Health Affairs, 25, no. 6 (2006): 1688-1699 doi: 10.1377/hlthaff.25.6.1688 © 2006 by Project HOPE	New Online   Senate Health Reform Bill   Rewarding Providers   Public Option Policy Brief   Health Reform & Abortion   Delivery System Reform

This Article Abstract Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Ziller, E. C. Articles by Yousefian, A. E. Search for Related Content PubMed PubMed Citation Articles by Ziller, E. C. Articles by Yousefian, A. E. Related Collections Access To Care Business Of Health Rural Health Care Health Spending Consumer Issues Insurance Market

DataWatch

Out-Of-Pocket Health Spending And The Rural Underinsured

Abstract

 
Multiple studies have documented higher uninsurance rates among rural compared to urban residents, yet the relative adequacy of coverage among rural residents with private health insurance remains unclear. This study estimates underinsurance rates among privately insured rural residents (both adjacent and nonadjacent to urban areas) and the characteristics associated with rural underinsurance. We found that 6 percent of privately insured urban residents were underinsured; the rate increased to 10 percent for rural adjacent and 12 percent for rural nonadjacent residents. Multivariate analyses suggest that rural residents’ underinsurance status is related to the design of the private plans through which they have coverage.

The purpose of this study is to estimate underinsurance rates among privately insured rural residents and to determine whether and how they differ from those of urban residents. By "underinsurance" we mean potential access barriers for people with private health insurance as a result of out-of-pocket costs for care. We also examine the characteristics associated with being underinsured and whether they differ for rural and urban residents. Finally, we attempt to explain rural-urban differences in underinsurance rates by controlling for socioeconomic, utilization, and health plan characteristics.

Prior research on rural health insurance coverage suggests that privately insured rural residents are at greater risk of being underinsured than their urban peers. Compared with those in urban areas, rural residents are, on average, lower income, in poorer health, more likely to be covered by a nongroup policy or to work for a small employer, and less likely to have coverage through a health maintenance organization (HMO).² Each of these characteristics has the potential to influence the use of services, out-of-pocket costs, and the extent to which paying for medical care is a hardship for privately insured rural families.

Study Data And Methods

Top Study Data And Methods Study Findings Discussion And Policy... NOTES

 
We examined annual out-of-pocket health care spending by U.S. residents under age sixty-five who were continuously insured by a private plan in 2001 or 2002. We categorized people as being underinsured based on their family health spending and income.

Data. This study used pooled 2001 and 2002 panels from the Medical Expenditure Panel Survey Household Component (MEPS HC) to create a file of 43,315 unique respondents under age sixty-five, of whom 23,314 (53.8 percent of the un-weighted sample) were privately insured for the entire year in which they were surveyed (either 2001 or 2002, depending on the respondent’s panel). Out-of-pocket expenditures were similarly calculated across the entire one-year period that corresponded to a respondent’s panel. Pooling was necessary to ensure that we had a sufficient rural sample to complete our analyses.

Selecting MEPS HC as our data source has advantages and limitations. Its sampling design and weighting methods increase the generalizability of our results to national rural and urban populations. Also, few other data sources contain the detailed information available through MEPS HC, particularly the spending data upon which our underinsurance definition depends. However, although the MEPS HC spending information includes family-level spending on deductibles and co-insurance, it lacks specific insurance plan data such as deductible levels and required copayments for services. As discussed below, this limitation makes it likely that our results underestimate actual underinsurance rates.

Defining "underinsurance." Researchers assessing the adequacy of private health insurance coverage have faced challenges in defining underinsurance.³ Generally, researchers have used either economic or experiential measures, or both. Economic measures typically consider actual or potential spending, either nominally or as a proportion of annual income.⁴ Experiential measures tend to emphasize reported financial barriers to care despite the presence of health insurance coverage.⁵

Following the example of previous research, we classified respondents as being underinsured if family out-of-pocket spending exceeded 10 percent of family income or (for families with incomes below 200 percent of the federal poverty level) if family out-of-pocket costs exceeded 5 percent of family income.⁶ Unlike prior research, however, our data limitations precluded us from using health plan characteristics such as deductible levels and coinsurance requirements as part of our definition. Thus, we did not capture those who were at risk of having high out-of-pocket spending but who did not use enough services over the course of the year to meet spending thresholds. Although this data limitation might have understated the underinsurance problem, it is unlikely to have greatly biased estimates of rural-urban differences in underinsurance rates. If bias exists, it is more likely to understate rural-urban differences than to overstate them, because rural residents tend to have higher rates of individual or small-group private coverage and, therefore, likely higher deductibles.⁷

Analysis. Rural or urban residence was our independent variable. We used modified Rural-Urban Continuum Codes to classify counties as urban, rural but adjacent to an urban county, or rural and not adjacent to an urban county. Differentiating between rural counties based on proximity to an urban place is critical for rural health services research because prior studies indicate that insurance coverage and access differ for rural residents in counties adjacent to a metropolitan area (sometimes referred to as "micropolitan" counties) versus those in nonadjacent counties.⁸

Our covariates included characteristics that prior research has linked to health insurance coverage in general, or to rural-urban differences in coverage, or both, such as employment, income, age, census region, minority status, and health status. We also created dichotomous measures of health care use, such as whether or not respondents had an emergency room (ER) visit, prescription drug use, or an inpatient hospital stay. Finally, we included nominal information about individuals’ private plans, including whether or not they covered prescription drugs, were an HMO, used a preferred provider list, or paid off-list providers, or if the respondent was covered by a nongroup plan at any time during the year.

We weighted our analyses to adjust for the complex sampling design of MEPS and for survey nonresponse, and to permit pooling of survey years. All statistical tests were completed using SUDAAN to adjust for clustering and to yield valid standard errors for the weighted data. Frequency differences were evaluated with chi-square tests of significance; means were compared through paired t-tests. Unless noted otherwise, only differences found to be significant at or below the .05 level are discussed in the text. To understand what factors are associated with being underinsured and whether rural-urban differences could be adequately explained by these factors, we fitted a series of logit models designed to predict rural underinsurance rates, controlling first for socioeconomic, then utilization, and finally insurance plan characteristics.

Study Findings

Top Study Data And Methods Study Findings Discussion And Policy... NOTES

 
Out-of-pocket spending. Among privately insured people with health care spending in 2001 or 2002, 17 percent of rural nonadjacent residents spent more than $1,000 out of pocket during the year, compared with 13 percent and 14 percent of urban and adjacent residents, respectively (p .001; data not shown). Total out-of-pocket spending for those who used health care services did not differ for rural adjacent and urban residents; however, rural nonadjacent residents spent more than $100 more in mean out-of-pocket dollars (Exhibit 1). Thus, while urban residents were responsible for 32 percent of their total costs, rural nonadjacent residents covered 39 percent of their total health care costs.

Rural-urban differences among the privately insured. To better understand the differences in out-of-pocket spending between privately insured rural (particularly nonadjacent) and urban residents, we compared them on characteristics most likely to affect use and costs (Exhibit 2). We found that privately insured rural residents in our study sample were generally a little older (particularly those adjacent to an urban area), had somewhat lower incomes, and were a bit more likely to have a chronic condition or to be in fair or poor health, compared with urban dwellers. Also, privately insured rural respondents, particularly in nonadjacent areas, were more likely than their urban peers to be self-employed or work for a small business.

Underinsurance rates. Unlike the out-of-pocket costs discussed above, which were based on individuals’ spending, the following sections describe aggregate family spending for everyone in the family with continuous private insurance. Relying only on individual spending could lead to an underestimate of families’ actual medical costs. Conversely, using only individuals’ incomes could lead to an overestimate of underinsurance if someone has limited personal income but could draw on the resources of other family members for medical care expenses.

Based on the definitions discussed above, 7.0 percent of all respondents were underinsured in 2001 or 2002, and underinsurance rates differed as residence became more rural.⁹ Although 6 percent of urban residents were underinsured, the rate increased to 10 percent for rural adjacent and 12 percent for rural nonadjacent residents. This pattern was essentially the same for those spending more than 10 percent of their income on medical care, or spending 5 percent at incomes below 200 percent of poverty.

Rural and urban underinsurance rates. Rural-urban differences in underinsurance persisted across all census regions, although the patterns differed slightly (Exhibit 3). For example, in the Northeast, Midwest, and West, nonadjacent underinsurance rates were about double urban rates. In the South, however, they were only about 65 percent higher.

Among adult workers, we found both employer and job characteristics to be associated with underinsurance, with some rural-urban differences (Exhibit 4). For example, while the self-employed had higher underinsurance rates than those working for a firm, differences among urban workers were minimal. Among rural adjacent workers, this gap widened; it was greatest among rural nonadjacent workers (Exhibit 4).

Explaining urban-rural differences. We estimated a series of three logit models to determine what characteristics are associated with rural-urban differences in the odds of being underinsured (Exhibit 5). First, we included basic socioeconomic and employment control variables, then utilization variables (second model), and finally health plan characteristics (third model).

Discussion And Policy Implications

Top Study Data And Methods Study Findings Discussion And Policy... NOTES

 
Study limitations. There are several important limitations to this study. Defining underinsurance remains an imprecise science. Ideally, we would have used data on deductibles and coinsurance to calculate the actuarial value of health plans relative to individual or family income; however, this was not possible with the data available through MEPS. Furthermore, using a ratio of spending to income means that rural-urban differences in underinsurance rates are driven by both out-of-pocket costs and differences in family income, and this relationship is difficult to disentangle. Finally, although the 2002 MEPS was the most recent data available at the time of our analyses, this four-year lag could limit the generalizability of our findings to current conditions. This is especially true given that recent consumer-directed health insurance initiatives have gradually been moving the health insurance market toward less comprehensive policies.

Policy implications. Despite having private coverage, those who use medical services pay a sizable portion of their own health care costs, particularly rural residents. One out of eight rural nonadjacent residents is underinsured (12 percent), compared with 10 percent of rural adjacent and 6 percent of urban residents. And, although we controlled for a number of statistically significant socioeconomic and utilization characteristics, we found that the odds of being underinsured remained 70 percent higher for rural nonadjacent than for urban residents. Only when plan characteristics were modeled did the risk of being underinsured cease to be significantly higher for rural nonadjacent residents. This suggests that the actuarial value of private health plans held by rural residents is lower than for urban residents.

Substantial policy attention has been paid to the adequacy of prescription drug coverage among rural seniors in light of Medicare Part D. However, less emphasis has been placed on access to prescription drugs for nonelderly rural residents. Our findings suggest that drug coverage is an issue for many rural residents, particularly those in nonadjacent areas, and may be a key factor in underinsurance. The impact of prescription drug coverage was statistically significant in our model, accounting for a 50 percent decrease in the odds of being underinsured.

It is likely that we underestimated the actual U.S. underinsurance rate. Without access to detailed health plan information, it is impossible to determine whether people might be inadequately covered if they were to develop an acute or chronic condition that required sizable medical spending. Also, because rural residents may be more likely than urban dwellers to have plans with higher deductibles or coinsurance for services, or both, our findings might understate the magnitude of difference in rates of underinsurance between rural and urban residents.

Implications for coverage expansion. This study suggests that policymakers concerned about access for uninsured rural residents should also focus their attention on the problem of underinsurance. Prior research indicates that underinsured people face barriers to care that are comparable to those the uninsured face, including delaying care, lack of confidence in the quality of their care, and debt collection for medical bills.¹⁰ Efforts to expand private coverage must take into account the impact that plan design and cost-sharing requirements have on rural family pocketbooks, and consequently on the use of medical care, particularly in households where someone has a chronic illness or other health problem.

Implications for providers. Being underinsured has implications not only for patients but also for providers. Given the higher public payer and uninsured patient base of many rural health providers, the rural health system infrastructure faces challenges to financial solvency. For example, one study found that Medicare and Medicaid account for about 60 percent of all rural hospital revenues.¹¹ Prior research has found that 46 percent of the underinsured are facing collection for medical debts.¹² Given these findings, the elevated underinsurance rate in rural areas suggests that rural providers might face additional financial challenges—even when their patients have private health insurance.

Role of small businesses. Because the rural economy is dominated by small businesses, the private coverage available to rural residents is intrinsically linked to small firms’ health insurance challenges.¹³ Research has found that although premiums for small firms are generally comparable to those for large firms, the quality of coverage differs greatly, with small firms having higher deductible and coinsurance requirements.¹⁴ Given our finding that 55 percent of privately insured rural nonadjacent residents work for a business with fewer than fifty employees (compared with 40 percent of their urban peers), improving costs in the small-group market will be critical for addressing rural uninsurance and underinsurance.

Efforts to increase health coverage in small businesses have led policymakers to consider a number of initiatives. For example, states have passed legislation to foster the creation of private health insurance purchasing alliances, permitting small businesses to group together and, in theory, negotiate better premium rates with insurers; to date, however, these alliances have met with only limited success.¹⁵ Other states, such as Maine’s Dirigo health reform effort, have used public dollars to pool and subsidize the coverage of small firms and their employees. Although the impact of Dirigo has not yet been fully evaluated, given the limited impact of association plans to date, public pooling efforts could be a sound policy option for achieving a more level purchasing field between small and large firms.

AS HEALTH CARE COSTS CLIMB, more employers intend to shift premium costs to their employees, and others will offer plans with greater cost sharing.¹⁶ Thus, the number of uninsured and underinsured Americans will likely increase. Given the generally lower incomes of rural residents, rural health researchers and policymakers must monitor the effect of changes in the private insurance market on those living in rural areas. In particular, future studies should seek to measure the degree to which underinsurance is affecting rural residents’ access to health care and rural providers’ financial survival.

Editor's Notes

 
Erika Ziller (eziller{at}usm.maine.edu) is a research associate at the Institute for Health Policy, Muskie School of Public Service, University of Southern Maine, in Portland. Andy Coburn is a professor and director of the institute, and Anush Yousefian is a research analyst.

The authors thank the federal Office of Rural Health Policy, HRSA, DHHS, for funding this study and Ray Kuntz, from the AHRQ Data Center, for his assistance with the MEPS data.

NOTES

Top Study Data And Methods Study Findings Discussion And Policy... NOTES

 

1. S.L. Larson and S.C. Hill, "Rural-Urban Differences in Employment-related Health Insurance," Journal of Rural Health 21, no. 1 (2005): 21–30[CrossRef][Web of Science][Medline]; and E.C. Ziller et al., Health Insurance Coverage in Rural America: A Chartbook (Washington: Kaiser Commission on Medicaid and the Uninsured, 2003).
2. Ziller et al., Health Insurance Coverage in Rural America; and M.M. Casey, "Rural Managed Care," in Rural Health in the United States, ed. T.C. Ricketts (New York: Oxford University Press, 1999), 113–118.
3. For a discussion of definitional issues, see R. Bashshur, D.G. Smith, and R.A. Stiles "Defining Underinsurance: A Conceptual Framework for Policy and Empirical Analysis," Medical Care Review 50, no. 2 (1993): 199–218.[Medline]
4. For example, the seminal article, P.J. Farley, "Who Are the Underinsured?" Milbank Quarterly 63, no. 3 (1985): 476–503, and its update, P.F. Short and J.S. Banthin, "New Estimates of the Underinsured Younger than Sixty-five Years," Journal of the American Medical Association 274, no. 16 (1995): 1302–1306, define underinsurance using a ratio of expected out-of-pocket expenses to annual income.[Abstract/Free Full Text]
5. Experiential measures include self-reported difficulty paying medical bills or forgoing care because of cost. See, for example, R.J. Blendon et al., "Paying Medical Bills in the United States—Why Health Insurance Is-n’t Enough," Journal of the American Medical Association 271, no. 12 (1994): 949–951[Abstract/Free Full Text]; K. Donelan, C.M. DesRoches, and C. Schoen, "Inadequate Health Insurance: Costs and Consequences," Medscape General Medicine 2, no. 3 (2000): E37; and Centers for Disease Control and Prevention, "State-specific Prevalence Estimates of Uninsured and Underinsured Persons—Behavioral Risk Factor Surveillance System, 1995," Morbidity and Mortality Weekly Report 47, no. 3 (1998): 51–55.
6. C. Schoen et al., "Insured but Not Protected: How Many Adults Are Underinsured?" Health Affairs 24 (2005): w289–w302 (published online 14 June 2005; 10.1377/hlthaff.w5.289).[Abstract/Free Full Text]
7. U.S. Government Accountability Office, Private Health Insurance: Declining Employer Coverage May Affect Access for Fifty-five- to Sixty-four-Year-Olds, Pub. no. GAO/HEHS-98-133 (Washington: GAO, June 1998).
8. Ziller et al., Health Insurance Coverage in Rural America; and B.A. Ormond, S. Zuckerman, and A. Lhila, "Rural/Urban Differences in Health Care are Not Uniform Across States," Assessing the New Federalism Series B, no. B-11 May 2000, http://www.urban.org/url.cfm?ID=309533 (accessed 21 August 2006).
9. As expected, this was lower than Schoen and colleagues’ estimate (12.3 percent). In addition to lacking deductible information, our study included only those with private health insurance, while Schoen and colleagues included people with public coverage. Additionally, since the most recently available MEPS panels were fielded in 2001 and 2002, versus the 2003 Commonwealth Biennial Survey, there may be some temporal factors at work. Finally, Schoen and colleagues included spending for all health care services (including dental care), while we limited our inquiry to services typically covered by private health insurance (ambulatory care, inpatient services, prescription drugs, and mental health services). Schoen et al., "Insured but Not Protected."
10. Ibid.
11. P.E. Mohr et al., "Vulnerability of Rural Hospitals to Medicare Outpatient Payment Reform," Health Care Financing Review 21, no. 1 (1999): 1–18.[Medline]
12. Schoen et al., "Insured but Not Protected."
13. K. McDaniel, "Small Business in Rural America," Main Street Economist, May 2001, http://www.kc.frb.org/RuralCenter/mainstreet/MSE_0501.pdf (accessed 21 August 2006).
14. J. Gabel and J. Pickreign, Risky Business: When Mom and Pop Buy Health Insurance for Their Employees (New York: Commonwealth Fund, April 2004).
15. S.H. Long and M.S. Marquis, "Have Small-Group Health Insurance Purchasing Alliances Increased Coverage?" Health Affairs 20, no. 1 (2001): 154–163.[Abstract/Free Full Text]
16. J. Gabel et al., "Health Benefits in 2005: Premium Increases Slow Down, Coverage Continues to Erode," Health Affairs 24, no. 5 (2005): 1273–1280.[Abstract/Free Full Text]

                      What's this?

This article has been cited by other articles:

				T. Somkotra and L. P. Lagrada Which Households Are At Risk Of Catastrophic Health Spending: Experience In Thailand After Universal Coverage Health Aff., May 1, 2009; 28(3): w467 - w478. [Abstract] [Full Text] [PDF]

				K. Voorhees, D. H. Fernald, C. Emsermann, L. Zittleman, P. C. Smith, B. Parnes, K. Winkelman, and J. M. Westfall Underinsurance in Primary Care: A Report from the State Networks of Colorado Ambulatory Practices and Partners (SNOCAP) J Am Board Fam Med, July 1, 2008; 21(4): 309 - 316. [Abstract] [Full Text] [PDF]