Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 87
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act 4 HIV Reporting Data and Title I and II Formulas Seventy percent ($1.3 billion in fiscal year [FY] 2002) of all Ryan White CARE Act (RWCA) funds are distributed under Titles I and II through explicit numerical formulas. These formula allocations are based on estimated living AIDS cases (ELCs),1 which are calculated using data from the AIDS case-reporting system2 (HRSA, 2002). Discussions, testimony, and Congressional committee reports related to the 2000 reauthorization raised questions concerning inequity in the allocations resulting from these formulas (U.S. Congress, 2000a,b). By 2000, concerns about inequity arose from the perception that the epidemic was not truly reflected by AIDS cases alone, and that the effect of addressing HIV disease in areas with emerging epidemics had been underestimated (U.S. Congress, 2000a,b). Jurisdictions were also concerned that they were not compensated for providing early access to care and treatment and thus preventing persons from progressing to AIDS (U.S. Congress, 2000a). The hold-harmless provisions, which were added in the 1996 reauthorization, also raised concerns about equity. Some were concerned that the hold-harmless provisions, which prevent a jurisdiction’s funding falling by 1 ELCs are calculated by applying annual survival weights to the most recent 10 years of reported AIDS cases and summing the totals. The Centers for Disease Control and Prevention (CDC) updates the survival weights every two years. CDC provides both the survival weights and the most recent 10 years of reported AIDS cases to the Health Resources and Services Administration (HRSA), which performs the award calculations. 2 See Chapter 3 for background on the AIDS and HIV case-reporting systems.
OCR for page 88
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act more than a set percent each year in order to help sustain needed health care infrastructure and continuity of services, seemed to accomplish their purpose at the expense of states and local areas believed to have younger epidemics and rising need (U.S. Congress, 2000a). Congress asked the General Accounting Office (GAO) to examine opportunities to enhance the equity of funding to RWCA grantees prior to the 1996 and 2000 reauthorizations. The GAO’s 1995 study found that CARE Act funding formulas led to disparities in per AIDS case funding that could not be completely explained by variations in service costs or the fiscal capacity of states and EMAs. In a 2000 report, the GAO again found large disparities across Eligible Metropolitan Areas (EMAs) and states in allocations per ELC. The GAO concluded that two formula features in particular, the hold-harmless provision and the “double-counting” of EMA cases in Title I and II formulas,3 contributed to these funding inequities. In particular, states with an EMA had up to 60 percent higher per case allocations than states without an EMA and the hold-harmless provision instituted in the 1996 reauthorization benefited only San Francisco (GAO, 2000). The GAO report further concluded that the formulas, which were based on living AIDS cases, did not reflect the changing nature of the HIV/AIDS epidemic and recommended the inclusion of HIV case data in the Title I and II formulas to more effectively target and deliver funding to persons in need of care. The GAO noted that, at a minimum, all states would have to report HIV cases to provide an equitable distribution of funds. At the time, only 60 percent of states had HIV reporting systems in place (GAO, 2000). Congress began the 2000 reauthorization with the expectation that HIV case-reporting data would be of value to the RWCA formula allocations, as well as to planning and evaluation efforts. The 2000 legislation specifies that, if appropriate, the Secretary of Health and Human Services (HHS) should incorporate cases of HIV disease in RWCA Title I and II funding formulas as early as FY2005 but no later than FY2007 (Ryan White CARE Act. 42 U.S.C. § 300ff-28 ). The reauthorization legislation authorized the Institute of Medicine (IOM) to assist the Secretary of HHS in assessing the readiness of states to produce accurate and reliable HIV case-reporting data, determine the accuracy of using HIV cases within the existing allocation formulas, and establish recommendations regarding the manner in which states could improve their HIV case-reporting systems (Ryan White CARE Act. 42 U.S.C. § 300ff-11 ).4 3 EMA cases are counted in both Title I and II formulas. 4 See Chapter 1 for legislative language relating to the Committee’s charge.
OCR for page 89
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act At the outset, the Committee recognized the difficulty of defining such terms as “sufficiently accurate,” “adequate,” and “reliable” to characterize the quality of HIV surveillance data. No absolute standards of accuracy, adequacy, or reliability exist; rather these standards and their definition will vary according to the purposes and tasks for which the data are used. Thus, evaluating HIV reporting systems for use in resource allocation formulas requires a different set of performance criteria than evaluating these data for public health purposes (e.g., epidemic surveillance, contact tracing, and partner notification). While CDC (1999) has established performance criteria5 for evaluating the latter functions, it was not the purpose of this Committee to assess these standards. Rather, the Committee defined criteria for “accuracy,” “adequacy,” and “reliability” in terms of over- and underfunding errors to RWCA grantees. In assessing whether the current surveillance systems provide HIV data that are “sufficiently accurate” for the purpose of formula grants, the Committee focused on the primary argument for including HIV case data in the formulas: that doing so would provide a better representation of HIV disease-related resource needs across jurisdictions and would thus more fairly channel scarce RWCA resources. Four conditions would need to be met for this to occur: First, the HIV reporting systems of all states would need to be capable of providing data that are used for the formulas. Second, the quality of HIV data across jurisdictions would have to be comparable. This means that if there are biases in HIV reporting, those biases should be of similar direction and magnitude across areas. Third, incorporating HIV case-reporting data in the formula would need to produce different and more accurate assessments of “relative disease burden” and resource needs than AIDS data alone. Fourth, including HIV data in the RWCA allocation formulas would have to result in material variation in the relative size of awards to states and EMAs and more equitable allocations. If formula provisions, in 5 CDC issued surveillance guidelines in December 1999 outlining the performance criteria for states’ HIV/AIDS surveillance systems. The criteria outlined in the document stated that the system had to be timely (66 percent of HIV/AIDS cases reported to the health department within 6 months of diagnoses), provide accurate case counts (≤5 percent mis-matched reports and ≤5 percent duplicate cases in the database), have complete ascertainment of mode of exposure to HIV (≥85 percent of reported cases, or representative sample, must have reported transmission mode after complete epidemiologic follow-up), and must include complete case reporting (≥85 percent of diagnosed cases reported to the health department). In addition, states must show that they can match to other databases of public health importance, follow up cases of public health importance, collect valid and reliable data for key data elements, and use data for public health planning (CDC, 2003i).
OCR for page 90
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act contrast, limit or dampen the influence of new HIV data, then the formulas may not result in more equitable allocations. The Committee sought evidence to determine whether each of these conditions would be met if reported HIV cases were used in the formulas. Specifically, the Committee examined: Whether state HIV reporting systems are capable of providing data for the formulas; Whether the quality of HIV data across jurisdictions is comparable; Whether the relative ranking of need among states and EMAs varies depending on whether HIV case data or AIDS case data are employed to measure disease burden; and Whether the RWCA formulas are sensitive enough to translate changes in input data into more equitable allocations. Each of these conditions proved difficult to verify. The evidence does not lead unequivocally to the conclusion that inclusion of HIV case-reporting data in the formulas would lead to a more equitable allocation of RWCA resources. Finally, the Committee concluded that it was beyond its capacity to evaluate the HIV case-reporting system of each state and territory. The Committee’s recommendations, however, could be used to make general improvements to HIV case reporting for allocating RWCA resources. CAPABILITY OF STATE HIV REPORTING SYSTEMS TO PROVIDE DATA FOR THE FORMULAS Three criteria are particularly relevant for evaluating the capability of HIV reporting systems to provide data for allocating resources under RWCA: Coverage: Does each state have an HIV reporting system? Maturity: Has the HIV reporting system of each state had sufficient time for full implementation? Full use of available data: Would the formulas use HIV reporting data from every state with a system of HIV reporting? Coverage For HIV case-reporting data to be used in RWCA funding formulas, all states and territories would need to report HIV cases. As of October
OCR for page 91
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act 2003, all states, territories, and cities except Georgia6 and Philadelphia, had implemented a confidential HIV case-reporting system (CDC, 2003a,h) (see Chapter 3, Table 3-1). Unlike AIDS case reporting, which uses a standardized name-based reporting system, states have adopted different procedures for reporting HIV cases (Chapter 3, Figure 3-1). As of October 2003, 34 states, the Virgin Islands, American Samoa, Puerto Rico, Northern Mariana Islands, and Guam had implemented the same confidential name-based reporting of HIV infection as is used for AIDS reporting and other reportable diseases and conditions. Eight states plus the District of Columbia use a coded identifier rather than the patient name to report HIV cases. Five states use a name-to-code system; initially, names are collected and then converted to codes by the local or state health department after any necessary public health follow-up. Connecticut conducts pediatric surveillance using name-based reporting but allows name or code reporting of adults/adolescents over 13 years of age. New Hampshire allows HIV cases to be reported with or without a name (CDC, 2003a,h). Of the 15 areas that use some form of code, only two use the same code. Maturity Case-reporting systems for new diseases take time to mature and become fully operational. For a system to operate well, physicians and other practitioners need to be educated about the need for new requirements for disease reporting. The burden of new reporting obligations can be increased by complex data requirements, such as the creation of encryption codes for patients in states with code-based reporting. For a disease like HIV infection, where physicians may have followed patients prior to the initiation of an HIV reporting requirement, practitioners will need to report a backlog of existing patients when HIV reporting is first implemented. This takes time and clinician effort, particularly in high-morbidity states that have only recently implemented HIV reporting, such as California. Even laboratories, which rely more heavily on information automation, will still require some time to completely develop, refine, and implement reporting procedures. Further, health departments need time to design and pilot test their surveillance systems for capturing and analyzing newly reportable disease data. While the Committee could not identify any standard criterion for 6 Since the release of this report, Georgia implemented name-based reporting and Philadelphia adopted code-based reporting. All areas of the United States now have HIV reporting.
OCR for page 92
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act how much time is needed to ensure that state reporting systems are fully operational, it was clear from discussions with surveillance experts from multiple states that it takes at least 18 months to several years after a reporting system is introduced for it to reach a reasonable level of completeness and timeliness (Birkhead, 2002; Kopelman, 2002). In a previous assessment by the GAO, CDC officials estimated that it would take 1 to 3 years for the backlog of HIV cases to be entered into a new reporting system (GAO, 2000). The GAO study compared the experience of states that had been reporting HIV for different periods of time and found: “The potential for lags in reporting the older cases was clear when we compared the experience of states that had been reporting HIV cases for different lengths of time. States with long reporting histories had many more HIV cases compared with their number of AIDS cases than did newly reporting states” and that “states that begin reporting more recently may continue for some time into the future to have a larger proportion of previously diagnosed but not reported cases” (GAO, 2000). Variations in the maturity of systems can create differential errors across states and EMAs. Immature systems capture a lower percentage of prevalent cases and are more likely to be missing key pieces of information. The HIV reporting systems of states are in various stages of maturity. Some, such as Minnesota and Colorado, implemented HIV reporting in the mid-1980s and have mature systems. Other states such as California and Pennsylvania adopted HIV reporting only recently and may require additional time to report the backlog of cases. Several of the factors determining system accuracy, such as the ability to follow up on a backlog of cases, depend on the capacity to conduct surveillance. As one indicator of capacity, the Committee examined federal and state funding for HIV/AIDS case reporting (Appendix B). This review identified important issues. First, state HIV/AIDS surveillance programs are largely dependent upon federal financing (Appendix B). In addition, neither federal nor state sources of program funding have changed appreciably from 1999 to 2002 when many states were implementing HIV case reporting. Although the 2000 reauthorization of RWCA authorized limited additional funds to help states implement HIV reporting systems (Ryan White CARE Act. 42 U.S.C. § 300ff-13 ), that funding has yet to be appropriated. Even though HIV and AIDS data may be perceived to be readily available for RWCA purposes at no additional cost, states must often include different or more-detailed data for RWCA planning than are provided in standard epidemiological reports. Such efforts can be costly, especially for states that have recently implemented HIV reporting and for states that lack adequate surveillance resources. States are facing financial crises, and state surveillance programs do not
OCR for page 93
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act anticipate additional state contributions (see Appendix B). Thus, it is important for Congress, HRSA, and RWCA grantees to recognize the burden imposed upon the state surveillance programs as they strive to meet the information needs of RWCA. Additional funds will be required to accommodate the additional informational burdens placed on states by the RWCA. Full Use of Available Data Currently, CDC accepts only name-based HIV reports in the national HIV reporting database, largely because algorithms have not been developed to unduplicate HIV data from the 15 code-based states and territories. Duplicate case reporting can occur for several reasons: case reports are completed by different practitioners at several different times (HIV diagnosis, AIDS diagnosis, and death); laboratories may send results from diagnostic and staging laboratory tests (CD4+ cell count or viral load) independently to the health department; and people may move and be reported in both their original and new state of residence (CDC, 2003b). CDC estimates that approximately 5 percent of HIV cases (from name-based reporting states only) in the national dataset represent duplicate reports. CDC suggests the potential for greater duplication grows as state HIV reporting systems mature and as people remain healthier longer owing to antiretroviral therapy (CDC, 2003b). Name-based reporting is cited as one way to facilitate elimination of duplicate reports (CDC, 1999). However, it is unclear if name-based reporting is intrinsically superior to code-based reporting for eliminating duplicate reports. Due to name variations, even name-based systems do not permit complete unduplication. In addition, code-based reporting systems were developed by some states after substantial political debate, and altering those systems would require significant legislative changes, time, and effort. For this reason, and because name-based reporting is not clearly superior to code-based reporting for the specific goal of accurately estimating the number of known cases for determining RWCA allocations, better methods for unduplicating reports for both code- and name-based reporting states need to be developed and implemented so allocation formulas can include data from all states. At the same time, the Committee recognizes that there are strong feelings, both pro and con, about the use of name-based reporting for other surveillance functions (Colfax and Bindman, 1998; Osmond et al., 1999; Hecht et al., 2000). The Committee did not take a position on these issues because its charge limits its scope of activity to reporting for allocation purposes.
OCR for page 94
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act Finding About States’ Capability to Provide Data for the Formulas Finding 4-1 While the Committee supports Congressional intent to incorporate data into the RWCA allocation formulas that reflect the evolving needs of the epidemic, the Committee finds that states’ HIV reporting systems are neither ready nor adequate for purposes of RWCA resource allocation. One state and one city have yet to implement HIV case reporting, states’ HIV reporting systems are in different stages of maturity across the United States, and the national HIV database does not include HIV cases from code-based reporting states. COMPARABILITY OF DATA QUALITY ACROSS JURISDICTIONS Even if states are capable of providing data on HIV cases, data would need to be of comparable quality across jurisdictions before they could be used in the RWCA formulas. Differences in the completeness and timeliness of data across jurisdictions have the potential to create significant biases in allocations. The greater the variability in the way HIV data are collected across states or EMAs, the greater are the chances for bias. The inclusion of data of varying quality across jurisdictions can decrease rather than increase the equity of resulting RWCA allocations.7 It is important to note that not all biases will adversely affect the fairness of resource allocation. Biases in prevalence that are consistent across states or EMAs will not affect allocations if the formulas depend only on the relative value of these measures across states or EMAs rather than the absolute value. For example, if all states underreported cases by 30 percent, then there would not necessarily be an effect on allocations, depending on the nature of the formula used (although such underreporting would be important for determining the gap between prevalent and diagnosed cases). By contrast, if variability across states in underreporting were large, such variability might have a major influence on allocations. Measuring or reducing biases may entail substantial cost. Such costs need to be weighed against the likely improvements in the allocation process. HIV data should be included in the formulas only if doing so enhances the equity of the resulting allocations. To examine these potential biases, the Committee reviewed published evaluations of the completeness and timeliness of HIV and AIDS case-reporting systems. These studies are summarized in Table 4-1. The Committee found that most evaluations have focused on the completeness of reporting—the degree to which all individuals with these conditions are 7 The concept of equity is discussed in Chapter 1.
OCR for page 95
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act reported to public health authorities. The accuracy of the elements of the report such as sex, primary risk factor, and place of residence has received less attention. To the extent that characteristics such as sex and primary risk factor are systematically misreported or underreported, estimates of the prevalence of HIV or AIDS based on those characteristics will also be biased. These types of biases would not necessarily affect allocations across states and EMAs if the proportion of such subgroups was similar across jurisdictions. Given the more recent advent of HIV case reporting, most published evaluations have focused on AIDS rather than HIV case reporting. The Committee found little information from existing evaluations about the completeness, accuracy, and timeliness of HIV case reporting, and about the variation in these factors for HIV reporting across states and EMAs. Because of the inadequacy of available information, the Committee could not fully investigate the potential influence of possible patterns of underreporting or reporting delays on resulting formula allocations. Completeness of AIDS and HIV Case Reporting Studies of the completeness of reporting compare AIDS and HIV case reports with independent data on AIDS or HIV cases that should have been reported—typically medical or administrative records or death certificates. The external source may also be incomplete, so capture– recapture methods may be required.8 Accuracy studies typically look at the correspondence between reported characteristics of cases that appear in both sources, so the results may not be representative of individuals who appear in only one or neither source. The Committee notes, however, that not all elements requested in case-reporting forms may be relevant to resource allocation. While certain data elements may be important in determining whether a state’s reporting system serves the state’s own case-finding purposes and identifies populations at risk, such information may not be relevant to resource allocation. AIDS case reporting is the most complete and highest quality of nearly any disease surveillance system (Doyle et al., 2002). AIDS case reporting 8 Capture–recapture methods can be used to adjust for incomplete ascertainment using information from two or more distinct sources. Capture–recapture estimates the size of a population (in this case, the number of cases) by making statistical assumptions about the proportion of individuals identified in various samples of the population (in the case of surveillance, reported from different sources) (Hook and Regal, 1995). The U.S. Census Bureau uses a similar method (dual-systems estimation) to estimate the U.S. population and population groups (NRC, 2001).
OCR for page 96
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act TABLE 4-1 Completeness and Timeliness of HIV and AIDS Case Reporting Study AIDS or HIV Purpose Methods AIDS STUDIES Scheer et al. (2001) AIDS To determine whether AIDS surveillance misses a substantial number of persons who die with unreported AIDS. Cross-sectional survey of decedents examined by San Francisco Medical Examiner. Decedents with positive or indeterminate HIV antibody test results were cross-referenced against the SF AIDS registry. Medical records of unreported cases reviewed to determine whether AIDS had been reported prior to death. Klevens et al. (2001) AIDS To assess the completeness, validity, and timeliness of AIDS surveillance system after the 1993 surveillance case definition change. In Louisiana and San Francisco, completeness was assessed by comparing the number of persons found in health facilities (hospitals, outpatient clinic, and private providers’ offices) to the number of cases reported to the AIDS surveillance registry. In Massachusetts, completeness assessed using capture–recapture method. Validity was assessed by comparing agreement of case report with medical record for same sites. Timeliness calculated using median delay from time of diagnosis to case report for same sites. Jara et al. (2000) AIDS To assess the completeness of AIDS case reporting in Massachusetts. To determine the effect of the 1993 AIDS case definition on the completeness of AIDS case reporting to the state registry and unreported case based on sex, race, and mode of transmission. Multisource capture–recapture using 1994 Massachusetts Uniform Hospital Discharge Data Set (UHDDS) and Medicaid claims was used to evaluate completeness of Massachusetts state registry.
OCR for page 97
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act Setting Results Conclusions San Francisco Diagnosis and reporting of AIDS 93% complete. HIV-infected decedents were more likely than uninfected to be men and <45 years old, less likely to be Asian/Pacific Islander or Native American, and more likely to have died of suicide or drug abuse/overdose. AIDS case reporting in San Francisco is highly complete. Current surveillance activities which identify cases from health care settings are appropriate. Louisiana, Massachusetts, San Francisco Completeness of case reporting in hospitals (≥93%) and outpatient clinics (≥90%); validity/concordance of info for sex was high (>98%), but lower for race/ethnicity (>83%) & mode of exposure to HIV (>67%); median reporting delay was 4 months, but varied by site from 3 to 6 months. Completeness, validity, and timeliness of AIDS surveillance system remain high after 1993 change in case definition. Massachusetts 92.6% (95% CI 91.6-93.5) complete using UHDDS; 94.5% (95% CI 93.7-95.3) complete using Medicaid claims dataset. Being unreported was significantly more likely in women than in men (OR = 1.72. 95% CI 1.20-2.46), and slightly more likely in IDU than in MSM (OR = 1.49, 95% CI 1.00-2.23). Completeness of state AIDS registry is high, but there are differences by gender and mode of transmission of HIV.
OCR for page 124
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act over, states whose ELCs are concentrated within an EMA benefit from current allocation rules. This is because reported cases within an EMA contribute to both a state’s Title I award and its Title II funding. States that lack an EMA face a corresponding disadvantage. Explicit set-asides for non-EMA states under Title II. Several features of the Title II award are designed to compensate jurisdictions that lack an EMA. These include the 20 percent of the Title II award that is reserved for non-EMA states and the minimum Title II base award of $500,000 per state. Both of these formula features would not be affected if HIV plus AIDS cases, instead of just AIDS cases, were used in the formulas. Set-asides for emerging communities. Title II Emerging Communities provisions, defined as cities with 500–1,999 reported AIDS cases in the most recent 5 years, expressly set funds aside for non-EMA localities. While the Emerging Communities provision may be an appropriate response to the geographic expansion of the epidemic, it will reduce the effect of including HIV data in allocation formulas. Finding 4-6 Several structural features of the Title I and Title II funding formulas—most notably the counting of EMA cases in both Title I and II state formula allocations, but also such measures as hold-harmless provisions and set-asides for emerging communities—have a large influence on resulting allocations. Such structural features may dampen the effect of variation introduced by the addition of HIV cases, and could obviate the potential benefits of adding HIV cases to the CARE Act allocation formulas. The Committee notes with concern that southeastern states appear to receive the smallest allocations per ELC under current allocation rules. Some of this disparity arises from the rurality of southeastern states. People living with AIDS in the Southeast are less likely to reside in EMAs than are their counterparts in other regions. Viewing combined Title I and Title II funding, southeastern states are thus less likely to benefit from counting of EMA cases in both the Title I and II formulas. Southeastern states might also benefit from changes in RWCA allocation formulas that consider HIV in addition to reported AIDS cases. However, the role of EMAs in RWCA formula allocations appears to matter more than alternative definitions of HIV burden in accounting for regional differences in per ELC funding. The Committee also notes the discordance between the intent of the RWCA formulas and their structure. RWCA is statutorily limited to acting as a payer of last resort, as it precludes expenditures for anything covered by other public or private insurance or benefit programs. Funds are intended for services to individuals who are low income, and unin-
OCR for page 125
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act sured or underinsured. Yet formula allocations are made purely on the estimated number of living AIDS cases. Insofar as the current RWCA case-reporting–based formula counts patients that have other sources of insurance or funding, it overestimates the number of cases that qualify for RWCA services, just as it may underestimate the needs of a particular jurisdiction with greater proportions of patients with HIV who are not included in the formula but who would qualify for RWCA services. The current formulas also do not account for variations in the costs of care or fiscal capacity across Title I and II jurisdictions. A 1995 GAO report cited similar concerns with the formulas and concluded that the equity of the formulas could be improved through the use of more appropriate measures of services costs and funding capacity of jurisdictions. Other formula-based programs, including Medicaid and the Substance Abuse and Mental Health block grants, consider costs and/or fiscal capacity (NRC, 2003). Finding 4-7 RWCA Title I and II formula allocations are determined by the ELC. Thus, they do not take into account factors defining those for whom such funds were intended, such as lack of insurance and special needs. That is, there are no provisions to estimate the number of persons in need of a “payer of last resort.” METHODS FOR IMPROVING DATA FOR THE FORMULAS The Committee believes that there are several ways to improve the overall quality and completeness of the HIV case-reporting system for allocating resources under RWCA. First and foremost is the need to include all reported HIV cases in the national database rather than only those reported from states with name-based reporting. Other sources of data, particularly from laboratories, and potentially from pharmacies and other drug providers, can also be more fully utilized to improve the completeness and comparability of HIV reporting systems.20 Twenty-nine states rely on electronic laboratory-based reporting of HIV test results as 20 Most states specifically require laboratories (as well as medical providers) to report cases of HIV or AIDS to the state health department pursuant to their state disease reporting laws or regulations. The Health Insurance Portability and Accountability Act allows covered entities to release this health information to state health departments in compliance with state disease-reporting laws. States that do not expressly require laboratories to report would have to enact a statute or issue a regulation (depending on the state’s statutory structure) to allow laboratories to report. An amendment to the state’s reporting statute or regulation could accomplish this rather simply and ensure that the laboratory is subject to
OCR for page 126
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act part of their HIV reporting systems; comparability in the completeness and timeliness of reporting could be enhanced if all states adopted electronic laboratory reporting. States can also boost their completeness of reporting by including laboratory reporting of other tests unique to HIV infection, such as plasma viral load, CD4+ cell counts, and phenotypic and genotypic resistance testing, as well as pharmacy records for antiretroviral drugs unique to HIV infection. Integrating laboratory-based and pharmacy-based reporting with the National Electronic Disease Surveillance System (NEDSS)21 may require additional research. Second, it is important to explore ways of accounting for differential bias in prevalence estimates due to migration. In addition to these concerns, it is important to evaluate other approaches, such as survey- and modeling-based approaches, for estimating both the overall burden of HIV and the differential disease burden among states and EMAs and to compare these estimates to those produced by case reporting. Such approaches have the potential of providing estimates that are more accurate, more timely, and more consistent across jurisdictions than complete enumeration. One such approach may rely on statistical models and make use of information from a variety of sources. An example of the use of a statistical model is backcalculation of HIV incidence and prevalence from data on AIDS incidence using estimates of the distribution of time from HIV infection to onset of AIDS (Brookmeyer and Gail, 1994). The usefulness of this method, however, has declined over time since the distribution of time from infection to AIDS has become less predictable with the advent of improved therapies. Other attempts to estimate HIV prevalence in specific metropolitan areas have made use of information about sizes of populations at risk and prevalence of HIV infection in those populations from a wide variety of sources (Holmberg, 1996). Holmberg and colleagues suggested as useful sources: the same requirements and privileges as other entities that are required to report, including the duty of the state health department to keep personally identifiable information, if any, confidential. This would allow the state health department to collect relevant data from laboratories in the same manner that it collects data from providers. Thus, statutes protecting the confidentiality of such information should not be a major impediment to collecting information from laboratories and pharmacies to improve the accuracy of HIV case data at the state level. 21 NEDSS is a project to integrate surveillance systems so that appropriate public health, laboratory, and clinical data can be transferred efficiently and securely over the Internet. NEDSS is designed to integrate and replace several current CDC surveillance systems, including the National Electronic Telecommunications System for Surveillance (NETSS), the HIV/AIDS reporting system, the vaccine preventable diseases and systems for tuberculosis (TB) and infectious diseases. See http://www.cdc.gov/nedss/ for more information on NEDSS.
OCR for page 127
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act specific studies of prevalence and transmission among people at risk of infection; information from sexually transmitted disease (STD) clinics, counseling and testing sites, and drug treatment centers; and information from other sources of population testing, such as the Veterans Administration and other medical centers or household-based surveys. The addition of information from population-based surveys could greatly increase the usefulness of such approaches. A previous IOM report (2001) recommended the use of sentinel surveillance in conjunction with population-based surveys as a way to estimate HIV incidence, but it could also be applied to obtain estimates of HIV prevalence. In this approach, one obtains prevalence data from targeted samples of special populations. For example, one might use information from blood donors, military recruits, and/or members of some special high-risk group. One then uses information from other, more representative samples, to estimate the prevalence of the characteristics that identify these special populations. For example, one could determine how likely different groups of people (defined by specific characteristics) in specific areas are to donate blood or join the military. Information from the targeted prevalence studies and representative surveys can then be used to develop estimates of the prevalence of HIV infection. Such an approach has the potential of providing estimates that are more accurate, more timely, and more consistent across jurisdictions than complete enumeration. However, there are also limitations to such approaches. For example, adequate data may not be available to produce accurate models and sampling can be both complex and expensive to implement. Many constituencies are also concerned about the confidentiality and ethics of surveillance surveys. Such a strategy should be reconsidered, with a review of the substantial technical, political, and ethical barriers to its implementation that were pointed out after the 2001 IOM report appeared. Both modeling and survey methods can also be used to estimate cases of undiagnosed HIV infection. Although the primary reasons to have accurate surveillance of the number of persons with known HIV infection are epidemiological, the total number of people with HIV infection is relevant to assessing the size of the population that is likely to need health care services, either currently or in the future, especially if a goal is to encourage everyone to enter care. The resource requirements for treating people who are in care vary widely depending on disease stage, and other factors, such as the patient’s own health insurance status. All HIV-infected persons—even those without clinical symptoms—require some services such as patient education and monitoring as well as treatment of primary infection and associated medical conditions. Most will require extensive medical intervention in the future even if they currently do not. Therefore, information on the total number of HIV infections is impor-
OCR for page 128
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act tant. Although the RWCA now focuses on providing care to individuals who have been diagnosed, if policy shifts to include more outreach to individuals, allocation decisions will benefit from information on the total HIV-infected population. Finding 4-8 The completeness of the data from existing HIV case-reporting systems can be improved by making changes, specifically counting all HIV cases that are reported to the national system rather than only those reported from states with name-based reporting, and more fully utilizing data from laboratories and other sources, such as pharmacies, to enhance the completeness of HIV reporting. Finding 4-9 Techniques exist to estimate the prevalence of HIV infection independently of the HIV case-reporting systems. Sample-based surveys and modeling approaches permit estimates of the total HIV-infected population, regardless of diagnostic status. Finding 4-10 A surveillance mechanism that provides information about the total population of persons with HIV infection, be they diagnosed or undiagnosed, is highly desirable. Knowing the size and distribution of the undiagnosed HIV-infected population is an important marker of success in providing care to all people with HIV. RECOMMENDATIONS The Committee’s recommendations are listed below. These recommendations should be implemented in a timely manner to provide evidence to either (1) justify inclusion of reported HIV cases in RWCA allocations formulas by FY2007, as contemplated by Congress, or (2) conclude that reported HIV cases do not result in more equitable resource allocation so that Congress can reconsider its recommendation prior to implementation in FY2007. Additional resources may be required to implement some of these recommendations. Recommendation 4-1 For at least the next four years, HRSA should continue to use ELCs in the RWCA Title I and II formulas. During that period, concerted effort should be devoted to improving the consistency, quality, and comparability of HIV case reporting. Specific attention should be paid to two, complementary approaches in this regard: (1) the attainment of coverage, maturity, and comparability standards and the development of de-duplication strategies that permit full use of all reported HIV cases; and (2) implementation of alternative strategies for estimating HIV cases, such as survey- or model-based estimation.
OCR for page 129
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act Recommendation 4-2 The following steps should be taken by states as quickly as possible to improve the consistency, quality, and comparability of HIV case reporting for RWCA allocation purposes. The CDC should accept reported HIV cases from all states. Until this occurs, large numbers of HIV cases will not be included in the national HIV reporting system, and there will be no reliable centralized way to use reported HIV cases to apportion CARE Act funds. CDC should work with all states to develop and evaluate methods for unduplicating HIV cases regardless of whether such cases are code- or name-based. The Secretary should provide CDC with the funding to provide the technical assistance to states necessary to support the integration of code with name-based data into the national HIV reporting database. Because of the importance of obtaining consistent data from all jurisdictions, CDC should include HIV reporting data from code-based states and estimate the degree of overcounting due to duplication while procedures and infrastructure for definitive unduplication are developed. CDC should collaborate with all states to periodically assess and compare the completeness and timeliness of their HIV reporting systems. The Secretary of HHS should provide additional funds to CDC to assist states in improving the completeness and timeliness and overall comparability of their HIV reporting systems. Enhancing electronic laboratory reporting in all states is critical in achieving this goal. Pharmacy-based surveillance, with a focus on the AIDS Drug Assistance Program (ADAP), is another potential source of information for enhancing completeness. Recommendation 4-3 CDC should obtain estimates of total HIV prevalence (including the undiagnosed population) and evaluate methods other than case reporting for use as an alternative or supplement in estimating HIV cases for RWCA Title I and II formula allocations, with advice and review by an independent body. This assessment should address the accuracy and costs of different strategies and should be repeated periodically. Recommendation 4-4 Prior to future reauthorizations of the CARE Act, the Secretary of HSS should initiate studies to improve the evidence base for understanding how well HIV case reporting and other methods for estimating HIV cases reflect the relative burden of disease and the relative resources necessary to respond to those needs in different areas. The Secretary should engage an independent body to
OCR for page 130
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act estimate the dollar allocations that would result for Title I and II grantees from alternative input data and alternative RWCA allocation formulas. Specifically: “What-if” assessments should be reported every five years on the range of each EMA’s and state’s RWCA formula allocation, depending on whether ELCs or total HIV cases are used as the measure of disease burden. Analyses should be conducted to estimate the dollar allocations that would result from modifying different structural elements of the formula, such as: Hold-harmless provisions, The eligibility requirements for becoming an EMA, The percentage set-aside in the Title II base award for non-EMA states (currently 20 percent), The minimum base Title II award (now $500,000 for states and $50,000 for territories), The eligibility criteria for becoming a Tier 1 and Tier 2 Emerging Community. Evaluate the extent of interregional variability in HIV epidemic maturity and its effect on relative resource needs. These activities should be repeated periodically. Recommendation 4-5 In keeping with the CARE Act’s intent as a payer of last resort, Congress should reevaluate the RWCA formulas to determine whether they allocate resources in proportion to the estimated number of individuals with HIV/AIDS who are uninsured or underinsured in states and EMAs. Readily available data on the insurance coverage of the general population may mirror insurance coverage of people with HIV/AIDS, but additional estimation will likely be required. REFERENCES Beltrami JF, Vermund SH, Fawal HJ, Moon TD, Von Bargen JC, Holmberg SD. 1999. HIV/AIDS in nonurban Alabama: Risk activities and access to services among HIV-infected persons. The Southern Medical Journal 92(7):677–83. Berk ML, Schur CL, Dunbar JL, Bozzette S, Shapiro M. 2003. Short report: Migration among persons living with HIV. Social Science and Medicine 57:1091–7. Birkhead, G. Panel discussion with committee members during public committee meeting, Washington, DC. May 15, 2002.
OCR for page 131
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act Brookmeyer R, Gail MH. 1994. AIDS Epidemiology: A Quantitative Approach. New York: Oxford University Press. Brookmeyer R, Liao J. 1990. The analysis of delays in disease reporting: Methods and results for the Acquired Immunodeficiency Syndrome. American Journal of Epidemiology 132(2):355–65. Buehler J, Berkelman R, Stehr–Green J. 1992. The completeness of AIDS surveillance. Journal of Acquired Immune Deficiency Syndromes 5:257–64. Buehler J, Frey R, Chu S. 1995. AIDS mortality project group. The migration of persons with AIDS: Data from 12 states, 1985 to 1992. American Journal of Public Health 85(11):1552–55. CDC (Centers for Disease Control and Prevention). 1992. 1993 Revised classification system for HIV infection and expanded surveillance. Morbidity and Mortality Weekly Report 41(RR17;001). Atlanta, GA: CDC. CDC. 1998. Evaluation of HIV case surveillance through the use of non–name unique identifies—Maryland and Texas, 1994–1996. Morbidity and Mortality Weekly Report 46(52 & 53):1254–72. Atlanta, GA: CDC. CDC. 1999. Guidelines for national human immunodeficiency virus case surveillance, including monitoring for human immunodeficiency virus infection and acquired immunodeficiency syndrome. Morbidity and Mortality Weekly Report 48(13):1–28. Atlanta, GA: CDC. CDC. 2003a. Current Status of HIV Infection Surveillance—April 2003. (Email communication, Patricia Sweeney, CDC, April 28, 2003). CDC. 2003b. Evaluation of the Performance of Integrated Human Immunodeficiency Virus (HIV)/Acquired Immunodeficiency Syndrome (AIDS Surveillance Systems). Update: Interstate Duplication Evaluation Project, April 2003. (Email communication, Patricia Sweeney, CDC, April 25, 2003). CDC. 2003c. Table 1—Status of CD4 and Viral Load Reporting—Draft April 2003 (Email communication, Patricia Sweeney, CDC, May 13, 2003). CDC. 2003d. Laboratory Reporting Issues Discussed with HIV/AIDS Surveillance Coordinators. Summary of Preliminary Findings for IOM Committee—Draft April 2003. (Email communication, Patricia Sweeney, CDC, April 28, 2003). CDC. 2003e. Preliminary Analysis: Timeliness of HIV/AIDS Reporting Among Adults/Adolescents. April 2002. (Email communication, Patricia Sweeney, CDC, June 26, 2002). CDC. 2003f. State Timeliness Results. Percentage of HIV Cases Reported Within 6 Months; Percentage of AIDS Cases Reported Within 6 Months. (Email communication, Patricia Sweeney, CDC, April 18, 2003) . CDC. 2003g. Number of AIDS Cases by State and EMA and Diagnosis Year, Adjusted for Reporting Delay. Data reported as of December 2002. (Email communication, Patricia Sweeney, CDC, June 20, 2003). CDC. 2003h. Update on current status of HIV Infection Surveillance: April 2003. (Email communication, Patricia Sweeney, CDC, October 13, 2003).
OCR for page 132
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act CDC. 2003i. Update on the Evaluation of the Performance of Integrated Human Immunodeficiency Virus (HIV)/Acquired Immunodeficiency Syndrome (AIDS) Surveillance Systems. (Email communication, Patricia Sweeney, CDC, March 4, 2003). Chamberland ME, Allen JR, Monroe JM, Garcia N, Morgan C, Reiss R, Stephens H, Walker J, Friedman SM. 1985. Acquired Immunodeficiency Syndrome in New York City: Evaluation of an active surveillance system. Journal of the American Medical Association 254(3):383–7. Colfax GN, Bindman AB. 1998. Health benefits and risks of reporting HIV-infected individuals by name. American Journal of Public Health 88(6):876–9. Conway GA, Colley-Niemeyer B, Pursley C, Cruz C, Burt S, Rion P, Heath CW. 1989. Underreporting of AIDS cases in South Carolina, 1986 and 1987. Journal of the American Medical Association 262(20):2859–63. Davis K, Stapleton J. 1991. Migration to rural areas by HIV patients: Impact on HIV-related healthcare use. Infection Control and Hospital Epidemiology 12(9): 540–3. Doyle T, Glynn K, Groseclose S. 2002. Completeness of notifiable infectious disease reporting in the United States: An analytical literature review. American Journal of Epidemiology 155(9):866–74. Elcock M, Simon T, Bradley GP, Copello AG, Klezer PJ. 1993. Active AIDS surveillance: Hospital-based case finding in a metropolitan California County. American Journal of Public Health 83(7):1002–5. Ellis M, Muschkin C. 1996. Migration of persons with AIDS—A search for support from elderly parents? Social Science and Medicine 43(7):1109–18. GAO (General Accounting Office). 1995. Ryan White CARE Act of 1990: Opportunities Are Available to Improve Funding Equity. GAO/T-HEHS-95-91. Washington, DC: GAO. GAO. 2000. Ryan White CARE Act: Opportunities to Enhance Funding Equity. GAO/ T-HEHS-00-150. Washington, DC: GAO. Georgia Division of Public Health, Epidemiology Branch, HIV/STD Epidemiology Section. 2003. Proposal for HIV Infection Reporting in Georgia. [Online]. Available: http://www.ph.dhr.state.ga.us/epi/hivreporting.shtml [accessed September 25, 2003]. Green TA. 1998. Using surveillance data to monitor trends in the AIDS epidemic. Statistics in Medicine 17(2):143–54. Greenberg A, Hindin R, Nicholas AG, Bryan EL, Thomas PA. 1993. The completeness of AIDS case reporting in New York City. Journal of the American Medical Association 269(23):2995–3001. Hardy A, Starcher T, Morgan W, Druker J, Kristal A, Day JKC, Ewing E, Curran J. 1987. Review of death certificates to assess completeness of AIDS reporting. Public Health Reports 102:386–9. Hecht FM, Chesney MA, Lehman JS, Osmond D, Vranizan K, Colman S, Keane D, Reingold A, Bindman AB. 2000. Does HIV reporting by name deter testing? MESH Study Group. AIDS 14(12):1801–8. Holmberg SD. 1996. The estimated prevalence and incidence of HIV in 96 large U.S. metropolitan areas. American Journal of Public Health 86(5):642–54.
OCR for page 133
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act Hook E, Regal RR. 1995. Capture-recapture methods in epidemiology: Methods and limitations. Epidemiologic Reviews 17:243–64. HRSA (Health Resources and Services Administration). 2002. Ryan White Care Act overview. HRSA Ryan White Care Act Information Sheets:1–2. Rockville, MD: HRSA. IOM (Institute of Medicine). 2001. No Time to Lose: Getting More from HIV Prevention. Washington, DC: National Academy Press. Jara MM, Gallagher KM, Schieman S. 2000. Estimation of completeness of AIDS case reporting in Massachusetts. Epidemiology 11(2):209–13. Kaiser Family Foundation. June 2003. Fact Sheet: HIV/AIDS Testing in the United States. [Online]. Available: http://www.kff.org/hivaids/6094-index.cfm. Kates J, Ruiz S. 2002. Sources of coverage and care for people with HIV/AIDS in the United States: Assessing coverage in southern states. Southern States Summit on HIV/AIDS and STDs. Menlo Park, CA: Kaiser Family Foundation. Klevens R, Fleming P, Li J, Karon J. 1997. Impact of laboratory-initiated reporting of CD4 T lymphocytes on U.S. AIDS surveillance. Journal of Acquired Immune Deficiency Syndromes & Human Retrovirology 14:56–60. Klevens R, Fleming PL, Gaines CG, Troxler S. 1998. Completeness of HIV reporting in Louisiana, USA. International Journal of Epidemiology 27:1105. Klevens RM, Fleming PL, Li J, Gaines CG, Gallagher K, Schwarcz S, Karon JM, Ward JW. 2001. The completeness, validity, and timeliness of AIDS surveillance data. Annals of Epidemiology 11(7):443–9. Kopelman, J. Panel discussion with committee members during public committee meeting, Washington, DC. May 15, 2002. Lam NS, Liu KB. 1994. Spread of AIDS in rural America, 1982-1990. Journal of Acquired Immune Deficency Syndromes 7(5):485–90. LaMendola B. 2002, August 5. South Florida Losing Out on AIDS Funding as More Patients Move There. South Florida Sun-Sentinel. LOCAL. P. 1A. Lee L, Harris N, Glynn K, Phelps R. 2002. Evaluating Integrated HIV/AIDS Surveillance: Part 1. Presentation at the 2002 National HIV/AIDS Surveillance Workshop. Lindan CP, Hearst N, Singleton JA, Trachtenberg AI, Riordan NM, Tokagawa DA, Chu GS. 1990. Underreporting of minority AIDS deaths in San Francisco bay area, 1985-1986. Public Health Reports 105(4):400–4. Marsh K, Morgan M, Bunch G, Costa S, Fleming P, Wortley P. 1999. Evaluation of Non-Name-Coded Identifiers in Los Angeles County and New Jersey. 1999 National HIV Prevention Conference. Meyer P, Jones J, Garrison C. 1994. Completeness of reporting of diagnosed HIV-infected hospital inpatients. Journal of Acquired Immune Deficiency Syndromes & Human Retrovirology 7(10):1067–73. Modesitt SK, Hulman S, Fleming D. 1990. Evaluation of active versus passive AIDS surveillance in Oregon. American Journal of Public Health 80(4):463–4. NRC (National Research Council). 2001. The 2000 Census: Interim Assessment. Washington, DC: National Academy Press. NRC. 2003.Statistical Issues in Allocating Funds by Formula. Louis TA, Jabine TB, Gerstein MA, Eds. Washington, DC: The National Academies Press.
OCR for page 134
Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act Osmond DH, Bindman AB, Vranizan K, Lehman JS, Hecht FM, Keane D, Reingold A. 1999. Name-based surveillance and public health interventions for persons with HIV infection. Multistate Evaluation of Surveillance for HIV Study Group. Annals of Internal Medicine 131(10):775–9. Pagano M, Tu X, De Gruttola V, MaWhinney S. 1994. Regression analysis of censored and truncated data: Estimating reporting-delay distributions and AIDS incidence from surveillance data. Biometrics 50:1203–14. Payne SM, Seage GR 3rd, Oddleifson S, Gallagher K, vanBeuzekom M, Losina H, Hertz T. 1995 Using administratively collected hospital discharge data for AIDS surveillance. Annals of Epidemiology 5(5):337–46. Rosenblum L, Buehler JW, Morgan MW, Costa S, Hidalgo J, Holmes R, Lieb L, Shields A, Whyte BM. 1992. The completeness of AIDS case reporting, 1988: A multisite collaborative surveillance project. American Journal of Public Health 82(11):1495–99. Scheer S, McQuitty M, Denning P, Hormel L, Stephens B, Katz M, Schwarcz S. 2001. Undiagnosed and unreported AIDS deaths: Results from the San Francisco medical examiner. Journal of Acquired Immune Deficiency Syndromes & Human Retrovirology 27(5):467–71. Schwarcz SK, Hsu LC, Parisi MK, Katz MH. 1999. The impact of the 1993 AIDS case definition on the completeness and timeliness of AIDS surveillance. AIDS 13(9):1109–14. Schwarcz S, Hsu L, Chu PL, Parisi MK, Bangsberg D, Hurley L, Pearlman J, Marsh K, Katz M. 2002. Evaluation of a non-name-based HIV reporting system in San Francisco. Journal of Acquired Immune Deficiency Syndromes & Human Retrovirology 29(5):504–10. Solomon L, Flynn C, Eldred L, Caldeira E, Wasserman MP, Benjamin G. 1999. Evaluation of a statewide non-name–based HIV surveillance system. Journal of Acquired Immune Deficiency Syndromes & Human Retrovirology 22(3):272–9. U.S. Congress, House of Representatives. Committee on Commerce. 2000a. Report on Ryan White CARE Act Amendments of 2000 (106-788). 106th Cong., 2d Sess. July 25, 2000. U.S. Congress, Senate. 2000b. Committee on Health, Education, Labor, and Pensions. Report on Ryan White CARE Act Amendments of 2000 (106-294). 106th Cong., 2d Sess. May 15, 2000. Verghese A, Berk SL, Sarubbi F. 1989. Urbs in rure: Human immunodeficiency virus infection in rural Tennessee. The Journal of Infectious Diseases 160(6):1051–5. Verghese A, Nabhan D, Escobedo MA, Guardado J, Guerra LG, Ho H, Casner P, Berk ML. 1995. Profile of HIV disease in an American border city. Southern Medical Journal 88(4):429–32.
Representative terms from entire chapter: