It is one thing to identify offenders with a high risk of reconviction, another to determine if psychological treatment will help them avoid it.
Various methods of judging the effectiveness of offender intervention programmes are either process (within treatment) measures and/or outcome (on completion of treatment) measures. This paper limits its definition of effectiveness to outcome: frequency and seriousness of reconviction(s) after a non-residential community-based sentence, or release from a community-based residential programme or prison.
The international literature reports hundreds of studies of the effects of intervention programmes on reconviction rates. The past decade and a half has seen a turn-around from the 'nothing works' philosophy of the 1970s and early 80s (Lipton et al 1975, Whitehead and Lab 1989), to recognition that programmes adhering to strict criteria are effective, some very effective.
Convictions have been reduced from 10 to 40 percent in treated offenders, compared to reconviction of non-treated controls (Losel 1993, Gensheimer et al 1986, Gottschalk et al 1987, Izzo & Ross 1994, Andrews et al 1990a,b, Lipsey 1992, Antanowicz & Ross 1994, Garret 1985, Gendreau & Ross 1979, 1987, McLaren 1992, Palmer 1991, Gendreau & Andrews 1990).
(NB: The numerous studies of effectiveness allow use of a technique known as meta-analysis, where one meta-analysis covers many separate studies. Only major meta-analyses are listed in this section, along with relevant individual studies they do not include eg. very recent reports, unpublished data, significant reports pre-dating use of metaanalyses.)
Rates vary depending on factors such as how reconviction rates are defined (see 2.3 above), type of programme, follow-up period, category of offence, offender characteristics, problems treated and programme integrity. The last is very significant, referring to criteria such as adequate time, resources and facilities; qualified, experienced and trained staff; appropriate assessment, selection and matching of offenders to programme; careful programme design and structured; use of cognitive-behavioural principles; consistent delivery of the intervention as designed, and so on. Programmes meeting these criteria have been consistently found to increase effectiveness up to 50 percent. Cognitive-behavioural methods have been shown to meet these criteria well. New Zealand is no exception. Psychological treatment has been shown to reduce reconviction rates by 30 percent, compared to non-treated controls, as well as reducing the frequency and seriousness of offences of those who are reconvicted (Love 1982, Bakker & Riley 1994).
The issue is no longer whether psychological treatment works, but: 'what treatment, by whom, is most effective for this individual, with that specific problem and under which set of circumstances?' (Paul 1979). Research is urgently needed to answer this question. It must address the many variables and their interactions, including personality types, causal attributions, alcohol and substance abuse, skill assets and deficits, learning style, response style, 'modus operandi,’ environmental factors, demographic variables and many others.
Treatment Failure
That psychological treatment based on cognitive-behavioural principles can reduce reconvictions by up to 40 percent is very encouraging. However, about 35 percent of treated people serving community-based sentences and about 38 percent of those released from prison are reconvicted. Treating them has been an inefficient use of resources.
There are two possible responses to this situation. The first is to research ways of identifying those who will not respond to treatment before it is offered them. The second is to develop effective treatments for these people. Given that the target population consists of the most frequent and serious offenders with high probabilities of being reconvicted, the second option is preferable.
Research into effective treatments is best integrated with clinical services delivery, and several strategies are available to help psychologists do this (Barlow et al 1984, Jacobson 1988, Kanfer 1990, Kazdin 1982, Yin 1989, Evans 1998).
The first option (avoiding treatment of those unlikely to respond) is not recommended, but would allow treatment of other offenders falling below the cut-off point, discussed in section 2 above.
Effectiveness and setting
An important issue is whether psychological treatment is equally effective with people in prison and those serving a community-based treatment programme, whether residential or not. The Psychological Service currently provides more treatment hours for the approximately 4,600 people in prison at any time than it does for the approximately 22,000 serving community-based sentences.
The meta-analytic literature evidences slightly greater treatment effectiveness for community-based programmes. At the same time, it unanimously asserts that programmes based on cognitive-behavioural principles are the most effective, and the greater the integrity of the programme the more effective it will be. So although community-based programmes overall are slightly more effective than prison-based programmes, the literature qualifies this conclusion by noting that for programmes based on cognitive-behavioural principles and delivered with a high level of integrity, setting has little influence on effectiveness. Psychological treatment based on cognitive-behavioural principles and delivered by properly qualified and experienced psychologists will have a high level of integrity and is, therefore, unlikely to be affected by the setting in which it is delivered.
Interestingly, a comparison of community-based residential and nonresidential programmes shows effectiveness varies according to offenders’ psychological characteristics. This emphasises the need for research into psychological variables associated with treatment effectiveness, as discussed above.
To some extent, debate about effectiveness of psychological treatment in residential as opposed to non-residential settings is becoming irrelevant. The literature and advancing knowledge of effective relapse prevention techniques suggest treatment should span both settings for optimum effectiveness. Not only should treatment initiated in prison continue after the inmate’s release into the community, but treatment begun in the community should continue in prison if the person reoffends and is imprisoned. The department's Integrated Offender Management Project should facilitate this.
Effectiveness, setting and cost
If treatment based on cognitive-behavioural principles and meeting integrity criteria is equally effective across settings, then efficiency demands that, given a cost difference in delivering treatment to prisons and community-based sentences, the cheaper setting should have priority. More people will be treated with the same amount of money.
There are no satisfactory cost-effectiveness studies that address this issue. There is unlikely to be much difference in the cost of hands-on delivery of psychological treatment to a community-based or prison setting — both require the same personnel and materials. Travel costs may vary a little, but probably not significantly. However, information on consent and adherence to treatment and group therapy is also relevant to cost.
Consent to treatment
Internationally, throughout the 'helping professions,' lack of consent to treatment is a major problem. In intervention programmes as diverse as alcohol treatment centres, psychological and psychiatric clinics, nonalcohol drug programmes and private practice, failure to keep assessment appointments, or, if that appointment is kept, failure to keep the initial treatment appointment, can be as high as 50 percent (Meichenbaum & Turk 1989). Whatever the contributing reasons for this, an important one must undoubtedly be that the person does not want treatment.
There is little information in the literature about consent to treatment in corrections populations. Psychological Service information suggests it is not an insignificant problem. About five percent of people in prison for whom assessment appointments are made, and almost half those serving community-based sentences, fail to keep them. About 20 percent of those assessed likely to benefit from psychological treatment do not go on to treatment.
Maori offenders need specific mention. There is evidence that psychological treatment is as effective for Maori as for Europeans, but proportionately fewer Maori actually undertake it (Bakker & Riley 1994). The Psychological Service is investigating why this is, and its Bi-Cultural Therapy programme should address it.
There are at least four responses to non-consent to treatment. One option is compulsory treatment, which brings up complex ethical, legal and moral issues. The feasibility of compulsory psychological treatment must also be considered, since treatment success depends on clients’ motivation and co-operation. What degree of compulsion should be instigated is a further issue. One view is that the nature of prison itself mitigates against fully informed consent to treatment, so there will always be an element of compulsion. A full discussion of compulsory treatment is beyond the scope of this report. For concise comment, see Rush (1992) and Johnston (1997).
If more people need treatment than the Psychological Service can deliver, one solution might be to ignore those not consenting and provide it for the next on the list, who may be below the cut-off score discussed in section 2. But this option wastes resources.
Another option would be to research psychological variables identifying 'non-consenters,' so they could be sifted out before resources were wasted on missed appointments. But this option would mean those most in need of psychological treatment to avoid further convictions would not receive it.
A better approach would be to use psychological techniques to motivate consent to treatment. There is a growing body of knowledge on techniques to do this, and it could help significantly (Meichenbaum & Turk 1987, Dowd & Milne 1986, Rollnick & Miller 1993, D'Zurilla 1986, Miller 1985).
Adherence to treatment
It is one thing for people to consent to treatment and start attending treatment sessions, another for them to keep on attending. Dropping out of treatment is a major problem in all helping professions, including psychology. The international literature reports from 20 to 50 percent of clients in community-based treatment settings of all kinds drop out of treatment. The figure can be as high as 70 percent for those undertaking psychotherapy (Meichenbaum & Turk 1987).
This is also true for New Zealand generally, and for its corrections population. Sixty percent of those starting treatment will not complete it (Bakker & Riley 1994). As with consent, more drop out from community corrections — 50 percent plus — than from prisons — less than 10 percent.
In general, inmates attend scheduled psychological treatment sessions consistently. This is because staff encourage them and ensure they attend; because others — friends, relatives, family, partners — also encourage them; because inmates perceive attending to be to their advantage eg. in parole decisions; because the unpleasant prison environment encourages inmates to decide not to return, and that treatment might help them avoid it.
People serving community-based sentences do not have these reasons for attending treatment sessions, or at least, not to the same extent. Many attractive activities compete with keeping treatment appointments, which can be difficult, disturbing and demanding. Treatment requires a commitment that many on community-based sentences find hard to make, so appointments are often missed to the point that treatment ceases altogether.
Options for dealing with dropping out resemble those discussed for gaining consent — that is, instigate compulsory treatment, research psychological variables to identify potential drop-outs before treatment begins, or use psychological techniques to minimise dropping out. As with consent, the best option is to improve and use psychological techniques to motivate people to stay in treatment. A growing body of information is available to achieve that (Jennings & Ball 1982, Taylor et al 1983, Worthington 1986, Baekeland & Lundwell 1975, Blackwell 1979, Brownwell et al 1986, Marlatt & George 1984, Marlatt & Gordon 1985, Primakoff et al 1986, Rogalski 1984, Rosenthal & Downs 1985).
Group therapy
Treating people in groups is obviously more efficient than treating them alone. There is also evidence that for some problems and for some psychological techniques, treatment can be delivered as effectively to groups of offenders as it can to individuals.
The prison population tends to be more homogenous than the community corrections population, making it easier to establish psychological therapy groups of individuals with similar offending backgrounds and problems. Even when such groups can be established for those serving community-based sentences, the reasons noted above often mean groups starting out with 10 or 12 members shrink to half a dozen or less after a few weeks. This weakens the group dynamics necessary for effective therapy.
Conclusions on effectiveness, setting and cost
Community-based treatment may cost more than treatment in prison (that is, the cost per completed treatment with an effective outcome: reduction in frequency and seriousness of reconvictions). This is because it is easier to set up and maintain therapy groups in prison than in the community, and because community-based treatments have a higher rate of 'no-shows' and drop-outs.
Costing out the differential between delivering effective psychological treatment to prison and community correction settings, after factoring in 'no shows,’ drop-outs, group therapy and effective outcome, would be a complex exercise. It is a fruitful arena for investigation, though, because of the need to share out resources between the two settings.
Effectiveness and special units
Overseas jurisdictions include prison units for special needs groups such as alcohol and drug abusers, the mentally ill and intellectually disabled, violent and sex offenders, and young offenders. The New Zealand Psychological Service has developed and implemented two special unit programmes for people imprisoned for sex offences against children, and a pilot programme for violent offenders.
These special units are an exception to the rule of targeting psychological treatment resources to offences with high base rates and high probabilities of reconviction. General base rates for sex offences against children are comparatively low — less than three percent of all convictions, excluding traffic offences (Spier 1996), as are reconviction base rates — about 20 percent (Bakker and Riley 1997). The general base rate for violent offences is also comparatively low — about 13 percent of all convictions in 1995, excluding traffic offences (Spiers 1996). But the reconviction base rate is high for some types of violent offences — 60 percent plus (Bakker and Riley 1997).
There is good evidence that psychological treatment delivered to groups of child molesters is very effective (Furby et al 1989, Marshall et al 1991, Marshall 1993, Quinsey et al 1993 Bakker et al 1997). To a lesser extent, the same is true for violent offenders (Novaco 1975, Meichenbaum & Turk 1976, Goldstein & Glick 1995, Raynor & Vanstone 1996, Bakker & Riley 1994). The seriousness of sex and violent offences, and the demonstrated effectiveness of psychological treatment in reducing reconvictions for them, is strong support for the use of psychological treatment in special units.
Special units have other benefits too. They operate as large group therapy programmes, and the efficiency of group treatment and high rates of consent and adherence for prison-based programmes have already been noted in relation to costs. Additionally, the comparatively high numbers treated, and concentration of professional staff working together on common problems, provides opportunities and incentives for refining and developing psychological assessment and treatment techniques. Disseminating this information throughout the Psychological Service benefits the assessment and treatment work of all staff.
The advent of special units raises the often-debated issue of whether psychological treatment should focus on the offence or the offender. New Zealand special units treatment seeks to change a specific offending behaviour, either sex offences against children or violent behaviour. Treating specific offences suggests a degree of offence specialisation in offenders, and assumes that if the offender stops committing this kind of offence, his probability of reconviction is reduced.
There is evidence that some offenders do show a degree of specialisation (Bakker & Riley 1997), but most commit a variety of crimes. This suggests psychological treatment should focus on the person rather than the crime.
The solution lies in thorough psychological assessment prior to and during treatment. This highlights what needs to be addressed by treatment, and the specific offending behaviour that must be directly dealt with. It also shows up what offence-related behaviours, personal characteristics and aspects of the offender's environment must be addressed. So it is not an either/or issue — effective treatment addresses both specific offending behaviour and other offender variables related to it.
