5
Estimating Resource Needs
The Ryan White Comprehensive AIDS Resources Emergency (CARE) Act (RWCA) attempts to direct funds to areas in the greatest need of financial assistance through several of its discretionary grant programs, including Title I supplemental awards, Title II AIDS Drug Assistance Program (ADAP) supplemental awards, and Title III and IV awards. In contrast to formula awards, which are based exclusively on estimates of living AIDS cases (ELCs), these grants attempt to take into account other factors affecting severity of need. The Health Resources and Services Administration’s HIV/AIDS Bureau (HRSA/HAB) defines severity of need as “the degree to which providing primary medical care to people with HIV disease in any given area is more complicated and costly than in other areas based on a combination of the adverse health and socio-economic circumstances of the populations to be served” (HRSA, 2003).
In the 2000 reauthorization, Congress asked the Institute of Medicine (IOM) Committee to examine “existing and needed epidemiological data and other analytic tools for resource planning and allocation decisions, specifically for estimating severity of need of a community and the relationship to the allocations process” (Ryan White CARE Act. 42 U.S.C. § 300ff-11 [2003]). The Committee focused its analysis on the application of severity-of-need criteria in determining Title I supplemental awards, the largest of these discretionary grant awards, because of the specific requests for assistance by Congress and HRSA/HAB in this area. Although this chapter does not discuss other discretionary grant programs that use
severe-need criteria in allocating resources, the Committee’s findings and recommendations may also be relevant to those programs. In the remainder of this chapter, the Committee uses the term “resource needs” instead of “severity of need” to reflect Congress’ interest in the relationship between need and resource allocation. The Committee uses the term “severity of need,” however, when referencing the specific severity-of-need component of the Title I application.
Congress specified that “[Title I] supplemental awards are to be directed principally to those eligible areas with ‘severe need,’ or the greatest or expanding public health challenges in confronting the epidemic” (U.S. Congress, 2000). Reflecting this notion, Congress increased the weight assigned to severity of need in determining the supplemental award from 25 percent to 33 percent in the 2000 reauthorization (Ryan White CARE Act. 42. U.S.C. § 300ff-13 [2003]). In determining severity of need, Congress directed HRSA/HAB to consider factors such as: “(I) STDs, substance abuse, tuberculosis, severe mental illness, or other co-morbid factors; (II) new or growing populations of individuals with HIV; (III) homelessness; (IV) current prevalence of HIV; (V) increasing need for HIV services including the relative rates of increase in the number of cases of HIV disease; [and] (VI) unmet need for services” (Ryan White CARE Act. 42 U.S.C. §. 300ff-13 [2003]). Congress further directed HRSA/HAB to “employ standard, quantitative measures to the maximum extent possible in lieu of narrative self-reporting when awarding supplemental awards” (U.S. Congress, 2000).
In addressing its charge, the Committee organized its work into the following tasks:
-
Developing a conceptual framework for factors affecting resource needs;
-
Defining criteria for assessing measures of resource needs;
-
Evaluating the process and data currently used to award Title I supplemental funds;
-
Proposing a new way of identifying predictors of resource needs; and
-
Making recommendations to evaluate and implement this approach.
Although HRSA/HAB uses explicit criteria to evaluate resource needs and allocate Title I supplemental grants, no consistent indicators are used to evaluate relative need, and much of the evaluation process is subjective. The Committee also found that the process for awarding Title I supplemental grants focuses on the characteristics of individuals, such as the prevalence of comorbid conditions that often accompany HIV disease,
and does not account for other important factors affecting resource needs, such as the cost of providing services and the availability of local resources. The Committee proposes a potential new approach to allocating Title I supplemental awards that is based on standardized, quantitative indicators of resource needs of different jurisdictions.
FACTORS AFFECTING RESOURCE NEEDS
A broad array of individual and social factors determines an area’s resource needs. The Committee groups these factors into three categories: disease burden, the costs of providing care, and available resources. Resource needs can be viewed as a product of disease burden and cost of care minus available resources:
Resource needs = (Disease burden * Costs of providing care) − Available resources
Table 5-1 provides some examples of the types of measures that could be used to assess resource needs.
TABLE 5-1 Examples of Measures of Resource Needs
Factors Affecting Resource Needs |
Example Measuresa |
Disease Burden |
|
Costs |
|
Resources |
|
aThis is not intended to be a comprehensive list. |
Disease Burden
Disease burden is commonly measured by the incidence or prevalence of a disease.1 Incidence refers to the number of new cases of a condition during a specified period of time in a population at risk for developing the disease. Prevalence can be assessed in terms of either point prevalence or period prevalence. Point prevalence is defined as the proportion of persons in the population with the condition at a specific point in time (Gordis, 1996). Period prevalence is the proportion of people who have had the disease at any time during a certain period (e.g., a calendar year). Some people may have developed the disease during that time while others may have had the disease and died during that period (Gordis, 1996). Incident, or new, HIV infection is the ideal measure for understanding the dynamics, spread, and success of prevention programs. Prevalent known HIV infection is appropriate for estimating the clinical burden to apportion care resources. Current RWCA allocation formulas are based on estimated AIDS prevalence, based on data from states through their AIDS case-reporting systems.
It is important to note that the medical and financial significance of an AIDS diagnosis has changed as treatments have evolved. Thus, unlike the early period of the epidemic when the effects of therapy were small and transient, many people who currently have a diagnosis of AIDS have responded well to highly active antiretroviral therapy and are now at relatively low risk for opportunistic infections and are able to maintain an active and productive life. Estimates of HIV cases using a uniform methodology are not available (see Chapter 4).
Costs of Care
Costs of care may be driven by several factors, including the complexity of a person’s medical condition. For instance, HIV-infected individuals who are at a later stage of disease generally require more resources for their care than HIV-infected people who are at earlier stages of disease (Bozzette et al., 2001).2 The costs of care also depend on the cost of obtaining and providing services, such as the prevailing wages of health care
workers and the local costs of medical supplies. To receive comparable care, a patient in a costly metropolitan area may require greater financial resources than one residing in a less costly locality. Even with comparable patients and costs, some areas may have fewer resource needs because they are more efficient at providing care.
Available Resources
Available resources or fiscal capacity across regions or states also affect resource needs. Such resource disparities are important in many policy arenas. Formula allocations for Temporary Assistance to Needy Families (formerly Aid to Families with Dependent Children) and other federal programs have long been designed to assist less affluent states (NRC, 2003).
Defining “available resources” equitably is an extremely difficult task, however. The purchasing power of a dollar varies greatly across areas and many resources that affect the difficulty and cost of providing HIV care are not measured. Some programs use per capita income as a proxy to account for such variations. For example, Medicaid and several other formula allocation programs use a formula that adjusts allocations based on the ratio between the state per capita income and the national per capita income when determining what proportion of state program expenditures will be reimbursed by the federal government (NRC, 2003). A 1975 study of alternative formulas for the General Revenue Sharing (GRS) program recommended inclusion of a poverty factor in the intrastate allocation and allocations on a per capita basis for governmental jurisdictions for which reliable estimates of income and poverty were available (NRC, 2001).
In the arena of health services, advantages that accrue from higher per capita income are partly offset by higher labor costs and by higher costs of other resources required in patient care. Medicare addresses these issues through the use of adjusted average per capita cost (AAPCC) in determining capitation rates in different medical markets (CMS, 1999; MEDPAC, 1999). This method, which relies upon historical average reimbursements, is imperfect and controversial. The AAPCC is based on past Medicare reimbursements rather than differences in input prices. Thus, the AAPCC methodology appears to penalize cities and states that have historically made most efficient use of medical resources (Society of Actuaries, 1997). Despite these limitations, data regarding regional variation in medical costs and prices could provide a useful complement to existing data in determining RWCA formula allocations and supplemental awards.
Similarly, the coverage of private and public health insurance pro-
grams is a major factor affecting states’ and Eligible Metropolitan Areas’ (EMAs) resource needs. Medicaid programs in particular—by far the largest payer of care for people with HIV/AIDS—vary substantially in the benefits they cover and their eligibility criteria across states. For example, in some states, being medically needy is not an eligibility criterion for Medicaid; many states also have limitations on Medicaid drug coverage (Kaiser Family Foundation, 2000). The relative “generosity” of Medicaid programs can greatly influence regions’ reliance on CARE Act funds, including its ADAP. All else equal, areas with a greater proportion of insured residents with HIV/AIDS should require fewer CARE Act funds than areas with a high proportion of uninsured patients with HIV/AIDS.
States further differ in the resources they devote to addressing the HIV/AIDS epidemic. For instance, some states have imposed one or more restrictions on ADAP, such as enrollment caps, limits on access to antiretroviral treatments, and expenditure caps (NASTAD et al., 2003). Furthermore, states vary a great deal in how much they contribute to ADAP programs. In some cases, lack of political will and emphasis on other priorities have contributed to these restrictions.
Title I supplemental awards, along with Title I and II formulas, do not take into account variations in the costs of providing care or other resources available to states and metropolitan areas. Including such information in allocation decisions could have a substantial impact on RWCA funding across states and EMAs. For example, an EMA in a state that has poor Medicaid coverage may choose to use more of its Title I funds on primary care services than an EMA in a state with more generous Medicaid coverage. The EMA in the state with poor Medicaid coverage will therefore have relatively fewer resources to devote to support services, since their RWCA funds must be used to cover basic primary medical care.
TITLE I SUPPLEMENTAL AWARD PROCESS
Congress divides Title I funds into two components, designating half for the formula-based awards and half for supplemental awards. After HRSA/HAB deducts funds for the Minority AIDS Initiative and the hold-harmless provision,3 approximately 80 percent of the supplemental award amount remains available for distribution among EMAs.
The Review Process and Scoring Guide
Each application for a supplemental award can receive a maximum of 100 points (Box 5-1). Applications submitted for fiscal year (FY) 2002 could achieve up to the following number of points in each category:
BOX 5-1
|
HRSA/HAB originally used an external review process to score Title I supplemental applications. However, beginning with the FY1999 review process, HRSA/HAB relied on Title I project officers as the primary reviewers, given their familiarity with grantees’ programs. Because RWCA operates on a 5-year budget period, early reviews of applications for supplemental funding set the standard for the remaining budget period. HRSA still uses an external review process for the first year of the budget cycle.
At least two HRSA/HAB project officers review and score each application; the scores are then averaged. A guide helps reviewers assign scores to each component but also states clearly that such guidance is not definitive: “Reviewers should use their own judgment and expertise in determining a final score” (HRSA, 2001b). Hence, the subjective scores can deviate significantly from empirical indicators of need. HRSA/HAB also uses an algorithm that may vary from year to year, which may reduce disparities among supplemental allocations (HRSA, 2001c). The detailed algorithm for determining final supplemental awards is not made public.
It is important to note that even though severity of need accounts for one-third of the total points, it may not account for one-third of the variation in total points. That depends on both the relative variation in severity-of-need scores and the relative variation in other components of
the application. If, for example, EMAs received identical scores for all items other than severity of need, 100 percent of the variation in scores—and thus in awards—would stem from severity of need. If, in contrast, all EMAs received similar severity-of-need scores, almost all the variation in scores and awards would stem from other components.
Severity-of-Need Component of the Application
Scoring of severity of need in the Title I application is based on three equally weighted components: (1) HIV/AIDS epidemiology; (2) comorbidity, poverty, and insurance status; and (3) assessment of populations with special needs.
HIV/AIDS Epidemiology
For this component, grantees supply data on AIDS incidence, AIDS prevalence, and HIV prevalence. Grantees also provide narrative detail on three issues:
-
Trends and compositional changes in caseloads, based on a comparison of the estimated number of people living with HIV, the number of people living with AIDS, and the number of new AIDS cases reported within the last 2 years.
-
The demographics of cases, based on populations in the EMA with disproportionately high HIV/AIDS prevalence compared with the general population.
-
The level of unmet need among populations who are underrepresented in the CARE-funded system, based on utilization data for all covered services (HRSA, 2001a).4
Comorbidity, Poverty, and Insurance Status
The EMA must also provide information on the incidence of six comorbid conditions: tuberculosis, syphilis, gonorrhea, intravenous drug use, other substance use, and homelessness. The EMA also reports the
number and percentage of residents with incomes below 300 percent of the federal poverty line during the prior fiscal year, and the number and percentage of residents without public or private health insurance. Applicants must describe the overall effect of these components on their populations, and explain how they affect the cost of service and the complexity of providing care.
Populations with Special Needs
In the final severity-of-need component, applicants respond to 10 questions regarding six special populations. These populations are youth 13–24 years of age, injection drug users (IDUs), substance users other than IDUs, men of color who have sex with men, white/Anglo men who have sex with men, and women of childbearing age (13 years of age and older). Applicants can also report on other populations they deem to have special needs. Applicants are requested to provide information on HIV and AIDS prevalence, trends, and service needs for each special population (Box 5-2).
BOX 5-2
|
CRITERIA FOR ASSESSING MEASURES OF RESOURCE NEEDS
The Committee reviewed all 51 Title I applications submitted by grantees in FY2002. In reviewing the applications, the Committee considered whether the measures used by different jurisdictions were important, scientifically sound, and feasible for national use. Specifically, the Committee identified nine criteria to evaluate current measures that are similar to those used by previous IOM and National Research Council (NRC) committees (NRC, 1997; IOM, 2001a) (Box 5-3).
Measures should reflect important resource needs. A measure’s importance encompasses its meaningfulness, the prevalence and seriousness of the needs being measured, the potential for changing the situation, and the overall impact of providing the resources under consideration.
BOX 5-3
Scientific soundness of the measure:
Feasibility of the measure:
|
-
Meaningfulness: There should be consensus among clinicians, patients, or policy makers that the measure reflects an important area of need.
-
Prevalence and seriousness of the problem: Measures should focus on major problems that affect a sizable proportion of RWCA clients. Measures should focus on aspects of need that are unmet or for which there is significant variation across grantees.
-
Potential for improvement: Measures should reflect areas of need that can be improved most by additional resources.
-
Potential impact: Considering the prevalence and seriousness of the problem and the potential for improvement, measures of resources that could have the greatest impact on persons living with HIV infection are desirable.
Measures should be scientifically sound. Scientific soundness entails three major components: reliability, validity, and evidence base.
-
Reliability: Reliability can be enhanced by using standard data collection methods across EMAs, collecting data in a way that minimizes manipulation, and employing a common definition of the population of interest and the time period.
-
Validity: Measures should capture what they purport to measure. Measures should make sense logically (face validity), should correlate well with other measures of resource needs (construct validity), and should capture meaningful aspects of resource needs (content validity) (IOM, 2001a). If one is developing predictors of needs, then the measures should have predictive validity. That is, one should be able to show that they predict measured needs. As indicated elsewhere in the report, when comparable assessments across regions are important in determining allocations and absolute levels are not, valid measures can have a bias as long as the bias (e.g., a given percentage of underreporting) is consistent across allocation regions.
-
Evidence base: Measures should have strong empirical support. For instance, indicators of resource needs should be empirically linked to needs or costs.
Measures should be feasible. Feasibility includes the availability of data and the cost and burden of measurement.
-
Availability of data: Data should be available at the appropriate level. For example, for assessments at the EMA level, each EMA should have the same data. Data should also be available in a timely manner and collected with reasonable periodicity.
-
Cost and burden of measurement: Data should be collected at a reasonable cost and should not impose an excessive burden on grantees. Measures based on data that are already being collected for other purposes, or that are publicly available are more feasible than measures that require new data collection.
Ideally, each measure of resource needs would satisfy all these criteria; in reality, few, if any, do. Thus, these criteria should not be viewed as absolute, but rather as guidelines for assessing the relative strengths and weaknesses of different measures.
ANALYSIS OF TITLE I SUPPLEMENTAL APPLICATIONS
This section summarizes the various data sources and resource needs measures used by grantees in the severity-of-need sections of their FY2002 Title I supplemental applications. The Committee found considerable variability in the actual data sources and measures, and the quality of those data sources and measures, used by grantees to describe their severity of need.
HIV/AIDS Epidemiology
The first component of the severity-of-need section requires grantees to report data on AIDS incidence, AIDS prevalence, and HIV prevalence. The Committee found significant variability in the data sources and the quality of the data used by grantees to describe the prevalence of HIV and AIDS in their areas:
-
41 percent (21) used CDC’s estimates of AIDS and/or HIV incidence and prevalence.5
-
25 percent (13) chose to use their own AIDS and/or HIV data and estimates (7 use state sources and 6 use local sources).
-
12 percent (6) used a combination of data from state and local health departments and CDC.
-
16 percent (3) applied their own adjustments to CDC’s estimates.
-
16 percent (8) did not explicitly state the source of prevalence data.
As discussed in Chapter 3, states have had AIDS case-reporting systems in place since the 1980s, and the overall completeness and quality of the data are very high. Thus, comparisons of AIDS incidence and AIDS prevalence (measured by newly reported AIDS cases or existing AIDS cases, respectively) across areas are possible. Estimates of HIV cases across EMAs using a consistent methodology are not currently available (see Chapter 4). The lack of complete coverage in states’ case-reporting systems and methods for de-duplicating code-based states, and variability in system maturity and the quality of reported cases, hinder the ability to compare estimates of HIV case reports. Furthermore, data on HIV cases represent only partial HIV prevalence because they include only individuals who chose to be tested at confidential testing sites. No data exist on the incidence of HIV cases; the HIV case-reporting system does not capture all new infections, only newly diagnosed infections (IOM, 2001b). The limited maturity of HIV case-reporting systems in some states means that it is virtually impossible for EMAs to describe trends in HIV prevalence and the demographic characteristics of their cases, unless they have conducted local studies.
The methodologies and data EMAs use to describe unmet need also vary substantially. Researchers at the University of California, San Francisco (Kahn et al., 2003), conducted a systematic review of existing methods of estimating unmet need for HIV care, including those used by all 51 EMAs in their grant applications.6 The researchers found that EMAs varied significantly in their definitions and measures of unmet need. Studies used different definitions for HIV care (for example, some included support services such as case management while others did not) and different samples (e.g., representative or convenience samples). The methods used by EMAs and others therefore varied widely in their usefulness in quantifying unmet need for HIV primary care. The most useful methods for estimating such unmet need were typically quantitative, although variability in the quality of the data sources and samples lessened the usefulness of some of these measures. Less useful methods for quantifying unmet need were generally qualitative studies that assessed population characteristics but not the size of the population, or focused on clinical outcomes.
Comorbidities, Poverty, and Insurance Status
The Committee also reviewed the second component of the severity-of-need section of the Title I supplemental award application, which requires applicants to document comorbidity, poverty, and insurance status (Table 5-2).
Comorbidity Factors
Applicants provided estimates of comorbid conditions, but the validity, reporting period, and definition of these estimates varied significantly. All EMAs but one reported data on tuberculosis, syphilis, and gonorrhea. The most commonly used indicator was prevalence rate, although only 40 percent specified the time period for which they were reporting. EMAs used several categories to indicate the prevalence of syphilis, including primary and secondary disease rates (5 EMAs), congenital syphilis rates (2 EMAs), and syphilis incidence rate (1 EMA). Most EMAs (66 percent) tap their state health departments for these data, whose reliability is unknown.
All but one EMA reported on IDUs. Localities used six separate indicators of IDUs, with the estimated absolute number being the most common measure (86 percent). Other EMAs used more specific indicators, such as the estimated number of crack and cocaine abusers (2 EMAs), estimated number of IDUs with HIV (1 EMA), and the number of clients in methadone clinics (1 EMA). State health departments provided 32 percent of EMAs with these estimates, while state and county substance abuse agencies provided the other EMAs with these data (16 percent and 8 percent, respectively).
Forty-eight of 51 EMAs reported on “other substance abuse”—relying on 17 different indicators. A majority of EMAs estimated the number of alcohol abusers (47 percent), cocaine and crack cocaine abusers (45 percent), or substance abusers (38 percent). State health departments most often provided this information, although EMAs also relied on local health departments, state substance abuse agencies, and the federal Substance Abuse and Mental Health Services Administration (SAMHSA).
All but one EMA reported on homelessness, and they used six different indicators. EMAs most often cited the estimated number of homeless people (58 percent), the yearly number of homeless (26 percent), or the daily number of homeless (24 percent). A majority of EMAs cited community-based organizations as the source of these data. However, different localities appear to use different methods to count homeless people and such variability can affect the data. For example, estimates of homelessness based on shelter counts are likely to be significantly smaller than
estimates based on population surveys that ask respondents whether they were homeless in the past month or year.
Although these six comorbid conditions (tuberculosis, syphilis, gonorrhea, injection drug use, other substance abuse, and homelessness) are of clear clinical and policy importance, their relationship to an area’s need for resources for HIV care is not well documented. Without an evidence base and explicit model connecting indicators to resource needs, assessing the appropriateness of the current comorbid measures is extremely difficult.
Even if one could link comorbid conditions with resource needs, information on the prevalence of co-orbidities in the HIV-infected population is lacking. As a result, many EMAs provide information on the prevalence of comorbidities among the general population. Reliably estimating the number of HIV-positive people with comorbid conditions within an EMA would be difficult. For instance, homelessness is difficult to estimate for the general population as no universally recognized quantitative measures exist, much less for the HIV-infected population. Many applications do not provide documentation that would enable reviewers to evaluate the reliability of these data.
Another challenge is that different levels of government with varying priorities collect data (e.g., on substance abuse, STDs, corrections, and mental health) that grantees are required to or choose to report in their applications. EMAs’ relationships with these data providers also vary widely: some EMA representatives reported good relationships with state correctional facilities and substance abuse agencies, for example, while others reported difficulty in obtaining data from them (Ryan White CARE Act 2002 Grantee Conference, August 22–23, 2003, Washington, DC: Meeting with Title I EMAs).
Poverty and Insurance Status
Fifty of the 51 EMAs reported information on poverty status. HRSA/ HAB asks grantees to provide the number of people at or below 300 percent of the federal poverty level, and 68 percent of EMAs provided such an estimate. Other EMAs estimated the number at or below 100 percent of the poverty line (24 percent), while a few used other thresholds, such as 133 percent and 200 percent of the federal poverty line. EMAs primarily used data from the U.S. Census, although some relied on data from state health departments and other state agencies.
HRSA/HAB requires applicants to estimate the number of people without insurance, including those without Medicaid coverage, and all 51 EMAs did so, with 92 percent using the definition provided by HRSA/ HAB. A few EMAs estimated the number of people living with HIV who
TABLE 5-2 Required Factors Used to Describe Severity of Need in FY2002 Title I Supplemental Applications
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
COMORBIDITY INDICATORS |
|||||
TB (Total EMAs Reporting = 50) |
|||||
Prevalence rate |
49 |
20 |
33 |
15 |
1 |
Incidence and prev rates |
1 |
1 |
|
1 |
|
Did not report |
1 |
|
|||
Syphilis (Total EMAs Reporting = 50) |
|||||
Total prevalence rate |
45 |
21 |
30 |
13 |
1 |
Primary & secondary rate |
5 |
2 |
3 |
2 |
|
Infectious syphilis |
2 |
1 |
2 |
|
|
Congenital rate |
2 |
1 |
1 |
1 |
|
Early latent rate |
1 |
1 |
1 |
|
|
Incidence rate |
1 |
1 |
|
1 |
|
Did not report |
1 |
|
|||
Gonorrhea (Total EMAs Reporting = 50) |
|||||
Prevalence rate |
49 |
20 |
33 |
14 |
1 |
Incidence rate |
1 |
1 |
|
1 |
|
Did not report |
1 |
|
|||
Injection Drug Use (Total EMAs Reporting = 50) |
|||||
Estimated number of IDUs |
43 |
4 |
16 |
3 |
1 (surveillance report) |
IDUs, crack, and cocaine abusers |
2 |
|
|||
Estimated number of uninfected IDUs |
1 |
|
1 |
||
Estimated number of IDUs w/HIV |
1 |
|
1 |
|
|
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
Number of clients in methadone clinic |
1 |
|
|
1 |
|
Estimated number of heroin users |
1 |
1 |
1 |
|
|
Did not report |
1 |
|
|||
Other Substance Abuse (Total EMAs Reporting = 48) |
|||||
Estimated number of alcohol abusers |
22 |
5 |
8 |
6 |
|
Estimated number of cocaine and crack cocaine abusers |
21 |
5 |
7 |
5 |
|
Estimated number of substance abusers (no clear definition) |
18 |
|
7 |
|
1 (SHAS) |
Estimated number using marijuana |
9 |
2 |
2 |
1 |
|
Use of illicit drugs |
6 |
1 |
5 |
|
|
Methamphetamine users |
6 |
1 |
|
3 |
|
Psychedelic/hallucinogen users |
4 |
1 |
|
|
|
Did not report |
3 |
|
|||
Estimated number needing treatment for SA problem |
3 |
|
1 |
1 |
|
Inhalant users |
3 |
1 |
|
||
Number of admissions for SA treatment |
2 |
|
1 |
1 |
|
Smoking/tobacco use |
2 |
|
|||
Binge drinkers |
2 |
|
|
1 |
|
Estimated number using stimulants |
2 |
|
1 |
|
|
Meth, PCP, benzo, |
1 |
|
Other State |
Other Local |
Other Federal or National Org |
Academic Research |
No Source |
4 (state drug abuse agency) |
1 (county drug agency) |
1 (SAMHSA) 1 (NHSDA) |
|
1 |
4 (state drug abuse agency) |
1 (county drug agency) |
3 (NHSDA) 1 (SAMHSA) |
|
|
4 (state drug abuse agency) |
|
4 (SAMHSA) 1 (NHSDA) |
1 |
|
2 (state drug abuse agency) |
1 (county drug agency) |
1 (SAMHSA) 1 (NHSDA) |
|
1 |
|
|
1 (NHSDA) |
|
|
2 (state drug abuse agency) |
|
1 (NHSDA) |
|
|
3 (state drug abuse agency) |
|
1 (NHSDA) |
|
|
1 (state drug abuse agency) |
|
|
|
|
2 (state drug abuse agency) |
|
1 (NHSDA) |
|
|
|
1 (NHSDA) |
|||
1 (NHSDA) |
||||
1 (NHSDA) |
||||
1 (SAMHSA) |
|
1 |
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
barbituates, tranquilizers, sedatives |
|
||||
Users of sedatives |
1 |
|
|||
Problem drinkers and nonnarcotic users in need of treatment |
1 |
|
1 |
|
|
Percent population using alcohol, marijuana, hallucinogenics/cocaine and inhalants |
1 |
|
|||
No definition included |
1 |
|
1 |
|
|
Homelessness (Total EMAs Reporting = 50) |
|||||
Estimated number of homeless |
29 |
|
|
3 |
|
Yearly number of homeless (cumulative) |
13 |
|
|
3 |
|
Daily number of homeless (point prevalence) |
12 |
|
|
1 |
|
Number homeless among PLWH/A |
2 |
|
1 |
|
|
Number of shelter beds provided (1 year) |
1 |
|
|||
Low-income and working poor families and individuals for FY2000 |
1 |
1 |
|
||
Not reported |
1 |
|
|||
POVERTY AND INSURANCE STATUS INDICATORS |
|||||
Insurance Status (Total EMAs Reporting = 51) |
|||||
Estimated # of people without insurance, including without Medicaid |
48 |
|
12 |
2 |
|
Other State |
Other Local |
Other Federal or National Org |
Academic Research |
No Source |
1 (state drug abuse agency) |
|
|||
|
|
1 (NHSDA) |
|
|
2 (Dept of Housing) |
18 (CBO) 3 (city agency) |
1 (HUD) |
2 |
|
1 (state DSS) |
6 (CBO) 3 (local DSS) |
|
||
1 (state DSS) |
6 (CBO) 2 (local DSS) 2 (city agency) |
|||
|
1 (local DSS) |
|||
|
1 (CBO) |
|||
1 (state DSS) |
|
|||
3 |
2 (community planning report) |
9 (census) 3 (CPS) 2 (KFF) 1 (HRSA) |
5 |
11 |
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
Estimated # of living HIV cases uninsured at time of HIV diagnosis |
2 |
|
1 |
|
|
Estimated # of living AIDS cases uninsured at time of AIDS diagnosis |
1 |
|
|||
Poverty (Total EMAs Reporting = 50) |
|||||
Number people at/below 300% FPL |
34 |
|
4 |
2 |
|
Number people at/below 100% FPL |
12 |
|
2 |
|
|
Number adults at/below 300%, 200%, 133%, and 100% of FPL |
2 |
|
1 |
|
|
Number people at/below 200% FPL |
1 |
|
|||
Number of children (<18) below poverty |
3 |
|
2 |
|
|
Not reported |
1 |
|
|||
CDC = Centers for Disease Control and Prevention CPS = Current Population Survey SHD = State health department DOL = Department of Labor LHD = Local health department HRSA = Health Services and Resources Administration SAMHSA = Substance Abuse and Mental Health Services Administration NIMH = National Institute for Mental Health |
Other State |
Other Local |
Other Federal or National Org |
Academic Research |
No Source |
|
|
1 (KFF) |
|
|
3 (state agency) |
|
|
|
|
2 (DOL) |
1 (community planning report) 1 (city agency) |
12 (census) 1 (CHSI) 1 (CPS) |
|
8 |
4 |
|
5 (census) 1 (KFF) |
|
|
|
|
1 (census) |
|
|
|
|
1 (census) |
|
|
|
|
|
1 |
|
CHSI = Community Health Status Indicators NIDA = National Institute for Drug Abuse DSS = Department of social services CBO = Community-based organization KFF = Kaiser Family Foundation BRFS = Behavioral Risk Factor Survey NHSDA = National Household Drug Abuse Survey SHAS = Supplement to HIV/AIDS Surveillance |
were uninsured at the time of HIV diagnosis (2 EMAs) or AIDS diagnosis (1 EMA). Many EMAs obtained their insurance data from the state health department (27 percent) or the U.S. Census (18 percent), although other EMAs used the Census Bureau’s Current Population Survey (CPS),7 the Kaiser Family Foundation’s State Health Facts,8 and other state agencies. Each source uses different definitions and measures of insurance status.
These data sources have several limitations. While the U.S. Census provides comprehensive data on poverty rates at the substate level, this information becomes less reliable as time from the decennial U.S. Census increases. The CPS provides more timely data than the decennial census, but does not provide data at the county level. One potential new source is the American Community Survey,9 which if funded as planned will provide annual data on poverty rates for states and areas with a population of 250,000 or more (NRC, 2000).
Data on insurance rates are not uniformly available at the substate level. While many states and areas have conducted surveys of uninsurance rates, these tend not to be comparable from state to state because of differences in sample size and methods. For comparisons of uninsured rates at
7 |
The CPS is a monthly survey of about 50,000 households conducted by the Bureau of the Census for the Bureau of Labor Statistics. The sample is representative of the civilian noninstitutional population. Estimates obtained from the CPS include employment, unemployment, earnings, hours of work, and other indicators. They are available by a variety of demographic characteristics including age, sex, race, marital status, and educational attainment. They are also available by occupation, industry, and class of worker. Supplemental questions to produce estimates on a variety of topics including school enrollment, income, previous work experience, health, employee benefits, and work schedules are also often added to the regular CPS questionnaire. See http://www.bls.census.gov/cps/cpsmain.htm for more information. |
8 |
The Kaiser Family Foundation has compiled an online resource containing state-level data on demographics, health, and health policy (including HIV/AIDS), including health coverage, access, financing, and state legislation. Data presented on State Health Facts Online are collected from a variety of public and private sources, including original Kaiser Family Foundation reports, data from public websites, and information purchased from private organizations. See http://www.statehealthfacts.kff.org/ for more information. |
9 |
The American Community Survey is planned as a large-scale, monthly sample survey of U.S. households similar to the census long-form version in content but is administered continuously (NRC, 2000). If implemented as planned (pending congressional funding), the annual sample size will include approximately 3 million addresses, and would provide the same sort of data as the census long form, updated every year. The survey will provide demographic, social, economic, and housing profiles annually for areas and subgroups with 65,000 or more people. For communities of less than 65,000, 3 to 5 years will be required to provide estimates similar in quality to those based on the census long form (http://www.census.gov/acs/www/, accessed July 8, 2003). |
the state level, the Census Bureau and others often calculate averages over 2 to 3 years so that all the estimates have similar power (IOM, 2003).
Optional Factors
Several EMAs submitted additional information to describe severity of need (Table 5-3). These optional factors included chlamydia prevalence (26 EMAs), mental illness (18 EMAs), hepatitis (17 EMAs), incarceration and probation (4 EMAs), domestic violence (2 EMAs), other STDs (2 EMAs), coccidiomycosis (1 EMA), and teenage childbearing rates (1 EMA).
For mental illness, applicants used 13 different measures. General prevalence of mental illness was the most common indicator (55 percent of EMAs), followed by the number of residents receiving mental health services (17 percent) and the estimated number of severely mentally ill (11 percent). Variations across states and localities in collecting and reporting these data call into question their reliability and validity.
The majority of EMAs obtained optional data on comorbid and other conditions from the state health department, while a few relied on data from CDC and other federal agencies such as the Substance Abuse and Mental Health Services Administration and the National Institute of Mental Health. Again, the lack of standardization of measures and inconsistency in the quality of the data makes comparison across areas very difficult.
Populations with Special Needs
EMAs included assessments of the following populations with special needs (Table 5-4):
-
Homeless (17 EMAs, or 33 percent of total EMAs)
-
African Americans (10 EMAs, or 20 percent)
-
Latinos (8 EMAs, or 17 percent)
-
Recently or soon-to-be released from jail or prison (7 EMAs, or 14 percent)
-
Rural individuals (6 EMAs, or 12 percent)
-
Incarcerated (5 EMAs, or 10 percent)
-
Mentally ill (5 EMAs, or 10 percent)
-
Immigrants/undocumented (3 EMAs, or 6 percent)
-
Haitians (2 EMAs, or 4 percent)
-
Deaf people (2 EMAs, or 4 percent)
-
Transgender people (2 EMAs, or 4 percent)
TABLE 5-3 Optional Factors Used to Describe Severity of Need in FY2002 Title I Supplemental Applications
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
OPTIONAL COMORBIDITY INDICATORS |
|||||
Chlamydia (Total EMAs Reporting = 26) |
|||||
Prevalence rate |
25 |
8 |
17 |
7 |
1 |
Rate among women |
1 |
|
|
1 |
|
Mental Illness (Total EMAs Reporting = 18) |
|||||
General prevalence of mental illness |
10 |
|
4 |
|
1 (BRFS) 1(SHAS) |
Number receiving mental health services |
3 |
1 |
2 |
1 |
|
Estimated total severely mentally ill |
2 |
|
1 |
1 |
|
Daily average in mental health hospitals |
1 |
|
1 |
|
|
Mentally ill chemically addicted |
1 |
|
|||
Estimated # PLWH with severe mental illness |
1 |
|
1 |
|
|
Multiply diagnosed (SMI, SA, HIV) |
1 |
|
1 |
|
|
Mental illness based on psychiatric hospital data |
1 |
|
|||
Estimated # with schizophrenia |
1 |
|
1 |
|
|
Estimated # with bi-polar |
1 |
|
1 |
|
|
Estimated # with major depression |
1 |
|
1 |
|
|
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
Estimated # with personality disorder |
1 |
|
1 |
|
|
Nonseverely mentally ill |
1 |
|
1 |
|
|
Hepatitis (Total EMAs Reporting = 17) |
|||||
Hep C prevalence rate |
5 |
4 |
3 |
2 |
|
Hep B and C |
4 |
4 |
2 |
1 |
|
Hep A, B, C separately |
3 |
|
1 |
2 |
|
Hep A and B |
1 |
1 |
1 |
|
|
Hep C incidence rate |
1 |
1 |
|
1 |
|
Hep C, chronic and acute |
1 |
1 |
|
1 |
|
Incarceration/Probation (Total EMAs Reporting = 4) |
|||||
Incarcerated |
3 |
|
|||
Number on parole or probation |
1 |
|
|||
Domestic Violence (Total EMAs Reporting = 2) |
|||||
Number of domestic violence victims |
1 |
|
|||
Case rate per 100,000 |
1 |
||||
Other Venereal Diseases (Total EMAs Reporting = 2) |
|||||
Herpes, Tricomonas, N.S. Uteritis, Escabiosis, V. Bacteria |
1 |
1 |
1 |
|
|
Herpes and other ulcerative STDs |
1 |
|
1 |
|
|
|
Measures Used |
Data Sources Used |
|||
Indicators |
# of EMAs Reporting the Measure |
For Defined Time Period |
SHD |
LHD |
CDC |
Coccidiomycosis (Total EMAs Reporting = 1) |
|||||
Prevalence rate |
1 |
|
|
1 |
|
Teen Births (Total EMAs Reporting = 1) |
|||||
Birth rate among 13–19 yo women (1) |
1 |
|
1 |
|
|
CDC = Centers for Disease Control and Prevention CHSI = Community Health Status Indicators CPS = Current Population Survey NIDA = National Institute for Drug Abuse SHD = State health department DSS = Department of social services DOL = Department of Labor CBO = Community-based organization LHD = Local health department KFF = Kaiser Family Foundation HRSA = Health Services and Resources Administration BRFS = Behavioral Risk Factor Survey SAMHSA = Substance Abuse and Mental Health Services Administrations NHSDA = National Household Drug Abuse Survey NIMH = National Institute for Mental Health SHAS = Supplement to HIV/AIDS Surveillance |
TABLE 5-4 EMAs Reporting on Populations with Special Needs FY2002
EMA |
Homeless |
African Americans |
Latinos |
Recently/Soon-To-Be Released from Jail or Prison |
SUBPOPULATIONS DESCRIBED IN TABLE 6 OF THE |
||||
SUPPLEMENTAL APPLICATION* |
||||
Atlanta, GA |
x |
|
|
|
Austin, TX |
|
|||
Baltimore, MD |
x |
|
||
Bergen-Passaic, NJ |
x |
x |
x |
x |
Boston, MA |
|
|||
Caguas, PR |
||||
Chicago, IL |
x |
|
x |
|
Cleveland, OH |
|
x |
|
|
Dallas, TX |
|
x |
|
|
Denver, CO |
|
|||
Detroit, MI |
||||
Dutchess County, NY |
||||
Ft. Lauderdale, FL |
x |
|
||
Ft. Worth, TX |
|
|||
Hartford, CN |
x |
|
|
x |
Houston, TX |
|
|
|
x |
Hudson County, NJ |
|
x |
x |
|
Jacksonville, FL |
|
|||
Kansas City, MO |
||||
Las Vegas, NV |
||||
Los Angeles, CA |
x |
|
||
Miami, FL |
x |
|||
Middlesex-Somerset, NJ |
x |
|||
Minneapolis-St. Paul, MN |
|
x |
x |
|
Nassau-Suffolk, NY |
|
|||
New Haven, CN |
||||
New Orleans, LA |
x |
|
x |
|
New York, NY |
x |
x |
||
Newark, NJ |
x |
|
||
Norfolk, VA |
|
|
|
|
Oakland, CA |
|
x |
|
|
Orange County, CA |
|
x |
x |
|
Orlando, FL |
|
|||
Philadelphia, PA |
||||
Phoenix, AZ |
|
x |
x |
|
Ponce, PR |
x |
|
|
|
Portland, OR |
|
|||
Riverside-San Bern., CA |
||||
Sacramento, CA |
Other special populations described by only one EMA include migrant workers, pediatric patients, people in surrounding counties, suburban residents, Native Americans, people 45 years and over, people who have relocated in the past 2 years, and people living in a particular part of a county.
Information on populations with special needs is difficult to compare, as the special population categories vary across EMAs, and as consistent data are not available on HIV cases across EMAs. In addition, no reliable estimates exist for the overall size of several of the special populations, such as homeless or transgender populations, much less for the proportion of those populations with HIV.
FINDINGS
Applicants for Title I supplemental awards are asked to supply a tremendous amount of information about the epidemiology of HIV infection and AIDS, the prevalence of various comorbid conditions, poverty and insurance status, and populations with special needs. The Committee found that many of the measures requested in the application did not meet scientific soundness standards of reliability, validity, or empirical
Rural Individuals |
Incarcerated |
Mentally Ill |
Immigrants & Undocumented |
Other |
|
Deaf; hemopheliacs Transgender |
|||
PLWH in S/E King Cnty |
||||
x |
|
|||
|
Migrant workers |
|||
x |
x |
|
||
|
Haitians |
|||
6 |
5 |
5 |
3 |
|
support. Many measures lacked standard definitions of the population of interest, time period, and measurement procedure. There was also enormous variation in the data sources used by grantees. Some measures also lacked validity. For example, the application asks grantees to provide HIV prevalence information for their area and for “special populations,” but consistent estimates of HIV cases are not available at the EMA level, much less for special populations such as homeless or transgenders. Furthermore, many measures (e.g., comorbidities) lack a scientific evidence base linking them to resource needs. While such measures may have face validity, in that they are logical and are viewed as important by policy makers and grantees, scientific evidence should link them to direct measures of resource needs (such as costs of care, unmet needs, needs for additional services).
Overall, there is little consistency in the factors described in applications, and even when EMAs do describe comparable factors, there is tremendous variability in the types of indicators they use. Thus, it is virtually impossible to make objective comparative assessments of relative needs across areas. Furthermore, except for some information on insurance status and poverty rates, the applications do not describe variations in the costs of providing care in different areas or on the availability of
other resources. HRSA/HAB provides a scoring guide for reviewers of applications, but it also emphasizes the importance of reviewer judgment. Perhaps as a result of the difficulty in assessing the variation in resource needs across regions, over and above those due to differences in the prevalence of AIDS, EMAs’ Title I supplemental awards are highly correlated with their Title I formula (base) award. In FY2002, for example, the correlation in per-ELC supplemental and base awards was 0.87, indicating a close correspondence between the base and supplemental awards.
Although the use of idiosyncratic indicators of need makes comparisons across EMAs nearly impossible, discussions with Title I representatives suggest that some EMAs find the process of compiling a supplemental application useful for their own local planning efforts (Ryan White CARE Act 2002 Grantee Conference, August 22–23, 2003, Washington, DC: Meeting with Title I EMAs). Thus, EMAs may find value in retaining some aspects of this process for local planning and evaluation. The Committee also notes that the special needs of jurisdictions may not be completely captured by quantitative measures, and that one possibility is to have grantees include a brief (e.g., two pages) summary of their special needs with their application.
Finding 5-1 Resource needs are determined by a complex array of factors, including disease burden, the costs of providing care, and available resources. These factors, for example insurance coverage or costs of care, vary widely across regions. RWCA formula allocations rely primarily on one measure of disease burden (i.e., ELCs) in determining awards, although this measure does not well reflect underlying variations in resource needs. The Title I supplemental award is the largest RWCA grant program that attempts to take into account other factors affecting the complexity and costs of care.
Finding 5-2 The current Title I supplemental award process, which is determined by competitive application, relies on nonstandard and unvalidated measures of local need. Simple, commensurable measures are preferable to complex idiosyncratic measures in allocating resources and their use would improve the award process and resulting allocations.
Finding 5-3 The current Title I supplemental application process is burdensome for grantees. Given the high correlation between grantees’ per-ELC supplemental and base awards, the effort required for grantees to complete the application seems unjustified.
One solution to these challenges would be to specify a set of direct or indirect measures of resource needs. An example of the former would be to interview patients and ask them about the kinds of services they need.
An example of the latter would be to develop predictors for areas that are likely to have extra resource needs, such as those with a high proportion of residents with incomes below the poverty level. These issues are discussed in the following sections.
AVAILABLE DATA SOURCES
There are numerous data that could be used to develop indicators of resource needs. CDC and HRSA are conducting a comprehensive review of available data sources as part of the development of guidelines for epidemiologic profiles (CDC, 2003a). The epidemiologic profile is intended to assist RWCA grantees and HIV prevention community planning groups in resource and program planning, evaluation, and allocation decisions (CDC, 2003a). This review will provide information about each data source, the relevant population of interest, its strengths and limitations, and its availability. The report will include information on data sources such as HIV/AIDS reporting and supplemental surveillance efforts, other disease-reporting systems (e.g., STDs, tuberculosis), census data, and vital records information.
The Agency for Healthcare Research and Quality (AHRQ) and HRSA have also compiled a number of potentially relevant variables from a collection of data sources as part of a joint Safety Net Monitoring Initiative in response to a 2000 IOM report, America’s Health Care Safety Net: Intact but Endangered. One of the recommendations of the IOM report is that “concerted efforts be directed to improving the Nation’s capacity and ability to monitor the changing structure, capacity and financial stability of the safety net to meet the health care needs of the uninsured and other vulnerable populations.” (p. 10)
AHRQ and HRSA compiled two data books to assist with this effort: one for county and metropolitan areas and one for states. The first book presents data from 90 metropolitan areas in 30 states and the District of Columbia, including 354 counties and 171 cities. The data describe the health care safety net where 80 percent of Americans with family incomes below the federal poverty line live. The second book has similar information for all 1,818 counties in these 30 states (nonmetropolitan and metropolitan counties) (AHRQ, 2003). Data have been compiled on a number of variables that are potentially relevant in determining an area’s need for resources. Examples of data collected include rates of uninsurance, Medicaid coverage, presence of a community health center, level of uncompensated care, physician supply per 100,000 population, as well as economic indices, population data, information on immigrant population, and sociodemographic factors.
CDC conducts a number of supplemental surveillance studies (Table
5-5) that could provide information on resource needs, as well as quality of care. CDC is currently in the process of developing a new Morbidity Monitoring System that will use interview and chart data and will allow CDC to collect HIV/AIDS data from a representative population sample. A meeting regarding the design of the new system is planned for early 2004. CDC has announced plans to discontinue the use of two of its supplemental data collection systems, the Adult Spectrum of HIV/AIDS Disease (ASD) and the Supplement to HIV/AIDS Surveillance Project (SHAS) beginning in mid-2004 in favor of this project (Personal communication, CDC, October 16, 2003).
HRSA has also recently undertaken a major effort to standardize the types of data collected from grantees. All grantees now report data to HRSA using the Ryan White CARE Act Data Report (CADR).10 The CADR asks grantees to provide information on hundreds of data elements, such as client characteristics, service provision, and the costs of providing care. Unfortunately, the amount of data requested might preclude precise estimates of the most important elements, and the cost data are not compiled in a way that could be used to estimate the relative costs of providing a comparable element of care in different areas. Nevertheless, forms such as the CADR, if simplified and designed to provide specific types of information, such as estimates of resource needs and quality of care, could help standardize the plethora of data now submitted and evaluated for Title I supplemental awards. Redesign and coordination of these efforts across CDC and HRSA would enhance the usability of these data in assessing needs for both care and prevention resources and quality of care.
Finding 5-4 Many publicly available data sources, including data routinely collected by HRSA/HAB and CDC, could be used to assess resource needs using indicators that are comparable across areas. Direct measures probably would yield the most valid measures of need, but would be more expensive and perhaps less feasible than indirect measures.
PROPOSED APPROACH
The Committee recommends several steps for addressing some of the limitations of the current Title I supplemental award approach and the desire of Congress for HRSA/HAB to develop more quantitative indicators of need. The first step is to identify direct measures or predictors of resource needs that meet acceptable standards of scientific importance,
soundness, and feasibility. The second step is to develop a more explicit definition of the factors that applicants should consider when defining their resource needs.
One strategy for identifying predictors of needs, sometimes referred to as social area analysis, attempts to relate the characteristics of geographically defined populations to variations in disease or service use (Shevky and Bell, 1955; Pittman et al., 1986; Kessler, 1998). In a social area analysis, one relates the characteristics of an area to the needs of individuals in that area (Shevsky and Bell, 1955; Pittman et al., 1986; Kessler, 1998). This technique is potentially useful when it is difficult or infeasible to routinely ascertain the variable of direct interest (e.g., individual needs) but when the characteristics of an area allow one to predict the distribution of the variable of interest. For example, the Alcohol, Drug Abuse & Mental Health Administration Reorganization Act (P.L. 102-321 [1992]) required states to provide estimates of the prevalence of serious mental illness in their applications for block grant funds (Kessler, 1998). Since such data are not routinely collected, it was necessary to develop a way of estimating prevalence. A technical expert committee developed prevalence estimates for sociodemographic subgroups that then could be applied to the population counts in each state to produce final state estimates. To do this, they used data from a mental health study to estimate prediction equations in which area characteristics were used to predict the prevalence of mental disorders.
To illustrate how this type of approach might be used to estimate resource needs in the context of RWCA allocations, the Committee developed an example of a model that related publicly available county characteristics to direct measures of need. The Committee selected several publicly available variables as example predictors of need. Although the Committee did not find any direct measures of resource needs, the HIV Cost and Services Utilization Study (HCSUS)—an interview study of a probability sample of noninstitutionalized HIV-infected U.S. residents—did collect data on reported needs and quality of care (Bozzette et al., 1998; Shapiro et al., 1999).11 Using these data, the Committee identified significant predictors for reported needs and an indicator of care quality, whether a patient received treatment with highly active antiretroviral
11 |
HCSUS (Bozzette et al., 1998; Shapiro et al., 1999) was an interview study of a national probability sample of noninstitutionalized persons with HIV infection. HCSUS interviewed a total of 2,864 patients. Several members of the Committee were members of the HCSUS consortium, so the Committee was able to access internal data allowing it to match 2,360 of the patients to their primary site of care in 82 counties. See Appendix D for more information. |
TABLE 5-5 Selected Supplemental Surveillance Studies (CDC)
Project |
Start Date |
Description |
Supplement to HIV/AIDS Surveillance Project (SHAS) |
1989 |
The SHAS was begun in 1989 to obtain increased descriptive information on persons over 18 years of age with newly reported cases of HIV infection and AIDS. Information is collected from persons reported as having HIV infection or AIDS using a standardized questionnaire administered by trained interviewers. The questionnaire consists of modules involving demographic-socioeconomic information; drug use history, both injected and noninjected; sexual behavior history, including information about STDs, and use of health care services; reproductive history and children’s health of women with HIV infection or AIDS, and information on disabilities treatment and adherence. This information supplements the data routinely collected through national HIV and AIDS surveillance and is used to improve our understanding of a variety of issues related to the epidemic of HIV infections for use by prevention programs. |
Survey of HIV Disease and Care (SHDC) |
2000 |
The SHDC is a population-based, medical record abstraction project which collects information from the medical records of HIV-infected persons. In the pilot phase of the project, the data elements collected have included: opportunistic illness diagnoses, prescription of prophylactic medications, and other prophylactic practices, such as influenza vaccination and TB skin testing. Prescription of antiretroviral therapies, laboratory markers of state of HIV disease, and comorbid conditions, such as homelessness, mental illness, and substance use. The outcome measures of the project are estimates of proportions of HIV-infected persons receiving a certain medication or with a certain clinical outcome, with confidence intervals. The SHDC uses a two-stage, cluster sampling methodology to allow the calculation of estimates of clinical endpoints generalizable to the population of HIV-infected persons in care; the cluster sampling also improves the efficiency of the study by limiting the number of medical facilities where medical record abstraction must be done. |
|
Funding Information |
|
Locations |
Year |
Total |
Arizona |
2000 |
$1,937,637 |
Atlanta, GA |
2001 |
$2,110,588 |
Austin, TX |
2002 |
$2,903,831 |
Chicago, IL |
|
|
Delaware |
||
Denver, CO |
||
Detroit, MI |
||
Hartford, CT |
||
Houston, TX |
||
Jacksonville, FL |
||
Jersey City, NJ |
||
Kansas |
||
Los Angeles County, CA |
||
Maryland |
||
Miami, FL |
||
Minnesota |
||
New Haven, CT |
||
New Mexico |
||
Philadelphia, PA |
||
Richland and Charleston Counties, SC |
||
Tampa, FL |
||
Washington |
||
Houston, TX |
2000 |
$243,061 |
Louisiana |
2001 |
$559,364 |
Maryland |
2002 |
$235,532 |
Michigan |
|
|
New Jersey |
||
Ohio |
||
Philadelphia, PA |
||
Puerto Rico |
||
Virginia |
||
Washington |
Project |
Start Date |
Description |
Survey of HIV Disease and Care Plus (SHDC+) |
2001 |
The purpose of this project is to estimate the proportion of HIV-related morbidity that results from the known behavioral determinants of access to care and adherence to medical treatment in association with clinical and laboratory indicators of HIV morbidity, selected sites participating in the SHDC interview patients included in the chart abstraction portion of the Survey. The interview project incorporates standard behavioral surveillance questions on HIV testing, risk, care-seeking, and other related behaviors. This project allows the evaluation of behavioral findings (including access to care, HIV and OI therapy adherence, client perception of value of therapies, disclosure of HIV infection, etc.) and clinical outcomes (virological and immunologic markers or occurrence of OIs, prescription of antiretroviral and OI therapies, etc.). |
Adult/Adolescent Spectrum of Disease (ASD) |
1990 |
The ASD is a national surveillance project which collects demographic, clinical, laboratory, surveillance, health care utilization, and other related data on HIV-infected persons 13 years of age and over through a broad range of participating facilities in 10 U.S. cities. The geographic diversity of participating sites is further enhanced by the diversity of race, sex, sexual orientation, and socioeconomic status of the participants. Cases are drawn from hospital inpatient and outpatient facilities, infectious disease practitioners specializing in HIV infection, private practice medical groups, HIV treatment facilities, and health maintenance organizations. Since the inception of ASD in 1990, over 43,000 patients have been observed. Currently, ASD follows HIV/AIDS patients accessing care at participating sites to retrieve clinical, treatment, and laboratory data at 6 month intervals, beginning with the induction of the patient into ASD and ending with the patient’s demise. The ASD project is currently the principal source of national surveillance data on HIV-related morbidity. |
Project |
Start Date |
Description |
HIV Testing Survey (HITS) |
1996 |
Some public health and community groups remain concerned that implementing HIV case reporting may deter some at-risk persons from seeking HIV testing. The primary objective of the HITS is to identify the reasons that persons at risk for HIV infection may seek or defer HIV testing and HIV-related health care, and the role state HIV testing and reporting policies play in the decision, to assess whether HIV case reports underrepresent some populations and to improve HIV prevention planning. Additional objectives for HITS are to evaluate the influence of recent events, such as availability of drug therapies and new testing methodologies, on persons’ decisions to seek HIV testing. HITS is an anonymous, cross-sectional survey of persons at risk for HIV. Project areas use a standardized protocol based on targeted venue-based sampling methods. Specific recruitment sites and methods will be developed locally, in order to provide a generalizable understanding of HIV testing patterns in at-risk racial/ethnic minority populations. |
|
Funding Information |
|
Locations |
Year |
Total |
Arizona |
2000 |
$199,999 |
California |
2001 |
$1,252,205 |
Colorado |
2002 |
$1,581,088 |
Florida |
|
|
Houston, TX |
||
Illinois |
||
Kansas |
||
Los Angeles, County, CA |
||
Louisiana |
||
Maryland |
||
Michigan |
||
Mississippi |
||
Missouri |
||
Nevada |
||
New Jersey |
||
New Mexico |
||
New York |
||
New York City, NY |
||
North Carolina |
||
Ohio |
||
Oregon |
||
Philadelphia, PA |
||
Portland, OR |
||
San Francisco, CA |
||
Seattle, WA |
||
Texas |
||
Vermont |
||
Washington |
Project |
Start Date |
Description |
Enhanced Perinatal Surveillance (EPS) |
1990 |
EPS activities include two main activities in addition to those activities that are considered core pediatric surveillance. Participating areas are expected to match birth registries to HIV/AIDS registries in order to improve ascertainment of mother-infant pairs, and to collect supplemental information on both mothers and infants from a variety of medical records, including mother’s prenatal care chart, labor and delivery chart, and the infant’s birth chart and pediatric chart. In areas where HIV infection is not reportable by name a hospital-based approach to identify mother-infant pairs is pursued rather than the population-based approach which is feasible in HIV-reporting states only. Twenty-six areas were funded with 1999 supplemental funds to participate in EPS. |
AIDS Progression Study (APS) |
2001 |
The APS was designed to understand the characteristics of people who are infected with HIV who progress to or die from AIDS and to explain why progression to AIDS occurs. In addition, this time-limited study examines reasons for progression from AIDS to death among deceased AIDS cases. Abstracted from medical records during the 12-month period preceding AIDS diagnosis, the data include patient characteristics, HIV/AIDS-related history, testing history, AIDS defining conditions, HIV exposure, and laboratory data. |
SOURCE: CDC, 2003b. |
|
Funding Information |
|
Locations |
Year |
Total |
Alabama |
2000 |
$1,866,553 |
California |
2001 |
$1,874,431 |
Chicago, IL |
2002 |
$1,814,324 |
Connecticut |
|
|
District of Columbia |
||
Houston, TX |
||
Los Angeles, CA |
||
Louisiana |
||
Maryland |
||
Massachusetts |
||
Michigan |
||
New Jersey |
||
New York |
||
New York City, NY |
||
North Carolina |
||
Ohio |
||
Pennsylvania |
||
Philadelphia, PA |
||
Puerto Rico |
||
South Carolina |
||
Tennessee |
||
Texas |
||
Virginia |
||
Boston, MA |
2000 |
$199,999 |
Chicago, IL |
2001 |
$319,593 |
Denver, CO |
|
|
Hartford, CT |
||
Los Angeles, CA |
||
San Francisco, CA |
therapy (HAART).12 Details of the sample and analyses are represented in Appendix D. Below we summarize the approach and results.
As examples of indicators of need, the Committee used the number of needs reported by the patients and whether the patients had been treated with HAART drugs. In the HCSUS interview, each patient was asked the following questions:
-
Did you need income assistance such as SSI, SSDI, AFDC, or health care benefits from Medicaid or the Veterans Administration in the last 6 months?
-
Did you need to find a place to live in the last 6 months?
-
Did you need home health care in the last 6 months?
-
Did you need mental health or emotional care or counseling in the last 6 months?
-
Did you need drug or alcohol treatment in the last 6 months?
Using these data, the Committee calculated a variable called “number of needs,” which is simply the number of these questions that the respondent answered affirmatively.
The Committee selected several county characteristics that it considered representative of the kinds of variables that are likely to be related to resource needs for HIV care.
Examples of indicators of medical resources are:
-
Total general practitioners in 1996 divided by the total population in 1990,
-
Total number of medical specialists in 1996 divided by the total population in 1990.
Examples of area sociodemographic characteristics are:
-
Percent of the population that was African American in 1990,
-
Percent of the population that is foreign born in 1990,
-
Percent of population that lives in urban areas in 1990,
-
Percent of population who live in poverty in 1990,
-
Percent of population who are college graduates in 1990.
12 |
At the time of the HCSUS study, the recommended therapy for HIV disease was HAART. Recommendations for therapy change over time are updated in the treatment guidelines published by the Department of Health and Human Services (http://www.aidsinfo.nih.gov/) and others. |
Using publicly available data (the U.S. Census and a compilation of information called the Area Resource File [ARF]), the Committee compiled data on several characteristics of each county.
The Committee then used the data to estimate models that assessed how well these area characteristics and regions predicted the number of reported needs and receipt of HAART. The models (Appendix D) indicate that the education in an area, the number of general practice physicians, and the number of specialists in an area are statistically significant predictors of the number of needs reported by individuals. Specifically, persons reported more needs if they lived in areas with fewer college educated persons, fewer general practitioners, and more medical specialists. The relationships with education and general practitioners seem reasonable. The relationship for medical specialists is counterintuitive, but it might be a reflection of an emphasis on more expensive care at the expense of more basic services. The strongest predictors of not receiving HAART therapy were living in a county with a high percent of African Americans, a high percentage of families below the poverty level, and an area with more general practitioners. Some of these effects imply striking differences. For example, the average percentage of persons with a college education in the counties studied was 16 percent, with a standard deviation of 5 percent. The coefficient in the model for percent of persons with a college education implies that if one went from a county with a percentage of college educated persons that is a standard deviation below the mean (11 percent) to a county that is a standard deviation above the mean (21 percent), the average need score, which has a range of 0 to 5 and a mean of 1.29 would increase by 0.33. This model illustrates an approach that could be used to select and calibrate variables that predict various types of resource needs.
This example has several limitations. For example, the HCSUS sample was not designed to support the analysis of county effects. The Committee selected variables that were readily available for modeling. Thus, the Committee mainly included variables representing the general availability of medical personnel in an area and its socioeconomic characteristics. Undoubtedly, other variables are more likely to be related to the needs of HIV persons. For example, the health provider shortage areas and medically underserved areas (designated by HRSA’s Bureau of Primary Care) may have particularly high resource needs. The predictive value of such variables needs to be tested. However, it is important to recognize that many of the measures that grantees now report on their supplemental applications might not be related to resource needs. None of the applications the Committee reviewed provided empirical support for the association between resource needs and a specific “need factor.”
An alternative approach would be to develop a set of direct indica-
tors. There currently are no such indicators that are comparable across regions. It might be possible, however, to coordinate and/or consolidate current efforts conducted by HRSA/HAB and CDC. One would need to assess the tradeoff between scientific accuracy and cost. Surveying a scientifically valid random sample of HIV-infected persons would probably produce the most accurate assessment of needs and would allow one to develop measures directly related to the conceptualization of need proposed by HRSA. However, such an approach would be difficult and expensive to implement. The HCSUS study was able to identify and survey a probability sample of HIV-infected persons in treatment, but it was a complicated and expensive study. Presumably this would be less so for entities with legal surveillance authority. An indirect modeling approach could use available measures. The limitations of such a model are that it might be a poor predictor of interregional variations in needs and might be relatively insensitive to changes over time, if the predictor variables are not updated with sufficient frequency.
RECOMMENDATIONS
Recommendation 5-1 HRSA should modify the Title I supplemental application process. The severity-of-need component of the Title I supplemental award should be based on two components:
-
Quantitatively defined need, based on a small number of measures that can be calculated by HRSA/HAB.
-
Locally defined need described in a short narrative by the applicant.
Recommendation 5-2 A predominance of the weight for determining Title I awards should be given to the quantitative measure of resource needs that reflect variations in costs of care and fiscal capacity across EMAs.
Recommendation 5-3 HRSA/HAB should evaluate the feasibility and usefulness of using social area indicator models based on publicly available data that are collected in standardized ways across jurisdictions, to estimate EMA-level resource needs for the Title I supplemental award. This approach also might be useful in assessing resource needs for other RWCA discretionary grant programs.
Such an evaluation would entail several steps:
-
First, HRSA/HAB should review with additional experts the po-
-
tential data sources and develop recommendations for additional measures to be considered. HRSA/HAB should determine the availability of data to support these measures. The potential measures and corresponding data sources should be evaluated according to their importance, scientific soundness, and feasibility.
-
Second, HRSA/HAB should determine an appropriate definition of resource needs that could be measured directly. Examples used in this report were needs and unmet needs reported by persons living with HIV infection and total costs of care. However, none of these measures captures the resource needs that are most appropriately provided by RWCA funds. Such a definition might take into account whether an individual had alternative sources of public or private health insurance, generosity of that insurance, and/or the cost of providing services in a given area. Thus, estimating need might involve determining both the needs of individuals living with HIV as well as the costs of meeting those needs in a particular area.
-
Third, HRSA/HAB should develop a strategy for directly determining need in an adequate number of areas so the relationship between social area indicators and actual need can be estimated. The Committee knows of no such measures now available and so this step will involve consultation with survey experts and statisticians skilled in developing and estimating such models.
-
Finally, HRSA/HAB should develop models and assess the association between social area indicators and direct measures of need. It likely will be not be feasible to collect enough data to do this at the EMA level, but models for county variability should be created and data from those models used to assess their ability to explain between-EMA variability.
It is important to evaluate whether a periodic direct assessment of needs or a model-based approach would be more feasible and useful. Resource needs change rapidly and many area-level predictors (e.g., census data) do not change frequently enough to capture such changes. Thus, it may be better to periodically review and update the quantitative indicators used to allocate Title I supplemental award funds.
Recommendation 5-4 The Secretary of Health and Human Services (HHS) should evaluate the cost and utility of redesigning and coordinating studies conducted by HRSA/HAB and CDC to assess the specific needs and circumstances of people living with HIV. These data can be used to estimate resource needs and as part of quality assessment activities. The Secretary of HHS should also assess the cost and utility of the indirect modeling approach described in Recommendation 5-3 for assessing regional variations in resource requirements.
REFERENCES
AHRQ (Agency for Healthcare Research and Quality). 2003. Fact Sheet: Safety Net Monitoring Initiative. [Online]. Available: http://www.ahrq.gov/data/safetynet/netfact.htm [accessed September 3, 2003].
Bozzette SA, Berry SH, Duan N, Frankel MR, Leibowitz AA, Lefkowitz D, Emmons CA, Senterfitt JW, Berk ML, Morton SC, Shapiro MF. 1998. The care of HIV-infected adults in the United States. HIV Cost and Services Utilization Study Consortium. New England Journal of Medicine 339(26):1897–904.
Bozzette SA, Joyce G, McCaffrey DF, Leibowitz AA, Morton SC, Berry SH, Rastegar A, Timberlake D, Shapiro MF, Goldman DP, and the HIV Cost and Services Utilization Study Consortium. 2001. Expenditures for the care of HIV-infected patients in the era of highly active antiretroviral therapy. New England Journal of Medicine 344(11):817–23.
CDC (Centers for Disease Control and Prevention). 2003a. Data for decision making: HIV epidemiologic profiles, HIV program evaluation. Developing Epidemiologic Profiles for HIV Prevention and Ryan White CARE Act Community Planning—Grantee Training Meeting. Atlanta, GA: CDC.
CDC. 2003b. Selected Programs Under Program Announcement 00005. (Email communication, Patricia Sweeney, CDC, May 2, 2003).
CMS (Centers for Medicare and Medicaid Services). 1999. Report to Congress: Proposed Method of Incorporating Health Status Risk Adjusters into Medicare+Choice Payments. [Online]. Available: http://cms.hhs.gov/researchers/reports/1999/RTC_RiskAdjusters.pdf.
Freedberg KA, Losina E, Weinstein MC, Paltiel AD, Cohen CJ, Seage GR, Craven DE, Zhang H, Kimmel AD, Goldie SJ. 2001. The cost effectiveness of combination antiretroviral therapy for HIV disease. New England Journal of Medicine 344(11):824–31.
Gordis L. 1996. Epidemiology. Philadelphia, PA: W.B. Saunders Company.
HRSA (Health Resources and Services Administration). 2000. Unmet Need Consultation Report. Meeting Sponsored by the Health Resources and Services Administration HIV/AIDS Bureau. Rockville, MD: HRSA.
HRSA. 2001a. FY 2002 Grant Application Guidance. The Ryan White Comprehensive AIDS Resources Emergency (CARE) Act: Title I HIV Emergency Relief Grant Program. Rockville, MD: HRSA.
HRSA. 2001b. FY 2002 Title I Application Supplemental Application Scoring Guide. Rockville, MD: HRSA.
HRSA. 2001c. A Primer on Title I and Title II Formula Allocation Calculations. Unpublished document.
HRSA. 2003. Appendix 1: Glossary of Terms. Ryan White Comprehensive AIDS Resources Emergency (CARE) Act: Needs Assessment Guide (2003 Version). Rockville, MD: HRSA.
IOM (Institute of Medicine). 2000. America’s Health Care Safety Net: Intact But Endangered. Lewin ME, Altman S, Eds. Washington, DC: National Academy Press.
IOM. 2001a. Envisioning the National Health Care Quality Report. Hurtado MP, Swift EK , Corrigan JM, Eds. Washington, DC: National Academy Press.
IOM. 2001b. No Time to Lose: Getting More from HIV Prevention. Ruiz MS, Gable AR, Kaplan EH, Stoto MA, Fineberg HV, Trussell J, Eds. Washington, DC: National Academy Press.
IOM. 2003. A Shared Destiny: Community Effects of Uninsurance. Washington, DC: The National Academies Press.
Kahn JG, Janney J, Franks PE. 2003. Interim Report: Measuring Unmet Need for HIV Care Project. Submitted to Office of Science and Epidemiology, HIV/ AIDS Bureau, HRSA, U.S. DHHS.
Kaiser Family Foundation. 2000. Financing HIV/AIDS Care: A Quilt with Many Holes. Capital Hill Briefing Series on HIV/AIDS. Kaiser Family Foundation.
Kessler RC. 1998. A methodology for estimating 12-month prevalence of serious mental illness. Mental Health, United States 1998. DHHS Pub. No. (SMA) 99-3285 ed. Washington, DC: Government Printing Office.
MEDPAC (Medicare Payment Advisory Commission). 1999. Appendix A: Determining Medicare+choice payment rates. Report to the Congress: Medicare Payment Policy. [Online]. Available: http://www.medpac.gov/publications/congressional_reports/Mar99%20AppA.pdf.
NASTAD, Kaiser Family Foundation, AIDS Treatment Data Network. 2003. Prepared by Davis D, Aldridge C, Kates J, Chou L. National ADAP Monitoring Project Annual Report. NASTAD, Kaiser Family Foundation, AIDS Treatment Data Network.
NRC (National Research Council). 1997. Assessment of Performance Measures for Public Health, Substance Abuse, and Mental Health. Perrin EB, Koshel JJ, Eds. Washington, DC: National Academy Press.
NRC. 2000. Small-Area Income and Poverty Estimates: Priorities for 2000 and Beyond. Citro CF, Kalton G, Eds. Washington, DC: National Academy Press.
NRC. 2001. Choosing the Right Formula. Jabine TB, Louis TA, Schirm AL, Eds. 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.
Pittman J, Andrews H, Struening E. 1986. The use of zip coded population data in social area studies of service utilization. Evaluation and Program Planning 9(4): 309–17.
Shapiro MF, Morton SC, McCaffrey DF, Senterfitt JW, Fleishman JA, Perlman JF, Athey LA, Keesey JW, Goldman DP, Berry SH, Bozzette SA. 1999. Variations in the care of HIV-infected adults in the United States: results from the HIV Cost and Services Utilization Study. Journal of the American Medical Association 281(24):2305–15.
Shevky E, Bell W. 1955. Social Area Analysis. Stanford: Stanford University Press.
Society of Actuaries. 1997. Medicare/Medicaid risk contracting—profitability. RECORD 22(2):1–20.
U.S. Congress. 2000. Congressional Record. 146(123):S10032.