The United States has seen major advances in medical care over the past decades, but access to care at an affordable cost is not universal. Many Americans lack health care insurance of any kind, and many others with insurance are nonetheless exposed to financial risk because of high premiums, deductibles, copays, limits on insurance payments, and uncovered services. One might expect that the U.S. poverty measure would capture these financial effects and trends in them over time. Yet the current official poverty measure developed in the early 1960s does not take into account significant increases and variations in medical care costs, insurance coverage, out-of-pocket spending, and the financial burden imposed on families and individuals. Although medical costs consume a growing share of family and national income and studies regularly document high rates of medical financial stress and debt, the current poverty measure does not capture the consequences for families’ economic security nor their income available for other basic needs.
In 1995, a panel of the National Research Council (NRC) in Measuring Poverty: A New Approach recommended a new poverty measure, which compares families’ disposable income to poverty thresholds based on current spending for food, clothing, shelter, utilities, and a little more. The panel also recommended that the federal government develop a separate measure of medical care risk that would track the economic risk to families and individuals of lacking adequate health insurance coverage.
The panel’s recommendations stimulated extensive collaborative research involving several government agencies on experimental poverty measures that led to a new research Supplemental Poverty Measure (SPM),
which the U.S. Census Bureau first published in November 2011 and will update annually. Analyses of the effects of including and excluding certain factors from the new SPM showed that, were it not for the cost that families incurred for premiums and other medical expenses not covered by health insurance, 10 million fewer people would have been poor according to the SPM, and the SPM poverty rate in 2010 would have been 3 percentage points lower (Short, 2011:Table 3a).
Yet, although the SPM subtracts out-of-pocket medical care costs in the calculation of disposable income, it does not directly measure the burden of out-of-pocket medical care expenses nor does it address the medical care economic risk to the population in terms of the adequacy of their health insurance coverage to pay for their expected health care needs. The implementation of the Affordable Care Act (ACA) provides a strong impetus to think rigorously about ways to measure medical care economic burden and risk. As new policies—whether part of the ACA or other policies—are implemented that seek to expand and improve health insurance coverage and to protect against the high costs of medical care relative to income, such measures will be important to assess the effects of policy changes in both the short and the long term on the extent of financial burden and risk for the population.
The U.S. Department of Health and Human Services (HHS) is responsible for carrying out the provisions of the ACA, which is intended to extend health insurance coverage to most Americans. To monitor the effectiveness of health care reform in reducing out-of-pocket medical care expenses for low-income families and children, HHS can make use of the new SPM, but the SPM does not fully address the medical care risk to the population in terms of the adequacy of their health insurance coverage to pay for their expected health care needs. HHS would also find useful a companion measure of medical care economic risk (MCER), which estimates the proportion of families and children who are at risk of incurring high out-of-pocket medical care expenses, including health insurance premiums, in relation to their resources. Such a measure would enable HHS to answer such questions as which groups face a greater likelihood of economic insecurity due to lack of or inadequate health insurance coverage.
In fall 2010, the Office of the Assistant Secretary for Planning and Evaluation in HHS requested the National Academies to convene an ad hoc panel of experts to
organize, commission papers for, and conduct a public workshop to critically examine the state of the science in the development and implementation of a new measure of medical care risk as a companion measure
to the new Supplemental Poverty Measure. The workshop will examine retrospective and prospective measures of medical care risk, defined as the risk of incurring high out-of-pocket medical care expenses (including insurance premiums) relative to income … and other related issues. Based on the workshop and its deliberations, the panel will prepare a report with findings and recommendations that will help the field to move forward to implement a new measure of medical care risk that will be valuable for monitoring the implementation of health care reform. The report will include a summary of the workshop and commissioned papers.
In response to this request, the National Research Council’s (NRC’s) Committee on National Statistics, in collaboration with the Board on Health Care Services of the Institute of Medicine, appointed a nine-member panel representing a range of expertise related to the scope of the study. The panel executed its charge through the conduct of a workshop, commissioning background papers, holding panel meetings, and reviewing research and other reports. The goal of the panel was to move forward toward developing measures to inform policy that are feasible to collect and estimate and that will monitor changes in medical care economic risk and burden as health care reform is implemented and other relevant public- and private-sector changes occur.
On the basis of the workshop discussions and its own review and deliberations of the issues, the panel developed conclusions and recommendations in five areas: (1) concepts of medical care economic burden and risk, (2) concepts of resources, (3) measurement of medical care economic risk, (4) data sources, and (5) development and implementation of the panel’s proposed measures. Recommendations in this summary are numbered by the chapter in which they appear in the body of the report.
CONCLUSIONS AND RECOMMENDATIONS
Concepts of Burden and Risk
There is a conceptual difference between medical care economic burden and risk, and the panel thinks that measures of both are needed to inform national and state policy and to assess economic trends. Burden is a retrospective measure that examines actual out-of-pocket spending for health insurance and medical care relative to a family’s available resources. Risk is a prospective measure that assesses the likelihood that a family’s future out-of-pocket medical care expenditures would be high or unaffordable relative to the family’s resources.
Measuring Medical Care Economic Burden
The panel proposes that a measure of medical care economic burden be estimated by the U.S. Census Bureau in conjunction with estimating the SPM. This would be done by comparing a family or individual’s actual out-of-pocket medical spending with resources available for medical care. (Chapter 2 provides details of the calculation, which involves taking a family’s SPM measure of resources, adding back its out-of-pocket medical spending, and subtracting its nonmedical needs as represented by the SPM poverty threshold for the family.) The difference would be expressed by the extent to which families and individuals who are already poor in terms of having insufficient resources for their nonmedical needs are moved deeper into poverty because of their medical costs and the extent to which those who are not poor are moved into poverty or below a low multiple of poverty, such as 100 percent or 250 percent. Estimates of these effects should be provided separately for health insurance premiums and other expenses for medical care and should also take account of important features of the new national health care policy, which include a major role for states going forward, premium subsidies and other features of affordability that are linked explicitly to multiples of the poverty thresholds, and continued policy differences by age. To inform policy, it is important that the SPM and the measure of medical care economic burden reflect trends in actual spending—not hypothetical spending. Thus, there should be no adjustment for underutilization of medical care in the definition of resources.
Recommendation 2-1: The panel recommends that the U.S. Census Bureau refine its Supplemental Poverty Measure (SPM) reports and tables to include the estimated effects of medical care economic burden on poverty by component, showing the effects of premiums separately from other out-of-pocket expenses. It further recommends that the SPM reports and tables include the estimated effects of medical care economic burden by region or state, recognizing that aggregation over time or by groups of states may be necessary to obtain reliable estimates.
Recommendation 2-2: The panel recommends that the U.S. Census Bureau examine medical care economic burden in its Supplemental Poverty Measure (SPM) reports and tables by providing estimates of the number of people who move from higher to lower multiples of the SPM poverty thresholds—including thresholds above and below the poverty level—because of their health insurance premiums and other out-of-pocket medical care costs.
Recommendation 2-3: The panel recommends that the U.S. Census Bureau report findings on medical care economic burden in its Supplemental Poverty Measure reports and tables separately for the populations under age 65 and ages 65 and older.
Recommendation 2-4: The panel recommends that the U.S. Census Bureau continue to use a definition of resources for the Supplemental Poverty Measure and estimates of medical care economic burden that incorporates estimates of actual out-of-pocket spending on health insurance premiums and other out-of-pocket expenses for medical care. The Census Bureau should not model potential spending for people lacking health insurance coverage.
Concepts of Resources
The choice of a measure of resources for use in measuring MCER1 will be tightly constrained by the choice of a survey to serve as home to a measure of MCER, and in this decision the measurement of medical care risk is likely to dominate the measurement of resources. Nevertheless, it is important to understand the key issues that exist in defining resources and the potential implications of including or excluding particular types of resources.
The resources available to families and individuals to meet their financial needs include not only income, but also assets—the product of families’ saving and investment activities over the life course. With regard to income, the panel encourages the Census Bureau to update its concepts and improve its measurement of money income (used in the official poverty measure) and disposable income (used in the SPM) in its household surveys, particularly self-employment income and new forms of retirement income that are neither regular flows nor lump sums, as traditionally understood.
In the context of how people pay for extraordinary and, especially, unexpected medical care expenses, the role of assets cannot be overlooked. To exclude all assets from the resources used to measure MCER, and in so doing make it a measure of income-related economic risk, ignores accumulating evidence on how families prepare for potentially high medical expenditures and how well they are able to absorb them. Consequently, the panel concludes that the resources component of a measure of MCER must take account of a portion of assets if the goal is to assess resources available to pay for medical care costs currently and over time. The panel further concludes that only financial assets that a family can access relatively quickly
1 The measure of resources for medical care economic burden is derived from the SPM as discussed above.
should be considered in determining the amount to be included and that assets of all family members should be used to determine family resources without regard to employment status or age.
Although the panel concludes that the calculation of an annuitized value from the family’s liquid assets is a compelling approach, there are operational issues that we could not examine. Consequently, the method for calculating the asset contribution to resources will need to be determined by the federal agency charged with producing the measure of MCER. The asset contribution derived in this manner should be added to disposable income to provide the measure of resources for evaluating MCER.
Recommendation 3-1: The panel recommends that the U.S. Census Bureau modify its concepts and measurement of money income and disposable income to better account for income flows from self-employment and from new forms of retirement income for use in measures of poverty and medical care economic risk and burden that are derived from its household surveys.
Recommendation 3-2: The panel recommends that, for measuring medical care economic risk, a portion of liquid assets be included in the resources of all persons, regardless of age or employment status. Only assets that the family or individual can access relatively quickly should be considered in determining the amount to be included—namely, financial assets held outside retirement accounts, the posttax value of assets held in retirement accounts, and, in principle, the amount received from a reverse mortgage (treating it as income rather than as an asset), acknowledging the limitations of existing data.
Recommendation 3-3: The panel recommends that the method for calculating the share of liquid asset contribution to resources for measuring medical care economic risk be determined by the federal agency charged with producing the measures and that the methodology be based on one of two options—either a fixed share of assets or an annuitized value. The share of liquid asset contribution derived in this manner should be added to disposable income to provide the measure of resources for evaluating medical care economic risk.
Measures of MCER and Recommended Approach
In addition to measuring retrospectively the financial burden from actual out-of-pocket medical care spending, the panel agrees with the 1995 NRC panel that it is important to develop a measure of medical care economic risk that can assess the exposure to, or potential for incurring, future
expenses. This is especially true because of the skewed nature of medical care costs. The panel considered various methods, including retrospective and prospective approaches, to constructing a measure of MCER as distinct from economic burden. The outcome of interest is a measure of risk, for example, the expected number (or fraction) of families and their individual members who, as a result of out-of-pocket spending for medical care services and premiums, would be in poverty or some multiple of poverty as defined by the SPM. For medical care risk to differ from the medical care burden of large expenditures, it must be based on the distribution of future out-of-pocket expenditures that an individual or household may face given their characteristics at some baseline point in time. Thus, it is a forward-looking or prospective measure as distinct from the burden measure, which is retrospective.
In order to understand the effects of financial exposure to medical care costs on available household income across the U.S. population, it is necessary to calculate the probability for families with particular characteristics of having out-of-pocket premiums and spending on medical care services greater than their resources available for medical care spending. Ideally, the calculation would reflect the actual terms of family members’ health insurance coverage, their age, gender, and health status, the income of the family, and the composition of the family for a large number of families. Practically speaking, it must be constructed on the basis of information that is available from the Medical Expenditure Panel Survey (MEPS) or the Current Population Survey Annual Social and Economic Supplement (CPS ASEC). Both surveys, however, have limitations in terms of relevant information collected, as discussed below. The trade-offs in the choice between these two surveys leads to a two-pronged strategy.
Although the concept of MCER is prospective, 1 year of retrospective cross-sectional data could be used to estimate it, which facilitates timeliness and makes it possible to use nonpanel data like the CPS ASEC. The retrospectively determined burden of out-of-pocket medical care spending for a given year can be used as a simple predictor of MCER in the following period. However, nonpanel data sources systematically exclude recent deaths and those who have entered institutions in the immediate past time period—two groups known to have high health care expenditures, so information about the impact of these transitions on out-of-pocket medical care spending will have to come from other sources. Another problem is that the characteristics that predict out-of-pocket medical care spending must logically be defined at the start of the year. So the groupings of individuals or families with similar characteristics predictive of expected medical care spending (called “risk cells” in this report) cannot be defined using current medical care spending because that would produce overly small amounts of observed variation in spending. Nonetheless, in the short
term, with the data now being collected, the CPS ASEC could be used to report the burden of out-of-pocket medical care spending retrospectively, roughly 10 months after the end of the calendar year for which income and spending are reported. Furthermore, with additional assumptions, the retrospective measure of burden could serve as a proxy for the prospective MCER.2
Then why continue to pursue construction of a prospective measure of MCER? With its richer data on health conditions, distribution of spending by service type, and 2-year panel, MEPS offers the opportunity to learn much more about the interplay of health status, health insurance, income, and out-of-pocket medical care spending with respect to family finances. Over the next several years, as the landscape of health insurance coverage in the United States undergoes substantial change, understanding the underlying drivers of any shifts in the impact of out-of-pocket medical care spending on family financial resources will be extremely important. With 2 years of data, one can use data on second-period expenses and base-period characteristics together with multivariate regression methods to estimate the probability that a family with given characteristics will have an expenditure large enough to push it to the poverty threshold.
However, the truly prospective measures that require 2 or more years of data run up against limitations in the available data sources (discussed in the “Data Sources” section); they also run up against the dearth of relevant literature on which to base prediction models. Although much is known about total health care expenditures, very little is known about family and individual covariates that predict family out-of-pocket medical care spending or the impact on family finances.
This situation dictates a research agenda to consider several possible alternative analyses to better understand these issues before making highly specific recommendations on a prospective measure of MCER. The results of these analyses can be used to inform the move from a purely retrospective approach based on burden to a more prospective approach. Research topics include the predictive value at the family level of out-of-pocket medical care spending in year 1 in relation to spending in year 2 and the stability of the relationship; the added predictive value of expanding the covariate list to include other family characteristics, such as the age, gender, and health status of members; whether to build a family model or an individual model that subsequently combines individual predictions for the family; because individual characteristics are the strongest predictors of future average expenditures, how to roll up individual predictions into a composite family measure that is predictive of future family out-of-pocket medical care
2 The measure of burden discussed here is similar to but not the same as that recommended for regular publication above, which adheres to the SPM definition of resources.
expenditures; how to combine distributions of expenditures for individual family members into the family’s distribution around its expected amount; and the pros and cons of regression methods versus cell-based approaches.
All of these topics require detailed information, not all of which is currently available. Moreover, in the absence of sufficient research on the distribution of out-of-pocket costs relative to SPM thresholds, it will be necessary to do that work empirically. For example, one would expect that a working poor family with one or more members in fair or poor health might have a substantial risk even without a hospitalization or high-cost drug regimen. An emergency department visit or a flare-up of a chronic condition might be enough to drop such a family below the threshold. For a middle-income family, however, it might take a larger health shock such as an uncovered hospital stay.
Recommendation 4-1: Given what limited work has been done in the field on issues in measuring medical care economic risk (MCER) prospectively, the panel recommends that appropriate federal agencies— the Agency for Healthcare Research and Quality, the Office of the Assistant Secretary for Planning and Evaluation, or both—perform a series of analyses using the Medical Expenditure Panel Survey to examine different prospective MCER measures.
Recommendation 4-2: The panel recommends that the results of the analyses from Recommendation 4-1 be used to inform the move from a purely retrospective approach based on burden to a more prospective approach for measuring medical care economic risk.
Data Sources for Developing and Producing an MCER
The data requirements for developing a measure of medical care economic risk are not the same as the requirements for producing a measure on a recurring basis. Development has more extensive data needs than production, but production requires annual data that are available on a timely basis from a large federal sample survey that represents the civilian noninstitutionalized population.
To develop a prospective measure of MCER requires longitudinal data, so that medical expenditures (and resources) observed prospectively over the course of a period—ideally a year—can be related to characteristics observed at the start of that period that are potentially predictive of medical expenditures. Actual out-of-pocket expenditures for premiums and other medical care expenses in the prior year may be the strongest predictor of expenditures during the current year, and although they are not a baseline characteristic per se, these expenditures ought to be included in the devel-
opment of a predictive model of prospective risk. Both the risk variables and the resources variables must be recorded at the person level, so that the variables in each case can be aggregated to the health insurance unit (for aspects of modeling risk) and family levels (for comparing risk with resources). Sufficient information on family relationships must be included to enable the membership of each health insurance unit and family to be identified.
The panel looked closely at three longitudinal surveys: MEPS, the Survey of Income and Program Participation (SIPP), and the Health and Retirement Study (HRS). None of the three surveys collects all of the variables that would be required to develop a prospective measure of MCER, as described in Chapter 4. Most notably, none of the three surveys collects a description of the services and treatments covered by each person’s health insurance plan, and none of them collects sufficient information with which to assess each sample member’s potential liability for out-of-pocket medical costs. MEPS collected detailed information on the health insurance plans of sample members in 1996 but has not done so again. Other survey-specific data gaps exist as well, which limit how fully each survey could support the modeling of MCER. These are discussed in detail in the report.
In summary, none of the surveys is nearly as strong as one would like in its measurement of key baseline characteristics. With its strong measures of chronic health conditions and very high quality expenditure data, MEPS is clearly superior to SIPP. The HRS could provide a supplemental data source for the one-fifth of households that fall into its universe of people over age 50. Estimates from the HRS could be used to validate the model estimates from MEPS for this segment of the population (or perhaps just the elderly).
Once a model of MCER has been developed, the estimates could be used directly (in MEPS), or the predictive model could be applied to another data set that provides measures of the relevant baseline characteristics. The latter approach offers a way to make the measurement of MCER more timely and to extend the measure to a larger and possibly more representative sample. For production, in addition to MEPS, the panel considered the CPS ASEC, the National Health Interview Survey, the American Community Survey, and the Consumer Expenditure series quarterly survey. The CPS ASEC is the source of both the official poverty measure and the SPM, to which the MCER measure is intended as a companion. Producing the two measures from the same survey would enable more direct comparisons than if the two were based on different surveys. The CPS ASEC is the only one of the surveys that can estimate disposable income currently, using imputations for taxes and commuting expenses, but it lacks a measure of liquid assets and has limited information on health conditions. MEPS cannot currently estimate disposable income because it does not collect or
impute such variables as taxes, commuting, and child care expenses. These components could be imputed to MEPS or added to the MEPS questionnaire in the future.
Although the panel favors a prospective measure of MCER over a retrospective measure, the more substantial data requirements of the prospective measure cannot be fully met with an existing survey. The MEPS longitudinal file comes closest to meeting these requirements, with the HRS providing a means to validate the results of MEPS modeling for older people.
Recommendation 5-1: The panel recommends that the development of a model for estimating a prospective measure of medical care economic risk be carried out with the Medical Expenditure Panel Survey (MEPS) longitudinal file. The panel also recommends that the Health and Retirement Study (HRS) be used to validate the results of the MEPS modeling for at least the elderly, if not the entire population over age 50, which the HRS sample represents.
Recommendation 5-2: The panel recommends that the Census Bureau and the Agency for Healthcare Research and Quality assess the merits of adding items to both the Current Population Survey Annual Social and Economic Supplement and the Medical Expenditure Panel Survey to at least partially address the most critical data limitations identified for measuring medical care economic risk.
Implementing Measures of MCER and Burden
Throughout its review and deliberations, the panel has aimed to develop rigorous yet practical approaches to defining and measuring the financial burden and risk associated with out-of-pocket medical care costs. Specifically, we focused on how exposure to medical care expenses can threaten families and individuals with being driven into poverty. Through commissioned papers, workshop presentations and discussions, and deliberations, we sought to bring to bear the latest research and data. We have also kept in mind what actually can be done by government agencies without major infusions of additional staff or funding.
For the introduction of a measure of MCER to be successful, clear lines of responsibility for its implementation must be established. The two federal agencies with the greatest expertise in the development and implementation of such a measure are the U.S. Department of Health and Human Services and the U.S. Department of Commerce. The panel thinks that a subcabinet-level coordinating group would help to ensure that a measure of MCER moves forward in its development and launch. This
coordinating group would provide guidance to the agencies producing the measure and suggest changes in methodology or appropriate data sets. The leadership of agencies with contributions to make to the construction and implementation of the measure could constitute such a group. The panel also suggests that one or two members of the coordinating group be chosen from outside government with relevant expertise in the measurement of poverty and financial burden of health care. Having one or more outside members would enhance the transparency and credibility of the process as well as provide the government with the latest thinking from the scholarly community outside the government.
Based on these findings and conclusions the panel provides the following recommendations for implementation:
Recommendation 6-1: Because technical and cross-departmental efforts such as the construction and maintenance of a measure of medical care economic risk (MCER) require both political and resource support, the panel recommends that the secretaries of the U.S. Departments of Health and Human Services and Commerce be jointly responsible for developing and reporting measures of MCER (and burden) on an annual basis with involvement of the U.S. Office of Management and Budget chief statistician. This effort should coincide with the production and release schedule for the Supplemental Poverty Measure.
Recommendation 6-2: The panel further recommends the creation of a medical care economic risk coordinating group composed of senior officials from the U.S. Department of Health and Human Services, the U.S. Census Bureau, and the U.S. Office of Management and Budget to provide oversight and make suggestions for needed improvements.
Recommendation 6-3: The panel recommends that funding for the current data collection efforts be maintained at a level to ensure that rigorous, accurate calculations of measures of medical care economic burden and risk can be made.