Impact of Migration on Estimates of the HIV/AIDS Population Out-of-Care

Background: Estimating HIV/AIDS cases not receiving medical care requires a denominator of cases resident in the region. These are typically based on surveillance data that does not account for migration after diagnosis.

Objective: To examine size and description of estimated out-of-care HIV/AIDS population in Oregon without and with adjustment for migration.

Methods: The unadjusted denominator of HIV/AIDS was defined as all cases diagnosed before 12/5/2006, residing in Oregon at diagnosis and living on 12/4/2007. Cases not residing in Oregon at time of diagnosis were excluded. To estimate emigration, Oregon HIV/AIDS Reporting System cases without evidence of recent care in Oregon were matched to the CDC national database to identify cases reported in other states after their Oregon case report. The adjusted population included both immigrant and ‘Oregon-native' cases minus emigrants. Out-of-care was defined as absence of CD4 or viral load test result during 12/5/2006 – 12/4/2007.

Results: The unadjusted number of prevalent cases was 4,736 and 32% of these was out-of-care. The adjusted number of prevalent cases was 4,038 and 20% of these was out-of-care. Without accounting for migration, no race/ethnicity or risk group was more likely to be out of care. After accounting for migration, Hispanics and African Americans were more likely than whites, and male injection drug users and heterosexuals more likely than gay men to be out of care.

Conclusion: Controlling for migration substantively changes estimates of both proportion out-of-care and distribution of that population and leads to markedly different conclusions about composition of the out-of-care population.