
3.
Centers for Medicare and Medicaid Services. The Merit-Based Incentive Payment System: MIPS Scoring Methodology Overview. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/MACRA-MIPS-and-APMs/MIPS-Scoring-Methodology-slide-deck.pdf. Accessed March 16, 2018.
4.
Joynt Maddox  KE.  Social and behavioral determinants of spending.  JAMA Intern Med. 2017;177(10):1431-1432. doi:10.1001/jamainternmed.2017.3325PubMedGoogle ScholarCrossref
5.
Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation. Report to Congress: social risk factors and performance under Medicare’s value-based payment programs. 2016. https://aspe.hhs.gov/pdf-report/report-congress-social-risk-factors-and-performance-under-medicares-value-based-purchasing-programs. Accessed March 27, 2018.
6.
Medicare Payment Advisory Commission. Moving Beyond the Merit-Based Incentive Payment System. Rep to Congr Medicare Paym Policy. 2018:445-468. http://www.medpac.gov/docs/default-source/reports/mar18_medpac_entirereport_sec.pdf?sfvrsn=0. Accessed June 7, 2018.
7.
Centers for Medicare and Medicaid Services. The Merit-Based Incentive Payment System: Resource Use Performance Category. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/MACRA-MIPS-and-APMs/Resource-Use-Performance-Category-slide-deck.pdf. Accessed March 16, 2018.
8.
Centers for Medicare and Medicaid Services. 2018 MIPS Cost Performance Category Fact Sheet. Qual Paym Progr. 2018:1-7. https://www.cms.gov/Medicare/Quality-Payment-Program/Resource-Library/2018-Cost-Performance-Category-Fact-Sheet.pdf. Accessed June 11, 2018.
9.
Centers for Medicare and Medicaid Services. 2016 Measure Information About the Per Capita Costs for All Attributed Beneficiaries Measure, Calculated for the 2018 Value-Based Payment Modifier Program. 2017:1-8. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/PhysicianFeedbackProgram/Downloads/2016-TPCC-MIF.pdf. Accessed March 16, 2018.
10.
Lledó  R, Martín  E, Jiménez  C,  et al.  Characteristics of elderly inpatients at high risk of needing supportive social and health care services.  Eur J Epidemiol. 1997;13(8):903-907. doi:10.1023/A:1007420911766PubMedGoogle ScholarCrossref
11.
Fong  TG, Jones  RN, Marcantonio  ER,  et al.  Adverse outcomes after hospitalization and delirium in persons with Alzheimer disease.  Ann Intern Med. 2012;156(12):848-856, W296. doi:10.7326/0003-4819-156-12-201206190-00005PubMedGoogle ScholarCrossref
12.
Linden  M, Horgas  AL, Gilberg  R, Steinhagen-Thiessen  E.  Predicting health care utilization in the very old: the role of physical health, mental health, attitudinal and social factors.  J Aging Health. 1997;9(1):3-27. doi:10.1177/089826439700900101PubMedGoogle ScholarCrossref
13.
McGinty  EE, Sridhara  S.  Potentially preventable medical hospitalizations among Maryland residents with mental illness, 2005-2010.  Psychiatr Serv. 2014;65(7):951-953. doi:10.1176/appi.ps.201300323PubMedGoogle ScholarCrossref
14.
Benzer  JK, Sullivan  JL, Williams  S, Burgess  JF.  One-year cost implications of using mental health care after discharge from a general medical hospitalization.  Psychiatr Serv. 2012;63(7):672-678. doi:10.1176/appi.ps.201100457PubMedGoogle ScholarCrossref
15.
Greysen  SR, Stijacic Cenzer  I, Boscardin  WJ, Covinsky  KE.  Functional impairment: an unmeasured marker of Medicare costs for postacute care of older adults.  J Am Geriatr Soc. 2017;65(9):1996-2002. doi:10.1111/jgs.14955PubMedGoogle ScholarCrossref
16.
McInerney  M, Mellor  JM.  Recessions and seniors’ health, health behaviors, and healthcare use: analysis of the Medicare Current Beneficiary Survey.  J Health Econ. 2012;31(5):744-751. doi:10.1016/j.jhealeco.2012.06.002PubMedGoogle ScholarCrossref
17.
Maeda  JLK, Henke  RM, Marder  WD, Karaca  Z, Friedman  BS, Wong  HS.  Association between the unemployment rate and inpatient cost per discharge by payer in the United States, 2005-2010.  BMC Health Serv Res. 2014;14(1):378. doi:10.1186/1472-6963-14-378PubMedGoogle ScholarCrossref
18.
Kind  AJH, Jencks  S, Brock  J,  et al.  Neighborhood socioeconomic disadvantage and 30-day rehospitalization: a retrospective cohort study.  Ann Intern Med. 2014;161(11):765-774. doi:10.7326/M13-2946PubMedGoogle ScholarCrossref
19.
Ash  AS, Mick  EO, Ellis  RP, Kiefe  CI, Allison  JJ, Clark  MA.  Social determinants of health in managed care payment formulas.  JAMA Intern Med. 2017;177(10):1424-1430. doi:10.1001/jamainternmed.2017.3317PubMedGoogle ScholarCrossref
20.
Nguyen  OK, Makam  AN, Halm  EA.  National use of safety-net clinics for primary care among adults with non-Medicaid insurance in the United States.  PLoS One. 2016;11(3):e0151610. doi:10.1371/journal.pone.0151610PubMedGoogle ScholarCrossref
21.
Gilman  M, Adams  EK, Hockenberry  JM, Milstein  AS, Wilson  IB, Becker  ER.  Safety-net hospitals more likely than other hospitals to fare poorly under Medicare’s value-based purchasing.  Health Aff (Millwood). 2015;34(3):398-405. doi:10.1377/hlthaff.2014.1059PubMedGoogle ScholarCrossref
22.
Chokshi  DA, Chang  JE, Wilson  RM.  Health reform and the changing safety net in the United States.  N Engl J Med. 2016;375(18):1790-1796. doi:10.1056/NEJMhpr1608578PubMedGoogle ScholarCrossref
23.
Andrulis  DP, Siddiqui  NJ.  Health reform holds both risks and rewards for safety-net providers and racially and ethnically diverse patients.  Health Aff (Millwood). 2011;30(10):1830-1836. doi:10.1377/hlthaff.2011.0661PubMedGoogle ScholarCrossref
24.
Center for Medicare and Medicaid Services. Technical Documentation for the Medicare Current Beneficiary Survey. In: Medicare Current Beneficiary Survey Data Tables. https://www.cms.gov/Research-Statistics-Data-and-Systems/Research/MCBS/Downloads/OLD/HHC2003appendixA.pdf. Accessed August 7, 2018.
25.
Health Resources and Services Administration. Area Health Resources Files (AHRF). Rockville, MD; 2014. https://datawarehouse.hrsa.gov/topics/ahrf.aspx.
26.
Chronic Conditions Data Warehouse. Condition Categories: 27 Chronic Condition Algorithms. 2016. https://www.ccwdata.org/web/guest/condition-categories. Accessed April 28, 2016.
27.
Pope  GC, Kautter  J, Ellis  RP,  et al.  Risk adjustment of Medicare capitation payments using the CMS-HCC model.  Health Care Financ Rev. 2004;25(4):119-141.PubMedGoogle Scholar
28.
Ivanova  JI, Mytelka  DS, Duh  MS,  et al.  Evaluating the utility of existing patient-reported outcome scales in novel patient populations with pancreatic cancer, lung cancer, and myeloproliferative neoplasms using Medicare current beneficiary survey data.  Patient. 2013;6(3):189-200. doi:10.1007/s40271-013-0018-zPubMedGoogle ScholarCrossref
29.
Meddings  J, Reichert  H, Smith  SN,  et al.  The impact of disability and social determinants of health on condition-specific readmissions beyond Medicare risk adjustments: a cohort study.  J Gen Intern Med. 2017;32(1):71-80. doi:10.1007/s11606-016-3869-xPubMedGoogle ScholarCrossref
30.
Johnston  L, Williamson  S. What Was the US GDP Then? Measuring Worth. 2017. https://www.measuringworth.com/datasets/usgdp/. Accessed March 16, 2018.
31.
Centers for Medicare and Medicaid Services. Chapter 9—Rural Health Clinics/Federally Qualified Health Centers. Medicare Claims Process Man. 2015:1-38. https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/clm104c09.pdf. Accessed June 14, 2018.
32.
Ash  AS, Ellis  RP, Pope  GC,  et al.  Using diagnoses to describe populations and predict costs.  Health Care Financ Rev. 2000;21(3):7-28.PubMedGoogle Scholar
33.
Davydow  DS, Hough  CL, Zivin  K, Langa  KM, Katon  WJ.  Depression and risk of hospitalization for pneumonia in a cohort study of older Americans.  J Psychosom Res. 2014;77(6):528-534. doi:10.1016/j.jpsychores.2014.08.002PubMedGoogle ScholarCrossref
34.
Sherwood  A, Blumenthal  JA, Trivedi  R,  et al.  Relationship of depression to death or hospitalization in patients with heart failure.  Arch Intern Med. 2007;167(4):367-373. doi:10.1001/archinte.167.4.367PubMedGoogle ScholarCrossref
35.
Rutledge  T, Reis  VA, Linke  SE, Greenberg  BH, Mills  PJ.  Depression in heart failure a meta-analytic review of prevalence, intervention effects, and associations with clinical outcomes.  J Am Coll Cardiol. 2006;48(8):1527-1537. doi:10.1016/j.jacc.2006.06.055PubMedGoogle ScholarCrossref
36.
Daiello  LA, Gardner  R, Epstein-Lubow  G, Butterfield  K, Gravenstein  S.  Association of dementia with early rehospitalization among Medicare beneficiaries.  Arch Gerontol Geriatr. 2014;59(1):162-168. doi:10.1016/j.archger.2014.02.010PubMedGoogle ScholarCrossref
37.
Schüssler-Fiorenza Rose  SM, Stineman  MG, Pan  Q,  et al.  Potentially avoidable hospitalizations among people at different activity of daily living limitation stages.  Health Serv Res. 2017;52(1):132-155. doi:10.1111/1475-6773.12484PubMedGoogle ScholarCrossref
38.
Barnett  ML, Hsu  J, McWilliams  JM.  Patient characteristics and differences in hospital readmission rates.  JAMA Intern Med. 2015;175(11):1803-1812. doi:10.1001/jamainternmed.2015.4660PubMedGoogle ScholarCrossref
39.
Greysen  SR, Stijacic Cenzer  I, Auerbach  AD, Covinsky  KE.  Functional impairment and hospital readmission in Medicare seniors.  JAMA Intern Med. 2015;175(4):559-565. doi:10.1001/jamainternmed.2014.7756PubMedGoogle ScholarCrossref
40.
Kim  DH, Schneeweiss  S, Glynn  RJ, Lipsitz  LA, Rockwood  K, Avorn  J.  Measuring frailty in Medicare data: development and validation of a claims-based frailty index.  J Gerontol A Biol Sci Med Sci. 2017;(December). doi:10.1093/gerona/glx229PubMedGoogle Scholar
41.
Segal  JB, Chang  H-Y, Du  Y, Walston  JD, Carlson  MC, Varadhan  R.  Development of a claims-based frailty indicator anchored to a well-established frailty phenotype.  Med Care. 2017;55(7):716-722. doi:10.1097/MLR.0000000000000729PubMedGoogle ScholarCrossref
42.
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
43.
United States Government Accountability Office. Medicare Advantage: Changes Improved Accuracy of Risk Adjustment for Certain Beneficiaries. 2011. https://www.gao.gov/assets/590/587519.pdf. Accessed March 30, 2018.
44.
Centers for Medicare and Medicaid Services. Note To: All Medicare Advantage Organizations, Prescription Drug Plan Sponsors, and Other Interested Parties. 2013. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Downloads/Announcement2014.pdf. Accessed March 30, 2018.
45.
Alzheimer’s Association. Re: Advance Notice of Methodological Changes for Calendar Year (CY) 2017 for Medicare Advantage (MA) Capitation Rates, Part C and Part D Payment Policies and 2017 Call Letter. 2016. http://act.alz.org/site/DocServer/Alzheimer_s_Association_Comments_on_2017_Advance_Notice_.pdf?docID=51379. Accessed March 30, 2018.
46.
Centers for Medicare and Medicaid Services. Home Health Value-Based Purchasing Model. https://innovation.cms.gov/initiatives/home-health-value-based-purchasing-model. Accessed March 30, 2018.
47.
Centers for Medicare and Medicaid Services. The Skilled Nursing Facility Value-Based Purchasing Program. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/Other-VBPs/SNF-VBP.html. Accessed March 30, 2018.
48.
Moseley  C, Shen  J, Cochran  C.  Mental health services availability and admission of the seriously mentally ill from the emergency department.  J Health Hum Serv Adm. 2008;31(3):292-308.PubMedGoogle Scholar
49.
Chronic Conditions Data Warehouse. Cognitive and Mental Health Conditions–Trends: Medicare Chronic Cond Charts. 2018. https://www.ccwdata.org/documents/10280/28579486/b2-cog-mh-2006-2015.jpg. Accessed June 6, 2018.
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
    Original Investigation
    November 2018

    Association Between Patient Cognitive and Functional Status and Medicare Total Annual Cost of Care: Implications for Value-Based Payment

    Author Affiliations
    • 1Department of Health Management and Policy, Center for Outcomes Research, College for Public Health and Social Justice, St Louis University, St Louis, Missouri
    • 2Department of Health Management and Policy, University of Kentucky, Lexington
    • 3Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, Georgia
    • 4Cardiovascular Division, Washington University School of Medicine in St Louis, St Louis, Missouri
    JAMA Intern Med. 2018;178(11):1489-1497. doi:10.1001/jamainternmed.2018.4143
    Key Points

    Question  Are patient cognitive and functional status and local area health care supply and economic conditions associated with outpatient clinicians’ performance on treated Medicare patients’ total annual cost of care (TACC)?

    Findings  This observational study of data from the Medicare Current Beneficiary Survey and Area Health Resources File found that patient depression, dementia, limitations in activities of daily living, and residence in areas of mental health care shortage or high unemployment were associated with substantially higher TACC, after applying standard Medicare risk adjustment methods. In a model adding these factors to risk adjustment, the TACC performance of safety-net clinicians was improved.

    Meaning  Assessing outpatient clinicians for the TACC of treated Medicare patients without considering other patient factors may inappropriately penalize safety-net clinicians who care for vulnerable patients.

    Abstract

    Importance  Medicare is moving toward value-based payment. The Merit-Based Incentive Payment System (MIPS) program judges outpatient clinicians’ performance on a measure of annual Medicare spending. However, this measure may disadvantage outpatient clinicians who care for vulnerable populations because the algorithm omits meaningful determinants of cost.

    Objectives  To determine whether factors not included in Medicare risk adjustment, including patient neuropsychological and functional status, as well as local area health resources and economic conditions, are associated with Medicare total annual cost of care (TACC), and evaluate whether accounting for these factors is associated with improved TACC performance by outpatient safety-net clinicians.

    Design, Setting, and Participants  In this retrospective observational study, we used the Medicare Current Beneficiary Survey (MCBS) to examine patient-reported neuropsychological and functional status and the Area Health Resources File to obtain information on local area characteristics. Included were Medicare beneficiaries with annual physician or clinic visits to outpatient safety-net (federally qualified health centers and rural health clinics) and non–safety-net clinics, contributing 76 927 person-years of data to the MCBS from 2006 through 2013. We used patient-level multivariable regression models to estimate the association between each factor and annual Medicare spending, and compared outpatient safety-net performance under current risk adjustment and after adding additional adjustment for these factors.

    Main Outcomes and Measures  Medicare TACC, measured as the total annual reimbursed amount per patient for Medicare Part A and Part B services, in all categories.

    Results  Our study included 111 414 unique identifiable physicians, and the final weighted sample included 213 904 324 patient-years (unweighted, 76 927 patient-years) from 30 058 unique patients, of whom 17 478 (58.1%) were women. The mean (SD) patient age was 71.84 (12.48) years. The mean TACC was $9117. Those with higher than mean TACC included beneficiaries with depression ($14 436), dementia ($18 311), and difficulty with 3 or more activities of daily living (ADLs, $19 113) or instrumental ADLs ($17 443). After adjusting for comorbidities, depression and dementia were still associated with $2740 (95% CI, $2200-$2739) and $2922 (95% CI, $2399-$3445) higher TACC, respectively. Difficulty with 3 or more ADLs ($3121 higher; 95% CI, $2633-$3609) or instrumental ADLs ($895 higher; 95% CI, $452-$1337) was also associated with higher TACC. Adding these neuropsychological and functional factors, as well as local residence area factors, to risk adjustment calculations reduced outpatient safety-net clinicians’ underperformance on Medicare TACC relative to non-safety–net clinicians by 52% (from 0.098 to 0.047 difference in the observed to expected ratio).

    Conclusions and Relevance  Neuropsychological and functional impairment are common in Medicare beneficiaries and are associated with increased annual Medicare spending. Failure to account for these factors may inappropriately penalize outpatient clinicians who care for these vulnerable groups, such as safety-net clinicians, for factors that are arguably beyond their control.

    Introduction

    Medicare is moving toward value-based and alternative payment models, in which outpatient clinicians are held accountable for quality and costs of care.1,2 One example is the Merit-Based Incentive Payment System (MIPS), which will adjust outpatient clinicians’ payments up or down by 4% in 2019 based on cost and quality performance.3 Outpatient clinicians covered by MIPS include all physicians, physician assistants, nurse practitioners, and certified nurse specialists who provide care to Medicare beneficiaries and are not already participating in qualified advanced alternative payment models (such as Medicare Accountable Care Organizations).3 As the MIPS start date approaches, there is concern that the program will inappropriately penalize outpatient clinicians for factors beyond their control, and assign penalties and bonuses based on these factors rather than on true differences in clinician performance.1,2,4,5 As a result, the Medicare Payment Advisory Commission recently recommended that Congress repeal MIPS and replace it with a voluntary performance program for outpatient Medicare clinicians.6

    The broadest cost measure in MIPS is the Total Per Capita Costs (TPCC) measure, which evaluates outpatient clinicians on annual costs of care for all attributed beneficiaries.7 Beneficiaries are attributed to outpatient clinicians based on the plurality of their evaluation and management (E&M) office visits in a year.8 To make comparisons among outpatient clinicians more equitable, the measure adjusts for medical risk using the Centers for Medicare and Medicaid Services (CMS) hierarchical conditions category (HCC) score, which includes medical conditions such as heart failure, diabetes, and kidney disease.9 The CMS-HCC risk adjustment also accounts for whether or not a patient is dually enrolled in Medicare and Medicaid, as a proxy for socioeconomic status. However, the TPCC measure is not adjusted for functional status and common neuropsychological comorbidities or for local area supply-side and economic condition variables. These unaccounted for patient risk factors are known to be associated with higher health care utilization and costs10-19 and are largely outside of outpatient clinicians’ control.

    In the presence of inadequate risk adjustment, the TPCC measure could lead to inappropriate penalties for outpatient safety-net clinicians, such as those working in federally qualified health centers (FQHCs) and rural health clinics (RHCs), who serve a high proportion of patients with neuropsychological and functional needs and living in underresourced areas.20,21 Penalties to such clinicians, often operating on already-thin margins, could be harmful.22,23 Therefore, understanding associations between these risk factors and spending, as well as their association with outpatient safety-net clinicians’ performance on treated patients’ annual Medicare spending, is crucial as value-based payment models proliferate more broadly.

    Therefore, we set out to answer 3 questions. First, are patient neuropsychological conditions and functional status associated with higher Medicare costs above and beyond the risk factors included in the standard CMS-HCC risk model? Second, are local area supply-side and economic factors associated with higher Medicare costs above and beyond the CMS-HCC model? Third, does adjusting for these additional patient risk factors have any association with the performance of outpatient safety-net clinicians—defined as FQHCs and RHCs—regarding patients’ Medicare total annual cost of care (TACC)?

    Methods
    Data and Study Sample

    We conducted a retrospective observational study using the Medicare Current Beneficiary Survey (MCBS) linked to respondents’ fee-for-service Medicare claims data for the 2006-2013 period. The MCBS is an annual nationally representative survey of the Medicare population with a rotating 4-year cohort design.24 We linked county-level health care supply and economic conditions from the Area Health Resources File provided by the Health Resources and Services Administration to respondents’ records.25

    Our sample includes all Medicare beneficiaries who would be eligible for inclusion in the Medicare annual costs measure and with available data in the MCBS. Because MIPS will require physician or clinic visits for patient attribution, we limited the sample to beneficiaries with at least a 12-month continuous enrollment in fee-for-service Medicare Part A and Part B, and at least 1 outpatient physician or clinic visit during the year, as identified by E&M visits (Berenson-Eggers Type of Service codes M1A, M1B, M4A, M4B, M5C, M5D, and M6 and revenue center codes 0521, 0522, and 0525) and positive covered dollar amounts. We used the MCBS cross-sectional survey weights to compute nationally representative estimates. These weights account for the selection probability of each person sampled and include adjustments for stratified sampling design, survey nonresponse, and coverage error.

    Cost-Associated Variables

    Our cost-associated variables were patient neuropsychological conditions, patient functional status, and local area supply-side and economic characteristics of patients’ residence (see eTable 1 in the Supplement). For neuropsychological conditions, we used patient self-report of having been diagnosed with depression, Alzheimer/dementia or a confirmatory diagnosis in their medical claims using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes defined by CMS26; Alzheimer/dementia is not included in the CMS-HCC risk model.27 In addition, although the CMS-HCC model includes acute major depressive, bipolar, and paranoid disorders, it does not include less acute but more common types of depression such as persistent depressive disorders and depressive disorders not classified as major depression. For functional status, we grouped counts of self-reported activities of daily living (ADLs) and instrumental ADLs (IADLs) that could be performed only with difficulty or not at all28 into no limitations (0), mild/moderate limitations (1-2), and severe limitations (3-6).29 For local area supply-side and economic characteristics, we used patient home addresses and linked them to the Area Health Resources File to obtain annual county level measures of availability of primary care or mental health professional facilities, rural-urban setting (metropolitan, micropolitan, rural), poverty rate, and unemployment rate.25

    Medicare TACC

    Our outcome variable was Medicare TACC measured as the total annual reimbursed amount per patient for Medicare Part A and Part B services. Following the MIPS methodology we bottom-coded patient costs at the 1st percentile value ($39) and top-coded at the 99th percentile value ($93 614).9 All costs were converted to 2013 US dollars using the gross domestic product deflator.30 Our analysis of Medicare TACC should be considered exploratory because, owing to sample size limitations, and unlike the official MIPS TPCC measure,9 we did not standardize costs to account for regional differences in the price of medical services paid by Medicare, nor did we “specialty adjust” to account for differential prices paid to specialists.

    Safety-Net vs Non–Safety-Net Patients

    We identified outpatient safety-net clinic patients as beneficiaries who visited a FQHC or RHC at least once during the year using revenue center codes beginning with “052” in their outpatient claims data.31 Non–safety-net patients were identified as all other beneficiaries in the sample.

    Statistical Analysis

    First, we computed descriptive statistics for our cost-associated variables and the CMS-HCC risk score. The HCC risk score (version V1210.70.F2) includes age, sex, original reason for Medicare entitlement, dual enrollment in Medicaid, institutionalization in long-term care, and a set of ICD-9-CM diagnosis codes.27,32 We used the Wald test to compare differences in proportions (or means) across the low-cost vs high-cost quintiles. In addition we computed overall mean annual TACC for each predictor variable category. We ran the CMS-HCC risk score model to estimate overall expected costs at the patient level and by cost quintile.27,32

    Next, we created 2 patient-level multivariable regression models estimating the association between each variable and TACC after applying the standard CMS-HCC risk-adjustment approach, which adjusts for HCC risk score, risk score squared, and end-stage renal disease (ESRD) status. Similar to the federal Marketplace approach and the model used by Ash et al,19 we used a “concurrent frame” model to predict TACC using variables measured during the same year. In the first model, we assessed the association of the neuropsychological and functional status variables with increased TACC. In the second model, we added local area supply and economic characteristics to the covariates that we included in the first model. Following a protocol outlined in prior research,19,27 we used simple least squares regression to estimate the average effect of our variables on TACC measured during the same year. Both models included year fixed effects to control for secular trend and adjusted for the complex survey design of the MCBS and intraperson correlation over time.

    Finally, we estimated outpatient safety-net vs non–safety-net clinicians’ performance on TACC under current risk adjustment and after adding additional adjustment for neuropsychological status, functional status, and local area variables to the risk models. All analyses were performed at the patient and not clinician level. We compared expected costs under these 3 modeling scenarios with observed costs and computed observed-to-expected ratios; a ratio greater than 1 implies that the risk model underpredicted actual costs; a ratio less than 1 implies that the risk model overpredicted actual costs.

    This study was approved by the St Louis University institutional review board, waiving participant written informed consent for deidentified retrospective data. We performed analyses using SAS software, version 9.4, and Stata version 14.

    Results
    Patient Sample and Characteristics

    Of 214 627 192 weighted patient-years eligible for inclusion in the study, we excluded 722 868 (0.3% of eligible patient-years) that were missing a valid US zip code or other key covariates (eFigure in the Supplement). Our final weighted sample consisted of 213 904 324 patient-years (unweighted, 76 927 patient-years; 30 058 unique patients; 111 414 unique identifiable physicians).

    The mean overall TACC was $9117 (Table 1). Beneficiaries in the highest-cost quintile (mean TPCC $34 494) were older and more often female, and were more likely to be dually enrolled in Medicare and Medicaid, institutionalized ($20 088), disabled ($9588), or have ESRD ($55 060). Individuals with depression ($14 436) or dementia ($18 311) and those with difficulty with 3 or more ADLs ($19 113) or IADLs ($17 443), living in a mental health care shortage area ($9233), having a high proportion of residents in poverty ($9569), or unemployed ($9658) had higher observed TACC.

    Association Between Neuropsychological Conditions, Functional Status, and Spending

    After applying the CMS-HCC risk adjustment, depression and dementia were still associated with $2740 (95% CI, $2200-$2739) and $2922 (95% CI, $2399-$3445) higher mean TACCs, respectively (Table 2). Difficulty with 3 or more ADLs was associated with a $3121 (95% CI, $2633-$3609) higher TACC, and difficulty with 3 or more IADLs was associated with an $895 (95% CI, $452-$1337) higher TACC.

    Association Between Local Area Factors and Spending

    After adjusting for the CMS-HCC risk score, we found that treated patients living in a mental health shortage area or an area with a high unemployment rate were associated with mean $517 (95% CI, $61-$972) and $455 (95% CI, $54-$857) higher TACCs, respectively (Table 3). In contrast, treated patients living in a micropolitan area or an area with a high proportion of residents in poverty were associated with $702 (95% CI, $399-$1005) and $407 (95% CI, $22-$792) lower TACCs, respectively.

    Adding Neuropsychological, Functional, and Local Area Factors to the Standard CMS-HCC Risk Adjustment Model

    Patients who utilized outpatient safety-net clinicians had a mean observed TACC ($9160) similar to patients who utilized non–safety-net clinicians ($9112) (Table 4). However, under the CMS-HCC risk adjustment model, the actual TACC of safety-net clinicians was 9% higher than the expected costs ($9160 vs $8413), whereas the actual TACC of non–safety-net clinicians was 1% less than expected ($9112 vs $9199).

    Adding patient neuropsychological and functional status variables to the CMS-HCC risk adjustment model increased the expected costs of outpatient safety-net clinicians from $8413 to $8722, whereas it decreased the expected costs of non–safety-net clinicians. Further adding residence in a mental health shortage area and county unemployment ranking to the model increased the expected cost of outpatient safety-net clinicians to $8791. As a result, adding these variables to the standard model led to a reduction in outpatient safety-net clinicians’ underperformance on TACC by 5 percentage points (a 52% relative improvement) and left non–safety-net clinicians performing at par.

    Sensitivity Analyses

    We performed 4 sensitivity analyses. First, we reestimated our models for community-dwelling patients only, excluding those with ESRD or who were institutionalized (eTables 2 and 3 in the Supplement); results were similar to those of our main analysis. Second, we tested the effect of including indicator variables for US Census Region in our safety-net to non–safety-net comparisons to control for regional differences in medical diagnostic coding intensity and the prices paid by Medicare for medical services (eTable 4 in the Supplement). Including these additional controls did not change our results. Third, we tightened our criteria for attributing patients as safety-net utilizers to 2 or more E&M visits to FQHCs or RHCs and then reestimated our safety-net to non–safety-net comparisons. Although those with 2 or more E&M visits to safety-net clinicians had higher observed and expected costs, there was no change in our findings showing an improved observed-to-expected cost ratio after adding cognitive, neuropsychological, and local area factors to the HCC risk model (eTable 5 in the Supplement). Fourth, we controlled for specialty physician involvement in beneficiaries’ E&M visits as a proxy for specialty adjusting TACC and then reestimated our safety-net to non–safety-net comparisons. Although safety-net clinicians had slightly lower expected costs, there was no change in our findings showing an improved observed-to-expected cost ratio after adding cognitive, neuropsychological, and local area factors to the HCC risk model (eTable 6 in the Supplement).

    Discussion

    In this nationally representative study of Medicare beneficiaries we found that neuropsychological and functional risk factors were common, and they were associated with substantially higher annual costs and poorer outpatient clinician performance on TACC. Adding these factors, along with local area variables, to the CMS-HCC risk adjustment model reduced outpatient safety-net clinicians’ underperformance on TACC relative to non–safety-net clinicians by 5 percentage points or 52%. This implies that failing to account for these factors could penalize outpatient safety-net clinicians under the MIPS program who care for patients who have greater neuropsychological and functional health needs and are located in underresourced areas.

    There is a growing body of literature showing that neuropsychological and functional risk factors have a significant effect on health services utilization and health care costs. Multiple studies find that depression33-35 and dementia10,36 are associated with hospitalization or rehospitalization. Importantly, even mild depression and anxiety, in addition to serious mental illness, may be drivers of readmissions and costly post–acute-care utilization.14 Many prior studies have found that functional status, as measured by deficits in ADLs and IADLs, is associated with higher admissions,37 readmissions,29,38,39 and post–acute-care costs15—even after adjusting for medical comorbidities and CMS-HCC risk factors. These findings also align with a recent Report to Congress on Medicare’s value-based purchasing programs, which finds that Medicare beneficiaries who are disabled have higher odds of hospital admission, even after risk adjustment.5

    Despite a robust literature on neuropsychological and functional status, most quality and cost measures do not include these factors in risk adjustment. The reasons are manifold but include both practical and philosophical limitations. Practically speaking, measuring functional status is difficult using claims alone; we relied on patient-reported function, which would be difficult to scale nationally. However, there are a number of validated claims-based indices of disability and frailty that may hold real promise for incorporating functional status information into claims-based quality metrics in the near term.40-42 The lack of inclusion of neuropsychological factors—particularly dementia—in risk adjustment models is more of a philosophical issue, as these diagnoses are easily identified in claims. The CMS has previously argued that including dementia in risk models would increase its inappropriate coding in physician claims because dementia is a “discretionary” diagnosis.43,44 Many groups, including the Alzheimer Association, have disputed this stance out of concern that failing to account for this important condition could lead to inadequate risk adjustment.45 Our findings suggest that their concerns are valid, and that failing to adjust for dementia could inappropriately penalize clinicians who care for a high proportion of patients with this devastating and costly disease, particularly as value-based payment programs spread into the outpatient and post-acute settings.46,47

    There is also prior work suggesting that local area health care supply and economic conditions have a significant effect on health care utilization above and beyond patient-level comorbidities. For example, seriously mentally ill patients who visit the emergency department are more likely to be admitted to the hospital if they live in a mental health professional shortage area.48 Higher area-level unemployment rates are associated with not only higher hospital admissions16 but also greater costs per hospitalization,17 and neighborhood disadvantage is associated with rehospitalization and higher health care costs.18,19 In the present study, we found that even county-level data on clinician supply and unemployment, which are very easily linkable, provided some additional information for risk adjustment and may be worth exploring further.

    We also found living in micropolitan or high poverty areas were significant negative predictors of costs—perhaps owing to less opportunity to access medical care in these communities. We did not include these factors in our augmented risk adjustment model out of a concern for fairness because including them would have the effect of lowering these patients’ expected costs and effectively taking money away from clinicians who serve patients living in micropolitan or high-poverty areas. This raises the philosophical issue of what to do when traditional markers of greater risk (eg, poverty) are associated with lower expected costs. Following Ash et al,19 we believe that considerations of fairness or equity in terms of real-world effect on clinicians who serve vulnerable patients should also guide risk adjustment decisions.

    Limitations

    This study has several limitations. First, we were unable to identify outpatient safety-net clinicians who were not FQHCs or RHCs. As a result, these unidentified outpatient clinicians’ patients were counted as non–safety-net utilizers; this would likely bias our results toward the null. Second, it is possible that safety-net clinicians have systematically lower diagnostic coding intensity and that this is what is driving their lower expected costs under the HCC model. We are unable to confirm this in our data. Third, we were only able to identify patients who participated in the MCBS; while survey and nonresponse weighting can significantly reduce bias related to this issue, our results may not generalize more broadly. Relatedly, it is possible that patients with the most serious forms of Alzheimer/dementia are less likely to participate in the MCBS. However, we found a prevalence rate of 10.1% in our study sample that is very similar to the prevalence rate of 11% reported for the entire Medicare fee-for-service population in the Chronic Condition Warehouse over the 2006-2015 period.49 Fourth, we measured TACC at a period prior to implementation of the MIPS or value-based payment modifier programs using a similar measure, and it is possible that patients or practices have changed since then. Repeating these analyses as MIPS is implemented will be important.

    Conclusions

    Neuropsychological and functional risk factors are common in Medicare beneficiaries, especially those cared for by outpatient safety-net clinicians. These risk factors exert an independent effect on annual Medicare spending that is largely beyond the control of outpatient clinicians and not accounted for by current Medicare risk adjustment methods. Consequently, Medicare’s MIPS scoring formula may inappropriately penalize outpatient safety-net clinicians. In the future, CMS could consider accounting for patient cognitive and functional status as well as local area health care supply and economic conditions in risk adjustment.

    Back to top
    Article Information

    Corresponding Author: Kenton J. Johnston, PhD, Department of Health Management and Policy, Center for Outcomes Research, College for Public Health and Social Justice, St Louis University, 3545 Lafayette Ave, Room 362, St Louis, MO 63104 (johnstonkj@slu.edu).

    Accepted for Publication: July 2, 2018.

    Published Online: September 17, 2018. doi:10.1001/jamainternmed.2018.4143

    Author Contributions: Dr Johnston had full access to all of 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: All authors.

    Acquisition, analysis, or interpretation of data: Johnston, Wen, Joynt Maddox.

    Drafting of the manuscript: Johnston, Joynt Maddox.

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

    Statistical analysis: Johnston, Wen, Hockenberry.

    Obtained funding: Johnston.

    Administrative, technical, or material support: Johnston.

    Conflict of Interest Disclosures: Dr Joynt Maddox does intermittent work under contract with the United States Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. No other disclosures are reported.

    Funding/Support: St Louis University purchased and provided access to the data used in this study. Dr Joynt Maddox is supported by grant K23-HL109177-03 from the National Heart, Lung, and Blood Institute (NHLBI).

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

    Additional Contributions: We thank Julia Clarke of St Louis University for providing assistance on the literature review. Ms Clarke’s contributions were completed as part of her research assistantship at Saint Louis University, for which she received compensation.

    References
    1.
    Buntin  MB, Ayanian  JZ.  Social risk factors and equity in Medicare payment.  N Engl J Med. 2017;376(6):507-510. doi:10.1056/NEJMp1700081PubMedGoogle ScholarCrossref
    2.
    Joynt  KE, De Lew  N, Sheingold  SH, Conway  PH, Goodrich  K, Epstein  AM.  Should Medicare value-based purchasing take social risk into account?  N Engl J Med. 2017;376(6):510-513. doi:10.1056/NEJMp1616278PubMedGoogle ScholarCrossref
    3.
    Centers for Medicare and Medicaid Services. The Merit-Based Incentive Payment System: MIPS Scoring Methodology Overview. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/MACRA-MIPS-and-APMs/MIPS-Scoring-Methodology-slide-deck.pdf. Accessed March 16, 2018.
    4.
    Joynt Maddox  KE.  Social and behavioral determinants of spending.  JAMA Intern Med. 2017;177(10):1431-1432. doi:10.1001/jamainternmed.2017.3325PubMedGoogle ScholarCrossref
    5.
    Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation. Report to Congress: social risk factors and performance under Medicare’s value-based payment programs. 2016. https://aspe.hhs.gov/pdf-report/report-congress-social-risk-factors-and-performance-under-medicares-value-based-purchasing-programs. Accessed March 27, 2018.
    6.
    Medicare Payment Advisory Commission. Moving Beyond the Merit-Based Incentive Payment System. Rep to Congr Medicare Paym Policy. 2018:445-468. http://www.medpac.gov/docs/default-source/reports/mar18_medpac_entirereport_sec.pdf?sfvrsn=0. Accessed June 7, 2018.
    7.
    Centers for Medicare and Medicaid Services. The Merit-Based Incentive Payment System: Resource Use Performance Category. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/MACRA-MIPS-and-APMs/Resource-Use-Performance-Category-slide-deck.pdf. Accessed March 16, 2018.
    8.
    Centers for Medicare and Medicaid Services. 2018 MIPS Cost Performance Category Fact Sheet. Qual Paym Progr. 2018:1-7. https://www.cms.gov/Medicare/Quality-Payment-Program/Resource-Library/2018-Cost-Performance-Category-Fact-Sheet.pdf. Accessed June 11, 2018.
    9.
    Centers for Medicare and Medicaid Services. 2016 Measure Information About the Per Capita Costs for All Attributed Beneficiaries Measure, Calculated for the 2018 Value-Based Payment Modifier Program. 2017:1-8. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/PhysicianFeedbackProgram/Downloads/2016-TPCC-MIF.pdf. Accessed March 16, 2018.
    10.
    Lledó  R, Martín  E, Jiménez  C,  et al.  Characteristics of elderly inpatients at high risk of needing supportive social and health care services.  Eur J Epidemiol. 1997;13(8):903-907. doi:10.1023/A:1007420911766PubMedGoogle ScholarCrossref
    11.
    Fong  TG, Jones  RN, Marcantonio  ER,  et al.  Adverse outcomes after hospitalization and delirium in persons with Alzheimer disease.  Ann Intern Med. 2012;156(12):848-856, W296. doi:10.7326/0003-4819-156-12-201206190-00005PubMedGoogle ScholarCrossref
    12.
    Linden  M, Horgas  AL, Gilberg  R, Steinhagen-Thiessen  E.  Predicting health care utilization in the very old: the role of physical health, mental health, attitudinal and social factors.  J Aging Health. 1997;9(1):3-27. doi:10.1177/089826439700900101PubMedGoogle ScholarCrossref
    13.
    McGinty  EE, Sridhara  S.  Potentially preventable medical hospitalizations among Maryland residents with mental illness, 2005-2010.  Psychiatr Serv. 2014;65(7):951-953. doi:10.1176/appi.ps.201300323PubMedGoogle ScholarCrossref
    14.
    Benzer  JK, Sullivan  JL, Williams  S, Burgess  JF.  One-year cost implications of using mental health care after discharge from a general medical hospitalization.  Psychiatr Serv. 2012;63(7):672-678. doi:10.1176/appi.ps.201100457PubMedGoogle ScholarCrossref
    15.
    Greysen  SR, Stijacic Cenzer  I, Boscardin  WJ, Covinsky  KE.  Functional impairment: an unmeasured marker of Medicare costs for postacute care of older adults.  J Am Geriatr Soc. 2017;65(9):1996-2002. doi:10.1111/jgs.14955PubMedGoogle ScholarCrossref
    16.
    McInerney  M, Mellor  JM.  Recessions and seniors’ health, health behaviors, and healthcare use: analysis of the Medicare Current Beneficiary Survey.  J Health Econ. 2012;31(5):744-751. doi:10.1016/j.jhealeco.2012.06.002PubMedGoogle ScholarCrossref
    17.
    Maeda  JLK, Henke  RM, Marder  WD, Karaca  Z, Friedman  BS, Wong  HS.  Association between the unemployment rate and inpatient cost per discharge by payer in the United States, 2005-2010.  BMC Health Serv Res. 2014;14(1):378. doi:10.1186/1472-6963-14-378PubMedGoogle ScholarCrossref
    18.
    Kind  AJH, Jencks  S, Brock  J,  et al.  Neighborhood socioeconomic disadvantage and 30-day rehospitalization: a retrospective cohort study.  Ann Intern Med. 2014;161(11):765-774. doi:10.7326/M13-2946PubMedGoogle ScholarCrossref
    19.
    Ash  AS, Mick  EO, Ellis  RP, Kiefe  CI, Allison  JJ, Clark  MA.  Social determinants of health in managed care payment formulas.  JAMA Intern Med. 2017;177(10):1424-1430. doi:10.1001/jamainternmed.2017.3317PubMedGoogle ScholarCrossref
    20.
    Nguyen  OK, Makam  AN, Halm  EA.  National use of safety-net clinics for primary care among adults with non-Medicaid insurance in the United States.  PLoS One. 2016;11(3):e0151610. doi:10.1371/journal.pone.0151610PubMedGoogle ScholarCrossref
    21.
    Gilman  M, Adams  EK, Hockenberry  JM, Milstein  AS, Wilson  IB, Becker  ER.  Safety-net hospitals more likely than other hospitals to fare poorly under Medicare’s value-based purchasing.  Health Aff (Millwood). 2015;34(3):398-405. doi:10.1377/hlthaff.2014.1059PubMedGoogle ScholarCrossref
    22.
    Chokshi  DA, Chang  JE, Wilson  RM.  Health reform and the changing safety net in the United States.  N Engl J Med. 2016;375(18):1790-1796. doi:10.1056/NEJMhpr1608578PubMedGoogle ScholarCrossref
    23.
    Andrulis  DP, Siddiqui  NJ.  Health reform holds both risks and rewards for safety-net providers and racially and ethnically diverse patients.  Health Aff (Millwood). 2011;30(10):1830-1836. doi:10.1377/hlthaff.2011.0661PubMedGoogle ScholarCrossref
    24.
    Center for Medicare and Medicaid Services. Technical Documentation for the Medicare Current Beneficiary Survey. In: Medicare Current Beneficiary Survey Data Tables. https://www.cms.gov/Research-Statistics-Data-and-Systems/Research/MCBS/Downloads/OLD/HHC2003appendixA.pdf. Accessed August 7, 2018.
    25.
    Health Resources and Services Administration. Area Health Resources Files (AHRF). Rockville, MD; 2014. https://datawarehouse.hrsa.gov/topics/ahrf.aspx.
    26.
    Chronic Conditions Data Warehouse. Condition Categories: 27 Chronic Condition Algorithms. 2016. https://www.ccwdata.org/web/guest/condition-categories. Accessed April 28, 2016.
    27.
    Pope  GC, Kautter  J, Ellis  RP,  et al.  Risk adjustment of Medicare capitation payments using the CMS-HCC model.  Health Care Financ Rev. 2004;25(4):119-141.PubMedGoogle Scholar
    28.
    Ivanova  JI, Mytelka  DS, Duh  MS,  et al.  Evaluating the utility of existing patient-reported outcome scales in novel patient populations with pancreatic cancer, lung cancer, and myeloproliferative neoplasms using Medicare current beneficiary survey data.  Patient. 2013;6(3):189-200. doi:10.1007/s40271-013-0018-zPubMedGoogle ScholarCrossref
    29.
    Meddings  J, Reichert  H, Smith  SN,  et al.  The impact of disability and social determinants of health on condition-specific readmissions beyond Medicare risk adjustments: a cohort study.  J Gen Intern Med. 2017;32(1):71-80. doi:10.1007/s11606-016-3869-xPubMedGoogle ScholarCrossref
    30.
    Johnston  L, Williamson  S. What Was the US GDP Then? Measuring Worth. 2017. https://www.measuringworth.com/datasets/usgdp/. Accessed March 16, 2018.
    31.
    Centers for Medicare and Medicaid Services. Chapter 9—Rural Health Clinics/Federally Qualified Health Centers. Medicare Claims Process Man. 2015:1-38. https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/clm104c09.pdf. Accessed June 14, 2018.
    32.
    Ash  AS, Ellis  RP, Pope  GC,  et al.  Using diagnoses to describe populations and predict costs.  Health Care Financ Rev. 2000;21(3):7-28.PubMedGoogle Scholar
    33.
    Davydow  DS, Hough  CL, Zivin  K, Langa  KM, Katon  WJ.  Depression and risk of hospitalization for pneumonia in a cohort study of older Americans.  J Psychosom Res. 2014;77(6):528-534. doi:10.1016/j.jpsychores.2014.08.002PubMedGoogle ScholarCrossref
    34.
    Sherwood  A, Blumenthal  JA, Trivedi  R,  et al.  Relationship of depression to death or hospitalization in patients with heart failure.  Arch Intern Med. 2007;167(4):367-373. doi:10.1001/archinte.167.4.367PubMedGoogle ScholarCrossref
    35.
    Rutledge  T, Reis  VA, Linke  SE, Greenberg  BH, Mills  PJ.  Depression in heart failure a meta-analytic review of prevalence, intervention effects, and associations with clinical outcomes.  J Am Coll Cardiol. 2006;48(8):1527-1537. doi:10.1016/j.jacc.2006.06.055PubMedGoogle ScholarCrossref
    36.
    Daiello  LA, Gardner  R, Epstein-Lubow  G, Butterfield  K, Gravenstein  S.  Association of dementia with early rehospitalization among Medicare beneficiaries.  Arch Gerontol Geriatr. 2014;59(1):162-168. doi:10.1016/j.archger.2014.02.010PubMedGoogle ScholarCrossref
    37.
    Schüssler-Fiorenza Rose  SM, Stineman  MG, Pan  Q,  et al.  Potentially avoidable hospitalizations among people at different activity of daily living limitation stages.  Health Serv Res. 2017;52(1):132-155. doi:10.1111/1475-6773.12484PubMedGoogle ScholarCrossref
    38.
    Barnett  ML, Hsu  J, McWilliams  JM.  Patient characteristics and differences in hospital readmission rates.  JAMA Intern Med. 2015;175(11):1803-1812. doi:10.1001/jamainternmed.2015.4660PubMedGoogle ScholarCrossref
    39.
    Greysen  SR, Stijacic Cenzer  I, Auerbach  AD, Covinsky  KE.  Functional impairment and hospital readmission in Medicare seniors.  JAMA Intern Med. 2015;175(4):559-565. doi:10.1001/jamainternmed.2014.7756PubMedGoogle ScholarCrossref
    40.
    Kim  DH, Schneeweiss  S, Glynn  RJ, Lipsitz  LA, Rockwood  K, Avorn  J.  Measuring frailty in Medicare data: development and validation of a claims-based frailty index.  J Gerontol A Biol Sci Med Sci. 2017;(December). doi:10.1093/gerona/glx229PubMedGoogle Scholar
    41.
    Segal  JB, Chang  H-Y, Du  Y, Walston  JD, Carlson  MC, Varadhan  R.  Development of a claims-based frailty indicator anchored to a well-established frailty phenotype.  Med Care. 2017;55(7):716-722. doi:10.1097/MLR.0000000000000729PubMedGoogle ScholarCrossref
    42.
    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
    43.
    United States Government Accountability Office. Medicare Advantage: Changes Improved Accuracy of Risk Adjustment for Certain Beneficiaries. 2011. https://www.gao.gov/assets/590/587519.pdf. Accessed March 30, 2018.
    44.
    Centers for Medicare and Medicaid Services. Note To: All Medicare Advantage Organizations, Prescription Drug Plan Sponsors, and Other Interested Parties. 2013. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Downloads/Announcement2014.pdf. Accessed March 30, 2018.
    45.
    Alzheimer’s Association. Re: Advance Notice of Methodological Changes for Calendar Year (CY) 2017 for Medicare Advantage (MA) Capitation Rates, Part C and Part D Payment Policies and 2017 Call Letter. 2016. http://act.alz.org/site/DocServer/Alzheimer_s_Association_Comments_on_2017_Advance_Notice_.pdf?docID=51379. Accessed March 30, 2018.
    46.
    Centers for Medicare and Medicaid Services. Home Health Value-Based Purchasing Model. https://innovation.cms.gov/initiatives/home-health-value-based-purchasing-model. Accessed March 30, 2018.
    47.
    Centers for Medicare and Medicaid Services. The Skilled Nursing Facility Value-Based Purchasing Program. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/Other-VBPs/SNF-VBP.html. Accessed March 30, 2018.
    48.
    Moseley  C, Shen  J, Cochran  C.  Mental health services availability and admission of the seriously mentally ill from the emergency department.  J Health Hum Serv Adm. 2008;31(3):292-308.PubMedGoogle Scholar
    49.
    Chronic Conditions Data Warehouse. Cognitive and Mental Health Conditions–Trends: Medicare Chronic Cond Charts. 2018. https://www.ccwdata.org/documents/10280/28579486/b2-cog-mh-2006-2015.jpg. Accessed June 6, 2018.
    ×