Health Affairs, 25, no. 4 (2006): 1103-1112 doi: 10.1377/hlthaff.25.4.1103 © 2006 by Project HOPE	New Online   House Health Reform Bill   Paying for Reform   Vetting AHIP's Report   HIV/AIDS Costs   Brief: Insurance Reform   HA Blog Top 10

This Article Abstract Reprint (PDF) Online Appendix 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 Martin, E. G. Articles by Paltiel, A. D. Search for Related Content PubMed PubMed Citation Articles by Martin, E. G. Articles by Paltiel, A. D. Related Collections Access To Care AIDS/HIV Business Of Health Chronic Care Politics Consumer Issues

Health Tracking

TRENDS

Fact, Fiction, And Fairness: Resource Allocation Under The Ryan White CARE Act

Abstract

 
Debate over the reauthorization of the Ryan White CARE Act (RWCA) raises questions of fairness, equity, and efficiency. Critics charge that the program targets disproportionate resources to historical urban epicenters at the expense of underserved areas of incident HIV infection. We used 1998–2004 data on RWCA allocations to examine these claims. We found that states’ concentration of AIDS cases within urban areas remains the dominant predictor of RWCA funding, although the impact of concentration declined after 2000. Other state characteristics, such as poverty rates or racial/ethnic diversity, play a much smaller role.

Since its inception, the RWCA has sought to balance diverse objectives, including responding to local economic needs, targeting funds to areas of highest HIV/AIDS prevalence, promoting equity, and increasing efficiency. The original 1990 legislation states the intent of Congress to target resources to areas with the greatest HIV/AIDS burden.² Pulling in another direction is the principle that the RWCA should serve effectively (although not legally) as a "payer of last resort." In July 2005 the Bush administration emphasized this objective, proposing resource allocation based upon "severity of need" and taking into account "not only HIV incidence, but levels of poverty, availability of other resources including local, state, and federal programs and support, and private resources."³

In response to these diverse and sometimes divergent objectives, the HIV/AIDS Bureau (HAB) of the Health Resources and Services Administration (HRSA) has established elaborate algorithms to allocate Title I and II resources. At the heart of these allocation formulas is the estimated living cases (ELCs) measure, an approximation of the number of people living with AIDS in a given area. ELCs are estimated by the Centers for Disease Control and Prevention (CDC), based on a lagged average of annual reported AIDS cases, weighted by estimated survival rates for AIDS patients identified in a given year.⁴

States and localities receive a core allocation of RWCA funds based largely on the number of ELCs. EMAs also compete for Title I supplements based upon specific local needs. Awards are further constrained by "hold-harmless" provisions, which prevent precipitous declines in RWCA allocations. To ensure that rural areas can compete with urban centers for resources, the 2000 reauthorization legislation earmarked 20 percent of Title II funds for non-EMA states. A few states that have very low estimated AIDS prevalence receive a minimum $500,000 to finance infrastructure.

Within each title, the result of these formula complexities is an allocation that is roughly proportional to AIDS burden but that has some flexibility to respond to local needs. Funding varies across jurisdictions, both in absolute terms (with more dollars going to states and EMAs that document the highest number of known AIDS cases) and when measured per unit of HIV/AIDS burden.⁵

Critiques of allocations. Stakeholders with different conceptions of efficiency, need, and fairness contest these allocations. Some argue that southeastern states and rural localities represent the leading edge of the HIV epidemic but are penalized by formulas that favor historical epicenters.⁶ Others assert that rural areas receive too little funding because the funding streams penalize states without EMAs.⁷ Still others defend current formulas by delineating "differences between the local, community-driven services provided in EMAs, and the state-level services funded by Title II" and by noting potential incentives to under-invest local resources that arise from the RWCA’s role as "funder of last resort."⁸

These conflicts extend to discourse on the share of RWCA funds allocated to populations marginalized by minority or socioeconomic status. Although African Americans constitute one-eighth of the U.S. population, they accounted for roughly 40 percent of cumulative AIDS cases and half of all incident cases by 2003.⁹ The National Minority AIDS Council has demanded that a "full complement" of resources be earmarked to communities of color.¹⁰ The Bush administration has echoed this view, advocating "more responsive" legislation for "minority communities who disproportionately suffer from the disease."¹¹

IOM reevaluation. As part of fiscal year 2000 reauthorization legislation, Congress directed the Institute of Medicine (IOM) to reevaluate RWCA formula allocations.¹² The IOM identified the presence of an EMA and the concentration of states’ cases in EMAs to be the most important predictors of allocations per ELC.

Our analysis. The IOM report served as the point of departure for our analysis. As did the IOM, we explored combined Title I and Title II formula allocations as a function of concentration of ELCs in EMAs, southeastern geography, racial distribution, and poverty rate. However, the IOM’s analysis was restricted to cross-sectional 2001 data; we used longitudinal data from 1998–2004. This longitudinal approach permitted analysis of funding trends and allowed us to assess the redistributional effects of the 2000 reauthorization. Our analysis also expanded upon the IOM’s two regression models, formally examining multicollinearity among predictor variables of policy concern. Freed from the IOM’s specific charge and institutional mandate, we examined political power as an explicit variable and explored changes in the RWCA allocation formulas as a means of remediation.

Our work represents a timely contribution to the upcoming congressional debate over reauthorization. HRSA administrator Elizabeth Duke has argued, "Currently, CARE Act money is distributed through formulas that use unreliable data, double count AIDS cases and include hold-harmless provisions. These formulas don’t even take into account the number of H.I.V. cases in a community."¹³ A 2005 New York Times editorial countered that the administration plans to "slash financial aid to the metropolitan areas that have done the most pioneering work in AIDS...and redirect the money to the states that have lagged behind."¹⁴ The San Francisco AIDS Foundation claimed that elimination of hold-harmless provisions would "undermine the continuity of care for thousands of San Franciscans living with HIV disease."¹⁵

With so many conflicting viewpoints engaged, our aim is to clarify public debate by examining RWCA funding allocations across time. We cannot reveal whether policymakers have properly allocated RWCA resources. We can, however, identify key parameters that explain how funds were actually allocated and establish whether existing allocations deviate in measurable ways from a dollars-per-ELC measure of equity.

Study Data And Methods

Top Study Data And Methods Study Results Discussion NOTES

 
We used publicly available data from the CDC, HRSA/HAB, and U.S. Census Bureau to evaluate state characteristics related to variation in funding from 1998 to 2004.¹⁶ We limited our analysis to RWCA funding in the fifty states and the District of Columbia. RWCA allocations to territories outside the United States raise issues beyond the scope of this paper. Also excluded were "emerging communities" awards, which represented less than 1 percent of total (Titles I and II) funds.¹⁷ We excluded data from odd years (1997, 1999, 2001, and 2003).¹⁸ Building on the IOM report, we considered the following independent variables: (1) concentration (the percentage of a state’s ELCs who reside in an EMA), which was set to zero for states without EMAs; (2) southeastern region, using coding consistent with the CDC’s geographic categorization; (3) poverty (percentage of state residents living below the federal poverty level); and (4) race/ethnicity (two separate variables—the percentage of self-identified African Americans and the percentage of self-identified Hispanics in a given jurisdiction).¹⁹ Deviating from the IOM report, we also included a measure of political power: the number of U.S. House members in 2000. Following the IOM’s analysis on interstate allocations, we could not include data on incident HIV cases—one measure of current "need"—because they were not available from all states, including several key states with high disease burden.

The state-specific trajectories of dollars-per-ELC funding revealed a marked shift in slope in 2000. We hypothesize that this corresponds to the new presidential administration and to the 2000 RWCA reauthorization. We used a "change point" specification to test formally for a change in time trends after 2000, which was not possible in the IOM model.²⁰

We used univariate and bivariate statistics to examine the distribution of covariates throughout the period and their relationship to one another. Extending the population-weighted linear regression framework used by the IOM, we used linear mixed effects (LME) models, estimated through restricted maximum likelihood estimation. These specifications included southeastern region, geographic concentration of AIDS cases within each state, race, and poverty as fixed effects and the state as the level of analysis. Each LME model included a random-effects intercept to account for variation in baseline values among the states. Time was used as a control variable in all models.

Study Results

Top Study Data And Methods Study Results Discussion NOTES

 
Forty-five percent (N = 23) of states included at least one EMA by the end of the observation period. (Two states—Nevada and Virginia—gained an EMA in 1999.) Exhibit 1 lists univariate statistics on each of the explanatory variables, for all states. It also presents results stratified by the presence of at least one EMA, since EMA states have combined Titles I and II funding.

Plots of the mean dollars-per-ELC values and state trajectories over time suggest greater proportional funding in states with at least one EMA and with larger Hispanic populations (Exhibit 2).²² (As noted in Exhibit 1, states with EMAs have significantly larger Hispanic populations, which could explain this correlation.) There do not appear to be strong effects based on geography, the proportion of African American residents, or state poverty rates.

Multicollinearity among variables for race, poverty, and geography, and associated over-fitting posed a challenge to our small data set (fifty-one jurisdictions over four time periods) in assessing the effects of individual variables, particularly when the concentration of ELCs and time trends explain much of the observed funding variation. In our first two regression specifications, we could not reject the hypothesis that geography, state patterns of race/ethnicity, and poverty rates jointly exert zero influence over RWCA funding allocations. A joint test of these effects was highly insignificant (p = .80 and p = .86, respectively).

Our models showed that after concentration was adjusted for, African American race was a significant predictor of funds: Each percentage-point increase in a state’s proportion of self-identified African American residents is associated with a $17 decrease in funds (Model 1). This covariate is statistically significant (p < .0001). However, the resulting effect is modest because the variation across states in the proportion of African American residents is relatively small. Holding all other variables fixed, the difference between states at the seventy-fifth versus twenty-fifth percentile of the distribution of percentage of African American residents is $237 per ELC—less than 5 percent of mean allocations. In contrast, the funding difference between states at the seventy-fifth versus twenty-fifth percentile of the distribution of concentration (for states with at least one EMA) is almost twice as large: $438 per ELC. When all states (with and without EMAs) are included, two states at the extreme ends of the interquartile range of concentration values are expected to differ in funding by $1,658 per ELC. When the District of Columbia was excluded from the regression, African American race was no longer a significant predictor of funds, which underscores the importance of outliers in our small data set.²⁴

We found some evidence that states with higher percentages of African American residents received a small relative increase of funds after the 2000 reauthorization, as measured by the time-interaction variables. However, this relative redistributive effect is small, when the coefficients are expressed as a fraction of the mean dollars-per-ELC allocation. The lack of statistical significance for geography held even after Florida was excluded, which arguably differs from other southeastern states in its urban concentration, HIV history, and demographics.²⁵ Despite legislative changes enacted in 2000 intended to channel more resources to areas of "need," the presence of an EMA remains the dominant predictor of dollars-per-ELC allocations.

The "political power" variable was statistically significant (p < .001) and positive but was small.²⁶ A five-member increase in a state’s delegation was associated with a marginal increase of $66 per case (results not shown).

Discussion

Top Study Data And Methods Study Results Discussion NOTES

 
Debate over reauthorization of the Ryan White CARE Act raises questions of fairness, equity, and efficiency. Our analysis yields several findings that inform this discussion.

Informative findings. First, the RWCA provides useful but modest overall subsidies to states and EMAs. Combined Title I and Title II funding provides approximately $5,000 annually per ELC. In nominal dollars, RWCA budgets increased modestly between 1998 and 2004. However, these budget increases failed to keep pace with the cost growth of overall medical services. As chronicled by the IOM and by others, sizable unmet need persists, even in states that receive comparatively generous RWCA formula allocations. Overall resource scarcity might receive less public attention than cross-state differences in RWCA funding, but national funding trends are no less important.

Second, we confirmed the IOM’s findings that the separate funding streams to states and EMAs, embodied in the dual Title I/Title II allocation, remains the dominant source of funding variation across jurisdictions. States that contain an EMA—particularly those whose ELCs are highly concentrated in these EMAs—receive markedly greater funding per case than other states. Persistent urban/rural discrepancies can be traced to the fact that ELCs within EMAs are counted both for Title I and Title II allocations. The U.S. Government Accountability Office (GAO) has described similar findings.²⁷ Critics describe this structure as "double counting" and discriminatory toward rural localities. Proponents of current policies counter by highlighting distinct goals of Title I and Title II funds.²⁸

Third, despite an elaborate application process for Title I funds, we found little evidence that funds are systematically allocated to advance broader political or substantive policy objectives. The majority of our time–main effects interaction terms are not significant, which suggests no proportional reallocation along these dimensions. Furthermore, where the time interaction terms are significant (for example, for concentration), the magnitude of the effect is small, which again suggests that proportional reallocation is not a major driver of the current distribution of funds.

We found little evidence that southeastern region, a state’s proportion of African American or Hispanic citizens, or state poverty levels notably influence RWCA allocations. Contrary to current discussion regarding geographic inequities, we found little evidence that southeastern states differ from the rest of the country in their dollars-per-ELC allocation.

Regression models showed a statistically significant decrease in funding among states with higher proportions of self-identified African American residents. However, because the range of values for the percentage of African American residents in a state is relatively small, corresponding absolute differences in funding for states of varying minority populations was modest.

Limitations. Our analysis has several limitations that must be considered in evaluating the results. First, our sample is inherently small. We could not disentangle the effects of many important but highly correlated variables. For example, states with greater political influence (as proxied by the number of U.S. Representatives) might strongly influence the interest-group politics of RWCA allocations. Yet state population is also strongly correlated with urbanicity and the presence of an EMA, which we found to be important predictors of per case funds.

Second, we used states’ aggregate Title I and Title II formula allocations as the key dependent variable. We did not explore the heterogeneity of services provided, policy objectives pursued, the availability of other resources, and government levels targeted under these two funding streams. This approach presumes that there is some fungibility of state and local resources available for HIV- and AIDS-related services and that it is appropriate to scrutinize the overall level of federal resources provided for these purposes.

Third, our analysis focused on Title I and Title II RWCA funding. We did not examine resource allocation within other RWCA program titles. In particular, Title III finances important early intervention services, often in non-EMA states. We excluded Title III from our analysis for several reasons. First, Title III funds are paid directly to providers rather than to states or localities. Title III is also tightly focused on early interventions, such as HIV counseling and testing, and adherence support. In contrast, Titles I and II support a broad range of activities, including AIDS Drug Assistance Programs, substance abuse and mental health treatment, nutritional counseling, and emergency housing assistance. Title III is also comparatively small (only about 12 percent of combined Title I and II spending).

Our dependent variable describes determinants of macrolevel federal allocations. As such, it does not address contemporary debates on whether the dual funding streams—in which EMA cases are tallied twice, for both Title I and Title II allocations—represent unjustified "double counting" or whether these funds are being used for distinct services. Furthermore, we could not quantitatively assess situations in which the RWCA successfully compensates jurisdictions that face costs or barriers beyond their immediate control.

More fundamentally, this outcome measure frames the analysis as a search for departures from a constant, per ELC allocation of funds. Conceptions of need and equity could reasonably be expanded to include within-state funding disparities or the presence of "hard to reach" or hidden populations. They could take account of variation across jurisdictions in HIV treatment and prevention costs as well as regional differences in the scope, range, and intensity of services provided. Small or fiscally constrained areas might define need in terms of additional resources required to bolster their public health infrastructures and improve the efficiency with which they can respond to the epidemic and deploy their RWCA funding.

Proportional allocation also enacts a weak measure of equity. Indeed, one should not be surprised or disturbed to find departures from a crude dollars-per-case standard. Such instances might reveal those situations where funding formulas expressly pursue important social objectives or where RWCA compensates jurisdictions that face costs or barriers beyond their immediate control.

Equity and fairness. Tensions among diverse and sometimes competing goals are unavoidable, given a decentralized structure of health coverage, intrastate differences in political values and resources, and the multiple purposes and values served by RWCA resources. It was not within the scope of this analysis to suggest how these tensions should be resolved. Our findings do help correct key misperceptions in public debate. Our analysis bolsters the IOM’s finding that conventional explanations of differences between jurisdictions in per ELC dollar allocations—hold-harmless provisions and the use of AIDS cases rather than HIV cases—are less influential than is often suggested in public debate. We found no evidence to support contemporary arguments that southeastern states receive less money per case. We found little evidence that the modifications to allocation formulas enacted in the 2000 reauthorization have noticeably redistributed funds in a systematic manner to achieve stated societal objectives.

Viewed in this light, it is worrisome that our analysis identified so few deviations from the dollars-per-ELC benchmark of equity. Despite the complexity of RWCA funding algorithms and supplemental application procedures, we could identify little systematic targeting of funds to achieve explicit goals beyond the allocations dictated by the pattern of ELCs within and across Title I and Title II jurisdictions. The very complexity of RWCA funding algorithms and supplemental formula allocations could undermine the transparency and consistency required to strongly target funds to specific jurisdictions or recipient populations at the implicit expense of others.

If policymakers wish to reduce funding gaps between EMA and non-EMA states, these could be closed with simple adjustments to current allocation formulas. Since the 2000 reauthorization, the Title II allocation formula has recognized this policy lever by earmarking 20 percent of Title II funds for non-EMA states. Exhibit 2 illustrates that this change was effective in reducing—but not eliminating—the funding gap.

Policymakers could also explicitly allocate RWCA funds to low-income jurisdictions using algorithms used in other federal programs. In allocating Medicaid resources, for example, the federal government explicitly subsidizes low-income states via federal medical assistance percentages. This measure is based on state per capita income and is highly correlated with both poverty rates and state fiscal capacity.²⁹ RWCA algorithms currently include no such adjustments.

Richard Zeckhauser has noted that the best tool of policy analysis is long division.³⁰ Despite many imperfections of a dollars-per-ELC equity standard, it provides a powerful point of departure in scrutinizing current policies in light of one’s considered judgments of equity, efficiency, local need, political bargaining, and other considerations that properly influence RWCA formula allocation. We hope that our analysis can assist such scrutiny in serving people living with HIV and AIDS.

Editor's Notes

 
Erika Martin (erika.martin{at}yale.edu) is a doctoral candidate in health policy and administration at Yale University in New Haven, Connecticut. Harold Pollack is an associate professor in the School of Social Service Administration and faculty chair, Center for Health Administration Studies, at the University of Chicago, in Chicago, Illinois. David Paltiel is an associate professor at the Yale School of Medicine.

The authors thank Colleen Barry, Paul Cleary, Joel Dubin, and Mark Schlesinger for their helpful comments on this manuscript. David Paltiel is supported by Grant no. R01-A015612 from the National Institute on Drug Abuse.

NOTES

Top Study Data And Methods Study Results Discussion NOTES

 

1. See Health Resources and Services Administration, HIV/AIDS Bureau, "Ryan White CARE Act," http://hab.hrsa.gov/history.htm (accessed 18 April 2006).
2. Ryan White Comprehensive AIDS Resources Emergency Act of 1990, P.L. 101–381 (18 August 1990).
3. U.S. Department of Health and Human Services, "Ryan White Care Act Reauthorization Principles," Fact Sheet, 27 July 2005, http://www.hhs.gov/news/press/2005pres/ryanwhite.html (accessed 27 October 2005).
4. Centers for Disease Control and Prevention, "AIDS Cases Reported to CDC, July 1989 through June 1999, and Estimates of the Number of Persons Living with AIDS, by State and Metropolitan Area, as of June 30, 1999," HIV/AIDS Surveillance Supplemental Report 6, no. 1 (2000), http://www.cdc.gov/hiv/stats/hasrsupp61.pdf (accessed 18 April 2006).
5. Institute of Medicine, Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act (Washington: National Academies Press, 2004).
6. "Politics and Policy—Ryan White Care Act: Coburn Criticizes Fund Distribution," American HealthLine, 24 May 2000.
7. E.M. Duke, "Guarding the Fight against AIDS" (Letter to the Editor), New York Times, 29 August 2005.
8. Sen. Hillary Rodham Clinton, "Senators Clinton, Lautenberg, and Others Fight to Preserve Funding for States with High HIV/AIDS Populations," Press Release, 23 June 2005, http://clinton.senate.gov/news/statements/details.cfm?id=239909 (accessed 27 October 2005).
9. CDC, "Cases of HIV Infection and AIDS in the United States, 2003," HIV/AIDS Surveillance Report no. 15 (2003), http://www.cdc.gov/hiv/stats/2003SurveillanceReport.htm (accessed 18 April 2006).
10. Henry J. Kaiser Family Foundation, "Daily HIV/AIDS Report: HHS Proposes Principles for Ryan White CARE Act Reauthorization; Focus Is on Need-based Grant Distribution," 28 July 2005, http://kaisernetwork.org/daily_reports/rep_index.cfm?DR_ID=31677 (accessed 27 October 2005).
11. DHHS, "Ryan White Care Act Reauthorization."
12. IOM, Measuring What Matters.
13. Duke, "Guarding the Fight against AIDS."
14. "Guarding the Fight against AIDS" (editorial), New York Times, 18 August 2005.
15. Kaiser Family Foundation, "Daily HIV/AIDS Report."
16. For further methodological detail, including graphical presentation of RWCA funding trends, see Technical Appendix at http://content.healthaffairs.org/cgi/content/full/25/4/1103/DC1.
17. Results from regressions that include emerging community funds are available from the authors but are virtually identical to those reported here. Contact erika.martin{at}yale.edu.
18. We describe the reasons for this in the Technical Appendix, cited in Note 16.
19. Ibid., regarding the CDC’s geographical categorization. Following the IOM report, we used dollars per ELC as our outcome. The accuracy of the ELC measure strongly influences the validity of the analysis; however, we deliberately used this case index to test how well the government adheres to its own defined standards of fairness.
20. Any changes from the 2000 reauthorization and administration change would likely be seen in 2001 or 2002. However, since we used data from even years only, we set 2000 as our change point.
21. See Appendix Exhibit 1 (cited in Note 16) for a state-specific list of the number of ELCs and total Title I and II awards in 1998 and 2004.
22. Graphical depictions of the numbers in Exhibit 2 (smoothed mean curves superimposed on state-specific trajectories) are available as Appendix Exhibits 2 and 3, cited in Note 16.
23. Separate regressions for states with and without Title II funding confirm conclusions from models presented here.
24. Results excluding the District of Columbia are available from the authors; see Note 17.
25. Ibid., for results excluding Florida.
26. Ibid., for results of the political power variable.
27. U.S. Government Accountability Office, Ryan White CARE Act: Factors That Impact HIV and AIDS Funding and Client Coverage, Pub. no. GAO-05-841T (Washington: GAO, 23 June 2005).
28. Clinton, "Senators Clinton, Lautenberg, and Others Fight to Preserve Funding."
29. DHHS, "Federal Financial Participation in State Assistance Expenditures; Federal Matching Shares for Medicaid, the State Children’s Health Insurance Program, and Aid to Needy Aged, Blind, or Disabled Persons for October 1, 2005 through September 30, 2006," Federal Register (24 November 2004), http://aspe.hhs.gov/health/fmap06.pdf (accessed 27 October 2005).
30. Richard Zeckhauser, Kennedy School of Government, class lecture, September 2002.

                      What's this?

This article has been cited by other articles:

				S. Kazi and S. Boroumand Assessing Quality of Primary Care Provided to the HIV-Infected Ryan White Population in the Baltimore Eligible Metropolitan Area American Journal of Medical Quality, November 1, 2008; 23(6): 484 - 491. [Abstract] [PDF]

				R. J. Buchanan and W. Hatcher Compassionate Conservatism: Federal Funding for the Ryan White CARE Act During the Bush Administration Am J Public Health, November 1, 2007; 97(11): 2013 - 2016. [Abstract] [Full Text] [PDF]