Advancing Health Equity: Principles to Address the Social Determinants of Health in Alternative Payment Models

As health care continues to transition to a value-based environment, there has been a growing call for the inclusion of social determinants of health (SDoH) as a criterion in advanced primary care delivery and value-based payment arrangements. Academic literature is beginning to show how significantly social determinants affect the health and well-being of patients.

The American Academy of Family Physicians (AAFP) defines SDoH as the conditions under which people are born, grow, live, work, and age. In their patient-centered practices, family physicians identify and address the SDoH for individuals and families, incorporating this information into the biopsychosocial model to promote continuous healing relationships, wholeperson orientation, family and community context, and comprehensive care. The AAFP supports the assertion that physicians need to know how to identify and address SDoH to be successful in promoting positive health outcomes for individuals and populations.1

Opportunities to Advance Value-Based Payment and Measurement Methodologies

Research on the impact of social risk factors on health status and outcomes, coupled with the movement toward value-based payment, has created new policy opportunities and imperatives to address SDoH. While value-based payment programs often require physicians to assess and address SDoH, these models do not adjust or account for how these factors affect outcomes and performance. However, federal and state policy makers and key stakeholders are examining the issue more closely to determine how alternative payment models (APMs) can account for SDoH, and support physicians in advancing quality and patient outcomes and reducing costs.

Activity at the federal level has increased in recent years as more information on the effect of value-based payments on different patient populations and physicians emerges. Early programs, such as the Hospital Readmissions Reduction Program (HRRP), highlighted the impact of financial penalties on safety-net providers who often care for populations that have higher levels of social risk factors. In December 2016, the U.S. Department of Health and Human Services (HHS) and the Assistant Secretary for Planning and Evaluation (ASPE) released a congressionally mandated report that focused on the connections between social risk factors and performance in Medicare value-based payment programs in order to align payments with the goals of these programs.2 Most recently, passage and implementation of the Medicare Access and CHIP Reauthorization Act (MACRA) has prompted the development and testing of new APMs, underscoring the need to understand and account for the role of SDoH in physician assessment and payment. In response to new research, the Centers for Medicare & Medicaid Services (CMS) is also examining ways to account for social risk factors and reduce health disparities in its quality measurement programs.3

As more AAFP members participate in APMs, key issues for the AAFP include data on the role of social risk factors in health outcomes, the impact of such data on assessing physician performance, and policy opportunities to improve payment and measurement methodologies. The AAFP has developed five principles regarding SDoH to guide its assessment of APMs and value-based payment initiatives. These principles ensure that SDoH are appropriately accounted for in the payment and measurement design of APMs so that practices have adequate support to improve quality and outcomes for all patients, eliminate health disparities, and reduce costs for the health care system.


Principle #1: APMs should support practices’ efforts to identify and address social determinants that are shown to impact health outcomes.

  • Payment risk-adjustment methodologies should include multiple variables representing social determinants linked to health outcomes.
  • Variables included in any payment methodology should be evidence-based and demonstrably representative of SDoH.

Social determinants of health are the conditions in which people are born, grow, live, work, and age. They are shaped by the distribution of money, power, and resources at global, national, and local levels and lead to differences in health status among patients with otherwise similar demographic and physical characteristics. Social determinants of health are multifactorial, and these factors have been shown to have “marked associations with risks for different illnesses, life expectancy, and lifetime morbidity.”4

The type and quantity of health care resources needed to care for patients varies according to their differences in health status. Therefore, payers use risk-adjustment methodologies to allocate financial resources to practices commensurate with the health status of their patient population. Practices incur costs to identify and address SDoH for their patient population, and some payers are beginning to incorporate variables representing SDoH into their riskadjustment methodologies. Adjusting payments based on one or more of these variables improves payment to practices and the consequent allocation of resources to identify and address SDoH in their patient population.5 Models that incorporate variables representing SDoH should leverage existing data collection mechanisms to minimize new burdens on practices.

As variables representing SDoH are incorporated more systematically into payment models, reliance on any single variable as a proxy for SDoH should be discouraged. For example, one study of commercially insured children and adolescents showed that higher socioeconomic background was associated with greater levels of health care spending.6 Consequently, including socioeconomic information in risk-adjustment algorithms could potentially direct funds away from physicians caring for children and adolescents who are from lower socioeconomic backgrounds and are at greater risk of poor health. Risk-adjustment methodologies should allocate more resources to those most disadvantaged by SDoH.

Principle #2: The incorporation of variables representing SDoH in APMs should be founded on evidence-based research methods.

Practices should use additional risk-adjusted payments that account for SDoH to address health disparities in their patient population by supporting increased access, expanded population-based services, referrals, and comprehensive care.

  • Evidence-based measurement of SDoH should be used by payers to provide resources to support additional services that patients with social risk may need.
  • Measurement for payment adjustments should be made at a standardized geographic level, such as a census tract, census block group, or primary care service area.
  • At a minimum, measurement should include the following: poverty, unemployment, household provider status, high-need age group (i.e., 17 years of age or younger; 65 years of age or older), education level, transportation, crowding, uninsured status, and race.
  • Performance measurement should be risk-adjusted for SDoH when there is a clear relationship between social risk and health outcomes.

In the United States, there is a desire to use measurement of SDoH to guide the implementation of interventions at the clinic/community level and appropriate allocation of resources. This guidance may be carried out through policies or payment adjustments that support practices’ efforts to improve outcomes for patients whose social risk factors increase their risk for poor health outcomes. The HHS formed a committee to explore options for the inclusion of social risk factors in Medicare value-based payment. The committee presented five reports over 15 months and released its final report in 2017, outlining four categories that could be used individually or combined to include social risk factors in payment methods:7

  1. Stratified public reporting of quality and outcome measures
    2. Adjustment of performance measure scores
    3. Direct adjustment of payment
    4. Restructuring of payment incentive design

Exploring the foundational aspects of these categories provides insight into the most effective methods for measuring and evaluating SDoH data to improve outcomes for patients and increase payments for clinicians. Many other nations use data to improve health outcomes at the community level. For example, the United Kingdom and New Zealand use data on material and social deprivation to create indices that measure socioeconomic variation across geographic areas and communities.8 The indices provide universal guidance to fundamental principles for population health, allowing the United Kingdom and New Zealand to more effectively provide services, allocate funding, guide research, and implement policy.

In the United States, efforts exist to collect and analyze SDoH data, although these data are not yet effectively integrated into policy, primary care delivery, and payment. These efforts build on work examining the geographic breakdown of health care services and population health outcomes. Prime examples of geographic stratification are Primary Care Service Areas (PCSAs) and the Public Health Disparities Geocoding Project, both of which use national data to evaluate health outcomes in the United States at a geographic level. PCSAs are based on ZIP code data and characterize utilization-based service areas that reflect the travel of Medicare beneficiaries to primary care clinics.9 They can be used to identify areas of health care underservice and related health outcomes. The Public Health Disparities Geocoding Project compared area-based socioeconomic measures at three different geographic levels: census block group level, census tract level, and ZIP code level.10 Analysis suggested that measuring “percentage of persons below poverty” at the census tract level could meaningfully augment U.S. public health surveillance systems to monitor socioeconomic inequalities in health.

One way to expand on the geographic analyses described above is to create an index that can easily be used to measure the social deprivation of specific areas. The Agency for Toxic Substances & Disease Registry’s (ATSDR’s) Social Vulnerability Index(svi.cdc.gov) and The Robert Graham Center for Policy Studies in Family Medicine and Primary Care’s Social Deprivation Index are examples of indices that, with more testing, could be valid measures to assess resource allocation in the United States.11,12 Comparison of the two indices reveals commonalities among the measures used for index development, including the following: poverty, unemployment, household provider status, high-need age group (i.e., 17 years of age or younger; 65 years of age or older), education level, transportation, crowding, uninsured status, and race. All measures suggested can be obtained from national, state, and/or local data sources and do not require additional reporting by the physician.

As noted, many efforts to measure SDoH exist in the United States, but few payers are using these data to adjust payment to practices and physicians. To date, two states—Minnesota and Massachusetts—are examining the role of SDoH in their Medicaid programs and trying to account for these factors. In 2015, Minnesota’s state legislature directed the Medicaid program to develop a payment methodology that increases payment to health care providers serving patients who have elevated social risk.13 Early work in the state focused on gathering qualitative information to identify which social risk factors are most predictive of poor health outcomes. A report to the Minnesota state legislature identified six social risk factors: substance use disorder, serious mental illness, housing instability, prior incarceration, deep poverty, and child protection involvement. These risk factors informed a new payment methodology. Participating practices receive a population-based per member per month payment based on the six social risk factors and a clinical/medical risk score that uses the Adjusted Clinical Groups model.

Massachusetts launched its Social Determinants of Health Model in 2016.14 The model uses enhanced risk adjustment to determine a per member per month payment for each participating program. Researchers in Massachusetts have found that this risk-adjustment model performs at the high end of the best-performing prospective models in Medicaid populations when used to predict cost for future years. Future work will include serious mental illness and substance use disorder measures to further increase the predictive power of the model.

These examples in Minnesota and Massachusetts show that measurement of SDoH is possible. They also show that inclusion of SDoH in payment and care delivery models presents an opportunity to increase payment to practices so that they can improve the quality of health care services to patients who have social risk factors and often experience poor outcomes. These developments underscore the opportunity to advance performance measurement and payment to address SDoH and broadly impact health disparities.

The National Quality Forum (NQF) and the National Academy of Medicine (NAM) have examined the inclusion of SDoH in performance measurement and payment methodologies. In 2015, the NQF convened a two-year trial period to examine whether performance measures should account for social risk factors to ensure fair, accurate comparisons of provider performance.15 According to the NQF’s final report, the trial period showed that adjusting measures for SDoH is “feasible, but challenging” and that SDoH should be included in risk adjustment of performance when there is a clear relationship between a social risk factor and a health outcome. The NAM produced a series of consensus-based reports on the issue, concluding that taking SDoH into account in quality measurement and payment design could improve quality, reduce costs, and address a range of health disparities.16

Principle #3: Health information technology (HIT) platforms should facilitate SDoH data collection from medical records and other sources to support improved clinical decision making, care coordination, quality measurement, and population health management.

  • Public and private payers should provide practices with relevant claims and enrollment data for use within the electronic health record (EHR).
  • SDoH data should be defined, stored, and transmitted according to a defined set of standards that place minimal burden on the practice.
  • Data should be accessible at the point of clinical encounter and easily integrated with other data from the EHR17.
  • Data should be readily transferable to other health care providers and community outreach programs.

The AAFP recognizes that practices will have to make changes to facilitate SDoH data collection. Any standards or changes that require the use of SDoH data should place minimal burden on practices. A significant amount of relevant data already exists outside the practice in various sources, such as all-payer claims databases, health plan enrollment, public social service case data, and self-reported patient data. If these “Big Data” sources are shared with practices, they have clinical utility for individual patient and population health interventions without the need for additional data capture by practices.

For data to be actionable and impactful, they must be easily accessible during the clinical encounter and within the patient record. In addition, payment models should include financial support to facilitate data collection and resource allocation to address SDoH within physician patient panels.

Patient data related to social and behavioral factors can improve clinical care and patient satisfaction, as well as being useful in research and public health initiatives. These data can support more personalized and effective clinical decision making. At the population level, information about patients’ social risk factors allows for more targeted care management and coordination programs.18 Payment methodologies should include standardized measures for storing and collecting all health data, including SDoH, for use in shared decision making, quality improvement, research, and population health management. Establishing the framework for a common data model that includes SDoH and using predictive analytics can potentially provide physicians with better information that leads to better patient outcomes and improved population health19.    

Standardized measures should be evidence-based and customizable to meet the needs of physicians and patients. In 2013, the Institute of Medicine (IOM) released an evidence-based set of core social and behavioral health domains and corresponding measures that should be included in all EHRs.20 The IOM and other health policy organizations continue to study and refine these measures to enhance the data available to improve health outcomes for individuals and communities. When leveraged correctly, EHRs can improve the integration of SDoH into clinical systems.21

Principle #4: To minimize administrative burden on providers and patients, SDoH data should be collected by leveraging existing mechanisms.

  • Variables representing SDoH in payment risk-adjustment formulas should use existing data collection infrastructure and mechanisms.
  • Use of SDoH within payment models and care delivery settings should be transparent and harmonized.

Primary care physicians must comply with a daunting regulatory framework that often puts administrative burdens on them. In 2017, nearly 40% of AAFP members reported participating with 10 or more payers in the past 12 months.22 Standardization is not required among public or private payers, so practices are forced to learn and navigate the rules and forms of each payer. They spend countless hours reviewing documents and checking boxes to meet the requirements of health insurance plans. This time could be better spent caring for patients.

The regulatory framework for practices has driven operating costs up and reduced face time with patients. Administrative and regulatory burden is one of the top reasons independent practices close and is a leading cause of physician burnout. Despite the good intent of underlying health care policies, the burden has expanded to an untenable level and is a significant barrier to achieving the Quadruple Aim.23 For these reasons, the addition of SDoH to APMs should aim to decrease burden on practices and minimize any new burden.

While EHRs have the potential to collect data on SDoH for various purposes and integrate them into clinical systems, a significant amount of relevant data already exists outside the practice in various sources, such as all-payer claims databases, health plan enrollment, public social service case data, and self-reported patient data. Measures must be clearly defined and standardized for all health data before the EHR is programmed to incorporate SDoH.24 Thus, payers and HIT vendors are encouraged to use a consistent definition of SDoH and harmonize the variables and measures used to represent SDoH in risk-adjustment methodologies. Payers should be transparent in their incorporation of SDoH in their risk-adjustment methodologies and update them regularly or when new evidence is developed.

Principle #5: To ensure APMs improve access, quality, and health equity, practices should receive appropriate resources and support to identify, monitor, and assess SDoH.

  • APMs that provide increased payment should also provide support to practices to allow for innovation in patient care delivery and beyond to meet the needs of patient populations.
  • Support provided to practices may include the following:

o   Connection to EHR vendors dedicated to supporting the identification, monitoring, and assessment of SDoH
o   Access to relevant data sources outside the practice
o   Peer-to-peer collaboration, such as an online platform or in-person learning sessions
o   Training and technical assistance (TA)
o   Feedback reports and coaching on how to understand and use SDoH data to improve care delivery and health outcomes for their patient population

To adequately address the social needs tied to their patient population’s health outcomes, practices need resources and TA, in addition to increased funding. A systematic review of six primary care initiatives, which were convened by either a state entity or CMS, found that providing TA for practices had a positive effect on most outcomes; the form of assistance provided varied by outcome.25 The Comprehensive Primary Care Plus (CPC+) model is the first APM to provide this level of TA to help participating practices meet its increased care delivery requirements. Current CPC+ practices receive TA through national and regional learning contractors and can connect with and learn from other CPC+ practices through an online platform. Other support provided by CPC+ includes the following: web-based learning sessions; in-person learning sessions; EHR affinity groups; and regional practice coaching and facilitation.

Practices participating in CPC+ also receive feedback reports comparing their demographic, cost, and utilization data with data from other CPC+ practices in the region. Much of the information provided in the feedback reports is pulled from claims data. In future endeavors, APMs should continue this trend and rely on variables from extant data sources, such as claims (including all-payer claims databases, where available), health plan enrollment, public social service case data, and self-reported patient data. In addition, payers should share such data transparently with practices so the data can be integrated into the EHR to help practices identify and address SDoH in their patient population.26 By providing appropriate resources and support, APMs can help physicians and medical homes learn best practices and use data-driven approaches to meet the needs of their patient population.

References:

  1. American Academy of Family Physicians. Social determinants of health. [Policy statement]. https://www.aafp.org/about/policies/all/social-determinants.html. Accessed March 5, 2018.
  2. U.S. 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 purchasing programs. Washington, DC: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation; 2016. https://aspe.hhs.gov/system/files/pdf/253971/ASPESESRTCfull.pdf(aspe.hhs.gov). Accessed March 5, 2018.
  3. Centers for Medicare & Medicaid Services. CMS Quality Measure Development Plan: supporting the transition to the Quality Payment Program. Baltimore, Md.: Centers for Medicare & Medicaid Services; 2017. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Value-Based-Programs/MACRA-MIPS-and-APMs/2017-CMS-MDP-Annual-Report.pdf(www.cms.gov).  Accessed March 5, 2018.
  4. Pan American Health Organization. Determinants of health. http://www.paho.org/hq/index.php?option=com_content&view=article&id=5165&Itemid=3745&lang=en(www.paho.org). Accessed March 5, 2018.
  5. 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.
  6. Chien AT, Newhouse JP, Iezzoni LI, Petty CR, Normand ST, Schuster MA.  Socioeconomic background and commercial health plan spending. Pediatrics. 2017;140(5):e20171640.
  7. National Academies of Sciences, Engineering, and Medicine. Accounting for social risk factors in Medicare payment. Washington, DC: The National Academies Press; 2017.
  8. Phillips RL, Liaw W, Crampton P, et al. How other countries use deprivation indices—and why the United States desperately needs one. Health Aff (Millwood). 2016;35(11):1991-1998.
  9. Goodman DC, Mick SS, Bott D, et al. Primary care service areas: a new tool for the evaluation of primary care services. Health Serv Res. 2003;38(1):287-309.
  10. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Race/ethnicity, gender, and monitoring socioeconomic gradients in health: a comparison of area-based socioeconomic measures—the Public Health Disparities Geocoding Project. Am J Public Health. 2003;93(10):1655-1671.
  11. Butler DC, Petterson S, Phillips RL, Bazemore AW. Measures of social deprivation that predict health care access and need within a rational area of primary care service delivery. Health Serv Res. 2013;48(2 Pt 1):539-559.
  12. Field K. Measuring the need for primary health care: an index of relative disadvantage. Applied Geography. 2000;20(4):305-332.
  13. State Health and Value Strategies. Medicaid and social determinants of health: adjusting payment and measuring health outcomes. https://www.statenetwork.org/wp-content/uploads/2017/07/SHVS_SocialDeterminants_HMA_July2017.pdf(www.statenetwork.org). Accessed March 5, 2018.
  14. State Health and Value Strategies. Medicaid and social determinants of health: adjusting payment and measuring health outcomes. https://www.statenetwork.org/wp-content/uploads/2017/07/SHVS_SocialDeterminants_HMA_July2017.pdf(www.statenetwork.org). Accessed March 5, 2018.
  15. National Quality Forum. Evaluation of the NQF’s trial period for risk adjustment for social risk factors. Washington, DC: National Quality Forum; 2017.
  16. National Academies of Sciences, Engineering, and Medicine. Accounting for social risk factors in Medicare payment. Washington, DC: The National Academies Press; 2017.
  17. Hripcsak G, Forrest CB, Brennan PF, Stead WW. Informatics to support the IOM social and behavioral domains and measures. J Am Med Inform Assoc. 2015;22(4):921-924.
  18. Hripcsak G, Forrest CB, Brennan PF, Stead WW. Informatics to support the IOM social and behavioral domains and measures. J Am Med Inform Assoc. 2015;22(4):921-924.
  19. AMA, IBM's Watson, and Others Partner to Pursue Interoperability. http://www.hcanews.com/news/ama-ibms-watson-and-others-partner-to-pursue-interoperability(www.hcanews.com). Accessed March 5, 2018.
  20. Institute of Medicine. Capturing social and behavioral domains and measures in electronic health records: phase 2. Washington, DC: The National Academies Press; 2014.
  21. Gottlieb LM, Tirozzi KJ, Manchanda R, Burns AR, Sandel MT. Moving electronic medical records upstream: incorporating social determinants of health. Am J Prev Med. 2015;48(2):215-218.
  22. American Academy of Family Physicians. 2017 Value-Based Payment Survey data brief. http://humananews.com/wp-content/uploads/2017/11/Data-Brief2017_Value-Base_FINAL4.pdf(humananews.com). Accessed March 5, 2018.
  23. American Academy of Family Physicians. Principles for administrative simplification. [Policy statement]. https://www.aafp.org/about/policies/all/principles-adminsimplification.html. Accessed March 5, 2018.
  24. Hripcsak G, Forrest CB, Brennan PF, Stead WW. Informatics to support the IOM social and behavioral domains and measures. J Am Med Inform Assoc. 2015;22(4):921-924.
  25. Centers for Medicare & Medicaid Services. Systematic review of CMMI primary care initiatives final report. https://innovation.cms.gov/Files/reports/primarycare-finalevalrpt.pdf(innovation.cms.gov). Accessed March 5, 2018.
  26. Blewett LA, Owen RA. Accountable care for the poor and underserved: Minnesota’s Hennepin Health model. Am J Public Health. 2015;105(4):622-624.

(April 2018 BOD)