Leah M. Marcotte, MD

 To understand physician perspectives about drivers of value within the 4 domains—quality, improvement activities, promoting interoperability, and cost—of the Merit-based Incentive Payment System (MIPS).

Those surveyed who believed that MIPS domain activities would improve value were asked about mechanisms—quality (structural quality, process quality, or outcome quality), patient experience, and/or cost—they believe would drive those improvements.

Of 1431 physicians, 51% responded. Most believed that value would be improved by activities in the 4 MIPS domains of quality (55%), improvement activities (70%), promoting interoperability (54%), and cost (71%). Process quality was the most frequently selected driver in the quality (77%), improvement activities (70%), and promoting interoperability (70%) domains.

 Among physicians who believed that MIPS domain activities would improve value, most believed that would occur predominantly through process improvement activities rather than other forms of quality, patient experience, or cost containment.

Implemented in 2017, the Medicare Merit-based Incentive Payment System (MIPS) represents the largest pay-for-performance program to date in the United States.

 Through MIPS, Medicare engages more than 1 million clinicians

 caring for Medicare beneficiaries around the country and uses financial incentives to encourage these clinicians to deliver higher-quality and more cost-conscious (ie, higher-value) care.

In the MIPS program, clinicians are evaluated in 4 domains: quality (eg, reporting and performance on clinical quality measures), improvement activities, promoting interoperability (eg, using health information technology to improve care), and cost. Performance in these 4 domains ultimately dictates how clinicians are affected by MIPS, with higher-performing clinicians receiving overall increases in professional payment rates from Medicare and lower-performing clinicians receiving overall decreases in payment rates.

Physicians must be engaged stakeholders for MIPS to succeed, and many physicians believe that core activities in the 4 domains can improve value (ie, improve quality while maintaining or lowering costs, or lower costs while maintaining or improving quality).

 However, value is inherently multifaceted, given its relationship to the underlying mechanisms of quality, cost, and patient experience. In turn, quality itself is multifaceted. Based on the framework articulated by Donabedian,

 quality can be defined in 3 categories: structural (ie, the setting or context of care), process (ie, the actions taken to provide care), and outcome (improvement in care outcomes).

Consequently, understanding physician perspectives about how MIPS domain activities specifically improve value (eg, through improvements in quality, patient experience, or cost efficiency) can illuminate beliefs that policy makers can emphasize in order to engage physicians in MIPS. This is critical insight, given the calls to reform the program and improve the experience of participating clinicians.

Our analysis uses data from a nationwide web-based survey implemented in collaboration with the American College of Physicians and fielded among 1431 internal medicine physicians between March 22 and May 7, 2017. The study methodology has been described in detail previously.

Survey responses were stratified by respondents’ age (< 50 years vs ≥ 50 years), gender (male vs female), individual compensation (cost incentive vs other [noncost] incentive vs no incentive), specialty (generalist [primary care, geriatrics, hospital medicine] vs specialist), and regional level of physician reimbursement (using zip code, respondents were designated to hospital referral regions and determined to be in “high-spending” [mean Medicare reimbursement greater than or equal to national mean] or “low-spending” [mean Medicare reimbursement less than national mean] regions).

We focused our analysis on responses to 4 questions (

), the results of which have not been previously published. Each of these key questions focused on a MIPS domain and was directed to respondents who indicated in a separate, preceding question that activities in that domain improved value. In each question, respondents who believed that value would be improved through activities in a given domain were asked which mechanisms—quality, patient experience, and/or cost—explained the improvement in value (ie, how the activities improved value). For instance, respondents who reported in a preceding question the belief that quality domain activities improved value were asked the following question (1 of the 4 key questions in our analysis): If physician efforts to report and perform well on clinical quality measures somewhat or significantly improve the value of patient care, which of the following do they impact?

Answer choices consisted of (1) setting, context, or infrastructure in which care is delivered; (2) actions that physicians take in the process of providing care; (3) patient outcomes; (4) costs of care; and (5) patient experience. Answer choices 1 through 3 were adapted from the framework articulated by Donabedian

 described previously. We categorized answer choices 1, 2, and 3 as structural quality, process quality, and outcome quality, respectively.

Survey responses were described using percentages. Chi-square tests and 

 tests were used to compare categorical and continuous variables, respectively. Significance was determined with 2-tailed tests at the 0.05 α level. Analyses were performed using SAS version 9.4 (SAS Institute). This study was determined exempt by the University of Pennsylvania Institutional Review Board.

Among 1431 physicians, 51% (726) responded. Respondents were 51% male with a median age of 48 years, and 71% were general internists (primary care, geriatrics, or hospital medicine). Compared with internal medicine physicians nationally, respondents were younger and more likely to be female.

 Approximately half reported that quality (52%) or productivity (51%) affected their compensation. Compared with nonrespondents, respondents were more likely to be male (51% vs 42%; 

 < .001) and older (mean age, 50 years vs 45 years; 

 < .001). Forty-three percent of respondents practiced in low-spending regions; 57% practiced in high-spending regions. Most respondents believed that value would be improved by activities in the 4 domains: quality (55%; n = 399), improvement activities (70%; n = 503), promoting interoperability (54%; n = 385), and cost (71%; n = 509). Ninety percent of respondents thought that value would be improved by activities in at least 1 of the domains, and 32% thought that value would be improved by activities in all 4 of the domains.

Physician perspectives regarding whether value would be improved by activities in each domain varied by age, gender, and compensation incentives but did not vary by clinical specialty type (generalists vs specialists) or regional spending (

). In particular, older respondents (50 years and older) were less likely than younger respondents to believe that value was driven through the improvement activities (65% vs 74%; 

 = .008), promoting interoperability (45% vs 61%; 

 = .0021) domains. However, there was no difference in age groups regarding beliefs about the impact of the quality domain on value (51% among older respondents vs 59% among younger respondents; 

 = .099). Female respondents were more likely than male respondents to believe that value was driven through the quality (61% vs 53%; 

 = .024), improvement activities (78% vs 68%; 

 = .012), and promoting interoperability (62% vs 49%; 

 = .0008) domains but not through the cost domain (74% vs 71%; 

 = .23). Compensation incentives—cost vs other noncost (ie, incentives related to quality, patient satisfaction, or productivity) vs none—were associated with different beliefs about whether the quality domain (57% vs 60% vs 47%, respectively; 

 = .03) or improvement activities domain (65% vs 76% vs 63%; 

Of respondents who believed that domain activities would improve value, process quality was the most frequently selected mechanism in the quality (77%; n = 306), improvement activities (70%; n = 354), and promoting interoperability (70%; n = 270) domains (

). Costs of care was the most frequently selected mechanism (86%) among the 509 respondents who reported the belief that cost domain activities would improve value. As mechanisms, patient experience and structural quality were selected by fewer than half of respondents across the 4 MIPS domains.

Within the quality domain, 246 (62%) and 306 (77%) physicians believed that patient outcomes and process quality, respectively, drove value, whereas 182 (46%) believed both mechanisms did. Similarly, 335 (67%) and 354 (70%) physicians believed that patient outcomes and process quality, respectively, drove quality in the improvement activities domain, with 237 (47%) believing both did. In the promoting interoperability domain, 270 (70%) physicians responded that process quality drove value, whereas only 206 (54%) believed that patient outcomes did so and 154 (40%) believed that both did so. In the cost domain, approximately half of physicians thought that patient outcomes and process quality drove value (253 [50%] and 266 [52%], respectively), whereas fewer (158 [31%]) believed that both mechanisms did so.

This analysis provides insight into physician perspectives about how MIPS can improve value. The results demonstrate that when physicians believe that MIPS domain activities positively impact value, they do so predominantly through process quality rather than other forms of quality, patient experience, or cost containment. These findings pose several implications.

First, they raise caution about the policy decision to shift the focus of MIPS away from process quality (eg, moving incentives or performance evaluation toward other, nonquality domain activities), as Medicare has begun to do. In all 4 domains, more respondents believed that their efforts to improve value through reporting and performance on measures in MIPS domains were more likely to affect processes of care than outcomes. One potential implication is that retaining emphasis on process improvement alongside clinical outcomes and other forms of quality may be important for maintaining physician engagement in MIPS.

Engagement has been a key policy opportunity area for the program, particularly given early concerns raised by the complexity and uncertainty about MIPS performance.5 Going forward, policy makers have an opportunity to address these issues in the development of MIPS Value Pathways

—new approaches that seek to use tailored sets of performance measures to increase MIPS engagement among clinicians from different specialties. In particular, policy makers could intentionally ensure that salient process quality measures are selected and emphasized for each Value Pathway. This focus on quality would also be consistent with evaluations of other pay-for-performance programs, which have been associated with improvement in process measures but not outcome measures.

Second, our findings highlight the importance of defining and building consensus about the meaning of “value” in payment reform. To date, many payment reforms have equated value to cost reduction and stable quality. Although this is one appropriate definition, others exist. In particular, an alternative approach—which is codified in federal statute and recognized by Medicare—conceives of value as improved quality with stable cost.

 This definition aligns with our finding that physicians believed improving quality would be a major driver of value in the context of MIPS.

Third, nearly half of physicians who believed that MIPS activities improve value noted improvements in patient experience as a key mechanism. This finding aligns with the priority that Medicare has placed on increasing patient experience and engagement in value-based payment reforms—for example, through other novel advanced alternative payment models such as direct contracting.

 Given Medicare’s goal to better align MIPS and alternative payment models,

 as well as the fact that patient experience and engagement are not currently incorporated into MIPS, there is an opportunity to better incorporate them into the program going forward. MIPS Value Pathways provide another potential way for policy makers to further emphasize patient experience: Although they prioritize patient-centered outcomes, policy makers could design them to directly call out patient experience measures.

Fourth, our findings highlight potential for varying levels of engagement in different subgroups of physicians. For example, in multiple domains, female physicians and those who were younger were more likely to believe that domain activities contribute to value of care. Medicare could capture these perspectives by ensuring that MIPS Value Pathways engage a wide range of clinicians, including individuals from these subgroups, in the development process.

Our study has limitations. First, the generalizability of our results is limited by response rate, as well as the potential for more familiarity with MIPS and more favorable views regarding health care reform among respondents vs nonrespondents.

 Second, our survey evaluated physician perspectives in the first year of MIPS, which may differ from those in subsequent years. Third, survey questions measured beliefs about clinical decisions rather than behavior. Fourth, our results reflect overall physician perspectives, but future work should evaluate perspectives stratified by clinician (eg, MIPS performance) and practice (eg, type, size) characteristics.

Among physicians who believed that MIPS domain activities would improve value, most believed that would occur predominantly through process improvement activities rather than other forms of quality, patient experience, or cost containment. Our results spotlight potential opportunities for policy makers to engage physicians as key stakeholders to improve and ultimately achieve the goals of MIPS.

 Department of Medicine, University of Washington School of Medicine (LMM, LZ, JML), Seattle, WA; Value & Systems Science Lab (LMM, LZ, JML), Seattle, WA; Corporal Michael J. Crescenz VA Medical Center (ASN), Philadelphia, PA; Department of Medical Ethics and Health Policy, Perelman School of Medicine (ASN), and Leonard Davis Institute of Health Economics (ASN, JML), University of Pennsylvania, Philadelphia, PA.

 Dr Navathe reports grants from Hawaii Medical Service Association, Anthem Public Policy Institute, Commonwealth Fund, Oscar Health, Cigna Corporation, Robert Wood Johnson Foundation, Donaghue Foundation, Pennsylvania Department of Health, Ochsner Health System, United Healthcare, Blue Cross Blue Shield of NC, and Blue Shield of CA; personal fees from Navvis Healthcare, Agathos Inc, Navahealth, YNHHSC/CORE, Maine Health Accountable Care Organization, Maine Department of Health and Human Services, National University Health System - Singapore, Ministry of Health - Singapore, Elsevier Press, Medicare Payment Advisory Commission, Cleveland Clinic, and The Analysis Group; equity from Embedded Healthcare and Navahealth; and other from Integrated Services Inc outside the submitted work. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Concept and design (LMM, ASN, JML); acquisition of data (ASN, LZ, JML); analysis and interpretation of data (LMM,ASN, LZ, JML); drafting of the manuscript (LMM, ASN); critical revision of the manuscript for important intellectual content (LMM, ASN, JML); statistical analysis (LMM, ASN, LZ); obtaining funding (ASN); administrative, technical, or logistic support (LMM, ASN); and supervision (ASN, JML). 

Leah M. Marcotte, MD, Department of Medicine, University of Washington School of Medicine, 1959 NE Pacific St, Seattle, WA 98105. Email: leahmar@uw.edu.

1. Participation options overview. CMS. Accessed January 3, 2021. https://qpp.cms.gov/mips/overview

2. CMS. 2017 Quality Payment Program reporting experience. Primary Care Collaborative. Accessed January 3, 2021. https://www.pcpcc.org/sites/default/files/resources/2017%20QPP%20Experience%20Report.pdf

3. Liao JM, Shea JA, Weissman A, Navathe AS. Physician perspectives in year 1 of MACRA and its Merit-based Payment System: a national survey. 

. 2018;37(7):1079-1086. doi:10.1377/hlthaff.2017.1485

4. Donabedian A. Evaluating the quality of medical care. 

. 1966;44(3, pt 2):166-206. doi:10.2307/3348969

5. Liao JM, Navathe AS. Implications of MIPS scoring on clinician engagement in value-based payments. 

. October 30, 2018. Accessed January 3, 2021. https://www.healthaffairs.org/do/10.1377/hblog20181029.492700/full/

6. Liao JM, Chen A, Navathe AS. The real opportunities of MIPS payment adjustments. 

. October 29, 2018. Accessed January 3, 2021. https://www.healthaffairs.org/do/10.1377/hblog20181026.940799/full/

7. Medicare reimbursements. Dartmouth Atlas Project. Accessed January 3, 2021. https://www.dartmouthatlas.org/interactive-apps/medicare-reimbursements/

8. Active physicians by sex and specialty, 2017. American Association of Medical Colleges. Accessed January 3, 2021. https://www.aamc.org/data-reports/workforce/interactive-data/active-physicians-sex-and-specialty-2017

9. Active physicians by age and specialty, 2017. American Association of Medical Colleges. Accessed January 3, 2021. https://www.aamc.org/data-reports/workforce/interactive-data/active-physicians-age-and-specialty-2017

10. MIPS Value Pathways (MVPs). CMS. Accessed January 3, 2021. https://qpp.cms.gov/mips/mips-value-pathways

11. Mendelson A, Kondo K, Damberg C, et al. The effects of pay-for-performance programs on health, health care use, and processes of care: a systematic review. 

. 2017;166(5):341-353. doi:10.7326/M16-1881

12. CMMI model certifications. CMS. Updated January 8, 2021. Accessed January 3, 2021. https://www.cms.gov/Research-Statistics-Data-and-Systems/Research/ActuarialStudies/CMMI-Model-Certifications.html

13. Primary Care First Model Options. CMS. Updated December 1, 2020. Accessed January 3, 2021. https://innovation.cms.gov/initiatives/primary-care-first-model-options/

Articles

A national survey of physicians elicits beliefs regarding how the Merit-based Incentive Payment System domains drive value.

Physician Perspectives: How the Merit-based Incentive Payment  System Improves Value

As a broadly defined care delivery strategy, care management can improve outcomes but has yielded variable results, particularly within value-based payment models.

One contributing factor is that care management encompasses different types of activities, functions, and goals.

A framework categorizing care management as disease management, utilization management, and/or health care navigation interventions can help providers and payers invest in, implement, and reap benefits from care management programs.

Emphasis on care management has become ubiquitous in the era of value-based payment. At the vanguard of the movement, policy makers such as Medicare have emphasized care management via a variety of initiatives, ranging from payment models that encourage longitudinal care management

 to billing codes that reimburse clinicians for coordinating the care of patients who have chronic conditions.

 Hospitals and physician groups increasingly view care management as a key area for investment and as a strategy for succeeding in a value-based environment,

 and there have been isolated reports of success within value-based payment models.

However, despite widespread enthusiasm, complex care management and coordination activities are not associated with differences in quality, utilization, or cost outcomes in prominent value-based payment models such as the Medicare Shared Savings Program (MSSP).

 One potential driver is the variability in how care management programs are defined and implemented. For instance, a survey of individuals leading care management programs demonstrated such a wide range of roles, program characteristics, patient selection methods, and measurement strategies that researchers faced challenges when attempting to classify programs.

In light of this variation, it is important for policy makers and clinicians to clearly define care management in order to evaluate its impact on patient care. Doing so is vital because care management is not a monolithic concept, but rather one that encompasses several distinct types of activities: those related to clinical conditions, health care utilization, and patient experience and access. Although these types of activities need not be mutually exclusive or at odds with each other, failing to distinguish between them when implementing care management programs carries risks.

Organizations may invest in programs that emphasize one activity while expecting the benefits of another, creating misalignment among organizational goals, program investments, and eventual results. For example, care management programs that emphasize disease management may improve quality performance (eg, increasing adherence to guideline-concordant processes for a given disease) while worsening cost performance by increasing utilization

 (eg, if patients who were previously nonadherent to guideline-concordant care begin receiving more of it).

Consequently, organizations that build disease management—heavy programs expecting them to drive performance in total costs of care reforms, such as accountable care organizations, may fail to see return on investment and arrive at the conclusion that “care management” does not work. This dynamic may exist in the MSSP, in which clinical condition–focused management initiatives have failed to contribute to cost savings.

 For patients, the end result of confusing the functions, benefits, and risks of different care management activities is that some individuals may receive care management services they do not need, while others do not receive services that they do need.

To address this issue and support organizational efforts to engage and succeed in value-based care, we offer a care management taxonomy based on the distinct activities of managing clinical disease (hereafter, 

), containing health care utilization (hereafter, 

), and improving patients’ navigation through the health care system (hereafter, 

). We contrast these 3 activities by focusing on differences in function, target populations, requisite competencies, and key risks and limitations (

Care management programs focused on disease management predate the era of value-based care.

 These programs target patients with chronic conditions, using medical evidence to create processes that improve patients’ disease self-management via steps such as medication adherence and lifestyle or behavioral changes (eg, related to diet and exercise).

The main function of disease management initiatives is to drive greater use of evidence- or guideline-based care. In turn, the requisite competencies for this type of care management activity consist primarily of the ability to translate clinical knowledge into care processes and engage patients in their own care (eg, via motivational interviewing, self-management support). From the perspective of organizational performance goals, this type of program can align with value-based care delivery by improving disease-specific metric performance. However, there is a dearth of evidence that disease management programs lead to appreciable reductions in total costs of care.

 Because disease management programs emphasize evidence- or guideline-based medicine, they are also limited by the lack of medical evidence in many areas of clinical care.

Another limitation of disease management programs is that they are not designed to address nonclinical determinants of clinical, quality, and cost outcomes. For example, hyperglycemia and hypoglycemia in insulin-dependent patients with diabetes can result from food insecurity—an issue that cannot be addressed by clinical strategies such as insulin titration. The clinical focus of these care management programs excludes social and other determinants of health.

Like disease management, utilization management is not a new concept. However, interest in these programs has increased significantly as more clinicians and organizations have gained the ability to collect utilization information (eg, through advances in data capture via the electronic health record and interoperability via admission/discharge/transfer, or ADT, messaging) and the incentives to reduce utilization under value-based payment and delivery reforms. Rather than maximizing guideline-concordant care, the core main function of this care management activity is to reduce unnecessary health care utilization, particularly high-cost utilization associated with emergency department (ED), hospital, and postacute care facility care.

Given this focus, utilization management programs target patients at risk for high utilization (sometimes called “super-utilizers”) and emphasize core competencies such as data analytics and robust clinical triage capabilities, which can be used to identify and intervene in the care of super-utilizers. The specific strategy for triage can vary, with some organizations embedding care managers within their primary care infrastructure to reduce utilization related to ambulatory care—sensitive conditions,

 whereas others station utilization managers in EDs to more directly curb inappropriate hospitalization.

The major limitation of utilization management programs is the risk of inadvertently deterring appropriate care. No clinical triage processes are perfect, and it can be challenging to prospectively determine when ED or acute hospital care is medically necessary. Moreover, these programs do not inherently distinguish between clinical and social drivers of utilization. As organizations seek to discern that for themselves, underemphasis on social drivers can lead to disappointing results, while overemphasis on super-utilizer patients can create underinvestment in programs targeted to average- or low-risk individuals—interventions that have similar if not greater ability to curb unnecessary utilization.

 Finally, although patient experience is a key element of value-based care, organizations that heavily emphasize utilization management may be perceived as “gatekeepers” that lack patient-centered processes.

As a method for delivering patient-centered care, health care navigation has received increased attention in the era of value-based payment. Organizations have created new roles for “health navigators”

 or “health coaches” who focus on helping to remove structural barriers for patients who have difficulties in accessing and/or coordinating care.

 As another example of health care navigation, primary care clinics have enabled direct referrals to community-based programs to address adverse social determinants of health.

 By improving individuals’ experience and/or ability to access care, navigation-focused care management programs can support success in value-based payment arrangements via greater patient engagement and continuity.

However, because navigation seeks to address individualized barriers to care, there are also risks to broad implementation within value-based payment initiatives. In reducing barriers to care, organizations may spur utilization that unintentionally counteracts efforts to manage costs of care. The potential benefits from patient navigation and engagement may also be undercut by features of payment model design. For instance, most heavily emphasize quality and cost outcomes over patient experience outcomes, reducing the degree to which improved navigation is reflected in payment model performance.

As a broadly defined care delivery strategy, care management can improve outcomes. However, the results to date have been variable, particularly within value-based payment models. One way to improve the design, implementation, and potential benefits of care management programs is to utilize a framework that distinguishes among different types of activities, functions, and goals encompassed by the concept of care management.

 Department of Medicine (LMM, JML) and Value and Systems Science Lab (LMM, JML), University of Washington School of Medicine, Seattle, WA; Leonard Davis Institute of Health Economics, University of Pennsylvania (JML), Philadelphia, PA.

 The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (LMM, JML); drafting of the manuscript (LMM); critical revision of the manuscript for important intellectual content (LMM, JML); administrative, technical, or logistic support (LMM); and supervision (JML).

Leah M. Marcotte, MD, Department of Medicine, University of Washington, 1959 NE Pacific St, Seattle, WA 98105. Email: leahmar@uw.edu.

1. Baron RJ. New pathways for primary care: an update on primary care programs from the innovation center at CMS. 

. 2012;10(2):152-155. doi:10.1370/afm.1366

2. Edwards ST, Landon BE. Medicare’s chronic care management payment—payment reform for primary care. 

. 2014;371(22):2049-2051. doi:10.1056/NEJMp1410790

3. Bindman AB, Blum JD, Kronick R. Medicare’s transitional care payment—a step toward the medical home. 

. 2013;368(8):692-694. doi:10.1056/NEJMp1214122

4. Lewis VA, Schoenherr K, Fraze T, Cunningham A. Clinical coordination in accountable care organizations: a qualitative study. 

. 2019;44(2):127-136. doi:10.1097/HMR.0000000000000141

5. 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. 

. 2017;36(5):876-884. doi:10.1377/hlthaff.2016.0922

6. Ouayogodé MH, Mainor AJ, Meara E, Bynum JPW, Colla CH. Association between care management and outcomes among patients with complex needs in Medicare accountable care organizations. 

. 2019;2(7):e196939. doi:10.1001/jamanetworkopen.2019.6939

7. Donelan K, Barreto EA, Michael CU, et al. Variability in care management programs in Medicare ACOs: a survey of medical directors. 

. 2018;33(12):2043-2045. doi:10.1007/s11606-018-4609-1

8. Aboumatar H, Naqibuddin M, Chung S, et al. Effect of a hospital-initiated program combining transitional care and long-term self-management support on outcomes of patients hospitalized with chronic obstructive pulmonary disease: a randomized clinical trial. 

 2019;322(14):1371-1380. doi:10.1001/jama.2019.11982

9. McWilliams JM, Chernew ME, Landon BE. Medicare ACO program savings not tied to preventable hospitalizations or concentrated among high-risk patients. 

. 2017;36(12):2085-2093. doi:10.1377/hlthaff.2017.0814

10. Piro L, Doctor J. Managed oncology care: the disease management model. 

11. Bertel O, Conen D. Individualized care in patients with chronic congestive heart failure. 

. 1987;9(suppl 2):S68-S72. doi:10.1097/00005344-198700002-00014

12. Barlow J, Wright C, Sheasby J, Turner A, Hainsworth J. Self-management approaches for people with chronic conditions: a review. 

. 2002;48(2):177-187. doi:10.1016/s0738-3991(02)00032-0

13. Taliani CA, Bricker PL, Adelman AM, Cronholm PF, Gabbay RA. Implementing effective care management in the patient-centered medical home. 

14. Shumway M, Boccellari A, O’Brien K, Okin RL. Cost-effectiveness of clinical case management for ED frequent users: results of a randomized trial. 

. 2008;26(2):155-164. doi:10.1016/j.ajem.2007.04.021

15. Marcotte LM, Reddy A, Liao J. Addressing avoidable healthcare costs: time to cool off on hotspotting in primary care? 

. 2019;34(11):2634-2636. doi:10.1007/s11606-019-05285-z

16. McGough P, Kline S, Simpson L. Team care approach to population health and care management. 

. 2017;22(2):93-103. doi:10.1108/IJHG-11-2016-0048

17. Lawson K. The four pillars of health coaching: preserving the heart of a movement. 

. 2013;2(3):6-8. doi:10.7453/gahmj.2013.038

18. Smith AM, Zallman L, Betts K, et al. Implementing an electronic system to screen and actively refer to community based agencies for food insecurity in primary care. 

. Published online October 12, 2019. doi:10.1016/j.hjdsi.2019.100385

To support effective care management programs in the context of value-based care, we propose a framework categorizing care management as disease management, utilization management, and care navigation interventions.

What We Talk About When We Talk About Care Management

The American Journal of Accountable Care. 2020;8(1):3-5. 

Over the past decade, accountable care organizations (ACOs) have expanded across payers, including Medicare, Medicaid, and commercial insurers. Participation among provider organizations has also increased over time, to the point where companies have emerged with the sole purpose of helping health systems succeed in prominent ACOs, such as the Medicare Shared Savings Program.

Although ACOs all possess a set of core features that define the payment model, differences across payer types and patient populations mean that no single strategy will be effective for all types of ACOs. Provider organizations would be best served considering and participating in ACOs after they have considered these differences and their implications for care delivery initiatives that can support ACO success. In this article, we leverage our perspective as physician leaders in an integrated delivery system that participates in multiple ACOs and discuss 3 factors—clinical complexity of ACO populations, social complexity of ACO populations, and ACO network configuration—to advance the discourse about payment model strategy and performance.

Certain ACO strategies are more or less likely to be successful based on the clinical complexity of the covered population. For instance, approximately two-thirds of Medicare beneficiaries have at least 2 chronic conditions,

 whereas the majority of beneficiaries in many commercial patient populations have no chronic medical diagnoses. This difference poses significant implications for how an ACO participant organization might approach 2 crucial types of activities: assessing patients’ clinical risk (ie, risk capture) and using risk information to implement care management services (ie, risk-stratified care management).

For Medicare beneficiaries, many high-cost conditions as defined by CMS’ Hierarchical Condition Categories (HCC) model are chronic and will remain stable year over year. As a result, providers can implement risk-capture activities within Medicare ACOs via approaches such as clinical decision support within the electronic health record to populate past diagnoses and prompt clinicians to affirm or update them.

In contrast, the lower chronic disease burden among beneficiaries in most commercial insurance plans undercuts the potential benefit of such approaches in commercial ACOs. In particular, the high-cost conditions for these patients, as defined by the HHS HCC model, will be acute (eg, trauma) and therefore less likely to be stable over time (ie, less likely to be relevant in even the next calendar year). Therefore, using clinical decision support prompts based on historical diagnoses can create more “noise” than “signal,” thus causing workflow inefficiencies for clinicians.

This logic also carries implications for risk-stratified care management within ACOs and, in particular, activities directed at high-cost patients. Differences in the clinical complexity of different ACO populations exacerbate a well-established issue with implementing risk-stratified care management: that there is inherent instability in high-cost populations, with individuals cycling in and out of these populations year to year.

Given this dynamic, care management interventions will adopt a different form within Medicare populations (eg, heavy focus on self-management of multiple chronic conditions) versus commercial populations (eg, preventive services to avoid catastrophic events such as motor vehicle accidents). Without these distinctions, provider organizations in ACOs may not achieve anticipated benefits from care management programs because of variation across different ACO populations.

Social factors play an outsized role in determining many health outcomes, serving as a larger driver than even healthcare factors.

 In turn, the extent of an ACO population’s social complexity carries major implications for provider organizations’ payment model strategy and performance.

For instance, because of their predictable coverage over an extended period, Medicare fee-for-service populations tend to be relatively stable in composition. In turn, condition-based care management can be used within Medicare ACOs to deliver guideline-concordant services that not only improve outcomes but also address utilization and costs as core ACO metrics. In contrast, this approach may be less effective in populations who are less stable over time and/or marked by higher social complexity, such as those in Medicaid. Many Medicaid beneficiaries have high costs as a result of social determinants (eg, housing and/or food insecurity) rather than or in addition to clinical determinants (eg, suboptimal medication titration). Delivering clinical interventions aimed at disease management fails to address nonhealthcare root causes of utilization and costs.

The extent of a population’s social complexity may determine how patients access care and, in turn, how organizations implement interventions to perform in ACOs. For example, comprehensive primary care—which focuses on activities such as chronic and acute care management, care coordination, and care continuity—has been cited as a critical strategy to cost containment within ACOs.

 However, the value of this approach is based on the assumption that patients’ primary healthcare access point is via a comprehensive and often brick-and-mortar primary care infrastructure. Although this may be true for certain groups, such as older Medicare populations, it may not be for others.

For instance, it may not be the norm for employed, less socially complex patients in commercial populations to access care in this way, raising the potential for organizations in commercial ACOs to better meet patient needs via technological approaches such as telemedicine, which can expand access while circumventing some costs of physical locations altogether. As another example, populations with high social complexity frequently access nonemergent care via the emergency department. Consequently, ACOs may be better off investing in fast-track triage and colocated urgent care and primary care facilities rather than comprehensive primary care or telemedicine infrastructure.

Just as ACO strategy should be dictated in part by differences in patients’ clinical and social complexity, it will also likely differ as a result of which providers are delivering care to patients (ie, the ACO network configuration). One size does not fit all because of differences in the networks that are needed in different ACOs.

For instance, network configuration can influence provider organizations’ approaches to managing clinical referrals among different types of clinicians—a core ACO strategy for maintaining a stable, engaged patient population and addressing their utilization, quality, and costs.

 For Medicare patients with multiple chronic conditions, ACOs may require strong networks of primary and subspecialty care within which patients can be referred and comanaged.

In contrast, for commercial populations with less need for tight primary-subspecialty coordination, it may be more strategic for ACO networks to ensure inclusion of providers who are delivering occupation-aligned services (eg, physical therapy providers for labor workers, on-site services that can be accessed at the site of work). Given the extent to which some Medicaid patients receive care from emergency departments, organizations interested in Medicaid ACOs would be well served by carefully considering whether and how to include freestanding and hospital-related sites within their networks.

Differences in network configuration can also affect organizational care standardization activities within ACOs. As methods for reducing unwarranted care variation, clinical pathways can be a promising strategy to implement within ACOs.

 However, whereas ACOs with networks emphasizing primary care providers may elect to develop and implement pathways around prevalent chronic conditions encountered in primary care (eg, diabetes, hypertension), ACOs with networks encompassing quaternary care sites may opt to instead prioritize subspecialty pathways.

Although several core features define ACOs as a payment model, several differences (eg, clinical and social complexity of patient populations, ACO network configuration) highlight why specific strategies may have variable effectiveness in different types of ACOs. Provider organizations would be well served by accounting for these issues as they consider ACO participation.

Department of Medicine, University of Washington School of Medicine (LMM, JML), Seattle, WA; Value and System Science Lab (LMM, JML), Seattle, WA; Leonard Davis Institute of Health Economics, University of Pennsylvania (JML), Philadelphia, PA.

Concept and design (LMM, JML); drafting of the manuscript (LMM); critical revision of the manuscript for important intellectual content (LMM, JML); administrative, technical, or logistic support (LMM); and supervision (JML).

Leah M. Marcotte, MD, Department of Medicine, University of Washington School of Medicine, 1107 NE 45th St, Ste 355, Seattle, WA 98105. Email: leahmar@uw.edu.

1. Powers BW, Mostashari F, Maxson E, Lynch K, Navathe AS. Engaging small independent practices in value-based payment: building Aledade’s Medicare ACOs. 

. 2018;6(1):79-87. doi: 10.1016/j.hjdsi.2017.06.003.

2. Multiple chronic conditions. CMS website. cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Chronic-Conditions/MCC_Main. Updated April 5, 2019. Accessed November 24, 2019.

3. Marcotte LM, Reddy A, Liao J. Addressing avoidable healthcare costs: time to cool off on hotspotting in primary care? 

. 2019;34(11):2634-2636. doi: 10.1007/s11606-019-05285-z.

4. Schroeder SA. We can do better — improving the health of the American people. 

. 2007;357(12):1221-1228. doi: 10.1056/NEJMsa073350.

5. McClellan M, McKethan AN, Lewis JL, Roski J, Fisher ES. A national strategy to put accountable care into practice. 

. 2010;29(5):982-990. doi: 10.1377/hlthaff.2010.0194.

6. Song Z, Sequist TD, Barnett ML. Patient referrals: a linchpin for increasing the value of care. 

. 2014;312(6):597-598. doi: 10.1001/jama.2014.7878.

7. Marcotte LM, Chen A, Hwang CS, Liao JM. Primary care—oriented pathways to support population-based payment models. 

. 2019;5(7):39-42. doi: 10.25270/jcp.2019.09.00092.

Differences in patients’ clinical and social complexity and accountable care organization (ACO) network configuration highlight why specific strategies may have variable effectiveness in different types of ACOs.

Leah M. Marcotte, MD

Articles

Physician Perspectives: How the Merit-based Incentive Payment System Improves Value

What We Talk About When We Talk About Care Management

One Size Will Not Fit All: Factors That Drive Strategy in Accountable Care Organizations