Patient Coverage Panel Report

(printer-friendly version - PDF, 300 kb)

Contents on this Page:

1. Summary of Panel Recommendations
2. Overview
   1. Purpose of Panel
   2. Cross-cutting Issues
3. Discussion of Variables
   1. Disease Progression
   2. Medicaid and ADAP Adequacy
   3. Pharmaceutical Data
   4. Social Area Indicator Analysis/MMP
   5. Substance Abuse
   6. Unmet Need
4. History of the Panel
   1. Members and Affiliations
   2. HSR/RTI Contact Information

Views expressed are those of the panel participants and do not represent official positions of the Federal Government

I. Summary of Panel Recommendations

The Patient Coverage Panel recommended including four variables in the severity of need (SON) index:

• Case fatality rate among reported living AIDS patients
• Medicaid adequacy
• Percentage of Federal poverty level (FPL) required for eligibility for the Medicaid Medically Needy program
• AIDS Drug Assistance Program (ADAP) adequacy.

The AIDS case fatality rate is intended to serve as a proxy indicator for severe cases of unmet need for primary medical care services. The Medicaid adequacy variable measures the ability of a State Medicaid program to meet the health care needs of patients with HIV/AIDS. The Medicaid Medically Needy program eligibility variable measures the presence or absence of such a program in a State and the relative generosity of its eligibility requirements. The ADAP adequacy variable measures the ability of a State ADAP program to meet the medication needs of patients with HIV/AIDS.

Panelists identified several variables that they considered important but that they placed on hold for future consideration because the data (1) was currently unavailable but likely to be available in the near future or (2) was currently available, but its validity and reliability could not be accurately assessed given the time constraints of this panel’s work.

• Medicaid enrollment
• Rapid progression to AIDS diagnosis
• Receipt of highly active antiretroviral therapy (HAART) (pharmaceutical data)
• Social Area Indicator Analysis based on the Morbidity Monitoring Project (MMP).

The panel identified two variables that they also deemed extremely important but could not determine an acceptable data source that would be reliably available in the near to moderately distant future.

• Unmet need for HIV primary medical care
• Unmet need for substance abuse treatment.

Finally, the panel considered, but did not recommend, several other variables for inclusion in the SON index because they (1) were correlated with one of the variables recommended for inclusion, (2) did not have a sufficient impact on the SON yet, or (3) could not be accurately measured by the publicly available data yet. Variables considered but not recommended for inclusion include:

• Phencyclidine (PCP) incidence
• Hospital discharge data
• The ADAP waiting list
• The Federal medical assistance percentage (FMAP).

^ back to top

II. Overview

A. Purpose of Panel

The Patient Coverage Panel was responsible for identifying variables that describe the degree of medical care currently infected HIV/AIDS patients can expect to have access to in the absence of the CARE Act program. The panel evaluated a wide range of variables, considering the rationale for each variable and whether adequate data existed to measure the concept and proposing potential uses for the measures in an SON index. To accomplish these goals, panel members identified and conducted four sequential tasks:

1. The panel identified an initial set of variables thought to be potential determinants or descriptors of HIV/AIDS patients’ existing health coverage and need for services.
2. After identifying the variables to be considered, the panel divided into the following six workgroups to discuss and evaluate the variables in greater depth:
   • Disease Progression: Considered the proxy indicators of unmet need for primary medical care.
   • Medicaid and ADAP Adequacy: Considered a way to measure differences in State Medicaid and ADAP programs.
   • Pharmaceutical Data: Considered the use of proprietary prescription data to measure receipt of HAART.
   • Social Area Indicator Analysis/MMP: Considered the use of a specific methodology for estimating indirect measures of resource needs through area characteristics.
   • Substance Abuse: Considered the potential data sources for estimating the unmet need for substance abuse treatment services among injection drug users.
   • Unmet Need: Considered ways to estimate the number of people living with HIV and AIDS (PLWHA) who were aware of their infection status but currently did not receive medical care.

(Note: The hospital discharge data variable was not discussed exclusively by any of the six workgroups listed above. Panel members did not complete a template for this variable given the unanimous agreement for removal by the group at the first initial meeting in October 2005. The decision to remove this variable from further consideration was based primarily on limitations of the data. Specifically, hospital discharge data are often incomplete and characterized by significant reporting delays, and the AIDS-specific discharge codes are not consistent across States.)

3. The workgroups completed an evaluation template for each variable to help assess the value of each variable and to develop a final set of recommendations for inclusion of variables in an SON index. The templates asked panelists to define each variable; articulate a clear rationale for its inclusion; identify specific data to measure the variable; assess the validity, reliability, and potential biases of each measure; and suggest whether the variable should be forwarded to the larger SON Panel for inclusion.
4. Panelists were then asked to score each variable considered by the six workgroups from 1 to 5, with 1 indicating a variable of the highest importance and 5 indicating a variable of the lowest importance, based on how well each variable measures the theoretical concept of SON. Panelists were also asked to consider that variables may covary or measure the same concept and were asked to prioritize similar variables as opposed to giving them all the same score. These scores were then compiled, and the averages were ranked (Table 1).

Table 1. Patient coverage variables considered and panelists’ priority score (variables forwarded to the full panel highlighted in yellow)

* Due to the challenging and political nature of the ADAP adequacy variable, the Medicaid and ADAP adequacy subgroup completed much of its work for this variable offline with consistent feedback from the larger group. The measure required significantly more time and effort than the other variables considered and, as a result, was not completed at the time the group made its formal rankings. However, once the subgroup completed its work, there was general consensus among the larger group for the recommendation of the variable.

Variables such as rapid progression to AIDS diagnosis and Social Area Indicator Analysis/MMP received high conceptual rankings by the panel, but were not forwarded to the full panel primarily due to the inadequacy of the data used to measure them.

^ back to top

B. Cross-cutting Issues

The panel also discussed several issues that cut across two or more areas of concentration:

• The panel weighed concerns about creating disincentives or perverse rewards (e.g., penalizing States that make significant financial contributions) through the inclusion of variables related specifically to Medicaid and ADAP against the need to provide health care for needy patients in States that may stint on care. The panel recognized the inherent difficulty in identifying a revenue-neutral way to ensure adequate care for all needy patients nationwide without, to some degree, penalizing areas that invest State resources in caring for HIV/AIDS patients.
• The panel discussed at length the need for a standardized measure of undiagnosed HIV patients that could be applied without State/grantee input (e.g., they wanted to avoid the scenario of “We can’t identify persons with HIV because we don’t have any money”).
• The panel considered assessing only Federal contributions to specific programs, such as ADAP and Medicaid, as opposed to the program’s entire funding including State and local contributions.

The panel recommended the inclusion of four variables. Many of the other variables the panel considered were recognized as being important to measuring SON but either lacked adequate data to be measured without bias or covaried strongly with variables forwarded for inclusion. These limitations are described in Section III.

Section III of this report is divided into six sections that correspond to the main areas of investigation by the panel. Sections are then subdivided into two subsections, the first discussing variables that were accepted by the entire group as potential elements for an SON index and the second describing variables that were not. The work on each variable is summarized briefly, followed by the completed evaluation template.

^ back to top

III. Discussion of Variables

A. Disease Progression

Variables Considered: Case fatality rate among reported living AIDS patients, PCP incidence, rapid progression to AIDS

The Disease Progression Workgroup defined variables that would serve as proxy indicators of unmet need for HIV primary medical care. The group argued that case fatality rates among reported living AIDS patients were an indirect measure of unmet need. They theorized that increased mortality among this population in certain States was indicative of a lack of access to and utilization of primary medical care, including antiretroviral drug therapy and opportunistic illness (OI) prophylaxis.

The Workgroup also considered, but (because of data limitations) did not forward for inclusion, two other variables that were also intended to measure poor people’s access to medical care services indirectly: the incidence of Pneumocystis carinii pneumonia (PCP) and rapid progression from initial HIV infection to symptomatic AIDS illnesses. Reporting of PCP cases to the Centers for Disease Control and Prevention’s (CDC) HIV/AIDS Reporting System (HARS) is only required when it constitutes the AIDS-defining condition; therefore, case counts reported in the surveillance data may underestimate incidence of PCP to an unknown and variable degree across jurisdictions. Because no national standards for HIV and laboratory reporting exist, it is impossible to measure the rapid progression from HIV to AIDS in a reliable and consistent way across jurisdictions.

1. Variables forwarded for consideration

Case fatality rate among reported living AIDS patients: The panel recommended a measure of case fatality rates among reported living AIDS patients as a proxy indicator for severe cases of unmet need for HIV primary medical care services. If people living with AIDS (PLWA) die at significantly higher rates in certain States than in others, this may be because of large differences in access to and utilization of primary medical care (e.g., antiretroviral [ARV] drugs, OI prophylaxis).

The panel recommended calculating the case fatality rate by dividing the number of deaths among reported, living AIDS cases by the total number of reported, living AIDS cases alive during the corresponding year. Furthermore, the panel recommended calculating this ratio individually for each of the 5 preceding years and then averaging the estimates for final input into the model. The variable (constructed using data obtained from HARS) represents the observed probability of death over a 1-year period, not considering the cause of death or the background mortality of an area. The Workgroup discussed the extent to which patients infected with AIDS die of unrelated causes and the variability of such background mortality by area and other relevant demographic and risk factor characteristics. The panel explicitly recommended using data from AIDS cases only in place of HIV and AIDS cases primarily because the inclusion of HIV cases would bias systematically States with less mature epidemics (e.g., more HIV-infected patients as compared to AIDS). States with emerging epidemics may experience the same degree of mortality among AIDS cases as compared to States with mature epidemics, but the case fatality rate would be lower due to the larger numbers of HIV cases included in the denominator.

Panel members considered incorporating relative survival analysis to adjust for the background mortality of an area. This common epidemiological approach uses age-, sex-, and race-specific data from published life tables to estimate the expected mortality rate patients would have experienced irrespective of the relevant condition. However, performing a relative survival analysis as discussed by the Workgroup would not be possible because of limitations of HARS data. Such an analysis requires person-level diagnosis data to estimate observed survival time. In addition, AIDS-specific population information is not available in the life tables.

CDC surveillance experts invited to speak to the panel cautioned against attributing deaths reported in HARS to AIDS. CDC surveillance experts noted that jurisdictional variations in death rates depended on the age of the epidemic in an area (as patients with HIV infections of longer duration have a higher probability of death and patients who have been on ARV therapy longer are more likely to develop resistance); the risk characteristics of the HIV-infected group for death from other causes, such as drug overdoses, homicide, suicide, and acute injury; and access to health care. They noted that these differential causes of mortality were probably impossible to separate given existing data.

Despite these limitations, Workgroup members recommended forwarding this variable because regardless of the source of the disparity (e.g., presence of mature epidemic, inadequate access and adherence to treatment, or a disproportionate probability of death from causes unrelated to HIV/AIDS), panel members felt that significantly higher mortality in an area reflected greater resource needs.

2. Variables considered, but not forwarded for consideration

PCP incidence: The panel considered, but did not recommend, the use of PCP incidence as a proxy indicator for severe cases of unmet need for primary medical care services in addition to low levels of HIV counseling and testing. PCP is an opportunistic infection that can easily be prevented with prophylaxis; comparatively increased incidence of PCP would indicate a severe lack of primary medical care. The panel hoped this measure could be created using information contained in the HARS database.
However, the Workgroup removed this variable from further consideration because reporting of PCP cases to HARS is only required when it constitutes the AIDS-defining condition. This reporting system means the case counts observed in the surveillance data are likely to significantly underestimate the true incidence of PCP. Panel members discussed incorporating a statistical adjustment for the systematic underreporting of PCP, but they determined that calculating such an adjustment across States would be difficult and burdensome without further studying the relative completeness and reliability of PCP reporting in general.

Rapid progression to AIDS: The Workgroup also discussed a measure of rapid progression to AIDS - defined as an AIDS diagnosis within 12 months of an HIV diagnosis - as a potential proxy indicator for severe cases of unmet need for primary medical care. Such rapid progression may indicate late testing, issues related to access to care, poor adherence to treatment regimen, or viral resistance. Following this assumption, the availability of testing, Public Health Service (PHS) standard of care, and the provision of treatment adherence monitoring are strongly influenced by the availability of resources in an area.

Panelists were concerned about the variable’s ability to accurately capture the concept of unmet need, and the measure proved difficult if not impossible to construct with existing data sources. The Workgroup noted that the variable could be highly correlated with access to primary medical care (e.g., insurance and poverty status), the availability and utilization of anonymous testing, and viral resistance and therefore might be a strong measure of need. The current variability in name-based HIV and laboratory reporting would make the variable impossible to estimate. In the absence of uniform HIV and laboratory reporting, panel members felt the constructed progression measure would be unacceptably unreliable, inconsistent, and incomparable across States. However, the Workgroup made a specific recommendation to reexamine this variable for inclusion in future SON indices once mature and reliable HIV reporting systems are in place.

^ back to top

B. Medicaid and ADAP Adequacy

Variables Considered: Medicaid adequacy and enrollment, ADAP adequacy, ADAP waiting list, FMAP

Access to and the quality of State Medicaid and ADAP programs potentially measure the “available resources” portion of the Institute of Medicine (IOM) Report’s (“Measuring What Matters: Allocation, Planning, and Quality Assessment for the Ryan White CARE Act”) SON algorithm. The availability of generous State Medicaid and large State contributions to CARE Act co-financed ADAP programs might substantially alleviate the burden on the CARE Act to provide services in an area because many of the medically needy who in other jurisdictions would need to turn to the CARE Act for assistance would be covered by the State. In States that cannot afford it, do not choose to allocate funds to support a generous Medicaid program, or do not make substantial State contributions to ADAP, many patients may suffer with inadequate medical care. The key issue for this panel to consider was how to allocate additional funds to needy patients in States with poor Medicaid programs without treating generous States unfairly (at best) or creating incentives that might lead generous States to reduce medical services for PLWHA (at worst). (Note that even in a State that provided very high-quality medical services for all its patients with HIV and AIDS, there would still be a role for the CARE Act because the Act covers several services that are outside the scope of Medicaid.)

The Workgroup and the overall panel had extensive discussions regarding the potential disincentives to States created by linking Medicaid adequacy to funding allocations. Specifically, the panelists did not want to penalize directly States that make generous financial contributions to State Medicaid programs. The panel discussed a number of ways to mitigate or avoid this issue, including adjusting for the FMAP, simply not recommending a Medicaid measure in any form, or applying a weight to the Medicaid adequacy variable once its impact is estimated. However, the panel ultimately moved forward with the variable given its central importance in the IOM Report’s definition of resource needs and its practical and significant impact on whether patients receive the medical care they need to survive.

To measure Medicaid, the Workgroup considered two types of variables: (1) those that measure access or enrollment in Medicaid and (2) those that measure the quality of HIV/AIDS services available to those Medicaid recipients who are enrolled. To indicate enrollment, the group investigated using actual estimates of patients with any HIV or AIDS diagnosis, or any prescription for ARVs identified in the medical claims data of each State’s Medicaid program in a given calendar year.

However, an initial assessment of such enrollment figures provided by the Centers for Medicaid/Medicare Services (CMS) indicated that such estimates are likely inaccurate at this time, for several reasons, but primarily due to the exclusion of Medicaid Managed Care enrollees in the data (the impact of which varied substantially across States). When the panel compared the Medicaid enrollment numbers to the total number of cases reported by HARS, there was an unacceptable and in some cases, unexplainable amount of variation in the measurement. For example, the percent of total reported cases enrolled in Medicaid ranged from less than 8 percent in one State to greater than 100 percent in four States.

Instead, on an interim basis the panel recommended using percent of FPL required for eligibility for the Medicaid Medically Needy program. For future indexes, the Health Resources and Services Administration (HRSA) should work to develop better measures of HIV or AIDS patients enrolled in State Medicaid programs.

To measure Medicaid adequacy, the group recommended an index variable created from one variable measuring the quantitative amount States spent per Medicaid recipient (adjusting for jurisdictional variation in labor costs) and a second variable measuring the types of special Medicaid programs available for PLWHA.

To measure ADAP adequacy, the panel recommended using the percent of FPL required for eligibility in the program. Although CARE Act Title I supplemental funds do not provide direct reimbursement for drugs directly (paid for primarily through Title II), the panel thought inadequacies in a State ADAP program could result in additional strains on an area’s Title I program.

1. Variables forwarded for consideration

Medicaid enrollment and adequacy: Panelists recommended using a Medicaid adequacy scale constructed using two data elements which both the relative benefits available to PLWHA and programs which enhance enrollment or eligibility.

For Medicaid enrollment, the panel recommended using the poverty level required to qualify for the State’s Medically Needy program. Specifically, in the absence of an actual count of patients with HIV and AIDS enrolled in Medicaid, the panel elected to measure Medicaid enrollment the percent of the FPL required of eligibility in the program. The variable was created using data collected annually by the Kaiser Family Foundation. The variable was collapsed categorically in the following manner with higher

Scores indicating more generous ADAP programs:

• 1 = No Medically Needy program in the State
• 2 = 1 to 59 percent FPL
• 3 = =60 percent FPL.

Future considerations regarding measuring Medicaid enrollment and adequacy -The panel would have preferred to estimate the actual number of patients with HIV and AIDS enrolled in Medicaid, adjusted for the adequacy of the Medicaid program, but found this impossible using currently available data. The panel recommended that HRSA continue to investigate estimate the actual number of patients with HIV and AIDS enrolled in Medicaid by using claims data. They also suggested that HRSA consider requiring require States to report the number of PLWHA enrolled in Medicaid, using State data systems. States could achieve this either by requiring Medicaid Managed Care vendors to document HIV- or AIDS-related encounters or, at the State or local level, by comparing the State Medicaid enrollment roster to the names of PLWHA in a jurisdiction. Jurisdictions then would report these numbers in aggregate to HRSA. The panel realized that such a recommendation would create burden for the States, so it also recommended that HRSA continue to work with CMS to investigate how to identify PLWHA in CMS claims data.

Measuring Medicaid Adequacy - The panel created a Medicaid adequacy adjustment factor using the following two data elements:

1. Average Medicaid expenditures per social security income (SSI) beneficiary
2. Presence and number of special Medicaid programs covering PLWHA (four programs total)
   • Targeted case management for people with AIDS
   • Home and Community Based Care Program (HCBC)
   • Pharmacy carve out in capitation rates
   • Differential rates for HIV (capitation, fee-for-service, or risk adjustment pool).

The data used to construct the adequacy adjustment factor were obtained using data from the U.S. Social Security Administration’s Office of Policy and the Infectious Disease Society of America.

The average expenditure component of the adequacy adjustment factor was adjusted for State-level differences in the wages of health care professionals common to HIV primary care programs. Following the recommendations of the Associated Costs Panel, the panel used average hourly wage data from the Occupational Employment Statistical Survey (OESS) conducted by the Bureau of Labor Statistics (BLS) and identified three major labor categories specific to HIV primary care programs:

• Family and General Practitioner (SOC code 291062)
• Registered Nurse (SOC code 291111)
• Licensed Practical and Licensed Vocational Nurse (SOC code 292061).

For each of the three labor categories, the State-level average hourly wage was divided by the national average to create an index normalized to one. The indices for each labor category were then averaged to produce an overall geographic wage index.

After adjusting for regional variations in labor costs, the variable was then collapsed categorically in the following manner based on the number of standard deviations from the mean national expenditure observed in the data:

• 1 = average expenditures more than one standard deviation below the mean
• 2 = average expenditures within one standard deviation above and below the mean
• 3 = average expenditures more than one standard deviation above the mean.

The panel assigned a specified number of points to each of the four identified benefit programs based on the program’s relative impact on the resource needs of needy patients. The point system was constructed such that the maximum number of points that could be earned was (e.g., all four programs present):

• 0 points = No programs
• 2 points = Targeted case management for people with AIDS
• 1 point = HCBC
• 4 points = Pharmacy carve out in capitation rates or differential rates for HIV (capitation, fee-for-service, or risk adjustment pool).

Finally, the overall Medicaid adequacy adjustment factor was calculated by individually dividing each element by the highest possible value (7 and 3 for the special programs and average expenditures variables, respectively), summing the two variables (range = 0-2), and then dividing by the highest combined possible value (2). The net effect of this process was to produce a variable ranging from 0 to 1, such that scores closer to 0 indicated State Medicaid programs with lower per-person expenditures and fewer special programs benefiting PLWHA.

The panel identified several issues related to the data:

• The scale measuring Medicaid Adequacy was based on expert opinion and should be considered as preliminary. HRSA should consider funding research to evaluate the impact of different levels of spending and programs on the adequacy of care for PLWHA, and to evaluate methods (potentially using passively collected data at the Federal level or by requiring reporting by grantees) to enumerate patients with HIV/AIDS enrolled in Medicaid. Although Medicaid adequacy is a crucially important variable in determining regional differences in SON, the subjective nature of the variable the panel was able to forward limit the potential weight this variable could be given in a resource allocation index.
• The average Medicaid expenditures per SSI beneficiary variable included information collected from all SSI beneficiaries, not specifically data from patients with HIV/AIDS. Although it is possible to construct data for patients with HIV/AIDS, the panel felt this more general SSI number would indicate better the State’s overall program generosity and would not be confounded by the variations in the stage of disease of State Medicaid programs.
• The effort required in annually updating the presence and number of the four identified special Medicaid programs benefiting PLWHA could be significant. To the panel’s knowledge, these data are not available in a central location, so each State’s Medicaid plan would need to be reviewed individually on an annual basis. To compensate for this data limitation, panelists recommend that this information be submitted directly by grantees and consolidated by the HRSA HIV/AIDS Bureau (HAB).