Substance Use Trends among Younger vs. Older Homeless Parolees

Abstract

This cross-sectional study of 540 homeless ex-offenders exiting prisons and jails assessed sociodemographic, childhood and drug-related differences. Older ex-offenders from prison were more likely to have been married, come from a two-parent family and used crack while younger ex-offenders were more likely to have used methamphetamine. Older ex-offenders from jail were more likely to be African-American, have children, and report a history of crack and injection drug use, while those younger were more likely to have engaged in binge drinking and be in a gang. Our findings showcase the need to understand unique correlates of younger and older incarcerated populations.

Introduction

In the United States, nearly 2.3 million men and women are in prison or jail⁽¹⁾ and over 5.1 million under community supervision⁽²⁾. In California, over 132,000 are in custody⁽³⁾. While illicit drug use is one of the major contributors to incarceration and homelessness^{(4, 5)}, rates differ depending on age of the offenders.

When compared to the older age groups, those 24 and younger are more likely to be incarcerated as compared to those 45 to 54 years of age⁽⁶⁾. Among the young subgroup in prison, slightly over two-thirds (66.2%) used drugs in the month before their offense, compared with 55% of those 35 to 44 years of age, 47.4% of those 45 to 54 and 19.2% of those 55 years and older⁽⁶⁾. Nearly one-third of those 24 and younger had an active drug involvement compared to 4.2% of those between 45 to 54 years of age in jail⁽⁷⁾.

Differences also exist between the types of drugs used and chronological age. In general, among offenders from prison, 59.8% of those aged 24 and younger versus 47.2% of those 45 to 54 meet the criteria for drug abuse and dependence⁽⁶⁾. In terms of methamphetamine, a higher percentage of use has been found among prisoners aged 24 and younger when compared to those 45 to 54 years (11% vs 7.7%)⁽⁶⁾. For convicted jail inmates, the highest percentage of active drug involvement was for those between 18–24 (29%), followed by those between 35–44 years of age (24.6%)⁽⁷⁾.

Drug use among populations being released from prisons and jails

Upon release from prisons and jails, those on supervised release often continue to have unmanaged drug use issues. In a national study of parolees, about 27% of those who were over 18 were classified as having a drug problem, and for those on probation, 30% were current illicit drug users⁽⁸⁾. Among homeless men on parole (N=157), for which this study is a larger cohort, the most frequently cited drug used was crack (90.5%), followed by cocaine (64.7%), hallucinogens (47.8%), methamphetamine (49.0%), and heroin (38.2%)⁽⁹⁾.

Age-related patterns of types of drugs used among homeless populations

One main outcome of the cyclical pattern of drug use and incarceration is homelessness⁽¹⁰⁾. In Los Angeles, nearly 34% of homeless persons have a substance use problem⁽¹¹⁾; yet, substance use rates differ depending on age. In one study of young homeless adults between 18–23 years of age, the most frequently cited drugs used was marijuana (72%), followed by opiates (13.5%), and LSD/hallucinogens (3.8%)⁽¹²⁾. Among homeless adults in San Francisco (mean age = 44.2), 27% of which had a history of incarceration, 53% had used alcohol in the past month, followed by crack cocaine (27%) and methamphetamine (11%)⁽¹³⁾. Among a sample of homeless adults in Los Angeles (N=664, mean age = 42), 23% used methamphetamine and 35.9% consumed alcohol at least four times a day for the past six months, while 32.9% had been binge drinkers for the past six months⁽¹⁴⁾.

There is a dearth of literature that focuses on age-related differences among homeless men on parole with a history of incarceration. Such information is critical as a new era of healthcare is burgeoning wherein the identification of patterns may provide guidance to providers who may be interested in a more holistic approach to care for patients with physical, mental health issues compounded with substance use. The purpose of this paper is to identify age-related correlates for homeless men on parole who have been discharged from California jails and prisons. Further, during the post-release time period, a scarcity of research addresses challenges for varied age groups who may have low skill levels, lack education, and lack positive social networks. Clarity of these characteristics may guide the development of interventions to address recidivism and drug use problems.

Conceptual Framework

This study was based on the Comprehensive Health Seeking and Coping Paradigm (CHSCP)⁽¹⁵⁾. Components of the framework were developed from the Lazarus and Folkman⁽¹⁶⁾ Stress and Coping Model and the Schlotfeldt⁽¹⁷⁾ Health Seeking Paradigm. The CHSCP has been applied to studies integrating major factors that impact coping and adjustment and subsequent long-term outcomes. Over the years, the CHSCP has guided studies on HIV and hepatitis risk, protective behaviors, and health outcomes among drug-abusing, homeless adults^(18–22) and among ex-offenders in particular⁽²³⁾. Proposed in the paradigm are the socio-demographic, social, psychological and behavioral factors that might relate to age-related differences in this targeted at-risk population. In addition to age as an antecedent variable, other factors may include education, and health history. As mediating variables, social factors may likewise be important. For example, among male parolees, lack of social support may be more prevalent in younger vs. older ex-offenders. Potential correlates of age in this paper are taken from factors that are associated with important health outcomes in the CHSCP.

Methods

Sample and Setting

As part of a larger randomized control trial, this cross-sectional convenience sample of 540 male parolees released from prisons and jails were eligible for the study if they: a) had entered the participating residential drug treatment (RDT) facility within a thirty-day period; b) had a history of drug use prior to their incarceration; c) were 18–60 years of age; and d) identified themselves as homeless on their prison exit form. Exclusion criteria included: 1) monolingual speakers of languages other than English; and 2) persons judged to be cognitively impaired by the research staff. The majority of participants were released from prisons in southern California and were accepted as residents of a RDT program in Los Angeles. In total, 22 parolees were screened out due to failure to meeting eligibility criteria. The sample was enrolled between February 2010 to July 2012.

Procedure

The study was reviewed and approved by the institutional review board of the University of California at Los Angeles (UCLA) and was made known by approved flyers posted at the RDT site. The trained research staff met privately with male parolees interested in learning more about the study, to provide further information and respond to questions. Group sessions were also provided by the staff from time to time. If interest continued, the parolee signed a consent form that was followed by the administration of a brief screening questionnaire to assess eligibility criteria. Parolees who met these eligibility criteria and continued to be interested in enrolling completed another informed consent and a baseline questionnaire was then administered, as well as a detailed locator guide. All respondents who completed the baseline questionnaire received $20 and were then enrolled in the study.

Instruments

Sociodemographic Information

These data were collected by a structured questionnaire assessing age, education, race/ethnicity, marital status, children, physical health and history of cognitive difficulties and gang membership. Childhood history included history of juvenile hall experiences, whether parolees had two parents while growing up and whether they had been abused as children. Health was measured by a single item asking men to rate their general health on a 5-point scale ranging from (1) poor to (5) excellent⁽²⁴⁾. Crime was assessed using a crime chart severity index, which is a modified version of the CJ DATS 464⁽²⁵⁾. The modified instrument was composed of 21 different types of crimes which range from 1 (most severe) to 4 (least severe). Men were asked how many times they had committed each one in their lifetime; severe violent crimes included assault, aggravated assault, battery, kidnapping, hostage taking, homicide, manslaughter, attempted homicide, and violent sex offences were coded as “1.” All other types of crimes were otherwise coded as having no history of violent crime (0).

Drug and Alcohol use Behaviors

A modified version of the Texas Christian University (TCU) Drug History form⁽²⁶⁾ was used to measure drug, and alcohol use behaviors. Through this form, information regarding the use of alcohol and seven drugs was collected. This form allowed us to review the use of these substances (i.e., by injection and/or orally) during the respondents’ lifetime and during the six-month period before the last incarceration. Anglin et al.⁽²⁷⁾ have verified the reliability and validity of this format. The alcohol measure was expanded to include binge drinking.

Mental and Cognitive Health

Depression was measured by asking respondents whether they had ever experienced serious depression for two weeks or more at a time not due to the effects of alcohol or drug use. Cognitive problems were measured by another “yes/no” item asking respondents whether they had ever experienced trouble understanding, concentrating or remembering not due to alcohol or drug use.

Sexual, Physical and Verbal Abuse

Sexual, physical and verbal abuse in childhood was measured by three “yes/no” items asking respondents whether they had been abused as children in those specific ways.

Data Analysis

Associations of age (18–25 vs. 26 or older; split by young adult vs. older) with sociodemographic and background characteristics and substance use lifetime and in the last six months preceding the last incarceration were assessed with chi-square tests or t-tests separately for men discharged from prisons and jails and for the total sample. Variables that were associated with age at the .10 level and characterized at least 5% of the sample were then used in a two-step logistic regression analysis to identify independent correlates of older age. Like the unadjusted analyses, the regression modeling was done for men exiting prisons, those exiting jails and the total sample. In the first step, only sociodemographic and background characteristics were used as predictors.

Variables that were not significant at the .10 level in the resulting logistic regression model were then removed one at a time beginning with the predictor with the highest p value. In the second step, lifetime substance use variables that were related to age in preliminary analysis were added t the model from the first step; if any of the substance use variables were not significant at the .10 level, they were removed one by one, resulting in a final model the demonstrated the effect of adding lifetime substance use measures to the sociodemographic and background variables. Final models in each step were assessed for multicollinearity and model fit was confirmed with the Hosmer-Lemeshow test.

Results

In total, 540 men released from prisons and jails and currently residing in a RDT program were enrolled in the study. The sample was predominantly at least 26 years of age (M = 40.1; range = 19–60.5) with roughly equal numbers discharged from prison and jail (Table 1). Nearly half the sample (46%) was African American, with a greater percentage in the older age cohort. The majority of the sample had never married; however, the older men released from both prisons and jails were more often married and more often reported having had children than their younger counterparts. No difference was found between age and place of incarceration (p > .20, data not shown). About a quarter of the sample reported fair/poor health, about one-third a history of depressive episodes and two-thirds a history of a violent crime. A significantly lower percentage of older age males were ever or currently involved in gangs than younger males.

Table 1.

Sociodemographic Characteristics (n=540)

	Discharged From
	Prison				Jail				Total Sample
	18–25 years (n= 23)		≥ 26 years (n = 237)		18–25 years (n= 34)		≥ 26 years (n = 246)		18–25 years (n= 57)		≥ 26 years (n = 483)
	N	%	N	%	N	%	N	%	N	%	N	%
Race/Ethnicity:								*				**
African American	11	47.8	130	54.9	7	20.6	103	41.9	18	31.6	233	48.2
White	1	4.4	34	14.4	6	17.7	47	19.1	7	12.3	81	16.8
Latino	10	43.5	60	25.3	20	58.8	82	33.3	30	52.6	142	29.4
Other	1	4.4	13	5.5	1	2.9	14	5.7	2	3.5	27	5.6
High School	10	43.5	146	61.6+	18	52.9	136	55.3	28	49.1	282	58.4
Never Married	22	95.7	148	62.7**	33	97.1	156	63.4***	55	96.5	304	63.1***
Children	8	34.8	153	65.1**	9	26.5	166	67.5***	17	29.8	319	66.3***
Fair/Poor Health	6	26.1	57	24.3	8	23.5	67	27.4	14	24.6	124	25.8
Cognitive Problema	6	26.1	91	38.4	9	26.5	99	40.2	15	26.3	190	39.3+
Depressionb	10	43.5	80	33.8	6	17.7	77	31.3	16	28.1	157	32.5
Violent Crime	16	69.6	168	70.9	20	58.8	171	69.5	36	63.2	339	70.2
Gang, Ever	14	60.9	113	47.9	23	67.7	132	53.7	37	64.9	245	50.8*
Gang, Now	8	34.8	34	14.4*	18	52.9	50	20.3***	26	45.6	84	17.4***
Childhood
Two Parents	1	4.4	107	45.2***	10	29.4	107	43.5	11	19.3	214	44.3***
Juvenile Hall	18	78.3	129	54.4*	20	58.8	152	61.8	38	66.7	281	59.2
Sexual Abuse	3	13.0	39	16.5	6	17.7	38	15.5	9	15.8	77	15.9
Physical Abuse	7	30.4	83	35.0	10	29.4	87	35.4	17	29.8	170	35.2
Verbal Abuse	14	60.9	108	45.6	16	47.1	116	47.2	30	52.6	224	46.4

⁺ p < .10 for age difference, chi-square test

^*p< .05 for age difference, chi-square test

^**p <. 01 for age difference, chi-square test

^***p < .001 for age difference, chi-square test

^aEver have trouble understanding, concentrating or remembering not due to substance use

^bEver have serious depression for two weeks or more at a time not due to substance use

In terms of childhood history, a greater percentage of older men than younger men who had been released from prison and in the entire sample reported coming from a two-parent family, while among those discharged from prisons, a greater percentage of older than younger men reported spending time in juvenile hall. While childhood abuse was prevalent across both age groups and incarceration facilities, no significant differences were found.

Bivariate analyses between age and substance use showed a number of significant correlates (Table 2). Younger age was associated with a greater likelihood of binge drinking ever and in the six months preceding the last incarceration, among those coming from jails, as well as the total sample, while younger parolees were more likely than older ones to have used marijuana in the six months before incarceration in both prisons and jails and older persons were more likely to have used crack ever and more recently.

Table 2.

Associations between Substance Use and Age

	Discharged From
	Prison				Jail				Total Sample
	18–25 years (n= 23)		≥ 26 years (n = 237)		18–25 years (n= 34)		≥ 26 years (n = 246)		18–25 years (n= 57)		≥ 26 years (n = n=483)
	N	%	N	%	N	%	N	%	N	%	N	%
Alcohol, ever	23	100.0	216	91.5	32	94.1	223	91.8	55	96.5	439	91.7
Alcohol, 6 monthsa	21	91.3	181	76.4	30	88.2	185	75.2+	51	89.5	366	75.8*
Binge Drinking, Ever	13	56.5	116	49.2	23	67.7	108	44.4*	36	63.2	224	46.8*
Binge Drinking, 6 months	11	47.8	92	38.8	22	64.7	87	35.4***	33	57.9	179	37.1**
Marijuana, Ever	22	95.7	190	80.5+	31	91.2	212	87.2	53	93.0	402	83.9+
Marijuana, 6 months	20	87.0	142	59.9**	30	88.2	154	62.6**	50	87.7	296	61.3***
Crack, Ever	5	21.7	159	67.4***	8	23.5	164	67.5***	13	22.8	323	67.4***
Crack, 6 months	3	13.0	111	46.8**	2	5.9	95	38.6***	5	8.8	206	42.7***
Cocaine, Ever	13	56.5	134	56.8	14	41.2	148	60.9*	27	47.4	282	58.9+
Cocaine, 6 months	10	43.5	66	27.9	10	29.4	52	21.1	20	35.1	118	24.4+
Methamphetamine, Ever	15	65.2	95	40.3*	23	67.7	141	58.0	38	66.7	236	49.3*
Methamphetamine, 6 months	14	60.9	67	28.3***	19	55.9	109	44.3	33	57.9	176	36.4**
Heroin, Ever	4	17.4	77	32.6	8	23.5	104	42.8*	12	21.1	181	37.8*
Heroin, 6 months	2	8.7	45	19.0	3	8.8	46	18.7	5	8.8	91	18.8+
Inhalants, Ever	5	21.7	33	14.0	7	20.6	49	20.2	12	21.1	82	17.1
Inhalants, 6 Months	3	13.0	4	1.7*	3	8.8	4	1.6*	6	10.5	8	1.7**
Hallucinogens, Ever	7	30.4	103	43.6	5	14.1	122	50.2***	12	21.1	225	47.0***
Hallucinogens, 6 months	3	13.0	36	15.2	1	2.9	26	10.6	4	7.0	62	12.8
IDUa Ever	4	17.4	74	31.2	7	20.6	101	41.1*	11	19.3	175	36.2**
IDU, 6 months	3	13.0	52	21.9	4	11.8	59	24.0	7	12.3	111	23.0+

^aUse of cocaine, heroin or methamphetamine by injection

⁺ p < .10 for age differences, chi-square test, or Fisher’s exact test

^*p< .05 for age differences, chi-square test, or Fisher’s exact test

^**p <. 01 for age differences, chi-square test or Fisher exact test

^***p < .001 for age differences, chi-square test, or Fisher’s exact test

Among those discharged from prisons and in the total sample, methamphetamine use was associated with younger age; relatively recent use of inhalants, through less prevalent, showed similar age patterns, regardless of facility. In contrast, for men discharged from jails and the total sample, lifetime use of hallucinogens and injection drug use was related to older age.

Considering only sociodemographic and background characteristics, logistic regression analysis among men discharged from prison indicated that having been married and having a two-parent family in childhood were associated with older age, whereas having been in juvenile hall was associated with younger age. Among men discharged from jail, having been married was again related to older age and African Americans were more likely to be older than men from other racial/ethnic backgrounds. Additionally, strong associations were found between older age and having children and younger age and being a gang member. For the sample as a whole, strong associations were found between older age, having been married, having children and having had a two-parent family in childhood; a somewhat weaker association was found between older age and African American race/ethnicity and a weak association was seen for older age and ever experiencing a cognitive problem. In contrast, being a gang member was associated with younger age.

When lifetime substance use was added to the regression models, crack use was strongly related to older age and methamphetamine use was related to younger age among men discharged from prison. The sociodemographic and background correlates were relatively unchanged except having been in juvenile hall lost significance due to the very strong crack effect. Crack use was also strongly related to older age among men discharged from jail; in this subsample, injection drug use (IDU) was also related to older age and binge drinking was related to younger age. Having been married lost some importance when substance use was considered, but associations between other predictors and age became stronger. For the total sample, crack use was strongly associated with older age; IDU and use of hallucinogens were also associated with older age, while methamphetamine use was associated with younger age. Notably, associations of older age with African American race/ethnicity and cognitive problems were lost when lifetime substance use was controlled.

Discussion

This study focused on correlates of age for drug-involved, homeless men recently released from prisons and jails. In the final regression model, for ex-offenders released from prison, those who were older were more likely to have been married and have come from a two-parent household. It is possible that findings relate to the fact that older ex-offenders had more time to get married and have children compared to those younger. Those older were also more likely to be African American. A strong association has also been found between poverty and incarceration⁽²⁸⁾; particularly among minority populations. As African Americans have been the main minority subgroup in the past, it stands to reason that African Americans are overrepresented among the older men who were incarcerated. Further, as we have fewer younger men in the sample, our lack of significant findings with younger subgroup may have led to these non-significant findings.

It is also noteworthy to point out that the numbers of minor children impacted by parental incarceration are over 8 million⁽²⁹⁾. While RDT programs are not currently resourced to integrate family reunification counseling, the fact that older ex-offenders are more likely to deal with these issues puts the onus of RDT programs to strengthen communication and family reunification challenges. Furthermore, ongoing research is needed to study the implications of parental incarceration and family destabilization on minors.

In our study, older ex-offenders were also more likely than their younger counterparts to have used crack. This can be explained by the fact that crack use was popular in the 1980s and 1990’s; thus, chronologically older individuals are more likely to have been introduced to crack. Older ex-offenders released from jail were also more likely to be African American, have children, and have been an IDU. According to Langan & Levin⁽³⁰⁾, jail inmates were more likely to be older than prison inmates and over half had used alcohol or drugs at the time of the offense. Specifically, 13.6% had used marijuana, followed by cocaine (10.6%)⁽³¹⁾.

In Step 1 of the regression model, a trend was apparent for older men released from jails to have cognitive problems. While our data do not reveal differences in education levels and cognitive problems may not be related to education levels, one possible explanation is that older inmates are more likely than young one to self-disclose sensitive information, such as having cognitive problems. Alternatively, this may reflect the effects of long-term drug and alcohol use on cognition⁽³²⁾.

Our findings also revealed among the jail subgroup, IDU was related to older age, while the younger men released from jail were more likely than their older counterparts from jail to have engaged in binge drinking. As both IDU and binge drinking have significant social and health-related consequences, RDT programs may wish to address these concerns by supplementing the psychological and health offering within their programs. Our findings which are specific for age groups and facilities may provide information which may serve to augment current offerings in RDT programs.

Younger inmates released from prison were more likely than their older prison counterparts to have a history of methamphetamine use. Data reveal that among younger offenders, methamphetamine is a considerable challenge, as nearly 23.5% of state prisoners had ever used methamphetamine and 6.1% had used it at the time of the offense⁽⁶⁾. Younger men who were released from jails were also more likely to have gang involvement as compared with older men exiting jails. For over two decades, a proliferation of youth gangs can be seen in the US; predictors of gang involvement include socioeconomic and neighborhood factors⁽³³⁾ along with crime. The lower rates of gang membership among older subjects are not surprising. Pyrooz et al⁽³⁴⁾ found that for most, gang membership occurs at a relatively young age and only lasts for a few years. Specifically, they found that half to two-thirds of gang initiates remain in gangs for one year, about one in five remain for two to three years, and between 16 percent to 28 percent remain in a gang for over three years.

Currently, RDT programs provide counseling and referral services to those affected by substance-related problems. Services may include individual and group counseling, self-help groups, etcetera. However, strengthening the mental health and physical health of the residents is not a strong focus. Our findings revealing age-related differences in drug use should inform RDT programs of differential focus during counseling that may be drug-specific. These findings may highlight the need for RDT counselors to understand unique correlates of both younger and older jail and prison populations.

Limitations

This study used a cross sectional design and was based in Los Angeles. Findings are further limited due to the convenience sample and thus cannot be generalized to other homeless parolees. Another limitation is the small sample size of younger men coming from prisons and jails. Nevertheless the findings suggest that RDT programs for parolees, as well as other interventions designed to assist these individuals, need to consider both age and type of incarceration.

Conclusions

Our study highlights the heterogeneity of sociodemographic and drug differences for men exiting California prisons and jails. For older ex-offenders, the findings point to intervention needs, including the need to aid in familial reunification during the post-release time period which may encourage a more streamlined transition from institutions to the community. For younger ex-offenders, who are often gang involved, the period of parole and residing in a RDT facility may present an opportunity to learn about healthy, supportive relationships and build positive social support. While family reunification may be outside the resources of a RDT facility, gaining cooperation of community resources that focus on these areas may be worthwhile. In terms of drugs used, for older men, RDT centers should focus on recovery from crack use, while for younger men, it is necessary that they focus on methamphetamine use. Further, nurses may be in a prime position to work with therapeutic communities to focus on pharmacologic and behavioral comprehensive interventions. This may be particularly critical with the Affordable Care Act shift towards treating substance use disorders via mainstream healthcare centers as opposed to drug-focused drug programs.

Table 3.

Logistic Regressions for Correlates of Older Age

	Step 1									Step 2
	Prison			Jail			Total Sample			Prison			Jail			Total Sample
Sociodemographics	Coeff	s.e.	p	Coeff	s.e.	p	Coeff	s.e.	p	Coeff	s.e.	p	Coeff	s.e.	p	Coeff	s.e.	p
African American (vs. other race/ ethnicity)				0.98	0.5	.045	0.81	0.3	.017				1.68	0.6	.004	0.55	0.5	.222
Ever Married	2.66	1.0	.011	2.42	1.1	.021	2.40	0.7	.001	2.85	1.1	.008	2.01	1.1	.061	2.44	0.8	.002
Children				1.54	0.5	.001	1.28	0.3	.001				1.76	0.5	.001	1.06	0.4	.005
Two-Parent Familya	2.92	1.0	.005				1.25	0.4	.001	3.25	1.1	.003				1.65	0.4	.001
Gang, Now				−1.40	0.4	.001	−1.23	0.3	.001				−1.03	0.5	.033	−1.05	0.4	.004
Juvenile Hall	−1.09	0.5	.046							−0.92	0.6	.121
Cognitive Problemb							0.65	0.3	.060							0.54	0.4	.185
Substance Usec
Binge Drinking													−1.55	0.5	.003	−1.14	0.4	.002
Crack										2.16	0.6	.001	1.75	0.5	.001	1.45	0.4	.001
Methamphetamine										−1.38	0.5	.012				−1.24	0.5	.006
Hallucinogens																0.88	0.4	.039
IDUd													1.55	0.6	.008	1.11	0.5	.016

^aIn childhood

^bEver had trouble understanding, concentrating or remembering not due to substance use

^cLifetime

^dInjection use of cocaine, heroin or methamphetamine