Factors Associated With Receiving Rapid HIV Testing Among Individuals on Probation or Parole

Abstract

Despite the strong correlation between HIV and corrections, testing and prevention efforts have largely been ignored among community corrections populations. The current study is a secondary analysis to compare characteristics of individuals under community corrections supervision who completed rapid HIV testing with those who refused such testing (N = 2,382) in Baltimore, Maryland, and Providence, Rhode Island. Results indicate that the following variables were significantly associated with the receipt of rapid HIV testing: being female (p = .024), Black race (p = .004), homeless (p = .016), early age of crime onset (p = .001), more drug use during the past 90 days (p = .033), and previously tested for hepatitis C virus/hepatitis B virus (p = .024). Such findings make it especially important that individuals under community supervision be linked with services in the community to ensure that HIV testing and health care planning occur simultaneously.

Introduction

Due to the drastic policy changes that began in the early 1970s requiring mandatory minimum sentences for state and federal crimes primarily related to drug offenses, there has been a dramatic surge in the number of individuals incarcerated in the United States over the past few decades (Travis, Western, & Redburn, 2014). A major unintended consequence of these policy changes has been the fiscal costs borne by states to incarcerate large numbers of people. Although the number of inmates in state and federal prisons has decreased slightly since 2008, because of increasing awareness of the economic burden associated with mass incarceration, and changes in laws, sentencing, and enforcement, the number of individuals coming under the supervision of community corrections (probation or parole) has risen significantly (Glaze & Kaeble, 2014; James, 2015). In 2013, approximately 4.8 million adults were under community corrections, equating to approximately 1 in 51 adults in the United States (Glaze & Kaeble, 2014).

Currently, more than 95% of inmates will be released and reenter society, with nearly 80% being released to parole supervision (James, 2015). The number of individuals under community supervision is expected to continue to increase in the coming years, expanding the opportunity for prevention and intervention strategies, especially in the public health realm (Robillard, Brathwaite, Gallito-Zaparaniuk, & Kennedy, 2011).

There is substantial evidence that supports the relationships between injection drug use, HIV, and criminal justice system involvement (Belenko, Langley, Crimmins, & Chaple, 2004; Clark et al., 2013; Inciardi, 2008; Leukefeld et al., 2009; Oser et al., 2006; Westergaard, Spaulding, & Flanigan, 2013). Specifically, the estimated prevalence of HIV infection among individuals incarcerated in the U.S. prison system has ranged from over 2 to 5 times higher than in the general population (Maruschak, 2002, 2009; Rosen, Schoenbach, & Wohl, 2008; Rosen et al., 2009; Westergaard et al., 2013). These statistics translate to one in seven individuals with HIV who will come into contact with the criminal justice system each year (Spaulding et al., 2009). Although HIV testing rates in federal and state prisons are generally high (71%; Centers for Disease Control [CDC], 2009; Maruschak, Berzofsky, & Unangst, 2015), testing rates in jails are not (19%; CDC, 2009).

This gap in testing makes it likely that many individuals who are infected and pass through corrections will not be offered HIV testing (CDC, 2009; Solomon et al., 2014). These individuals may not know their HIV+ status or be aware of their potential for HIV transmission (CDC, 2006). Also, lack of access to health care services (including screening and early diagnosis) contributes to the likelihood that they will receive treatment only once the disease has already progressed to a more advanced stage of HIV infection (Obermeyer & Osborn, 2007; Westergaard et al., 2013). The CDC strongly encourages more frequent HIV testing of drug-dependent persons within the criminal justice system because HIV-infected individuals are less likely to spread HIV once they know their status and they are more likely to seek medical treatment, which lowers the potential for HIV transmission while simultaneously reducing the individual’s risk of HIV-associated morbidity and mortality (Belani et al., 2012; CDC, 2010).

Despite the strong correlation between HIV and corrections, testing and prevention efforts have largely been ignored among community corrections populations (Belenko et al., 2004; Martin, O’Connell, Inciardi, Surratt, & Beard, 2003; Oser et al., 2006; Westergaard et al., 2013), with even less attention focused on their HIV risks and behaviors. Although viewed as a lower HIV risk group than prisoners, those under community corrections sanctions are actually at higher risk of HIV transmission, as they typically live in neighborhoods characterized by high rates of joblessness, drug use, and poverty (Belenko et al., 2004). Research has consistently identified the small window of time upon release as being an exceptionally high-risk period for engaging in not only substance use but also increased and risky sexual behavior, two of the most prominent risk factors for HIV transmission (Green et al., 2013; Leukefeld et al., 2009). Individuals on probation and/or parole have more opportunities to engage in HIV risky behaviors as compared to individuals who are incarcerated in prison or jail, and they are also negatively impacted by many barriers to successful reentry including unemployment (Inciardi, 2008), poverty (Inciardi, 2008), homelessness (Inciardi, 2008), lack of adequate health care (Marauschak et al., 2015; Rich et al., 2015), housing uncertainty, poor retention in substance abuse treatment, returning to neighborhoods marked by drugs and violence, untreated mental illness (Whetten et al., 2005), sharing of injection equipment, unsafe sex (Inciardi, 2008; Rosenbaum, 1997) with multiple and high-risk sex partners (Braithwaite, Stephens, Treadwell, Braithwaite, & Conerly, 2005; Clarke, Stein, Hanna, Sobota, & Rich, 2001; Conklin, Lincoln, & Tuthill, 2000; Margolis et al., 2006), and other related factors, all of which make HIV transmission more likely (Green et al., 2013). Additionally, factors related to being on probation or parole may increase HIV risk behaviors, particularly engagement in behaviors associated with injection drug use. For instance, a fear of being charged with a parole/probation violation or being subject to having their homes or persons searched at any time by law enforcement officers may inhibit people on parole/probation from accessing syringe exchange services (Blankenship, Smoyer, Bray, & Mattocks, 2005; Green et al., 2013).

The current study is a secondary analysis to compare characteristics of individuals under community correction supervision who completed rapid HIV testing with those who refused such testing.

Methods

The parent study and methods were previously described in detail (Gordon, Kinlock, McKenzie, Wilson, & Rich, 2013). In brief, the larger study involved a two-group randomized controlled trial in which 2,382 male and female individuals on probation or parole in Baltimore City, Maryland (n = 1,182) and Providence and Pawtucket, Rhode Island (n = 1,200) were recruited on-site at community supervision offices to complete an assessment and then were randomly assigned to one of two treatment conditions: (1) on-site rapid HIV testing conducted by research staff co-located for the purposes of this study at the probation/parole office or (2) off-site referral for rapid HIV testing at a community health center or HIV testing clinic. Participants were assigned, within gender, to one of the two testing conditions, using a random permutation procedure, with an equal chance of being assigned to either condition. The protocol was approved by the institutional review boards at both research locations and by an external data and safety monitoring board. The study was also approved by the Maryland Department of Public Safety and Correctional Services and the Rhode Island Department of Corrections. This secondary analysis includes all participants completing baseline assessments regardless of acceptance to the condition of testing site.

Outcome Measures

The dependent variable was the receipt of free rapid oral swab HIV (Oraquick) testing (yes or no). Independent variables were identified and organized into five domains collected at baseline as follows:

1. Demographic/background: Self-reported information was collected about the participant’s race (Black or other), age in years, gender, insurance coverage (yes or no), and whether they were homeless (yes or no).

2. Criminal justice: Self-reported measures of legal history included their supervision status (probation, parole, or both), the number of days under supervision during the past 90 days, the number of probation/parole visits during the past 90 days, age of first arrest and first incarceration, lifetime arrest and incarceration, crime severity (violent, nonviolent), criminal activity (number of days self-reported committing crime during the 90 days prior to baseline assessment), and incarceration history.

3. Substance abuse: Drug use was assessed by self-reported drug use during the past 90 days, alcohol use past 90 days, heroin use (ever), IV drug use (ever), and participation in substance abuse treatment (ever).

4. Risky sexual behaviors and self-reported sexual behavior: Assessed by asking if they had sex without a condom and their number of sexual partners during the past 90 days.

5. Previously tested for sexually transmitted infections: Assessed as been tested by a health professional (ever) for HIV (yes or no) and hepatitis B and C (yes or no).

Statistical Analysis

The relationship between the independent variables and the dependent variable, receipt of rapid HIV testing, was determined at a bivariate level for each independent variable using t-test for continuous or χ² for categorical/dichotomous variables. Second, logistic regression analysis was used to determine which independent variables were related to the dependent variable being received HIV testing (yes or no). Project site (i.e., Baltimore or Providence) was not used as a control variable because we wanted to determine the overall impact of the model in the entire sample. Moreover, site differences were presented in the primary outcome paper (Gordon et al., 2013); however, major differences are reported below.

Results

Participant Characteristics

Site differences.

There were significant differences between Baltimore and Providence/Pawtucket on a number of baseline characteristics (all ps < .05 ). Baltimore had a higher percentage of African Americans (87.1% vs. 20.7%); study participants were older (M = 40.6 vs. 36.7), more likely to be injection drug users (IDU) (25.5% vs. 17.5%), had more incarceration days during the past 90 days (M = 21.4 vs. 8.1), more crime days in the past 90 days (M = 6.3 vs. 1.9), a higher percentage were on parole (25.3% vs. 3.6%), were more likely to be covered by health insurance (63.1% vs. 48.1%), and to be homeless (19.1% vs. 10.3%). Providence/Pawtucket had a higher percentage of Hispanic/ Latino individuals (23.7% vs. 1.7%), reported more days of drug use during the past 90 days (M = 19.9 vs. 15.1), and were more likely to be legitimately employed (20.5% vs 11.1%; data not reported in table).

Demographic/background.

Among the 2,382 individuals who participated in the study, 882 (37%) completed the baseline assessment and were tested for HIV. The 1,500 who were not tested included 1,100 (73.3%) who completed baseline assessments but refused randomization (see Gordon et al., 2013, for a further description of the study), 366 (24.4%) who agreed to testing but were randomized to off-site testing and did not followup on the referral, and 34 (2.3%) who agreed to testing and were randomized to on-site testing but did not complete testing. It should be noted that relatively few participants were tested off-site (Baltimore, n = 32, 18.2% vs. Providence, n = 14, 8.1%). As shown in Table 1, participant characteristics associated with getting tested for HIV included being a female (41.1% of females vs. 36.2% of males were tested; p < .05 and being Black (40.1.% of Blacks vs. 33.7% of other race were tested; p < .05). Furthermore, those who were homeless were more likely to be tested compared to those who were not homeless (46.9% vs. 36.4; p < .001; p < .001).

Table 1.

Baseline Characteristics of Individuals on Probationer/Parole Participating in a Study of Rapid HIV Testing in Baltimore Maryland and Providence/Pawtucket Rhode Island.

Variable	Testeda	Not Tested	Sig.
Demographic/background
Age, M (SD)	38.8 (11.6)	37.8 (11.4)	.250
Gender, n (%)b
Male	709/1,962 (36.2)	1,252/1,962 (63.8)	.035
Female	172/419 (41.1)	247/419 (58.9)
Race, n (%)			.001
Black	499/1,245 (40.1)	746/1,245 (59.9)
Other	383/1,137 (33.7)	754/1,137 (66.3)
Insurance			.057
Yes	371/1,175 (31.6)	804/1,175 (68.4)
No	505/1,184 (42.6)	679/1,184 (57.4)
Homeless			.000
Yes	227/484 (46.9)	257/484 (53.1)
No	507/1,392 (36.4)	885/1,392 (63.6)
Criminal justice
Supervision status, n (%)c			.000
Probation	674/1,937 (35)	1,263/1,937 (65.2)
Parole	170/360 (47.2)	190/360 (52.8)
Both	37/83 (44.6)	46/83 (55.4)
Supervision days, M (SD)	531.9 (945.3)	697.0 (1,238.6)	.000
Parole/probation visits, M (SD)	5.0 (6.5)	4.3 (6.5)	.086
Age of first crime, M (SD)	15.4 (7.0)	18.0 (8.7)	.000
Age of first arrest, M (SD)	19.0 (7.2)	19.0 (7.7)	.830
Age of first incarceration, M (SD)	20.1 (8.4)	21.0 (9.5)	.077
Lifetime arrests, M (SD)	14.1 (16.1)	13.6 (15.7)	.488
Lifetime incarcerations, M (SD)	9.0 (10.1)	8.6 (11.4)	.419
Crime severity, M (SD)d	5.5 (1.2)	5.3 (1.3)	.174
Criminal activity, M (SD)e	3.1 (13.8)	3.2 (13.9)	.733
Incarcerated, M (SD)e	15.0 (27.6)	13.4 (26.6)	.088
Substance abuse
Drug use, M (SD)e	19.3 (32.2)	16.6 (30.8)	.002
Alcohol use, M (SD)e	10.1 (20.5)	10.6 (21.4)	.324
Heroin use (ever), n (%)			.000
Yes	434/1,011 (42.9)	577/1,011	(57.1)
No	448/1,371 (32.7)	923/1,371 (67.3)
IV drug use (ever), n (%)			.068
Yes	211/529 (39.8)	318/529 (60.2)
No	671/1,853 (36.2)	1,182/1,853 (63.8)
Substance abuse treatment (ever), M (SD)	3.7 (16.1)	3.6 (9.7)	.625
Risky sex behaviors
Sex without a condom, M (SD)e	21.3 (60.1)	27.8 (96.9)	.011
Sex partners, M (SD)e	2.5 (15.1)	1.6 (6.4)	.001
Previous testing for sexually transmitted infections
HIV, n (%)			.216
Yes	826/2,243 (36.8)	1,417/2,243 (63.2)
No	54/133 (40.6)	79/133 (59.4)
Hepatitis B, n (%)			.043
Yes	468/1,294 0	826/1,294 ()
No	310/758 0	448/758 0
Hepatitis C, n (%)			.019
Yes	621/1,708 (36.4)	1,087/1,708 (63.6)
No	95/378 (25.1)	283/378 (74.9)

Note. N = 2,382. Not all ns will add up to total N for tested versus not tested, as data were missing (not reported).

^aIncludes both testing on-site at probation and parole and off-site at community health centers.

^bTwo participants reported being transgender.

^cTwo participants (I in tested and I in not tested) reported they did not know.

^dCrime Severity Scale is based on a I (nonviolent) to 7 (violent–physical harm) scale.

^eIndicates past 90 days.

Criminal justice.

Supervision status (p < .001) was significantly associated with being tested, with participants on parole compared to probation more likely to get tested (47.2% vs. 35.0%). Number of days currently on supervision (calculated from consent to date supervision started) was also significant (p < .001). Those individuals who were tested had fewer days of supervision compared to those who were not tested (M = 532 vs. M = 697). Finally, younger age of first crime was also significantly associated with being tested (M = 15.4, tested vs. 18.0, not tested; p < .01).

Moreover, because a number of variables were expected to be highly correlated, we examined them in a correlation matrix and found that lifetime arrests and incarcerations were highly correlated (r = .69); therefore, we made a methodological decision to drop arrests from the model, as incarceration could lead to greater risk among this population.

Substance abuse.

Participants were significantly more likely to be tested if they indicated drug use during the past 90 days (M = 19.3, tested vs. M = 16.6, not tested; p < .01) and those who reported lifetime heroin use were more likely to be tested (42.9% vs. 32.7%; p < .001).

Risky sexual behaviors.

Both sex without a condom (M = 21.3, tested vs. M = 27.8, not tested; p < .05) and with multiple sexual partners (M = 2.5, tested vs. M = 1.6, not tested; p < .05) were statistically significant. Participants who had more sexual activity without a condom were less likely to be tested. Moreover, participants with at least 2.5 sexual partners during the past 90 days were more likely to be tested than not tested.

Previous testing for sexually transmitted infections.

Both hepatitis B (p < .05) and hepatitis C (p < .05) were statistically significant. The variables were highly correlated with one another so we created one variable: tested for hepatitis B and C (p < .05), which was included in the logistic regression model.

Logistic Regression

Table 2 presents the Wald statistic, odds ratios (OR), 95% confidence intervals, and significance levels for the independent variables in the logistic regression analysis. Results indicate that the following variables were significantly associated with receiving rapid HIV testing: being female (p = .024), Black race (p < .004), homeless (p = .016), early age of crime onset (p = .001), more drug use during the past 90 days (p = .033), and previously being tested for hepatitis B virus/ hepatitis C virus (HBV/HCV; p = .024).

Table 2.

Results of Logistic Regression Analyses of Those Probationers and Parolees Who Received Rapid HIV Testing.

Variable	Wald	OR	95% CI	p
Gender (female)	5.089	1.501	[1.055, 2.137]	.024
Race (Black)	8.516	.661	[0.500, 0.873]	.004
Homeless (yes)	5.754	.703	[0.527, 0.938]	.016
Supervision status (probation)	0.016	1.044	[0.541, 2.013]	.898
Supervision days	1.425	1.000	[1.000, 1.000]	.233
Age of crime onset	11.283	.965	[0.945, 0.985]	.001
Drug use past 90 days	4.563	1.004	[1.000, 1.008]	.033
Heroin use (yes)	2.867	.801	[0.620, 1.036]	.090
Sex partners past 90 days	1.269	1.006	[0.995, 1.018]	.260
Sex without a condom	0.077	1.000	[0.999, 1.001]	.781
Previously tested for HBV/HCV	5.099	1.418	[1.047.1.919]	.024

Note. OR = odds ratio; HBV/HCV = hepatitis B virus/hepatitis C virus; 95% CI = confidence interval.

Discussion

This secondary analysis indicates that those who complete testing are at high risk of HIV transmission and are more likely to be drug users, to be younger when they committed their first crime, and to have previously been tested for HBV/HCV. Those who are more likely to be tested are Black, female, and homeless. Our findings are similar to what Belenko, Langley, Crimmins, and Chaple (2004) found indicating that probationers and parolees are being tested, possibly because they realize they are at risk. According to our findings, it appears that most high-risk people under community supervision were getting tested even though a relatively low proportion (37%; 882 of 2,382) of the sample got tested. Our results were similar to what Desai and Rosenheck (2004) found in their study of homelessness in that more risk-related variables such as non-White, longer term homelessness, and drug problems were positive predictors of HIV testing.

Furthermore, people reporting to probation/parole offices may be concerned about receiving HIV testing because of confidentiality concerns. For example, in some states, parole/probation officers are required to contact their county’s health department and their client’s physician to report positive HIV results, not only for data collection purposes but also for safety and health reasons (Davis, 2004; New York State, 2012). These parole/probation officers may also play a part in informing partners and spouses of the client’s HIV status if they have not already been informed (Davis, 2004). Nevertheless, findings from the present study indicate that, if given the opportunity, some individuals currently under community corrections supervision are willing to get HIV tested. Moreover, it is noteworthy that a large percentage of individuals refused rapid HIV testing for the following reasons: (1) were tested within the past year, (2) did not have the time to get. tested, (3) had only one sexual partner, and (4) just did not want to be tested. Therefore, it suggests that additional strategies need to be implemented to encourage those people to get tested.

Such findings make it especially important that individuals under community supervision be linked with services in the community to ensure that HIV testing and health care planning occur simultaneously. HIV is a significant health problem in the correctional system, and this makes for ample opportunity to both test inmates for HIV status and potentially treat them. Moreover, given the increased risk of HIV among probationers/parolees, the opportunity should not be missed to test and treat such individuals while they are under community supervision. Our study demonstrates that despite the barriers that may exist, including probation/parole confidentiality concerns and potential violations (whereby probationers/parolees might be concerned about sharing risky behavior activity with officers or seeking treatment for fear of being violated), many individuals under community supervision are willing to get tested while under community supervision (Gordon et al., 2013).

Limitations

There are several limitations to this study. Some participants may have undergone testing off-site and the results were not captured. However, in all locations, we maintained consistent communication with the staff members of each of the community testing sites. Additionally, more individuals passed through the community corrections offices in which we worked than were approached by research staff. This introduced the possibility of selection bias. Also, we offered $20 for interviews only and not for testing, which may have increased the likelihood of testing, although they were not paid for testing. Although participants were randomly assigned to corrections office versus community testing, it cannot be precisely determined whether the overall sample may be representative of the overall community corrections population. Furthermore, data were missing for a number of variables. However, despite these limitations, this secondary analysis provides important descriptive data on characteristics of those under community supervision who are, or who are not, willing to receive rapid HIV testing.

Conclusion

Much has been written about the benefits of HIV testing and linkage to care in prison and jail settings (Solomon et al., 2014; Westergaard et al., 2013); however, the community corrections population is often neglected. Our work provides evidence that HIV testing at parole and probation offices is feasible (Gordon et al., 2013). The results from the current report extend the previously mentioned findings, suggesting that community corrections clients with more severe histories of drug use and criminal activity may be more likely to realize the benefits of HIV testing. Moreover, those populations that are disenfranchised and vulnerable, such as Blacks and the homeless, need to be engaged, tested, and linked to treatment. With that said, we still need to develop interventions and strategies to make sure the large proportion of those who refuse testing and are at high risk for HIV get tested.