Violence and Rehabilitative Process

Prison Substance Misuse Programs and Offender Rehabilitation

Sharon Casey and Andrew Day

School of , Deakin University, Australia

This article considers the nature of the association between substance misuse and criminal
behaviour and how this might inform the development of prison-based substance misuse
treatment programs. The literature on what is known about the effectiveness of prison-based
treatment is reviewed and the implications for correctional practice considered. It is con-
cluded that prison-based substance misuse treatment should be considered a critical compo-
nent of rehabilitation programming and that justice outcomes are likely to be improved when
a number of program features are incorporated.

Key words: substance use; substance misuse; treatment; offender rehabilitation; program
effectiveness.

The use and abuse of illicit drugs by

offenders is a major concern for all of those

who work in criminal justice systems. In

Australia, reviews by Makkai and Payne

(2003) and Johnson (2004) revealed that

approximately half of all prisoners were

likely to meet the criteria for substance use

dependency. The majority of offenders who

acknowledge using drugs also see a direct

connection between their substance use and

their offending (Levinthal, 2011; Liriano &

Ramsay, 2003; Makkai & Payne, 2003;

Correctional Services Commissioner, 2002).

While the debate continues about whether

there is in fact a causal link between drug use

and crime (see below), a recent meta-analysis

conducted by Bennett, Holloway, and

Farrington (2008) showed the extent of the

association between the two, with the odds of

offending between 2.8 and 3.8 times greater

for drug users than non-drug users. Of partic-

ular note was their finding that these odds dif-

fer as a function of drug type: six times

higher for crack; three times higher for her-

oin; and one and a half times higher for mari-

juana use. It is hardly surprising then that

substance misuse has been identified as an

important dynamic risk factor (Andrews &

Bonta, 2010) and that substance misuse treat-

ment programs are now routinely offered to

prisoners across Australia (Heseltine, Sarre,

& Day, 2011).

It has been suggested, however, that the

provision of effective treatment services

to prisoners is constrained by a number of

issues. This includes the financial burden of

providing treatment to the large number

of offenders who present with substance use

problems, the complexity of problems that

are typically associated with substance use

disorders, and the particular challenges that

are commonly associated with successfully

reintegrating this offender group back into

the community (see Visher, Kachnowski, La

Vigne, & Travis, 2004). The aim of this arti-

cle is to consider how prison substance

Correspondence: Andrew Day, School of , Deakin University, Waterfront Campus, Geelong
3220, Victoria, Australia. Email: [email protected]

� 2013 The Australian and New Zealand Association of Psychiatry, and

Psychiatry, and , 2014

Vol. 21, No. 3, 360–369, http://dx.doi.org/10.1080/13218719.2013.818519

mailto:[email protected]

http://dx.doi.org/10.1080/13218719.2013.818519

misuse treatment programs might be designed

if their effectiveness is to be maximised. In

our view, this evidence-based approach

should provide the starting point from which

organisational constraints on service delivery

can be addressed. Before doing so, it is first

important to consider the ways in which sub-

stance misuse can be related to criminal

behaviour, and to outline the different types

of treatment that are available in Australian

prisons.

The Relationship Between Substance

Misuse and Offending

Explanations for the links between substance

misuse and crime generally fall within three

broad theoretical assumptions: first, that a

direct causal relationship exists whereby

illicit substances lead to offending (or, alter-

natively, offending leads to the consumption

of illicit substances); second, that there is an

indirect relationship produced by one or more

shared variables; and, third, that there is no

causal relationship but, rather, a relationship

exists solely on the basis that each behaviour

represents unique aspects of a broader deviant

lifestyle. Perhaps the most recognised theory

describing the direct mechanisms by which

substance misuse leads to crime is the tripar-

tite framework proposed by Goldstein (1985),

which includes the psychopharmacological,

economically compulsive, and the systemic

violence models. The psychopharmacological

model emphasises the physical and psycho-

logical impact of licit and illicit substances

on the individual’s cognitive systems (e.g.,

decreasing inhibitions, increasing aggression,

substance-induced psychoses). Crime, partic-

ularly violent crime, can be a consequence

of the short- or long-term effects of alcohol

or drug use (i.e., intoxication leads to

impaired judgement which, in turn, can result

in antisocial and/or criminal behaviour; see

Hagga
�
rd-Grann, Hallqvist, La

�
ngstr€om, &

M€oller, 2006). Menard and Mihalic (2001)
have also proposed that some offenders use

drugs and/or alcohol to ‘celebrate’ the

successful commission of a crime. According

to the economically compulsive model,

because drug use cannot be supported

through legitimate means, users engage in

criminal activities (particularly acquisitive

crime) to fund their substance use. Finally,

Goldstein’s systemic violence hypothesis

states that, because there are no legal sanc-

tions or guidelines which govern the drug

market (dealing or use), individuals involved

in this enterprise protect themselves and/or

further their interests through violence or

threats of violence.

Theories that imply an indirect causal

relationship between substance use and crime

suggest that the relationship is derived from

the presence of factors shared in the aetiology

of all antisocial behaviours, including sub-

stance misuse and offending (see Dembo,

Williams, Wothke, & Schmeidler, 1994).

Substance misuse and crime therefore emerge

in the same contextual milieu and have the

same antecedent variables (e.g., poor social

support systems, difficulty in school, deviant

peer group membership). Common variables

can be psychological, sociological and/or

environmental factors. Mental illness is

another factor with the potential to influence

the drug–crime relationship. In their research,

Swartz and Lurigio (2007) found that the

relationship between serious mental illness

and crime for non-violent and drug offences

was almost entirely mediated by substance

use; while the mediation was not as strong for

violent crime, substance use was nonetheless

implicated. The third group of theories views

the relationship between crime and drug use

as spurious, the result of co-existing problem-

atic behaviours. For example, ‘lifestyle’ or

‘systemic’ explanations view drug use and

crime as intrinsic elements of a broader devi-

ant lifestyle (see White & Gorman, 2000).

Prison Substance Misuse Treatment

Current approaches to treating substance mis-

use in prisoner populations can be broadly

divided into four categories: harm reduction

Prison Substance Misuse Treatment 361

programs, which seek to enhance awareness

of high-risk behaviours (overdose, blood-

borne infection and other disease transmis-

sion) and the physiological effects of

substance use (including pharmacotherapy);

psycho-educational programs, which seek to

improve the individual’s understanding and

awareness of the link between substance mis-

use and offending and enhance motivation to

enter more intensive programs; therapeutic

programs, which are generally of a higher

intensity and involve participation in group

treatment programs (which consider issues

such as understanding substance use and

offending, developing mechanisms to cope

with cravings and withdrawal, developing

alternative behaviours, managing emotions,

enhancing problem-solving and communica-

tion, and developing relapse prevention

plans); and prison-based therapeutic commu-

nities, the most intensive form of program, in

which participants are separated from the

prison culture and immersed in a dedicated

therapeutic environment.

A review of Australian prison-based pro-

grams that were available in 2009 illustrated

the level of diversity that exists in service

provision across jurisdictions (Heseltine

et al., 2011). Since the publication of

this report other treatment approaches have

been developed, most notably a dedicated

substance misuse treatment prison, the

Compulsory Drug Treatment Correctional

Centre (CDTCC) in New South Wales. In

addition, therapeutic communities also oper-

ate in the ACT (the Solaris therapeutic com-

munity) and in NSW (the Ngara Nura

program at Long Bay prison). The CDTCC

warrants specific consideration as it was

established by an Act of Parliament to ‘target

a hard-core group of offenders with long-

term drug addiction . . . who have failed to
enter or complete other voluntary or court-

based treatment programs’ (Hansard, 2004,

col 8769). The objectives of the legislation

are: (a) to provide a comprehensive program

of compulsory treatment and rehabilitation

under judicial supervision for drug dependent

persons who repeatedly resort to criminal

activity to support that dependency; (b) to

effectively treat those persons for drug depen-

dency, eliminating their illicit drug use while

in the program and reducing the likelihood of

relapse on release; (c) to promote the re-

integration of those persons into the commu-

nity; and (d) to prevent and reduce crime by

reducing those persons’ need to resort to

criminal activity to support their dependency

(Compulsory Drug Treatment Correctional

Centre Bill 2004 s. 106Ba-d). Thus, while the

majority of Australian offenders are not

legally mandated to enter treatment,
1
CDTCC

prisoners are legally obliged to attend offend-

ing behaviour programs.
2

Evidence of Program Effectiveness

To date, there have been five published sys-

tematic reviews which have investigated the

effect of drug treatment programs on criminal

behaviour. Of these, four (Marsch, 1998;

Mitchell, MacKenzie, & Wilson, 2006;

Pearson & Lipton, 1999; Prendergast, Podus,

Chang, & Urada, 2002) were undertaken in

the US; the fifth (Holloway, Bennett, &

Farrington, 2006) was conducted in the UK.

While the Marsch (1998) study focused only

on the effects of methadone maintenance pro-

grams (noting a positive reduction in crime

for 17 of the 24 studies examined), the stron-

gest evidence to emerge from the reviews in

terms of program effectiveness (i.e., reduc-

tions in drug use and/or recidivism) has been

for evaluations of prison therapeutic commu-

nities (TCs). Lipton, Pearson, Cleland, and

Yee (2002) summarised this evidence in their

examination of 35 outcome studies (involving

nearly 1000 participants) which revealed a

significant (albeit modest) effect on recidi-

vism rates for offenders undertaking treat-

ment in TC facilities when compared with

untreated or ‘treatment as usual’ groups. The

most recently published systematic review

and meta-analysis, conducted by Holloway,

Bennett, and Farrington (2008), was based

on evaluations from the United Kingdom

362 S. Casey and A. Day

(n ¼ 3), Europe (n ¼ 1) and the United States
(n ¼ 24) for methadone treatment, TCs, post-
release supervision, and drug courts, and was

primarily concerned with heroin, crack and

cocaine misuse. The overall (combined)

effect was significant and showed a reduction

in re-offending of 29%, although this

was found to vary as a function of treatment

type. Therapeutic communities (60%) again

showed the strongest reductions. The odds

ratio for post-release supervision and metha-

done maintenance reached significance using

fixed effects,
3
but not with random effects,

while drug courts were not found to be

effective.

Although outcomes from the initial evalu-

ations of TCs have been promising, studies

included in the meta-analyses described

above had short follow-up periods (typically

less than three years). The findings from

more recent outcome studies with longer fol-

low-up periods have been more equivocal.

For example, whereas Inciardi, Martin, and

Butzen (2004) found inmates were signifi-

cantly more likely than non-inmates to be

drug- and arrest-free five years post-release,

participants from the Wexler, Melnick,

Lowe, and Peters (1999) study showed more

favourable outcomes at 12 and 24 months

post-release, but no difference at 36 months

(even when treatment group members were

provided with aftercare following release).

This same pattern of convergence was noted

by Prendergast, Hall, Wexler, Melnick, and

Cao (2004): five years post-release there were

no significant differences on any outcome

measures (i.e., re-incarceration, heavy drug

use, employment) for the randomly assigned

treatment and control groups. In the most

recently published evaluation from the United

States, Zhang, Roberts, and McCollister

(2011) examined five-year outcomes (return

to prison rates) for a sample drawn from TC

programs in Californian prisons (25% of

whom received aftercare) and a case-matched

Californian prison comparison group. Twelve

months post-release, 52.7% of all participants

had returned to prison. For those in the TC

group, there was a small non-significant dif-

ference in the return to prison rate between

those who received aftercare (47.5%) and

those who did not (57.1%). At one year, the

return to prison rate for the combined TC

group (54.7%) was marginally higher than

the comparison group (51.9%), and by five

years the rate was almost identical (TC

group ¼ 72.4%; comparison group ¼ 72.2%).
Given that participants returned to prison at

the same rate and were arrested at the same

rate for similar offences, the authors con-

cluded that the findings failed to support the

efficacy of TCs as implemented in

Californian prisons.

For non-residential treatment, the meta-

analysis conducted by Pearson and Lipton

(1999) revealed treatment effectiveness varied

as a function of program type. For example,

whereas group counselling did not result in sta-

tistically significant reductions in recidivism,

cognitive behavioural therapy (CBT) was

shown to produce more encouraging results.

The superior efficacy of CBT-based programs

was later supported in the Lipton et al. (2002)

meta-analysis (involving 68 treatment evalua-

tions and more than 10,000 participants). As

compared to behavioural programs (i.e., those

without the cognitive element; n ¼ 23) which
produced a positive mean effect size of 0.07,

cognitive behavioural programs (i.e., with a

cognitive element; n ¼ 44) produced a positive
mean effect size of 0.14. Looking at specific

programs, an evaluation by the Correctional

Service of Canada of the Offender Substance

Abuse Pre-release Program (OSAPP; Light-

foot, 2001) revealed that one year post-release,

program completers had significantly fewer

violations of their release conditions or new

offences serious enough to warrant readmis-

sion to prison. The strongest effects were noted

for offenders with greater levels of substance

abuse severity (measured by psychometric test

scores) and those with less extensive criminal

histories, with significant reductions in read-

missions for offenders who entered commu-

nity-based programs post-release. Positive

outcomes were also noted on the evaluation of

Prison Substance Misuse Treatment 363

another Canadian program, the High Intensity

Substance Abuse Program (HISAP; Grant,

Kunic, MacPherson, McKeown, & Hansen,

2003) delivered to prisoners with severe addic-

tion disorders. At six month follow-up, HISAP

participants as compared to matched controls

were less likely to be readmitted to custody

(26% vs. 32%) or have their conditional release

revoked as a result of new offences (4% vs.

8%). It was estimated that the release-custody

days saved by HISAP participants remaining

longer in the community resulted in an average

saving of nine days or C$1224 per offender

who participated in the program. Finally,

Hollin and colleagues (Hollin, Palmer,

McGuire, Hounsome, Hatcher, Bilby, & Clark,

2004) evaluated the effectiveness of probation

treatment programs in the United Kingdom

and reported that the reconviction rates of par-

ticipants (n ¼ 457) who completed this pro-
gram was 37% for completers compared to

82% for offenders identified as in need of treat-

ment but who did not start, and 78% for

non-completers of the program. Removing

non-completers from the analysis revealed a

significantly lower rate of reconviction for

treated as compared to non-treated offenders.

Pharmacotherapy is one of the primary

forms of substance abuse treatment for offend-

ing and non-offending populations. The main

forms of substitution therapy are methadone

maintenance treatment and buprenorphine,

both of which have been reported as effective

in reducing opiate and cocaine use in rando-

mised clinical trials (e.g., Marsch, Chutuape

Stephens, Mudric, Strain, Bigelow, & Johnson,

2005; Montoya, Schroeder, Preston, Covi,

Umbricht, Contoreggi, Fudala, Johnson, &

Gorelick,, 2004). The impact on offending is,

however, less promising. A recent systematic

review conducted by Elgi, Pina, Skovbo

Christensen, Aebi, and Killias (2009) high-

lights that while methadone maintenance may

reduce involvement in criminal activity, these

reductions have been noted only during metha-

done maintenance when compared to pre-

treatment levels of offending. Based on

their findings, these authors concluded that

methadone maintenance should not be a pre-

ferred treatment option. Their review also

found that buprenorphine failed to significantly

reduce criminal behaviour, although naltrexone

(an opioid antagonist rather than a substitution

treatment) significantly reduced criminality

more than either behaviour therapy or counsel-

ling. More recently, heroin-assisted treatment

(i.e., a controlled dosage adapted to the user’s

needs) has also been considered as a substitu-

tion therapy. A randomised control study con-

ducted in Germany by L€obmann and Verthein
(2011) examined recidivism outcomes (self-

reported and police data) for heroin-assisted

treatment versus methadone maintenance in a

large sample (n ¼ 1015) of severely dependent
opiate users with extensive criminal histories.

The 12-month prevalence rate one year after

admission into the treatment program revealed

a significant reduction in drug and property

offences for those in the heroin-assisted group

as compared to those on methadone mainte-

nance based on both self-report and police

data. While police data also revealed a signifi-

cant decline in the number of violent and fraud

offences for the heroin-assisted group, this dif-

ference was not significant for self-reported

crime (a finding the authors considered might

be due to social desirability effects, memory

deficits, or low detection rates).

Implications for Practice

The preceding body of evidence outlining

program outcomes does allow some conclu-

sions to be drawn about effective practice;

namely, that substance misuse treatment can

have positive effects on reducing re-

offending. Typically, these programs, partic-

ularly those delivered in prisons, treat sub-

stance users with similar risk profiles and

similar treatment needs. However, there is

sufficient weight in the evidence to support a

claim of heterogeneity in the presentation of

substance users that should be considered

when developing treatment approaches if

optimal effectiveness is to be achieved. This

364 S. Casey and A. Day

idea has been generally accepted in violent

and sexual offender treatment services.

It is, for example, well-established that

higher risk offenders are more likely to be

successfully rehabilitated (Lipsey, 2009).

This is not, however, a principle always

adhered to in Australian treatment programs.

For example, an important aspect of the com-

pulsory treatment order that underpins the

CDTCC (Birgden & Grant, 2010) is that risk

and need are based on a legal rather than an

actuarial assessment of risk. Just under half

of those who enter the CDTCC meet the crite-

ria for inclusion in the high and medium/high

risk category, while an almost equal propor-

tion fall at the lower end at low and low/

medium. As the case management model

advocates higher intensity treatment for

higher risk offenders and less intensive (or no

treatment) for lower risk offenders, the

impact of such intensive treatment for those

at the lower end of the risk severity contin-

uum becomes a critical issue.

Although structured approaches to risk

assessment have been developed to assess the

risk of re-offending in both sexual offenders

(e.g., STABLE 2007, Hanson, Harris, Scott,

& Helmus, 2007; Static 99, Hanson & Thorn-

ton, 2000; Sex Offender Risk Appraisal

Guide [SORAG], Quinsey, Harris, Rice, &

Cormier, 1998), and violent offenders (e.g.,

Violence Risk Appraisal Guide [VRAG],

Quinsey et al., 1998; Violent Offender Risk

Assessment Scale [VORAS], Howells, Watt,

Hall, & Baldwin,1997; Violence Risk Scale

[VRS], Wong & Gordon, 1999–2003), there

are few tools currently available that can reli-

ably predict the likelihood of recidivism for

an offender with a substance misuse issue.

Why this is the case is not clear. It may be a

function of criminological theories of crime

which tend to portray substance use as a

byproduct of offending behaviour rather

than as causal in nature (see above, and

Levinthal, 2011) or the tendency to place sub-

stance use within a medical model where

treatment involves various forms of pharma-

cotherapy. One validation study by Kelly and

Welsh (2008) identifies the Level of Service

Inventory-Revised (LSI-R, Andrews &

Bonta, 1995) as a possible option in this

regard, but while this is one of the best vali-

dated measure for assessing general risk and

need in offender populations (see Andrews &

Bonta, 2000) there have been few studies

which have used this instrument with sub-

stance-using offenders. Moreover, despite

strong evidence for the predictive validity of

the LSI-R for general recidivism, the effect

size for both violent and sexual reoffending is

significantly lower (see Hanson & Morton-

Bourgon, 2009; Ogloff & Davis, 2005;

Simourd, 2004). Nonetheless, Kelly and

Welsh (2008) examined outcomes for 276

offenders who completed an intensive

(1300 hour) in-prison treatment program

over a period of 12 months and who were fol-

lowed up in the community for a period of at

least 14.9 months. Both total LSI-R scores

and the Drug and Alcohol Problem subscale

scores were found to be significant predictors

of the likelihood of re-incarceration.

The specific needs of the individual, par-

ticularly those who fall into the moderate/

high and high categories of the LSI-R, require

more detailed assessment to establish the

functions of the substance misuse for the indi-

vidual and extent to which substance misuse

can be considered to be part of broader anti-

social patterns of behaviour. Jolley and Kerbs

(2010) have recently described how this

might be achieved for in-prison substance

programs, in a model that is informed by the

principles of effective intervention and

framed within a model of service delivery

based on an enhanced Risk-Needs-

Responsivity perspective. The model shows

how the offender moves from an initial

(standardised) risk assessment which exam-

ines his or her propensity for future relapse

and recidivism; this assessment would serve

to place offenders into programs at an appro-

priate level of service intensity (i.e., service

structure, dosage, duration, services). After

administration of the risk assessment tool

(e.g., the LSI-R), offenders are assessed for

Prison Substance Misuse Treatment 365

specific criminogenic needs (i.e., factors that

drive their propensity for criminal activity).

Following risk and needs assessment,

offenders are provided with evidence-based

programs driven by organisational demands,

with the provision of any additional programs

driven by an offender’s needs, which would

include their learning styles, motivations for

treatment, personality traits, strengths, co-

occurring bio-psycho-social needs, and their

needs based on stage-based models of recov-

ery. These processes are tracked by a range

of program integrity indicators and those that

assess the extent to which treatment is

matched to the offender’s specific needs.

A meta-analytic review conducted by

Dowden and Brown (2002), which examined

the extent to which substance misuse factors

predict recidivism, found that combined alco-

hol and drug problems were the strongest pre-

dictors of recidivism (followed by drug

misuse, parental substance misuse and alco-

hol misuse). What this suggests is that not

only should substance misuse be regarded as

a criminogenic need, but also that some types

of substance misuse are more criminogenic

(or higher risk) than others and that the

strength of the drugs–crime connection varies

by type of drugs used. This claim is supported

by the associations noted by Bennett et al.

(2008) described above (i.e., the higher rates

for both crack and heroin users). Research

such as this has two implications for practice:

first, better outcomes are more likely to be

achieved by attending to those categories of

drug users who have the greatest odds of

offending; and, second, that offending behav-

iour focused treatment should target the spe-

cific type of substance that is implicated.

The implications of this review for cor-

rectional practice can be summarised in the

following way:

� Substance misuse treatment can have
positive effects on reducing re-offending

and, given the high prevalence of

substance misuse in correctional popula-

tions, should be considered to be a

core component of any rehabilitation

strategy.
� Prison therapeutic community models

of substance misuse treatment can be

considered to be the treatment of

choice, although there is also a need

for group-delivered cognitive behav-

ioural programs and pharmacological

treatments.
� The most significant reductions in re-

offending are likely to occur under the

following conditions:
� when efforts are made to establish

the nature of the relationship

between the misuse of licit and/or

illicit substances and criminal

behaviour, and this information is

used to guide program content;
� when risk of re-offending is for-

mally assessed and programs are

targeted towards higher risk

offenders;
� when pharmacological (substitu-

tion) treatment is offered as an

adjunctive treatment and sub-

stance misuse treatment is fol-

lowed up by intensive post-

release support and supervision

services; and
� when program content takes

account of how the strength of the

drugs–crime connection varies by

the type of drug(s) used.

Although features of these recommenda-

tions are present in each of the approaches to

program delivery that have been developed in

Australian jurisdictions, they are unlikely to

all be present in each state or territory. There

is also a need to improve the quality of

assessment processes that are currently used

including, for example, the development and

validation of specific risk assessment meas-

ures to: (a) predict substance misuse related

to re-offending; and (b) formally assess the

functions of substance misuse and how these

are related to the offending behaviour.

Finally, there may be a need to develop

366 S. Casey and A. Day

substance-specific treatment program mod-

ules. The conclusion of this paper is that sub-

stance misuse treatment should be given a

high level of priority if the rehabilitation

goals of correctional agencies are to be

realised.

Notes

1. Although most experience some level of for-
mal or informal coercion to participate in
rehabilitation programs (e.g., parole decisions
are often dependent on the offender having
successfully completed a rehabilitation
program).

2. The sentencing court makes a referral to the
relevant Drug Court if the offender meets five
‘eligibility’ criteria: (1) sentenced to impris-
onment with an unexpired non-parole period
of 18 months to three years; (2) convicted of
at least two offences in the previous five years;
(3) not convicted of specified offences such as
drug trafficking, sexual assault, and murder;
(4) reside in the broader Sydney region; and
(5) be over the age of 18.

3. In a fixed effect (FE), each effect size is
weighted by the inverse of its variance
(1/VAR) and an assumption is made that the
observed effect size falls within the range of
normal sampling error. For a random effects
(RE) model, each effect size is weighted by
the inverse of its variance, plus an additional
factor; the assumption is that the variance
associated with each effect size is based on
sampling error and a second component that
reflects random variations between the stud-
ies, such as differences in procedures and
settings.

References

Andrews, D.A., & Bonta, J. (1995). The level of
service …