Harzke AJ, Ross MW, Scott DP. Predictors of postrelease primary care utilization among HIV-positive prison inmates: a pilot study

University of Houston, Houston, Texas, United States
AIDS Care (Impact Factor: 1.6). 06/2006; 18(4):290-301. DOI: 10.1080/09540120500161892
Source: PubMed


The primary aims of this exploratory pilot study were (1) to determine the proportion of a sample of HIV-positive inmates utilizing primary care after recent release, and (2) to identify variables associated with utilization of primary care at the time of a post-release interview. Sixty HIV-positive, male and female state prison inmates were interviewed approximately three months prior to release, and 30 were interviewed again between seven and 21 days after release. Variables associated with having utilized primary care at the time of a post-release interview (chi(2) p-values < 0.20) included: taking anti-HIV medications at the time of release, no alcohol use since release, living in the same place as before incarceration and rating of housing situation as 'comfortable' or 'very comfortable'. For exploratory purposes, these variables were entered into a logistic regression model. The model correctly classified 80% of cases overall. Future studies are required to ascertain whether these results would obtain with a statistically adequate sample size.


Available from: Amy Jo Harzke, Mar 23, 2016
Predictors of post-release primary care utilization among HIV-positive
prison inmates: A pilot study
, M. W. ROSS
, & D. P. SCOTT
WHO Center for Health Promotion and Prevention Research, School of Public Health, University of Texas Health Science
Center at Houston, Houston, Texas and
Sage Associates, Inc. Houston, Texas, USA
The primary aims of this exploratory pilot study were (1) to determine the proportion of a sample of HIV-positive inmates
utilizing primary care after recent release, and (2) to identify variables associated with utilization of primary care at the time
of a post-release interview. Sixty HIV-positive, male and female state prison inmates were interviewed approximately three
months prior to release, and 30 were interviewed again between seven and 21 days after release. Variables associated with
having utilized primary care at the time of a post-release interview (x
p -valuesB/0.20) included: taking anti-HIV
medications at the time of release, no alcohol use since release, living in the same place as before incarceration and rating of
housing situation as ‘comfortable’ or ‘very comfortable’. For exploratory purposes, these variables were entered into a
logistic regression model. The model correctly classified 80% of cases overall. Future studies are required to ascertain
whether these results would obtain with a statistically adequate sample size.
The disproportionately high burden of HIV/AIDS
among inmates in US prisons is well documented
(Hammett et al., 2002; Kassira et al., 2001;
Maruschak, 2004; Sabin et al., 2001). The Bureau
of Justice Statistics (BJS) recently reported the
overall prevalence of confirmed AIDS cases in US
prisons (0.48%) in 2002 was nearly 3.5 times the
prevalence in the general population (0.14%). BJS
data indicated 2,053 females and 19,297 males in
state prisons were confirmed HIV-positive in 2002,
representing 3% and 1.9% of female and male
inmates, respectively (Maruschak, 2004).
The number of HIV-positive prisoners released
annually, although not precisely known, may be as
many as half the number of HIV-positive prisoners in
custody on a given day (Hammett et al., 2002). The
immediate post-release period has been anecdotally
described as a critical period for ensuring continuity
of care (Hammett et al., 2001; Rich et al., 2001).
HIV-positive prison inmates are released with 7
days of anti-HIV medications and, in that time frame
often must secure or re-establish reimbursement
sources for primary care and medication. HIV-
positive releasees, like many other inmates, may
face the challenges of meeting basic subsistence
needs (e.g., housing, clothing, food) and resisting
relapse into substance abuse. However, published
studies focusing specifically on HIV-positive relea-
sees are few and provide limited information about
factors potentially affecting post-release continuity of
care (Conklin et al., 1998; Rich et al., 2001; Richie
et al., 2001; Springer et al., 2004; Stephenson et al.,
Two studies have documented that a significant
proportion of HIV-positive prison releasees, who
were treated with antiretroviral therapy and achieved
undetectable viral loads while incarcerated, showed
substantial rebounds on viral load tests when re-
incarcerated (Springer et al., 2004; Stephenson
et al., 2005). These studies suggest HIV-positive
releasees may be vulnerable to discontinuities in
care. These studies, however, did not analytically
explore the potential causes for this vulnerability.
Differences between releasees who experienced viral
rebound versus those who experienced sustained
therapeutic effects were not examined.
Reports from on-going evaluation studies of three
reintegration/transitional assistance programs (Con-
klin et al., 1998; Rich et al., 2001; Richie et al.,
2001) have indicated high rates of post-release
primary care utilization among HIV-positive pro-
gram participants. These studies, however, have
focused primarily on specific program components
as independent variables and have given only limited
Correspondence: Amy Jo Harzke, Centre for Health Promotion and Prevention Research, School of Public Health, University of Houston-
Texas, PO Box 20036, Suite 2570C, Houston, Texas 77030, USA. Tel: /1 (713) 500 9975. Fax: /1 (713) 500 9750. E-mail:
AIDS Care, May 2006; 18(4): 290/301
ISSN 0954-0121 print/ISSN 1360-0451 online # 2006 Taylor & Francis
DOI: 10.1080/09540120500161892
Page 1
analytical attention to factors potentially mediating,
modifying or confounding program effects, such as
participants’ background characteristics (e.g., age,
gender, race/ethnicity) and their specific needs (e.g.,
stage of disease, substance abuse, mental health
issues) and available resources (e.g., health insur-
ance, housing, transportation). These factors have
been associated with various measures of health care
utilization in HIV-positive persons (Andersen et al.,
2000; Cunningham et al., 1999; Heckman et al.,
1998; Hellinger et al., 2004; Shapiro et al., 1999)
and have been shown to influence effects of ancillary
services on entry or retention in primary care among
HIV-positive, non-incarcerated persons (Ashman et
al., 2002; Chan et al., 2002; Conover & Whetten-
Goldstein, 2002; Lo et al., 2002; Messeri et al.,
2002). Additionally, only one of these evaluation
studies included a control group (Richie et al.,
2001). This control group was comprised of HIV-
positive inmates who elected not to enroll in the
program, but the potential effects of volunteer bias
were not addressed. Taken together, these evaluation
studies suggest certain program components might
support continuity of care for some HIV-positive
releasees, but these studies provide little information
about which HIV-positive releasees may be more or
less likely to benefit and who may be at greatest risk.
This purpose of this exploratory pilot study was
to learn about HIV-positive releasees and factors
potentially affecting their post-release health care
utilization. The specific aims were to determine what
proportion of HIV-positive male and female state
prison inmates used primary care after recent release
and to identify variables associated with utilization of
primary care at the time of a post-release interview
/21 days after release). The results of the study
may ultimately assist correctional health care provi-
ders and public health leaders in identifying HIV-
positive releasees at greatest risk for discontinuities
in care and in designing more effective and efficient
transitional assistance programs.
Participants were from two large state prisons (one
for males, one for females) in the Southwestern
USA. The facilities housed only inmates with
sentences of two years or less. Criteria for participa-
tion were as follows: at least 18 years of age, serving a
sentence in one of the study sites, diagnosed HIV-
positive by a physician (prior to or during their
incarceration), English-speaking and willingness to
sign an informed consent form. The sample was
purposefully constructed to include at least 40%
Pre-release data were collected between July 2002
and January 2003. Prison medical staff identified
potential participants. A liaison from the prison
medical system visited each of the facilities (six visits
to male facility, four visits to female facility) and
individually approached every HIV-positive inmate
on the unit on that day about participating in a
pre-release interview. All who were approached gave
consent for participation in the pre-release interview.
Interviews were scheduled within two weeks of
obtaining consent. Interviews were conducted ap-
proximately three months prior to release. The first
author conducted the interviews (60
/90 minutes) in
private offices within the medical areas at the study
When each pre-release interview was completed,
the participant was asked if he or she agreed to be
interviewed again after their release. After obtaining
consent, the interviewer and participant scheduled
a tentative meeting time. The participant was
provided with the interviewer’s contact information
(a cell phone number, specifically designated for the
project). The participant was instructed to confirm
their appointment with the interviewer upon release.
The participant provided phone numbers of places
where they might be staying immediately after
release. If the participant did not contact the
interviewer within the first seven days after release,
the interviewer attempted to contact the participant
by telephone (one to two times per week). If direct
contact was made with a participant within three
weeks, the participant was considered eligible for the
post-release interview until six weeks post-release
and considered lost to follow-up thereafter. The first
author conducted most post-release interviews (45
60 minutes) at the main county HIV clinic and two
community-based agencies serving people with HIV/
AIDS. A small number of post-release interviews
were conducted by telephone because participants
returned or relocated to regions outside the metro-
politan area. Participants received $15 upon com-
pletion of the post-release interview.
Study procedures were approved and monitored
by a university institutional review board and by the
research, evaluation and development unit of the
state department of criminal justice.
Interview schedules
In both interviews, a semi-structured interview
format was utilized that combined both qualitative
and quantitative questions in multiple domains. The
present study was limited to analysis of quantitative
items. The pre-release interview schedule included
125 quantitative items addressing the following
domains: demographic information, socioeconomic
status, social support, criminal history, medical
Primary care utilization among HIV-positive inmates 291
Page 2
history (related to HIV and Hepatitis B or C, where
applicable), medical care, medications, payment
sources, case management/provision of referrals or
information, mental health history and substance
abuse history.
The post-release interview included 80 quantita-
tive items. These items addressed aspects of partici-
pants’ post-release experience with respect to the
following domains: transportation, housing, employ-
ment, income, social support, sexual and sexually
protective behaviors, alcohol or illicit substance use,
sources of assistance (for housing, transportation,
food and other basic needs), medication adherence
and access to medical care, psychiatric care and
other supportive services.
Items in the pre-release interview relevant to
demographics, criminal history and substance abuse
history were adapted from a study of incarcerated
women by Mullen et al. (unpublished, personal
communication). All other items were developed
specifically for this study.
Descriptive statistics were calculated, using means
and standard deviations for interval data and fre-
quencies and proportions for categorical data. To
investigate potential biases associated with differen-
tial attrition, chi-square (x
) tests of independence
were conducted to assess the associations of variables
measured at pre-release with completion of the post-
release interview. Chi-square tests of independence
were also conducted to identify variables measured
at post-release that were associated with having
utilized medical care by the time of the post-release
interview. Post-release interviewees were considered
‘in care’ if they had been seen by a primary HIV-care
provider or if they had an appointment to be seen by
such a provider. Chi-square values obtained for all
/2 tables were adjusted using the Yates’ correction
for continuity (which produces an estimate similar to
Fisher’s exact test). Categories were collapsed for
some variables to avoid inappropriately small cell
sizes (e.g., 20% or more cells with expected values
/5), but only if it made conceptual sense to do so.
Variables with x
p-values of B/0.20 and appropriate
cell sizes were considered noteworthy.
For each variable identified as associated with
primary care utilization at the time of the post-
release interview (x
p-valuesB/0.20 and appropriate
cell sizes), crude odds ratios with 95% confidence
intervals (Fisher’s exact) were calculated for ease of
interpretability and for use in subsequent power
analyses and sample size calculations. These vari-
ables were also simultaneously entered into a logistic
regression model to explore their relative predictive
strength and significance and the overall ability of
the model to predict primary care status (in care
versus not in care). All calculations and analyses
described above were conducted using STATA 8.0.
Post-hoc power analyses and sample size calcula-
tions were conducted for each potential predictor
variable (Power and Sample Size Calculations, ver-
sion 2.1.31; Dupont & Plummer, 1990). Power
analyses assumed a retrospective case-control study
and specified a
/0.05 and Fisher’s exact test. The
ratio of post-release participants not in care to those
in care was assumed to be m , the ratio of controls to
cases. The percentage of those not in care who
reported a characteristic of interest was assumed to
be p
, the probability of exposure among controls.
Sample size calculations specified a
/0.05 and
/0.80 and considered multiple values for odds
Participant background characteristics
Consent was obtained from 66 eligible prison
inmates and 60 completed the pre-release interview.
Six eligible inmates were not interviewed due to
being released prior to the scheduled interview or
being repeatedly unavailable due to their ‘in-house’
work schedule or off-site medical appointments.
Detailed pre-release data are shown in the Appendix,
stratified by interview completion status.
The initial sample was comprised predominantly
of African-Americans (65%) and heterosexuals
(60%), with slightly more men (55%) than women.
A majority indicated less than 12 years of formal
education, incomes of less than US$15,000 annually
and unemployment or inability to work for more
than one year. More than two-thirds (71.7%) did not
live in their own home or apartment at any time
during the year prior to incarceration, and more than
a quarter (26.7%) perceived themselves as homeless.
The initial sample reported high rates of substance
use in the year prior to incarceration (83.3% crack
cocaine, 80% alcohol, 36.7% marijuana use). More
than 80% would spend 12 months or less in state
prison for their current incarceration.
The average duration of HIV infection was 5.7
years (mean
/4.5 years; range/0.08/17 years).
Among those able to report results from their
most recent viral load test (n
/34), the mean value
was 58,270 (SD
/141,610; range/0 /750,000).
Among those able to report results from their
most recent CD4
/ count (n /46), the mean value
was 403.41 (SD
/288.07; range/15/1,134). Two-
thirds (66.7%) reported ever having taken anti-HIV
medications. Excluding those diagnosed during their
concurrent incarceration (n
/9, all males), nearly
40% of those initially interviewed had not received
292 A. J. Harzke et al.
Page 3
regular HIV-related care in the year prior to incar-
ceration (operationalized as getting bloodwork and/
or seeing a physician at least every three months).
About half (48.3%) had public insurance or reim-
bursement for health care. More than a third (35%)
received their HIV-positive diagnosis while in a
correctional facility. Forty-five percent reported co-
infection with Hepatitis C and 45% reported ever
being treated for a psychological or emotional
Thirty participants completed the post-release inter-
view. Five participants were interviewed beyond the
21-day period due to scheduling difficulties. Exclud-
ing these five, post-release interviews were con-
ducted 11 days after release on average (mean
11.5, SD /4.93).
Compared to those who were lost to follow-up,
more post-release participants were: male, lived in
their own house or apartment in the year before
incarceration, had annual household incomes in
categories ranging from $10
/34K, had two or
more ‘close’ family members, had a spouse or main
partner (versus being separated, divorced or wi-
dowed), had CD4
/ counts of less than 200 and
rated prison health care as ‘excellent’ or ‘good’.
More non-completers had taken medications for
psychiatric conditions other than depression, such
as schizophrenia, bipolar disorder and post-trau-
matic stress disorder. No significant differences
were found between post-release interview comple-
ters and non-completers with respect to substance
abuse or criminal history.
Participant post-release characteristics
Most post-release participants relied on family and
friends for post-release assistance. Almost half were
transported by family (n
/11) or friends (n/2) to
their destination city upon release and the remainder
utilized prison-provided transportation. Most stayed
with family members (n
/16) or friends (n /6)
immediately after release, while others stayed in
halfway houses, shelters or motels. Two participants
indicated experiencing one or more nights without a
place to sleep, but nearly a third (n
/9) considered
themselves homeless. None were employed. Twelve
(40%) reported no income source of any kind. The
majority of those with any income source relied on
their spouses or family members (n
/10). Seventeen
participants reported receiving post-release trans-
portation assistance from family or friends and eight
from community-based agencies or clinics. Almost
half (n
/14; 10 in care) reported receipt of other
assistance for meeting basic needs (e.g., clothes,
hygiene packs). Twelve participants (five in care) had
used alcohol, two (in care) had used crack cocaine
and two (not in care) had used marijuana since
Eighteen post-release participants (12 in care)
were taking daily medications (anti-HIV medica-
tions, prophylaxis for opportunistic infections or
medications for other conditions) just prior to
release and received a ten-day supply upon release.
Of these, six reported obtaining more medications
before running out and another six reported they
would obtain medications before running out. All 12
of these participants indicated having some type of
reimbursement source for their medications. The
remaining six participants reported they had run out
of medications for one or more of the following
reasons: attending to basic survival needs, no money
or reimbursement source, not knowing how to get
more medications, hospitalization and stopped tak-
ing due to side effects. The range of missed doses
was from three days to more than two weeks. All 18
participants reported high levels of adherence while
they were taking medications post-release (i.e.,
before running out).
Primary care utilization
Eighteen post-release participants (60%) were in
care at the time of the post-release interview. Of
these, 14 had previously been patients of their post-
release provider. Of those considered in care, nine
had been seen by a provider and nine had an
appointment. Seven were seen by their provider
within the first seven days after release. Similarly,
six of the eight in care participants who saw a social
service provider did so within seven days after
release. The two post-release participants taking
psychiatric medications just prior to release had
accessed both psychiatric care and HIV primary
Variables associated with primary care utilization after
Variables associated with being in care at post-
release interview are shown in Table I. In a logistic
regression model comprised of these four variables,
the medications variable was not significant at the
/0.2 level, no alcohol use was significant at the
/0.1 level and the two housing-related variables
(i.e., living in same place as before incarcerated and
rating of housing situation as ‘very comfortable or
‘comfortable’) were significant at the B
/0.05 level.
The model correctly classified 80% of cases overall
(83.3% of those in care, 75% of those not in care),
with an omnibus x
value of 17.39 (p/ 0.002),
value of 0.44 (maximum likelihood) and
Primary care utilization among HIV-positive inmates 293
Page 4
a two log-likelihood value of 22.99. Exclusion of the
medications variable modestly improved sensitivity
for classifying those not in care (77.8%) but did not
change the percentage of cases correctly classified
Post-hoc power analysis and sample size calculation
Power analyses indicated a power (1
/b) of 0.45,
0.46, 0.53 and 0.69, respectively, for taking anti-
HIV medications at the time of release, no alcohol
use since release, housing comfort and living in the
same place as prior to incarceration. For the latter of
these variables, a sample size of 469 would be
sufficient to capture an odds ratio
/2. This sample
size would be adequate for logistic regression analy-
sis including these four variables.
Excluding newly diagnosed cases, nearly 40% of
those interviewed prior to release (n
/60) had not
received regular HIV-related care in the year prior to
incarceration, reflecting poor rates of primary care
utilization compared to HIV-positive persons more
generally (Andersen et al., 2000; Shapiro et al.,
1999). This is not unexpected, given that the ‘free
world’ lives of pre-release interviewees appeared to
be affected by multiple factors previously shown to
affect health care utilization negatively in non-
incarcerated HIV-positive persons: many pre-release
participants had low levels of income and education,
inadequate or marginal housing and substance abuse
and mental health problems (Andersen et al., 2000;
Cunningham et al., 1999; Shapiro et al., 1999;
Taylor et al., 2004); less than half of pre-release
participants had insurance or reimbursement for
health care (Shapiro et al., 1999); and less than
half reported ever having a case manager or case-
worker to assist in accessing reimbursement sources,
housing assistance, substance abuse treatment and
mental health care (Ashman et al., 2002; Conover &
Whetten-Goldstein, 2002; Lo et al., 2002; Messeri
et al. , 2002).
HIV-positive releasees’ utilization of primary care
after recent release was similarly poor, with only
60% of post-release participants (n
/30) being in
care within 21 days after release. Moreover, our data
may over-estimate the proportion in care. Those lost
to follow-up were more likely to be female, to be
separated, divorced or widowed, to have fewer close
family members, to have taken medication for a
major non-depressive, psychiatric illness and to have
not lived in their own home or apartment in the year
prior to incarceration. Given the prior evidence for
unfavorable patterns of health care utilization among
HIV-positive women (Andersen et al., 2000; Hellin-
ger & Encinosa, 2004; Shapiro et al., 1999) and
HIV-positive persons with inadequate housing (Cun-
ningham et al., 1999) and mental health issues
(Palmer et al., 2003; Taylor et al., 2004; Tucker et
al., 2003), it is likely that the proportion of those
utilizing primary care post-release would have been
less encouraging if those lost to follow-up had been
located and interviewed.
Our results suggest that housing stability, housing
comfort and lack of alcohol use may be especially
strong predictors of post-release primary care
utilization. These findings are commensurate with
Table I. Variables associated with primary care utilization at post-release interview.*
In care
Not in care
n Odds ratio
95% confidence
interval (exact) p (2-sided)
Currently living in same place as before incarcerated
Ye s
8 1 8.80 0.84
/425.15 0.05
10 11
Comfort level in current living situation
Very comfortable/Comfortable
12 4 4 0.68
/25.34 0.13
Less than comfortable to Very uncomfortable
Substance use
No alcohol use since release
Ye s
13 5 3.64 0.61
/22.37 0.14
Transitional medical care
Taking anti-HIV medications at the time of release
Ye s
11 3 4.71 0.76
/34.92 0.07
*Table presents odds ratios and confidence intervals for variables initially identified as associated with post-release primary care utilization
through x
tests (x
p -valuesB/0.20).
294 A. J. Harzke et al.
Page 5
previous research demonstrating the association of
inadequate or marginal housing with poorer health
care utilization (Cunningham et al., 1999) and the
association of housing assistance with entry and
retention in primary care among HIV-positive per-
sons (Conover & Whetten-Goldstein, 2002; Lo et
al., 2002; Messeri et al., 2002). Our findings appear
consistent with previous studies indicating an asso-
ciation of alcohol use with less favorable rates of
primary care utilization and antiretroviral adherence
(Conigliaro et al., 2004; Tucker et al., 2003; 2004).
In a larger sample, the inclusion of these variables in
logistic regression analyses may potentially achieve
correct classification of health care utilization for
80% of those followed-up.
This study had several limitations. The small
sample size limits generalizability of findings and
calls into question the precision and validity of the
analytic results. However, analyses were intended to
be exploratory and power calculations aid in the
interpretation of results. The small sample size also
motivated us to limit assessment of variables asso-
ciated with being in care to those measured at the
post-release interview. However, post-release vari-
ables measured may be more amenable to interven-
tion than many of the background, pre-release
characteristics measured. A larger sample would
permit modeling of both background and post-
release variables, which may be particularly impor-
tant for assessing post-release behaviors possibly
influenced by prior experience (e.g., 14 of the
18 in care had previously been patients of their
post-release provider). Variables were categorized or
re-categorized in some cases to permit selected
statistical analyses and misclassification of categories
is possible. Differential attrition may potentially bias
analyses of post-release data. This was addressed
partially through comparison of those who com-
pleted the post-release interview with those who did
not. The study’s restricted time frame, single post-
release data collection point and single measure of
health care utilization (i.e., primary care), limits
what knowledge may be gleaned about health care
utilization among HIV-positive releasees. Finally, the
validity of self-reported data in incarcerated popula-
tions is unknown.
These limitations not withstanding, this pilot
study contributes to current knowledge about HIV-
positive releasees and the factors that might support
or hinder post-release primary care utilization during
the immediate post-release period. Our findings are
consistent with prior studies suggesting poor health
care utilization in this population and the impact of
housing circumstances and alcohol on health care
utilization among HIV-positive persons. These find-
ings, although preliminary, provide a useful starting
point in the development of a set of indicators for
assessing inmates’ risk of discontinuities in HIV-
related care after release. Larger studies with multi-
ple data collection points and outcome measures are
needed. Such studies hold the promise of aiding
correctional healthcare providers and public health
leaders in understanding HIV-positive releasees’
patterns of health care utilization over time, in
identifying those most vulnerable and, ultimately,
in designing more effective and efficient transitional
assistance programs tailored to the unique needs and
resources of inmates returning to the community
and seeking HIV-related treatment.
This research was supported by the 2002 National
AIDS Foundation Challenge Award and by Health
Resources Services Administration, grant #U69
HA00062-02. Ms Harzke is supported by a Beha-
vioral Science Education Cancer Prevention and
Control grant from the National Cancer Institute/
NIH, #2R25CA57712. The Texas Department of
Criminal Justice supported this research under
Research Agreement #212-RM02. The opinions
expressed herein are solely those of the authors and
do not necessarily represent the position of the Texas
Department of Criminal Justice or other supportive
agencies or institutions.
Andersen, R., Bozette, S., Shapiro, M., St. Clair, P., Crystal, S.,
Goldman, D., et al. (2000). Access of vulnerable groups to
antiretroviral therapy among persons in care for HIV disease in
the United States. Health Services Research , 35 , 389
Ashman, J.J., Conviser, R., & Pounds, M.B. (2002). Associations
between HIV-positive individuals’ receipt of ancillary services
and medical care receipt and retention. AIDS Care , 14 (Suppl.
1), S109
Chan, D., Absher, D., & Sabatier, S. (2002). Recipients in need of
ancillary services and their receipt of HIV medical care in
California. AIDS Care , 14 (Suppl. 1), S73
Conigliaro, J., Madenwald, T., Bryant, K., Braithwaite, S.,
Gordon, A., Fultz, S.L., et al. (2004). The veterans aging
cohort study: Observational studies of alcohol use, abuse and
outcomes among human immunodeficiency virus-infected
veterans. Alcoholism: Clinical and Experimental Research , 28 ,
Conklin, T.J., Lincoln, T., & Flanigan, T.P. (1998). A public
health model to connect correctional health care with commu-
nities. American Journal of Public Health , 88 , 1249
Conover, C.J., & Whetten-Goldstein, K. (2002). The impact of
ancillary services on primary care use and outcomes for HIV/
AIDS patients with public insurance coverage. AIDS Care,
14 (Suppl. 1), S59
Cunningham, W.E., Andersen, R.M., Katz, M.H., Stein, M.D.,
Turner, B.J., Crystal, S., et al. (1999). The impact of
competing subsistence needs and barriers on access to medical
care for persons with human immunodeficiency virus receiving
care in the United States. Medical Care , 37 , 1270
Dupont, W.D., & Plummer, W.D. (1990). Power and sample size
calculations: A review and computer program. Controlled
Clinical Trials , 11 , 116
Primary care utilization among HIV-positive inmates 295
Page 6
Hammett, T.M., Harmon, P., & Rhodes, W. (2002). The burden
of infectious disease among inmates and releasees from correc-
tional facilities, 1997. American Journal of Public Health , 92 ,
Hammet, T., Roberts, C., & Kennedy, S. (2001). Health-related
issues in prisoner re-entry. Crime and Delinquency, 47 , 390
Heckman, T.G., Somlai, A.M., Peters, J., Walker, J., Otto-Salaj,
L., Galdabini, C.A., & Kelly, J.A. (1998). Barriers to care
among persons living with HIV/AIDS in urban and rural areas.
AIDS Care , 10 , 365
Hellinger, F.J., & Encinosa, W.E. (2004). Antiretroviral therapy
and health care utilization: A study of privately insured men
and women with HIV disease. Health Services Research , 39 ,
Kassira, E.N., Bauserman, R.L., Tomoyasu, N., Caldeira, E.,
Swetz, A., & Solomon, L. (2001). HIV and AIDS surveillance
among inmates in Maryland prisons. Journal of Urban Health ,
78 , 256
Lo, W., MacGovern, T., & Bradford, J. (2002). Association of
ancillary services with primary care utilization and retention for
patients with HIV/AIDS. AIDS Care , 14 (Suppl. 1), S45
Maruschak, L.M. (2004). HIV in prisons and jails, 2002. Bureau
of Justice Statistics Bulletin , NCJ 205333 ,1
Messeri, P.A., Abramson, D.M., Aidala, A.A., Lee, F., & Lee, G.
(2002). The impact of ancillary HIV service on engagement in
medical care in New York City. AIDS Care , 14 (Suppl. 1), S15
Palmer, N.B., Salcedo, J., Miller, A.L., Winiarski, M., & Arno, P.
(2003). Psychiatric and social barriers to HIV medication
adherence in a triply-diagnosed methadone population. AIDS
Patient Care and STDs , 17 , 635
Rich, J.D., Holmes, L., Salas, C., Macalino, G., Davis, D.,
Ryczek, J., & Flanigan, T. (2001). Successful linkage of medical
care and community services for HIV-positive offenders being
released from prison. Journal of Urban Health , 78 , 279
Richie, B.E., Freudenberg, N., & Page, J. (2001). Re-integrating
women leaving jail into urban communities: A description
of a model program. Journal of Urban Health , 78 , 290
Sabin, K.M., Frey, R.L., Horsley, R., & Greby, S.M. (2001).
Characteristics and trends of newly identified HIV infections
among incarcerated populations: CDC HIV voluntary counsel-
ing, testing and referral system, 1992
/1998. Journal of Urban
Health , 78 , 241
Shapiro, M.F., Morton, S.C., McCaffrey, D.F., Senterfitt, J.W.,
Fleishman, J.A., Perlman, J.F., et al. (1999). Variations in the
care of HIV-infected adults in the United States: Results from
the HIV cost and services utilization study. Journal of the
American Medical Association , 281 , 2305
Springer, S.A., Pesanti, E., Hodges, J., Macura, T., Doros, G., &
Altice, F.L. (2004). Effectiveness of antiretroviral therapy
among HIV-infected prisoners: Reincarceration and the lack
of sustained benefit after release to the community. Clinical
Infectious Diseases , 38 , 1754
Stephenson, B., Wohl, D., Golin, C.E., Tien, H.C., Stewart, P., &
Kaplan, A.H. (2005). Effect of release from prison and re-
incarceration on the viral loads of HIV-infected individuals.
Public Health Reports , 120 ,84
Taylor, S.L., Burnam, M.A., Sherbourne, C.D., Andersen, R., &
Cunningham, W.E. (2004). The relationship between type of
mental health provider and met and unmet mental health needs
in a nationally representative sample of HIV-positive patients.
Journal of Behavioral Health Services and Research , 31 , 149
Tucker, J.S., Burnam, M.A., Sherbourne, C.D., Kung, F.Y., &
Gifford, A.L. (2003). Substance use and mental health
correlates of non-adherence to antiretroviral medications in a
sample of patients with human immunodeficiency virus infec-
tion. American Journal of Medicine , 114 , 573
Tucker, J.S., Orlando, M., Burnam, M.A., Sherbourne, C.D.,
Kung, K.Y., & Gifford, A.L. (2004). Psychosocial mediators of
antiretroviral non-adherence in HIV-positive adults with sub-
stance use and mental health problems. Health Psychology, 23 ,
296 A. J. Harzke et al.
Page 7
Appendix. Pre-release interview participant background characteristics by interview completion status.
Completed pre-release
interview only (N
n (%)
Completed pre- and
post-release interviews (N/30)
n (%)
14 (46.7) 19 (63.3)
16 (53.3) 11 (36.7)
19 (63.3) 20 (66.7)
9 (30) 8 (26.7)
2 (6.7) 2 (6.7)
/29 years
6 (20) 4 (13.3)
30 /39 years
13 (43.3) 13 (43.3)
/49 years
6 (20) 8 (26.7)
50/ years
5 (16.7) 5 (16.7)
Sexual orientation/preference
Heterosexual/ ‘straight’
17 (56.7) 19 (63.3)
Homosexual/ ‘gay’ or ‘lesbian’
4 (13.3) 4 (13.3)
6 (20) 5 (16.7)
None of the above/Refused
3 (10) 2 (6.7)
Socioeconomic indicators
Less than 12 years formal education
18 (60) 20 (66.7)
High school diploma
12 (40) 10 (20)
Participated in GED classes
18 (60) 14 (46.7)
Participated in vocational training
16 (53.3) 9 (30)
One or more college courses
6 (20) 10 (30)
Employment (in year before incarceration)*
Employed for wages
10 (33.3) 10 (33.3)
Self-employed (legal)
1 (3.3) 4 (13.3)
Self-employed (illegal)
5 (16.7) 2 (6.7)
Unemployed more than one year
15 (50) 15 (50)
Unable to work
6 (20) 7 (23.3)
Income (total household)
Less than $10K annually
14 (46.7) 4 (13.3)
/14,999 K annually
1 (3.3) 11 (36.7)
$15 /24,999 K annually
4 (13.3) 4 (13.3)
/34,999 K annually
1 (3.3) 4 (13.3)
$35 K or more annually
7 (23.3) 3 (10)
Don’t know
3 (10) 4 (13.3)
Housing (in year before incarceration)*
Own house or apartment
5 (16.7) 12 (40)
Someone else’s house or apartment**
15 (50) 17 (56.7)
Hotel/motel/shelter/on the streets
14 (20) 2 (6.7)
Perceive self as homeless
14 (46.7) 2 (6.7)
Perceived quality of housing/living situation
Very comfortable/Comfortable
11 (36.7) 17 (56.7)
Less than comfortable, but okay
11 (36.7) 9 (30)
Not comfortable/Very uncomfortable
8 (26.7) 4 (13.3)
Financial insecurity (inability to pay)
14 (46.7) 13 (43.3)
House payment or rent
10 (33.3) 9 (30)
Utility bills
6 (20) 10 (33.3)
Doctor visits
5 (16.7) 4 (13.3)
Prescribed medication
5 (16.7) 3 (10)
2 (6.7) 1 (3.3)
None of the above
12 (40) 13 (43.3)
Perceived standard of living
Very comfortable/ ‘well-off
0 (0) 0 (0)
Living Comfortably
10 (33.3) 11 (36.7)
Just getting by
15 (50) 13 (43.3)
Nearly poor
0 (0) 2 (6.7)
5 (16.7) 4 (13.3)
Primary care utilization among HIV-positive inmates 297
Page 8
Appendix (Continued )
Completed pre-release
interview only (N
n (%)
Completed pre- and
post-release interviews (N/30)
n (%)
Medical history
Year of HIV diagnosis
8 (26.7) 9 (30)
7 (23.3) 8 (26.7)
10 (33.3) 9 (30)
1986 /1990
4 (13.3) 4 (13.3)
1985 or earlier
1 (3.3) 0 (0)
Received HIV diagnosis in a correctional
7 (23.3) 14 (46.7)
Received HIV diagnosis during current
3 (10) 6 (20)
Received AIDS diagnosis
11 (36.7) 9 (30)
Results of most recent CD4
/ count
Less than 200
3 (10) 8 (26.7)
200 /499
11 (36.7) 9 (30)
500 or more
10 (33.3) 5 (16.7)
Don’t know
6 (20) 8 (26.7)
Results of most recent viral load
100K or more
3 (10) 3 (10)
2 (6.7) 4 (13.3)
5 /19K
4 (13.3) 2 (6.7)
Less than 5K, but not undetectable
2 (6.7) 3 (10)
6 (20) 5 (16.7)
Don’t know
13 (43.3) 13 (43.3)
Received diagnosis of Hepatitis B
2 (6.7) 4 (13.3)
Received diagnosis of Hepatitis C
13 (43.3) 14 (46.7)
Received Hepatitis diagnosis in a
correctional facility
6 (20) 6 (20)
Received regular** medical care in
previous year***
15 (50) 16 (53.3)
Hospitalized in previous year
5 (16.7) 8 (26.7)
Ever taken anti-HIV medications prior
to incarceration
19 (63.3) 21 (70)
Reimbursement source(s) for medical care*
County (i.e. Harris County Gold Card
or other county program)
8 (26.7) 6 (20)
5 (16.7) 5 (16.7)
0 (0) 1 (3.3)
Veterans’ Administration
2 (6.7) 2 (6.7)
Insurance through spouse’s
1 (3.3) 0 (0)
Insurance through own employment
or self-paid insurance
1 (3.3) 2 (6.7)
Out of pocket
0 (0) 1 (3.3)
3 (10) 5 (16.7)
None of the above
10 (33.3) 8 (26.7)
Don’t know
2 (6.7) 1 (3.3)
Case management and referrals
Ever had a case manager or case
15 (50) 12 (40)
Received referrals or other assistance
in accessing reimbursement
7 (23.3) 9 (30)
Received information regarding how
to access care once released
13 (43.3) 7 (23.3)
Perceived quality of health care while incarcerated
4 (13.3) 9 (30)
26 (86.7) 21 (70)
Perceived quality of health
13 (43.3) 13 (43.3)
298 A. J. Harzke et al.
Page 9
Appendix (Continued )
Completed pre-release
interview only (N
n (%)
Completed pre- and
post-release interviews (N/30)
n (%)
17 (56.7) 17 (56.7)
Mental health history
Mental health treatment
Ever treated by psychologist/
psychiatrist for a psychiatric/
emotional problem
15 (50) 12 (40)
Ever received mental health care in a
correctional setting
5 (16.7) 3 (10)
Currently taking psychiatric
5 (16.7) 2 (6.7)
Ever taken medication for a psychologi-
cal or emotional problem
17 (56.7) 15 (50)
71 (23.3) 10 (33.3)
Bipolar disorder
4 (13.3) 3 (10)
Anxiety disorder
5 (16.7) 1 (3.3)
2 (6.7) 1 (3.3)
Other or unable to name diagnosis
6 (20) 4 (13.3)
Perceived current mental health status
11 (3.3) 10 (3.3)
19 (63.3) 20 (66.7)
Criminal history
Number of incarcerations at state jail or prison
16 (53.3) 12 (40)
7 (23.3) 11 (36.7)
4 (13.3) 4 (13.3)
Four or more
3 (10) 3 (10)
Length of current state prison incarceration
/6 months
8 (26.7) 11 (36.7)
/12 months
18 (60) 13 (43.3)
13 /18 months
3 (10) 6 (20)
19 /24 months
0 (0) 0 (0)
25 months or more
1 (3.3) 0 (0)
Total lifetime incarceration
/6 months
5 (16.7) 4 (13.3)
7 /12 months
4 (13.3) 6 (20)
13 /18 months
3 (10) 2 (6.7)
/24 months
4 (13.3) 5 (16.7)
/36 months
6 (20) 6 (20)
/48 months
0 (0) 3 (10)
49 /60 months
0 (0) 0 (0)
61 months or more
8 (26.7) 4 (60)
Charge(s) related to current incarceration
0 (0) 1 (3.3)
Drug-related (list various options)
17 (56.7) 20 (66.7)
3 (10) 2 (6.7)
2 (6.7) 1 (3.3)
Parole or probation violation
3 (10) 4 (13.3)
Burglary, larceny, breaking and
0 (0) 0 (0)
1 (0) 0 (0)
0 (0) 0 (0)
0 (0) 0 (0)
Contempt of court
0 (0) 0 (0)
Disorderly conduct, vagrancy, public
0 (0) 0 (0)
Driving while intoxicated
1 (3.3) 0 (0)
Major driving violations
0 (0) 0 (0)
Theft (including by check, credit card,
fraud, etc.)
2 (6.7) 1 (3.3)
0 (0) 0 (0)
Primary care utilization among HIV-positive inmates 299
Page 10
Appendix (Continued )
Completed pre-release
interview only (N
n (%)
Completed pre- and
post-release interviews (N/30)
n (%)
1 (3.3) 1 (3.3)
Lifetime charges (i.e., ‘ever charged’)
7 (23.3) 7 (23.3)
Drug-related charges
24 (80) 26 (86.7)
7 (23.3) 3 (10)
12 (40) 10 (33.3)
Parole or probation violation
16 (53.3) 18 (60)
Burglary, larceny, breaking and
4 (13.3) 5 (16.7)
5 (16.7) 1 (3.3)
6 (20) 3 (10)
1 (3.3) 0 (0)
0 (0) 0 (0)
Homicide or manslaughter
0 (0) 0 (0)
Contempt of court
3 (10) 1 (0)
Disorderly conduct, vagrancy, public
6 (20) 10 (33.3)
Driving while intoxicated
4 (13.3) 6 (20)
Major driving violations
1 (3.3) 5 (20)
Theft (including by check, credit card,
3 (10) 5 (20)
2 (6.7) 3 (10)
6 (20) 7 (23.3)
Substance abuse history
Substances used in year prior to present incarceration
24 (80) 24 (80)
10 (33.3) 12 (40)
Cocaine (smoke crack)
26 (86.7) 24 (80)
Cocaine (snort)
3 (10) 4 (13.3)
Cocaine (inject)
4 (13.3) 3 (10)
Heroin (snort)
1 (3.3) 0 (0)
Heroin (inject)
2 (6.7) 0 (0)
3 (10) 1 (3.3)
3 (10) 2 (6.7)
3 (10) 1 (3.3)
Pain killers
3 (10) 6 (20)
1 (3.3) 0 (0)
Frequency of use
11 (36.7) 10 (33.3)
/6 times per week
3 (10) 3 (10)
/3 times per week
5 (16.7) 1 (3.3)
Once a week
4 (13.3) 5 (16.7)
/3 times per month
1 (3.3) 1 (3.3)
Once a month
0 (0) 3 (10)
Less than once a month
0 (0) 1 (3.3)
Frequency of use
4 (13.3) 6 (20)
/6 times per week
0 (0) 1 (3.3)
2 /3 times per week
1 (3.3) 0 (0)
Once a week
0 (0) 0 (0)
/3 times per month
2 (6.7) 0 (0)
Once a month
0 (0) 1 (3.3)
Less than once a month
1 (3.3) 4 (13.3)
Frequency of use
Cocaine (smoke crack)
11 (36.7) 10 (33.3)
/6 times per week
5 (16.7) 6 (20)
/3 times per week
3 (10) 2 (6.7)
Once a week
0 (0) 2 (6.7)
300 A. J. Harzke et al.
Page 11
Appendix (Continued )
Completed pre-release
interview only (N
n (%)
Completed pre- and
post-release interviews (N/30)
n (%)
/3 times per month
3 (10) 4 (13.3)
Once a month
1 (3.3) 0 (0)
Less than once a month
2 (6.7) 0 (0)
Ever received substance abuse treatment in
Residential treatment
12 (40) 18 (60)
Halfway house
10 (33.3) 8 (26.7)
Outpatient treatment
6 (20) 11 (36.7)
Correctional facility
14 (46.7) 17 (56.7)
9 (30) 6 (20)
Number of times in residential treatment
5 (16.7) 9 (30)
4 (13.3) 4 (13.3)
Three times
1 (3.3) 4 (13.3)
Four or more times
2 (6.7) 1 (3.3)
Length of stay in residential treatment*
Less than 14 days (or ‘less than two
1 (3.3) 2 (6.7)
/29 days (‘more than two weeks,
less than a month’)
1 (3.3) 5 (16.7)
30 days (‘a month’)
5 (16.7) 7 (23.3)
/59 days (‘more than a month, but
less than two’)
0 (0) 1 (3.3)
60 days (‘two months’)
0 (0) 2 (6.7)
/89 days (‘more than two months,
less than three’)
0 (0) 0 (0)
90 days (‘three months’)
2 (6.7) 4 (13.3)
/119 days (‘more than three
months, less than six’)
1 (3.3) 1 (3.3)
120 days or more (‘six months or
2 (6.7) 4 (13.3)
Unspecified/missing values
2 (6.7) 2 (6.7)
Social support
Primary relationship status
Living with spouse/main partner
8 (26.7) 7 (23.3)
Spouse/main partner, not living
8 (26.7) 3 (10)
2 (6.7) 8 (26.7)
Single/No steady partner
12 (40) 12 (40)
Number of close friends
5 (16.7) 4 (13.3)
6 (20) 4 (13.3)
7 (23.3) 9 (30)
Three or more
12 (40) 13 (43.3)
Number of close family members
8 (26.7) 4 (13.3)
8 (26.7) 5 (16.7)
3 (10) 10 (33.3)
Three or more
11 (36.7) 11 (36.7)
*More than one answer possible.
**‘Regular’ care was operationalized as getting bloodwork and/or seeing a physician at least every three months.
***Nearly two-thirds of those who received care went to a single public HIV clinic.
This cannot be interpreted in a straightforward manner as responses may be contingent upon individual differences in time between
interview and release date.
Frequency of use reported only for most commonly used drugs (e.g., alcohol, marijuana and crack cocaine).
Language in quotation marks indicates actual language used during the interview.
Primary care utilization among HIV-positive inmates 301
Page 12
Page 13
  • Source
    • "For HIV-infected inmates leaving prison, continuity of HIV care is only part of community reentry (Rich et al., 2011). Most returning inmates face numerous reintegration challenges, including financial burdens and disorganized lifestyles , such as homelessness, poverty, mental illness, substance abuse, lack of medical coverage, unemployment, inadequate social support, and physical health issues (Baillargeon et al., 2009; Harzke, Ross, & Scott, 2006; Keuroghlian et al., 2011; Mellins, Kang, Leu, Havens, & Chesney, 2003; Springer, Azar, & Altice, 2011). These challenges often present conflicting demands, and priorities often shift depending on immediate needs. "
    [Show abstract] [Hide abstract] ABSTRACT: HIV risk is disproportionately high among incarcerated individuals. Corrections agencies have been slow to implement evidence-based guidelines and interventions for HIV prevention, testing, and treatment. The emerging field of implementation science focuses on organizational interventions to facilitate adoption and implementation of evidence-based practices. A survey of correctional agency partners from the Criminal Justice Drug Abuse Treatment Studies (CJ-DATS) revealed that HIV policies and practices in prevention, detection, and medical care varied widely, with some corrections agencies and facilities closely matching national guidelines and/or implementing evidence-based interventions. Others, principally attributed to limited resources, had numerous gaps in delivery of best HIV service practices. A brief overview is provided of a new CJ-DATS cooperative research protocol, informed by the survey findings, to test an organization-level intervention to reduce HIV service delivery gaps in corrections.
    Full-text · Article · Oct 2013 · Journal of Correctional Health Care
  • Source
    • "In a smaller study of 30 HIV-infected recently released prisoners, only 18 (60%) were enrolled in HIV primary care 21 days after incarceration. Enrollment in primary care was associated with abstinence from alcohol though this association did not reach statistical significance, likely owing to the small sample size [36]. Other published research on reduced uptake in HIV care after prison release has not assessed the important contribution of SUDs to discontinuous care [22, 23]. "
    [Show abstract] [Hide abstract] ABSTRACT: The criminal justice system bears a disproportionate burden of the HIV epidemic. Continuity of care is critical for HAART-based prevention of HIV-related morbidity and mortality. This paper describes four major challenges to successful management of HIV in the criminal justice system: relapse to substance use, homelessness, mental illness, and loss of medical and social benefits. Each of these areas constitutes a competing priority upon release that demands immediate attention and diverts time, energy, and valuable resources away from engagement in care and adherence to HAART. Numerous gaps exist in scientific knowledge about these issues and potential solutions. In illuminating these knowledge deficits, we present a contemporary research agenda for the management of HIV in correctional systems. Future empirical research should focus on these critical issues in HIV-infected prisoners and releasees while interventional research should incorporate evidence-based solutions into the criminal justice setting.
    Full-text · Article · Jul 2011 · AIDS research and treatment
  • Source
    • "Studies have found that 26 % of HIV-infected inmates reported unprotected sexual activity within 45 days after release (Stephenson et al., 2006). Forty percent of HIV-infected prison releasees had not accessed medical care by six weeks post-release (Harzke, Ross, and Scott, 2006) and decreased CD4 lymphocyte counts and increased plasma HIV RNA levels are common in releasees (Spring, et al., 2004; Stephenson et al., 2005). Depression has been linked to disease progression in HIV-infected individuals, probably through multiple mechanisms, including poor medication adherence and risky sexual behaviors, however depressive symptoms have been found to be linked to disease progression even after controlling for medication adherence and clinical and demographic factors (Boarts, Sledjeski, Bogart, and Delahanty, 2006; Evans et al., 2002; Ickovics et al., 2001; Lima et al., 2007; Olatunji, Mimiaga, O'Cleirigh, and Safren, 2006; Perdue, Hagan, Thiede, and Valleroy, 2003; Reynolds et al., 2004; Vanable, Carey, Blair, and Littlewood, 2006). "
    [Show abstract] [Hide abstract] ABSTRACT: High rates of both HIV and depression are seen in prison populations; depression has been linked to disease progression in HIV, risky behaviors, and medication non-adherence. Despite this, few studies have examined HIV-infected inmates with depression. We therefore conducted an exploratory study of a sample of HIV-infected inmates in North Carolina prisons (N=101) to determine what proportion of this sample screened positive for depression and whether depression was associated with different pre-incarceration characteristics or post-release needs. A high proportion of HIV infected inmates (44.5%) screened positive for depression. Depressed inmates were significantly more likely have low coping self-efficacy scores (180 vs. 214), to report having had resource needs (OR=2.91) prior to incarceration and to anticipate needing income (OR=2.81), housing (OR=4.07), transportation (OR=9.15), and assistance with adherence (OR=8.67) post-release. We conclude by discussion the implications of our findings for prison based care and effective prison release planning for HIV infected inmates.
    Full-text · Dataset · Apr 2010
Show more