Retention in care among adults infected with HIV

Background: Adults infected with HIV live longer when they receive regular medical care, yet many adults are not retained in care. This study compares two methods of operationalizing retention and evaluates the contribution of individual characteristics on likelihood of retention.

Methods: A chart review was performed for a random sample of 212 active clients of an HIV primary care clinic. Data on clinic utilization, sex, age, race, mental illness, substance use, AIDS diagnosis, language, and housing were collected for each participant. Clinic utilization data was used to classify participants for retention and to score them by frequency of visits.

Results: Binary logistic regression analysis was compared to ordinal logistic regression analysis, using distinct retention measures as outcome variables. In the binary model, those with a history of substance use were about half as likely to be retained (p=.030) and those with an AIDS diagnosis were more than 2 times as likely to be retained (p=.015). These characteristics were also significant predictors in the ordinal model, with the additional finding that those reporting primary language other than English were more likely to have higher retention scores (p=.015).

Conclusions: Individual characteristics are associated with retention in care. An ordinal score for retention can identify unexpected predictors, but a dichotomous measure of retention may be easier to use and interpret for providers involved in retention-related quality improvement activities.