BR Phillps, Veterans Affairs San Diego Health Care System; V McWhorter, Veterans Affairs San Diego Health Care System; SA Bozzette, Veterans Affairs San Diego Health Care System; S Chang, Veterans Affairs Palo Alto Health Care System

Objectives: For more than a decade, data on veterans with HIV Disease in VA care has been stored in a central registry, the Immunology Case Registry. While data is automatically downloaded nightly for patients already on the registry, initial entry is a manual process and thus subject to error. This paper evaluates the extent to which VA patients with HIV Disease are omitted from the registry and describes the characteristics of these patients. This analysis helps to assess the extent to which veterans with HIV Disease are not in regular care for that disease, potentially due to disparities in access to services, and the extent to which VA facilities are not receiving complex-care funding for veterans receiving antiretroviral agents.

Methods: The method is to identify veterans with characteristics indicative of HIV Disease on other national VA databases and to compare the population of patients so identified to the population on the registry. We developed a conservative algorithm to identify patients with HIV Disease based on the presence of:

1) inpatient and outpatient records (on Austin Automation Center databases) with diagnoses of AIDS, HIV Disease, or highly specific opportunistic complications of HIV Disease; or 2) pharmacy dispensing records (on Pharmacy Benefits Management database) for antiretroviral agents, excluding cases of post-exposure prophylaxis and treatment for hepatitis. Descriptive statistics were used to compare omitted patients with those on the registry, with t-tests and Chi-squared tests to identify characteristics with statistically significant differences.

Results: For fiscal years 1998 and 1999, a total of 5,874 veterans not listed on the registry had at least one inpatient or outpatient record with a diagnosis of HIV, AIDS, or an opportunistic complication highly specific to HIV Disease. These omitted veterans received an average of 11.0 outpatient visits during the period, compared with 15.5 for the veterans on the registry. In addition, preliminary analysis indicates that another 202 veterans not on the registry (and with no records with the diagnoses of interest) received antiretroviral agents from the VA during fiscal year 1999.

Conclusions: Under the conservative assumption that fifteen percent of the identified cases represent coding errors or the rare instances in which these opportunistic complications are not associated with HIV Disease, we estimate that at least 5,000 veterans with HIV Disease and under VA care are omitted from the registry. The omitted cases represent about 20 percent of the veterans with HIV Disease under VA care.

Impact: Many of the omitted veterans received a substantial amount of VA outpatient care, and it is likely that many were receiving antiretroviral agents. While further analysis of their use of antiretroviral agents is underway, VA facilities likely would have been entitled to tens of millions of additional dollars in complex-care payments if these veterans had been included on the registry. Additional complex-care payments would have totaled over $7 million just for the 200 cases with antiretroviral agents identified to date. These additional funds might have been used to improve access to care and the quality of care for HIV Disease.