Is Medical Debt a Social Determinant of Health?

Two major frameworks exist for social determinants of health. First, the World Health Organization (WHO) lists ten social determinants of health: social gradient; stress; early life; social exclusion; work; unemployment; social support; addiction; food, and; transport¹. Second, the Centers for Disease Control (CDC) offer a five-part framework: economic stability; education; social and community context; health and healthcare, and; neighborhood and built environment. Both frameworks complement one another and enhance our understanding of potential social determinants of health.

Health policy here in the United States needs to be informed at both international and national levels. The WHO framework helps us to explore social determinants comparatively from an international perspective, while CDC’s framework is tailored to localized contexts unique to the USA. However, neither framework explicitly lists medical debt as a social determinant of health. Were medical debt included in the CDC model, it might be listed in at least two of the five major categories: economic stability, and health and healthcare. However, neither of these broad level categories include medical debt as a subset.

Medical debt is particularly unique when compared to factors listed in either the CDC’s or WHO’s frameworks. In particular, medical debt can be inordinately impacted by a single catastrophic cost health event². Also, medical debt is a social determinant that can impact individuals from almost all income ranges. Finally, medical debt, even when it is not a single catastrophic event, is impacted by healthcare inflation, which far outpaces general inflation of other costs of living (3.00% vs. 1.88% from 2011 to 2016, respectively)^3,4. Based on these distinctions, we believe medical debt should be explored as a potential social determinant of health, thus informing data-driven interventions in healthcare.

One point-of-entry for exploring this phenomenon is Medicaid. While implementation of the Affordable Care Act (ACA) led to increased health coverage for nearly 32 million Americans⁵, the umbrella of ACA still left over 28 million without insurance⁶. Millions of Americans claim inadequate coverage of services or exorbitant premiums². To truly understand the financial burden of healthcare, one should not only account for insurance premiums, but also out-of-pocket costs and monetary funds owed for previous services⁷. While insurance coverage can contribute to medical debt, one must also appreciate that provider’s sky-rocketing costs for procedures and prescriptions exacerbate this problem². TransUnion’s Healthcare Report in 2016 stated that 77% of bills greater than $500 went unpaid in 2016, and over 51% of these patients owed more than $1000 to their providers⁸.

One of the tenets of ACA was the expansion of Medicaid programs via federal funding. This allowed state-wide Medicaid programs to increase health insurance to individuals previously not qualified. Medicaid expansion led to a substantial 5.3% decrease in medical debt to 16.0% nationally in individuals under the age of 65 who had trouble paying medical bills⁹. Some may argue insurance coverage has resolved the issue of medical debt. However, it has yet to be seen what happens to individuals, who despite gaining insurance under Medicaid expansion, must account for medical debt due to the program’s eventual phase out. Despite Medicaid expansion’s merits or limitations as a solution to medical debt, implementation of Medicaid expansion provides a proxy for debt reduction, thus exploring its association with health outcomes. Admittedly, there are multiple factors impacting health outcomes for patients receiving Medicaid. This challenge emphasizes further the need for fuller datasets that allow Big Data to statistically parse out the complex interdependence of determinants of health outcomes.

Albeit inconclusive, some evidence suggests medical debt may manifest as a social determinant of health. Medical debt can be considered a risk factor for reduced health access and poorer health status in individuals, with these individuals rationing medical care^4,10. In one study, 24% of individuals with medical debt abstained from going to their original site of treatment, with 18% delaying care when necessary, and 10% only visiting Emergency Departments to obtain care¹¹. Furthermore, an individual under financial duress may decide to forgo medical treatment due to social pressures such as pride and embarrassment⁴.

A national survey found that to pay medical debt, approximately 44% of individuals with outstanding medical debt used a majority of their savings⁴, depleting precious resources they need for food, rent, prescriptions, etc.⁷ The linkage between self-reported health status and healthcare usage in a setting of low financial resources can lead to increased medical debt⁷. Another study reported that when individuals utilized home equity to defray incurred medical expenses, medical bills led to foreclosure of homes 70% of the time and mortgage defaults in 23% of cases¹². Yet another study found that households with medical debt over two percent of their income was the greatest single variable associated with bankruptcy¹³. Additionally, over 38.2% of debtors lost over two weeks of income due to illness and 6.8% of individuals lost wages due to care for an ill family member, diminishing their ability to pay their medical debt¹⁴. Some individuals frequently utilized payday loans, credit cards, and other forms of credit to pay off medical bills, further incurring financial debt¹⁵.

Ideally, the best approach to understanding medical debt would be to directly mine data of individual patient records, regardless of Medicaid eligibility, and analyze variance to explore associations between debt and health outcomes. However, challenges to this approach are neither technical nor methodological in nature, but instead organizational. 

Big Data is defined as having the “four V’s” – tremendous Volume, Velocity of creation, generated in a Variety of ways, and Veracity of data^16,17. Note, however, the scope of Big Data is dependent on the discipline or field in which it is created. Computational experts have created incredible warehouses of healthcare data recorded at the individual patient level, and analytics experts continue to create cutting edge methodologies for analyzing and visualizing these massive datasets or Big Data. Yet organizationally, the scientific community needs to embrace further Big Data’s complexity by addressing issues like transparency that might be mistakenly viewed as banal. The challenge for Big Data in healthcare, is much more constrained by organizational complexity than by technological efficiency¹⁸. Therefore, simply getting access to data may prove to be more challenging than analyzing the data.

While the CDC’s Behavioral Risk Factor Surveillance System (BRFSS)¹⁹ provides access to a wealth of publicly available data, these data pose limitations. One major limitation is that they understandably do not share data in which individual patients serve as the unit of record. Therefore, analyzing variance for how a phenomenon like medical debt manifests at the individual level is, by design, impossible. A second limitation of these data is that for rural, less populated areas, results are based on estimations rather than direct observations. While methodologically sound, errors inherent in such estimations are particularly problematic for phenomena like medical debt, which may be more pronounced in poorer, rural areas. Medical debt’s disproportionate effect on rural communities should be a cause of great concern for the medical profession. Medical debt’s potential association with patient health outcomes calls for Big Data research within the medical community, with goals of elucidating data driven solutions that can improve patient care outcomes in rural and urban communities.

In a research environment where it may be easier to isolate an individual’s genetic code than it is to access data on their level of medical debt, Big Data needs to provide information that is relevant²⁰ and more importantly, actionable. Admittedly, optimization of Big Data for patient medical debt research will take years. In the interim, however, we believe a sensible adjacent possible²¹ step is to explore health outcomes data prior and subsequent to Medicaid expansion via publicly available and/or institutionally owned Big Data. Health outcomes in management of chronic conditions such as assessing the level of Diabetes control by understanding access to care, number of visits to health professional, proximity to care, and cost of insulin and diabetic drugs are a few of the many variables that can be explored through Big Data. Medicaid Expansion offers the scientific community an opportunity to use an exact cutoff date to explore whether Medicaid Expansion affects medical debt reporting and if health outcomes are positively or negatively impacted. Recognizing the potential complexity of this phenomenon as more than just solely financial, a patient’s medical debt may have as much bearing on their health outcomes as other social determinants identified by either the CDC or WHO.

---

References

1) Marmot, M. (2005). Social determinants of health inequalities. The Lancet, 365(9464), 1099–1104. https://doi.org/10.1016/S0140-6736(05)71146-6

2) Banegas MP, Guy GP, de Moor JS, et al. For Working-Age Cancer Survivors, Medical Debt And Bankruptcy Create Financial Hardships. Health Affairs. 2016;35(1):54-61. doi:10.1377/hlthaff.2015.0830

3) All items less food and energy in U.S. city average, all urban consumers, not seasonally adjusted. (n.d.). [United States Department of Labor]. Retrieved May 14, 2018, from https://data.bls.gov/timeseries/CUUR0000SA0L1E?output_view=pct_12mths

4) Medical care in U.S. city average, all urban consumers, not seasonally adjusted. (n.d.). [United States Department of Labor]. Retrieved May 14, 2018, from https://data.bls.gov/timeseries/CUUR0000SAM?output_view=pct_12mths

5) Kinney, E. D. (2010). For Profit Enterprise in Health Care: Can it Contribute to Health Reform? American Journal of Law & Medicine, 36, 405–435.

6) Key Facts about the Uninsured Population. Retrieved March 21, 2018, from https://www.kff.org/uninsured/fact-sheet/key-facts-about-the-uninsured-population/

7) Seifert, R. W., & Rukavina, M. (2006). Bankruptcy Is The Tip Of A Medical-Debt Iceberg. Health Affairs, 25(2), w89–w92. https://doi.org/10.1377/hlthaff.25.w89

8) 2016 TransUnion Healthcare Report. (n.d.). Retrieved November 15, 2017, from https://www.transunioninsights.com/studies/healthcare/

9) Cohen, R. A., & Zammitti, E. P. (2017). Problems paying medical bills among persons under age 65: Early release of estimates from the National Health Interview Survey, 2011–June 2017. National Health Interview Survey Early Release Program, 11.

10) Wiltshire, J. C., Elder, K., Kiefe, C., & Allison, J. J. (2016). Medical Debt and Related Financial Consequences Among Older African American and White Adults. American Journal of Public Health, 106(6), 1086–1091. https://doi.org/10.2105/AJPH.2016.303137

11) Doty, M. M., Edwards, J. N., & Holmgren, A. L. (2005). Seeing red: Americans driven into debt by medical bills.

12) Lichtenstein, B., & Weber, J. (2016). Losing Ground: Racial Disparities in Medical Debt and Home Foreclosure in the Deep South. Family & Community Health, 39(3), 178–187. https://doi.org/10.1097/FCH.0000000000000108

13) Domowitz, I., & Sartain, R. L. (1999). Determinants of the Consumer Bankruptcy Decision. The Journal of Finance, 54(1), 403–420.

14) Himmelstein, D. U., Thorne, D., Warren, E., & Woolhandler, S. (2009). Medical Bankruptcy in the United States, 2007: Results of a National Study. The American Journal of Medicine, 122(8), 741–746. https://doi.org/10.1016/j.amjmed.2009.04.012

15) Bickham, T., & Lim, Y. (2015). In Sickness and in Debt: Do Mounting Medical Bills Predict Payday Loan Debt? Social Work in Health Care, 54(6), 518–531. https://doi.org/10.1080/00981389.2015.1038410

16) McAfee, A., & Brynjolfsson, E. (2012). Big Data: The Management Revolution. Harvard Business Review, 1–9.

17) The FOUR V’s of Big Data. (n.d.). Retrieved March 2, 2018, from http://www.ibmbigdatahub.com/sites/default/files/infographic_file/4-Vs-of-big-data.jpg

18) Singer, P. W. (2009). Wired for War: The Robotics Revolution and Conflict in the Twenty-first Century. Penguin.

19) CDC - Behavioral Risk Factor Surveillance System. https://www.cdc.gov/brfss/index.html. Published August 24, 2017. Accessed November 20, 2017.

20) Silverman, E.K. Big Data and Network Medicine in COPD. In: COPD. Springer; 2017:321-332.

21) Johnson, S. (2010). Where Good Ideas Come from: The Natural History of Innovation (Hardcover). New York, NY: Riverhead Books.