In this investigation of 1,750 HIV-infected inmates released from the nation's largest prison system, we found that only 20% enrolled in outpatient care by 30 days, and only 28% did so by 90 days. Although descriptions of several programs developed to enhance linkage to medical care and other community services for recently released HIV-infected inmates have been published,12–14
our study is the first—with the exception of a small pilot study11
—to examine the association between inmate characteristics and clinical factors with the likelihood of initiating HIV clinical care after release.
Recent reports indicate that while most HIV-infected prison inmates adhere to ART and have positive -treatment results during incarceration, these effects are generally not sustained when inmates are released into the general community.8,9,17
Several studies have found that a substantial proportion of released inmates who are subsequently reincarcerated exhibit higher viral loads and lower CD4 counts upon their return to prison compared with when they were released.8,9
In a retrospective study of 292 HIV-infected -prisoners who received ART during incarceration, Springer et al. reported that the mean CD4 lymphocyte count decreased by 80 lymphocytes/μL and the mean viral load increased by 1.14 log10
copies/milliliter (mL) (p<0.001) between the time of the prisoners' release and subsequent reincarceration.8
Stephenson et al. retrospectively matched 15 HIV-infected prisoners receiving ART who were released and then reincarcerated within a mean of nine months with 30 inmates receiving ART who remained incarcerated during that time period.9
At the conclusion of the study period, the median change in plasma HIV RNA level in the reincarcerated inmates was 1.29 log10
copies/mL compared with –0.03 log10
copies/mL in the group that remained incarcerated.
In an earlier study of 2,115 released inmates who received ART while in the Texas prison system, we found that only 18% of the cohort filled a prescription for ART medications within 30 days of release, and only 30% did so within 60 days of release.10
Although our previous study provided additional evidence that release from prison is associated with decreased adherence to ART regimens, it did not examine the proportion of released HIV-infected inmates—including those not receiving ART in prison—who established timely linkage with HIV-related clinical outpatient care or the specific factors associated with successful linkage to care. Our current investigation was designed to address both of these issues, which have important public health implications but have received little attention among investigators.
We found that inmates who were ≥30 years of age were twice as likely to enroll in a community-based HIV clinic at both 30 and 90 days after their release. This finding is generally consistent with studies that have shown positive associations between age and either linkage to18
or retention in19,20
HIV care. Several factors may underlie this finding. First, it is possible that the lifestyle adaptations for successful initiation of outpatient care may be more difficult for younger adults. In particular, adults younger than 30 years of age are reported to have higher rates of substance abuse than older adults, which may play a role in their failure to establish timely HIV-related care.21
Older age may also be associated with increased recognition of mortality and, therefore, greater motivation to adhere to treatment guidelines. Alternatively, better treatment adherence among older adults may be partly attributable to a survivor effect, whereby those who more thoroughly comply with HIV treatment recommendations may outlive those who are less adherent.22
Our study also showed that inmates who were on ART at the time of their release had higher rates of enrolling in post-release outpatient care at both time points. It is possible that these inmates had been in treatment longer and, as a result, were more knowledgeable about their condition. In this case, taking steps to increase patients' knowledge of their condition by providing HIV education and adherence counseling while they are incarcerated would improve the likelihood of continuity of HIV care following release from prison.
We also found that inmates with schizophrenia had a greater likelihood of enrolling in an HIV clinic within 30 days of their release; however, there was no association between this condition and enrollment at the 90-day time point. Moreover, none of the other psychiatric disorders we examined were associated with linkage to HIV care at either time point. Because optimal management of schizophrenia requires frequent follow-up care, it is possible that those inmates with this condition may have established a linkage to the medical community before their incarceration and thus had a greater likelihood of accessing HIV care shortly after their release. It is noteworthy, however, that inmates with diabetes—a condition that also requires frequent monitoring—did not exhibit a greater likelihood of enrollment in post-release care. In designing a study of 1,209 patients who presented to an HIV/AIDS clinic in Alabama, Krawczyk et al. also postulated that an existing connection to the medical community as a result of a preexisting condition would reduce delays in accessing HIV care;23
however, using bivariate analyses, the investigators found that a history of diabetes was associated with delayed access to HIV treatment. In contrast, multivariate regression analyses showed that a history of mental illness was associated with an increased likelihood of timely access to HIV care. Additional studies are needed to determine whether preexisting chronic disorders are useful, valid predictors of successful linkage to HIV care.
Notably, inmates who were released on parole were slightly more likely to enroll in HIV care at both time points, but this association did not reach statistical significance. A recent study showed that, after adjusting for potential confounders, inmates who were released on parole were more likely than those who had a standard, unsupervised release to fill an ART prescription within 30 and 60 days after release.10
Taken together, these findings suggest that the various requirements of parole (e.g., mandatory recurring visits with parole officers and participation in mental health or substance abuse treatment) may slightly improve adherence to HIV-related medical care and treatment; however, further studies are needed to determine the overall impact of parole supervision on adherence to HIV-related medical care and treatment.
The results of our study suggest that several relatively inexpensive (approximately $45,000 per year) and simple discharge-planning interventions may have a modest effect on improving linkage to health care after release from prison. We found that HIV-infected inmates who received assistance with completing consent forms for the release of personal health information to HIV community clinics and who had a copy of their medical records sent to the clinic before their first appointment were more likely to initiate HIV outpatient care, at both time points, compared with inmates who did not receive such services. Given the retrospective design of this study, however, this finding should be interpreted with some caution.
Several potential limitations could have influenced our findings. First, it is possible that some members of the study cohort received outpatient care from a clinic not included in the CPCDMS database. However, because the majority of HIV-infected TDCJ inmates have no private or public health insurance at the time of their release from prison, and because almost all are from disadvantaged socioeconomic backgrounds, it is highly unlikely that a significant number would have received care from a private clinic in Harris County. Additionally, all members of the study cohort were referred to one of the four Harris County-based Ryan White Act-funded clinics on the basis of their documented permanent address and at their own request after discussing referral options with the TDCJ HIV discharge planning coordinator. Because no more than 5% of former inmates released from TDCJ move from the county to which they were originally released during the first year, only a small proportion of our study cohort who sought HIV care would fail to be captured in the CPCDMS for the follow-up period.
Another potential limitation was that it was difficult to determine the extent to which underlying selection bias might have contributed to the higher rates of outpatient treatment initiation among inmates who received enhanced discharge planning. Although HIV discharge planning coordinators did not target specific clinical or demographic inmate subgroups to receive application assistance, it is possible that unmeasured behavioral characteristics may have resulted in inmates either seeking or being selected for such assistance; however, our analyses showed that all demographic and clinical characteristics were evenly distributed across the two subgroups. In particular, we observed no statistically significant differences in CD4 lymphocyte counts or viral loads across the two discharge planning subgroups. Additionally, our use of multivariate modeling permitted simultaneous adjustment for several potential confounding factors. Finally, our findings are highly dependent on accuracy of data entry into the TDCJ and CPCDMS databases. Although both TDCJ and the Harris County Health District utilize universal and standardized medical screening procedures as well as standardized and validated data entry procedures, it is possible that some patients were misclassified or that some data were entered incorrectly.