Scottish Government

Chapter 2: Methodology

Introduction

A combination of quantitative and qualitative research methods was employed in this study. This included the analysis of Transitional Care monitoring data; surveys of prisoners four and seven months following release; in-depth interviews with ex-prisoners in three case study areas; and interviews with prison and community based staff associated with Transitional Care.

Analysis of monitoring data

Monitoring data collected by Cranstoun between October 2002 and April 2004 was analysed to obtain a profile of identified needs of ex-prisoners, their engagement with Transitional Care services and the extent to which links were being made with relevant community-based services. The community-based Transitional Care workers recorded the monitoring data in Transitional Care Action Plans and Logs that were returned to Cranstoun Drug Service Head Office to be entered into an Access database 12 weeks after ex-prisoners' release. The log was completed for those clients who were assessed by Cranstoun. Clients were assessed at either their case conference prior to release from prison, or at their first appointment with Transitional Care on release. This meant that the log was completed for both those clients who attended an appointment with a Transitional Care worker, as well as for those who did not attend Transitional Care on their release from prison.

Between October 2002 and April 2004 information on clients' living arrangements, employment status, ethnicity, local authority area, penal establishment and their needs in relation to six key domains was collected by Transitional Care staff via the Action Plan and monitoring log. The domains of need initially identified were:

• health generally and in relation to drug and alcohol use
• housing
• benefits and finance
• education and training
• employment
• social issues in relation to family members, children and diversionary activity programmes.

However, since Transitional Care was an evolving service its focus was subsequently revised to prevent overlap with and duplication of other services. The monitoring log was consequently amended to take account of the reduced number of domains that provided the focus for Transitional Care.

The electronic formatting of the log proved problematic in conversion to a statistical analysis software package ( SPSS). This resulted in a number of incomplete cases being omitted from the final data set. The resultant sample consisted of 4794 cases, 292 of whom were repeat attenders. However, there appeared to be inconsistencies in how the logs were completed by different sub-contracted agencies (see also Cranstoun Drug Services, 2003), making some data difficult to interpret. Moreover changes to the design of the monitoring log over time, reflecting its adaptation to the evolving service, meant that some data were available only for a sub-sample of cases. For example, whether or not clients had received a case conference while in prison was only recorded from October 2003. Further significant revisions were introduced from May 2004. The action plan and monitoring log became more concise with a number of sections omitted, including the key domains on benefits and finance, education and training, employment and social issues ⁵. Because this produced a distinctively different data set, analysis has concentrated upon the data collected in the 'first phase' of Transitional Care until April 2004.

The design of the log did not allow a breakdown of needs in relation to the first, second, third or fourth appointment, preventing any assessment of whether a client had the same or different needs over the course of appointments. It should also be noted in this report that the term 'action achieved ⁶' is used in relation to the needs of ex-prisoners and the extent to which they have been met. This means only that an appointment to attend an existing community service provider had been given to the client within the 12-week time frame. It does not necessarily mean that the client had met a representative from that service during that 12-week period following release nor that they would be actively supported by that service.

Analysis of changes in Christo scores

The Transitional Care action plan and log also recorded the client's Christo Inventory scores at each appointment. The Christo Inventory is an audit/evaluation tool completed by the Transitional Care workers that gives a professional indication of a client's drug and/or alcohol use through assigning scores in relation to involvement with crime, drug/alcohol use, general health, psychological well being, living situation, occupational activities and support. The lower the score the less severe the client's problems would appear to be in relation to their drug and/or alcohol use. For evaluative purposes the analysis of Christo Inventory scores has to be limited to those clients who attend at least two appointments. In the present study, Christo score were compared over subsequent appointments for those who attended three Transitional Care appointments. The relevant data were available in respect of 292 ex-prisoners.

Professional interviews

Semi-structured interviews were conducted in spring 2003 with thirty-seven staff involved in the Transitional Care initiative. The interviews explored respondents' views about the operation of Transitional Care in its first 12-18 months with a view, partly, to identifying practical issues that might need to be addressed to improve the effectiveness of the service. For this reason interviews with staff not only elicited data pertaining directly to the research aims but also elicited data on issues that were considered important by the staff involved. The sample consisted of:

Ten case workers employed by Cranstoun Drug services and responsible for implementing Transitional Care assessments for all short term and remand prisoners identified as having substance misuse issues. These caseworkers were selected from five prison establishments.

Fourteen Transitional Care workers employed by sub-contracted agencies: APEX, SACRO, Drugs Action, Dundee Cyrenians, PARC, Realise Community Care Project, Molendinar Drug Service and Cumbernauld and Lanarkshire Counselling services. These workers were responsible for implementing the Transitional Care Plan for clients within a 12-week period after liberation. One or two workers from each agency were included.

Seven Managers of the sub-contracted agencies.

Five senior personnel from Cranstoun Drug Services and the Scottish Prison Service Addiction Services employed to operationalise Transitional Care.

The interviews, mostly lasting between 30 and 70 minutes, were tape recorded and fully transcribed. Prior to analysis all interviews were coded and made non-identifiable to protect the identity of respondents.

In addition, information was sought, in autumn 2002, from each of the local Transitional Care schemes about organisational arrangements and service provision. Twenty forms designed to elicit relevant information were completed: 6 by telephone interview with a researcher and 14 by workers themselves.

Survey of ex-prisoners following release

A survey of prisoners at four and seven months post-release was undertaken by TNS Social Research ⁷. The cohort of prisoners was recruited from those leaving prison over a period of approximately 15 months from November/December 2002 to the end of March 2004. Fieldwork was completed for the four-month survey between March 2003 and August 2004, and for the seven-month survey between July 2003 and December 2004.

Recruitment

While in prison, at the point at which they were offered Transitional Care, prisoners were asked whether they would also be willing to take part in the research. Those consenting signed a consent form and were asked to provide contact details of where they could be contacted post-release. The difficulties of maintaining contact with a 'chaotic' sample of substance misusers and offenders are well known. For this reason, in addition to the address and phone number of the place the prisoner thought they would be staying post-release, we asked for one further address and two further telephone numbers of relatives or friends through whom they might be contactable. However, the quality and amount of contact details proved to be variable. Most prisoners provided one address and telephone number at best. As an incentive to take part and to thank them for their time, respondents who completed an interview or questionnaire were sent a £10 postal order.

Four-month quantitative interviews

Around 16 weeks post-release, prisoners were contacted by trained interviewers from TNS Social Research. Where possible, interviews were conducted over the telephone, but where no telephone number was available or telephone contact could not be made for some other reason, attempts were made to conduct the interview face-to-face in the respondent's home. In most cases, where an interview was not obtained, it was because no contact was made.

The interview was designed to take around 20-25 minutes to complete. However, in most cases, interviewers found that it took around 30-40 minutes because respondents were keen to talk and to provide additional details of their experiences and difficulties since leaving prison. The main topics covered were:

• experience and perceptions of the Transitional Care service
• needs on leaving prison (housing, education/training/employment, benefits or money, health/substance use, issues to do with partner/children/family)
• current health
• substance use since leaving prison
• offending behaviour since leaving prison
• current accommodation and economic activity.

The questionnaire was designed by the research team and adapted slightly following two series of pilot interviews. Most of the questions were closed (with a small number of 'other'/open-ended questions). Questions on health symptoms, substance use and offending were adapted from the Maudsley Addiction Profile instrument for treatment outcome research. ⁸

Seven-month quantitative interviews

At around seven months post-release, prisoners were sent a self-completion questionnaire. The questionnaire was sent to all contact addresses available for the prisoner - obtained from the original consent form or from additional contact information obtained at the four-month interview stage. The covering letter made clear that only one questionnaire need be completed and any duplicates received were discarded by the researchers.

The questionnaire was designed to provide data on outcomes comparable with that obtained at the four-month stage. It covered:

• current needs (housing, education/training/employment, benefits or money, health/substance use, issues to do with partner/children/family)
• current health
• current substance use
• current offending behaviour
• current accommodation and economic activity.

The questionnaire was essentially a slightly shorter and simplified version of the four-month questionnaire. As with the four-month questionnaire, the seven-month questionnaire was designed by the research team and adapted slightly following a series of pilot interviews.

Prisoners were sent seven-month questionnaires regardless of whether an interview had been successfully obtained at the four-month stage. This means that for some of the sample we have data for both the four-month and seven-month stages, for some we have four-month data only and for some we have seven-month data only (Table 1). Whether or not the prisoner attended any Transitional Care appointments post-release was based on monitoring data from Cranstoun. Where this was not available ⁹, the data were based on the prisoner's self-report at the four-month interview. The characteristics of the sample are shown in Tables 1 - 3 in the Appendix.

Response rates

A total of 698 contacts were issued to interviewers for the four-month interviews. Interviews were successfully achieved with 175 individuals, representing a response rate of 25%. The vast majority of the unsuccessful interviews were because we failed to make any contact with the individual (n=435, 62%). Only 3 (0.4%) individuals were contacted but refused to take part. In two cases (0.3%), someone else in the household refused permission to speak to the respondent and refused to pass on a message. In seven cases (1%), a relative informed us that the individual had died of an overdose. In 76 cases (11%) we were informed by a friend or relative that the individual was in prison or in custody (in addition, a proportion of the 435 with whom we failed to make any contact are likely to have been in prison/custody).

Seven month questionnaires were sent to 667 individuals (fewer than the number of four month contacts because we were not able to send questionnaires to those from whom we only had a telephone number but no address). We received 222 completed questionnaires, representing a response rate of 33%.

Three categories of ex-prisoner emerged from the fieldwork:

• those who declined contact with Cranstoun staff on release
• those who had contact (however brief) with Cranstoun staff on release
• those who expected to be contacted by Cranstoun staff on release, but heard nothing.

Levels of analysis

Much of the analysis has been based on comparisons between those who attended Transitional Care appointments and those who did not, at three different levels:

• Differences at the four-month stage between those who attended Transitional Care and those who did not. This analysis is based on the 4 month data only (n=175 for all respondents, n=49 for those attending Transitional Care, n=126 those not attending Transitional Care).
• Differences at the seven-month stage between those who attended Transitional Care and those who did not. This analysis is based on the 7 month data only (n=222 for all respondents, n=45 for those attending Transitional Care, n=128 those not attending Transitional Care. There were 49 cases where attendance was unknown).
• Where we have both four-month and seven-month data for an individual (n=77), any change between the two stages has been analysed to identify any differences between those who attended (n=21) and did not attend (n=56).

Caveats

Because of the relatively small number of quantitative interviews and the fact that 83 of the total 320 cases could not be linked to Cranstoun monitoring data (required, for example, for information on geography, nature of assessed needs on release, number of appointments attended) there is little analysis of sub-groups. Much of the sub-group analysis has therefore had to be based simply on whether the respondent had any contact with the Transitional Care service post-release. Further, those who did not attend any Transitional Care appointments cannot be seen as a 'control group' - they may well have different characteristics from those who did attend. There is no analysis by sex because there were only 14 females in the four month data and 23 in the seven month data. Only 10 of the four month sample and 15 of the seven month sample were non-white, so there is no analysis by ethnicity. Neither is there any regional analysis (the highest number of cases from any area was 18, in Highland), nor any analysis by the establishment the individual was liberated from (the highest number from any one establishment was 24 from Polmont Young Offenders' Institution).

Moreover, those interviewed may not be fully representative of all those who signed up for Transitional Care. It is likely that those able to provide better contact details, and with whom we subsequently managed to make contact, were in a more stable situation and were perhaps the less 'difficult' cases. This may overstate how effective the service might be for the whole of the target population.

In-depth interviews were conducted with ex-prisoners in three local authority areas to provide a more detailed assessment of the use and perceived effectiveness of Transitional Care services. The purpose of the interviews with ex-prisoners was to elicit their views about Transitional Care services and to identify any problems they had faced since their return to the community. The sample of ex-prisoners was recruited from the population returning to each area, all of whom had Transitional Care assessments conducted while in prison. This included both those who had made use of the Transitional Care service on their return to the community, those who wanted to but could not and those who had not wanted to. This allowed an exploration of the reasons why individuals drop out of Transitional Care arrangements and an examination of the experiences of those who remained in contact with the service. Interviews were conducted after the 12 week Transitional Care period, offering respondents the opportunity, where appropriate, to reflect upon their experiences of Transitional Care and its immediate consequences.

The in-depth interview sample was drawn from three areas representing different socio-geographic characteristics and different arrangements for the provision of Transitional Care. To minimise difficulties in contacting respondents, the sample was drawn from among those ex-prisoners who had participated in the questionnaire survey. Contact details were identified and potential respondents were contacted by letter and/or telephone and invited to attend for interview at a location suitable for them (for example local agency office, café or to coincide with an agency appointment). If the individual was known to be in custody, arrangements were made to interview them there. In addition to pursuing contact details, Transitional Care staff were invited to assist in the identification of suitable respondents. Thirty seven respondents participated in the in-depth qualitative interviews. Almost two thirds came from Glasgow (see Table 2) and most (23/37) were 21 years of age or older. Despite attempts to boost the numbers of young offenders and women, only two women were successfully recruited into the sample (out of a total of 7 'possible' cases for whom contact data were available).

The sample of young male offenders was boosted by drawing on respondents in Aberdeen City (1), Clackmannanshire (1), Midlothian (1), North Ayrshire (2) and Stirlingshire (1).

Twenty-two respondents were interviewed in the community ¹⁰, nine were interviewed in prison and six were interviewed over the phone.

The in-depth interviews explored the following issues:

• difficulties encountered since leaving prison
• relevance of the Transitional Care services accessed
• accessibility of the Transitional Care services accessed
• effectiveness of the Transitional Care services accessed
• perceived gaps in service provision
• further offending and drug use and how it compares with the period before their last prison sentence
• perceived impact of Transitional Care services on offending and drug use
• likelihood of continuing to access relevant services in the longer term.

In the case of clients who had dropped out of Transitional Care services, the interviews also explored their reasons for ceasing to make use of the service and their use of other services since being released from prison. All interviews were tape-recorded with the permission of respondents and identifying information was removed from the subsequent transcripts.

The majority of respondents were released from prison between December 2003 and March 2004, with almost half of these being released in February 2004. With the exception of five interviews undertaken in the winter of 2003/04, all interviews were conducted in June and July 2004. Thus, the vast majority of respondents were interviewed at least four months following release. Thirty-one of the sample had been convicted (one had been recalled from parole), whilst the remaining six had been on remand. Lengths of sentence varied from one month to approximately three years. The four Dumfries and Galloway respondents were held in HMP Dumfries, the seven Fife respondents were held in Low Moss, Saughton, Perth and Polmont, and the 20 Glasgow respondents were held in Barlinnie, Polmont, Cornton Vale and Greenock. Being under 21 years of age, all the respondents from the other local authorities were held in Polmont.

Even in the qualitative interviews there tended to be little detailed discussion regarding Cranstoun or Transitional Care arrangements at interview, not least because most respondents seemingly had no direct contact with Transitional Care on release from prison. Only seven had apparently received a service on release while 15 had declined the offer of a service and 15 indicated that they received no further contact from Transitional Care workers on leaving prison. Many respondents found it difficult to recall what services were offered to and taken up by them, partly because of the time that had elapsed since they left prison or because they experienced lapses of memory or concentration as a result of being on drugs at the time of interview. For illustrative purposes brief cases studies are presented as an Appendix to this report.