Censoring of HIV Viremia Data of Reincarcerated Individuals: A Response to Wohl et al
The recent publication of Wohl et al1 describes a randomized controlled trial of the imPACT intervention, which had the laudable aim of preventing loss of HIV viral suppression in persons leaving prison by facilitating care access after release. The objectives of the imPACT intervention are important, as engagement of vulnerable persons in care will be key to lowering community viral load. Nonetheless, we question the authors' characterization that their comprehensive motivational intervention to prevent the loss of HIV viral suppression failed. The investigators dropped from analysis the experiences of participants who were reincarcerated, 16.5% (63/381) of all subjects. We believe that this exclusion does not replicate real-world trajectories of persons with histories of incarceration and could skew the results of the study.
We strongly agree that, among those not reincarcerated, the large proportion of participants who experienced an increase in viral load following community reentry in both study arms was disappointing with respect to public health goals. Prevention of transmission, via continued viral suppression after release, is critical to treatment as prevention efforts in the community. We believe, however, that the exclusion of participants who were reincarcerated during the course of the study unfairly links maintenance of viral suppression with overcoming sociostructural barriers to staying out of jail or prison. Our primary question concerns the viremia outcomes in those reincarcerated, who comprised half of those dropped from the analysis for study noncompletion (63/125).
Given that presence in the community is necessary for engagement in care at a community site—a theoretical mediator of maintaining viral suppression—we understand how exclusion of those behind bars would seem logical at first glance. We argue, however, that this methodology mixes criminal justice outcomes and medical outcomes. Recidivism is common among persons living with HIV (PLWH) after release from corrections. Our experiences following a cohort of 89 HIV-infected persons upon release from jail in Atlanta, GA, showed that 56.2% returned to a correctional setting at least once in the year after release (Fig. 1). Furthermore, members in our cohort returned to jail up to 8 times in the year after the initial release and spent between none and 91% of their time back in jail (mean, 14.8%).2 Medical care, including HIV care, while incarcerated is a constitutionally guaranteed right.3 Thus, jail could be an unfortunate but stable medical home for some persons, even after an initial release.
An earlier study of predictors of reincarceration among PLWH released from the Texas prison system found that antiretroviral therapy was statistically significantly associated with a decreased risk of reincarceration. In contrast, male sex, African American race, psychiatric disorder, and release on parole were statistically significantly associated with an increased risk of reincarceration.4 Although imPACT study participants were presumably on antiretroviral therapy at the time of their release, baseline characteristics also show that they were majority male, majority Black, had histories of substance use, and one-third had high or very high psychological distress. We assume, but cannot confirm, whether those dropped from the study match these characteristics. These characteristics indicate a nontrivial risk of reincarceration among imPACT participants regardless of their receipt of an intervention with medically aimed outcome objectives.
We advocate disentangling failure to attain a criminal justice objective from failure to reach a medical goal. Efforts should be made to continue the delivery of interventions aimed at care engagement regardless of current residence (jail or nonjail) of study participants. Real strides in reducing recidivism come not from medical interventions but from prison programs, such as adult basic, secondary and postsecondary education, which a recent meta-analysis shows to reduce recidivism up to 43%.