Date of Award

Fall 12-2018

Degree Name

Doctor of Philosophy (PhD)



Second Advisor


Third Advisor



Background: HIV viral suppression reduces the likelihood of transmission to just 5%,1 demonstrating the critical link between care and prevention, yet only 51% of people living with HIV (PLWH) were retained in care and 46% achieved viral suppression in the Houston area (2012), with the out-of-care (OOC) population increasing between 2008-2011.2,3

Methods: Record-search investigations across HIV surveillance and care data systems determined if potentially OOC persons referred to the Houston Health Department (HHD) from 2013-2015 (N=1287) qualified for public health follow up. A portion were randomized to a non-intervention group (n=200) to assess the program effectiveness. Participants without a disposition (n=381) were assigned to a service linkage worker (SLW) for assistance with returning to care. Multiple logistic regression assessed (1) differences between follow-up populations (2) associations of persons relinked to care (3) statistical yield of the SLW intervention. Firth’s penalized likelihood approach analyzed rare events where applicable.

Results: The majority of PLWH presumed to be OOC failed to qualify for follow up primarily due to recent evidence of care (n=552, 67.3%) or having moved out of jurisdiction (OOJ) (n=131, 16.0%). Participants referred by Disease Intervention Specialists (DIS) or Texas jurisdictional health departments (incoming OOJ) (aOR:4.057, CI: 2.270-4.250; referent: provider), the Enhanced HIV/AIDS Reporting System (eHARS) surveillance referral source (aOR:2.054, CI:1.590-2.653; referent: provider), and having had an unsuppressed viral load at last report (aOR:1.368, CI:1.058-1.769; referent: viral suppression) had greater odds of qualifying for follow up while persons diagnosed with HIV longer (aOR: 0.968, CI:0.950-0.986) had lower odds. Even after exhausting HHD resources to identify persons for SL outreach (n=381, 35.1%), most persons were still unable to locate (n=157, 41.2%) and few relinked to care (n=31, 8.1%). Of those located (n=193), the majority self-reported already being in care (n=90, 46.6%) or refused SLW services once successfully contacted (n=59, 30.6%). DIS/incoming OOJ referral source (aOR:7.242, CI:2.603-20.343; referent: provider), surveillance referral source (aOR: 2.722, CI: 1.011-7.186; referent: provider), and 7+ client phone calls (aOR: 3.879, CI: 1.359-12.770; referent: two or fewer) were significantly related with returning to care via SLW. DIS/incoming OOJ referral source (aOR=3.489, CI:1.609-7.919; referent: provider) and 7+ client phone calls (aOR=2.341, CI: 1.130-5.003; referent: two or fewer) were associated with greater odds of successful SLW contact while persons with a last reported viral load that was unsuppressed (aOR:0.587, CI: 0.355-0.967; referent: viral suppression) and incrementally higher number of client field visits had lower odds [(three or more, aOR:0.017, CI:0.005-0.048; two, aOR:0.024, CI:0.007-0.065; one, aOR:0.074, CI: 0.022-0.0195) referent: zero]. Participants were about half as likely to return to care by an SLW vs. those who returned to care on their own (aOR: 0.459, CI:0.180-1.098), although results were insignificant (P-value=0.089).

Conclusions: HDs contribute key information about the OOC population, but these systems fall short in timeliness and completeness when it comes to producing the data needed to successfully contact and relink PLWH, requiring extensive resource management. More information is needed about the reasons for falling OOC to better address their specific needs and the dynamics influencing the fluid nature of HIV care.