Individual_Placement_&_Support_for_Unemployed

SESAMI Social inclusion through Employment Support for Adults with Mental Illness Final Report October 2006 Website: http://www.sesami.org.uk Contents Page Acknowledgments Summary 1 2 3 4 5 6 7 8 9 10 11 12 Introduction Policy, practice and research context Methods Fidelity to IPS and influencing factors Participants’ background characteristics Outcomes Predictors of moving into work Client and staff perceptions of what works The experiences of people in work Developing an evidence-based theory Discussion Conclusions and recommendations 2 3 9 11 15 21 26 29 35 38 54 83 93 103 105 109 120 121 131 References Appendix 1: Inputs and costs Appendix 2: Nearness to Labour Market Scale Appendix 3: Outcome statistics Appendix 4 Predictor statistics 1 Acknowledgements Our grateful thanks to our six partner agencies for their support and advice throughout the SESAMI study. Also to all the clients of the six agencies who took part in interviews and gave us so much valuable information. We hope the results of the study will contribute to the development of employment support services for all those who need to use them in the future. The study was supported by the Higher Education European Social Fund and we are grateful for the opportunity provided to carry out the SESAMI study. 2 Summary 1. Introduction SESAMI (Social inclusion through Employment Support for Adults with Mental Illness) was a two-year study designed to contribute to knowledge about how people with severe mental health problems can be helped to find and keep open employment. The study comprised five main strands: 1. Assessment of the extent to which partner agencies were working in line with the international evidence base and exploration of factors influencing this. 2. Identification of the outcomes achieved for their clients by partner agencies over a period of 12 months. 3. Identification of predictors for moving into work. 4. Exploration of clients’ and service providers’ views about what works in supporting clients towards employment. 5. In-depth study of the perceptions and experiences of those who had moved into work. 2. Policy, practice and research context Enabling people with severe mental health problems to find and keep work is important in the context of the government’s welfare reforms and drive to promote social inclusion. The guidance issued to health and social services regarding the development of socially inclusive vocational services recommends implementation of the Individual Placement and Support (IPS) approach. Although there is considerable research evidence in favour of this approach, most of it comes from North America and little is known about how it can be transferred to the UK. The aim of the SESAMI study was therefore to contribute to knowledge about this by evaluating the approaches being taken in the UK, in the light of what we know from North America about what helps people with mental health problems to get jobs. 3. Methods Identifying outcomes and predictors Structured interviews 12 months apart with 155 agency clients , asking about work, self esteem, hope, life satisfaction, service use and costs. Statistical analysis Exploring perceptions of what works Open questions about what works at the beginning of the interviews with agency clients. Supplementary questions in the fidelity interviews with staff Thematic analysis Exploring perceptions and experiences of people in work In-depth interviews with 20 SESAMI participants Assessing fidelity and influences Semi structured interviews with 21 partner agency staff based on Supported Employment Fidelity Scale Scoring of responses + thematic analysis Thematic analysis 3 4. Fidelity to IPS and influencing factors On the Supported Employment Fidelity Scale only one agency was implementing the IPS approach well, three achieved a ‘fair implementation’ score and one was rated as not implementing IPS. One agency was decentralised to such an extent that the fidelity scale was inapplicable. Constraints that were hindering more widespread implementation revolved around funding criteria, agency values and their competing organisational policies. 5. Participants’ background characteristics Ninety men and 65 women took part in interviews. 84% described themselves as White British, 9% as Asian, 2% as Black British and 3% as from ‘other ethnic groups’. The average age was just over 42. Twenty percent had a higher education degree and 6% had no educational qualifications. The most frequently reported qualifications were technical, including National Vocational Qualifications. Eighty-eight percent (137) lived in their own home, and the majority (73%) were satisfied with the place where they lived. Almost half of the participants (45%) lived alone All participants had severe mental health problems, and had experienced these for an average of 13.5 years (range 1-30 years). Almost all (92%) had used mental health services within the past 2 years. 70% had at some time been admitted to psychiatric hospital, 25% in the previous two years. Eighty eight participants (57%) were unemployed at their first interview, 45 (29%) were in paid employment, 14 (9%) were in a work experience placement and eight (5%) were in sheltered work. Two unemployed people had never worked. The others had worked for between one month and 34 years, with an average of 5.63 years. They had been in contact with the employment support agency for an average of 21 months. 6. Outcomes Twenty two unemployed participants moved into work during the course of the study, and three of the 14 who had been in a placement at time 1 moved into work. Six people left paid employment, and seven who were in a work placements at time 1 became unemployed. Of those in work at time 1, 88% were working one year on. The average number of hours worked increased from 25 to 30 hours per week over the 12 months. Those working at their second interview were in a wide range of occupations, with only 10% in entry level jobs such as catering, cleaning and gardening. People in higher level occupations including administrators, technicians and skilled tradespersons were more satisfied with their job and satisfaction was also associated with higher weekly earnings. 4 Being in work, and particularly getting a job after being unemployed, was associated with greater satisfaction with daily activities and social life. People moving into work had a significantly greater increase in self-esteem than the other participants. 7. Predictors of moving into work There was no difference in the proportions of men and women moving into work. Nor did age, number of years with mental health problems or severity of problems have any effect on whether participants moved into work. However, people from Black and minority ethnic groups were less likely to move into work and the longer participants had been out of work, the less likely they were to move into work. In addition:  The more obstacles to work people perceived, the less likely they were to move into work, especially if they were afraid of being financially worse off.  People were more likely to move into work if they had recently visited a job centre or looked at job adverts. Our partner agencies’ fidelity to the IPS approach did not affect whether people moved into work. Participants’ satisfaction with the employment support they received decreased over the 12 months and the decrease was significantly greater amongst clients of the agency that was not implementing IPS. However, this did not affect whether people moved into work. 8. Client and staff perceptions of what works Clients’ perceptions of the effective ingredients of employment support revolved around three main themes:  Feeling supported (motivation and encouragement, confidence building, staff being available for support and developing a good relationship)  Practical help (financial, being accompanied to interviews etc., being helped with job preparation, job searching and the application and recruitment process)  A client centred approach (supporting at the client’s pace, appropriate job matching, links with mental health services and employers). Partner agency staff also emphasised the importance of confidence building, being there to provide support, relationships with clients, practical assistance, a client centred approach and links with mental health services. 9. The experiences of people in work The in-depth interviews with people in work focused on their experiences of employment support and of finding and keeping a job. Positive experiences of employment support included:  A wide range of support  Continuity of support worker  A relaxed, informal relationship  Confidence building 5  Mediation with the employer. Less positively, many people did not receive follow up support in work and most thought this would have been useful. Where it was provided it was appreciated. The two interview participants from the agency that did not implement IPS were less satisfied with the support they had received. Both felt they had been exploited to some degree, as cheap labour or as a means of the agency meeting its own targets, and one was critical of her worker’s lack of mental health knowledge. Although most people had thought disclosing their mental health problems to a potential employer would be a barrier to work, this did not prove to be the case when people did disclose. Help with framing mental health problems positively to an employer was appreciated. Obtaining occupational health clearance was an issue for three people. This is a potential barrier to work that has not been identified in previous research. Difficulties at work included monotonous jobs and issues stemming from people’s mental health problems. Nevertheless, people described many benefits to working, including helping them to deal with their mental health problems, improving their self-esteem, inspiring optimism about the future and having a steady income. 10. Developing an evidence-based theory On the basis of the results outlined above we were able to begin to develop an evidence-based theory about what works in enabling people with severe and enduring mental health problems to find and keep a job, how it works and in what contexts. The results point to a number of active ingredients that operate during supported employment. Previous research on vocational rehabilitation suggests that mechanisms which might increase the likelihood of people with severe mental health problems obtaining employment include encouraging the development of self-confidence and helping people to identify and articulate their employment goals. Clients’ responses about what they found helpful about supported employment, and the views of practitioners (ESWs) about what they were trying to achieve, suggested that this was often happening. While these processes did not always produce the intended outcome (employment), they appeared to result in intermediate outcomes such as improved confidence, hope, social networks and job searching skills. There is also some quantitative evidence for this in that scores on scales to measure hope and self-esteem between the first interview (T1) and second (T2) around 12 months later increased significantly among those who remained unemployed, though not to the extent of those who gained employment. This improvement in service users’ overall role functioning lends support to a recovery orientated concept of supported employment where clients’ potential for gaining meaningful employment is increased. Clients’ and practitioners’ views also gave some insight into the contexts in which the mechanisms might be most effective. In particular interviewees gave meaning to the term ‘a client-centred approach’. To them this meant receiving the right amount of support at a pace that suited them. It seems likely that the IPS approach has been 6 found effective in supporting people with severe mental health problems back to work because many of its features are conducive to providing the kind of encouraging, appropriate support that SESAMI findings suggest are the key. However, wider contextual factors, such as pre-programme characteristics of the clients, or of the external environment, may prevent the intended outcomes from being achieved. 11. Discussion and conclusions For the first time, the SESAMI study has been able to profile a sizeable group of people with severe mental health problems at the stage where employment support was likely to make a critical difference to their outcomes. To our knowledge, it is also the first time that the Supported Employment Fidelity Scale has been used in the UK. This proved a useful means of assessing fidelity to the IPS approach. Although the size of our overall sample was substantial, the numbers we were able to include in the analyses of predictors for moving into work were small, reducing the likelihood of results reaching statistical significance. Nevertheless, the study yielded results that we believe should inform the further development of supported employment: 1 We have demonstrated that people with severe mental health problems are capable of working in the full range of occupations and of sustaining part-time and full time employment in the open labour market. The recent guidance for commissioning vocational services outlines how services can incorporate the principles of supported employment into NHS mental health care. The guidance advocates the Individual Placement and Support (IPS) approach. Our research indicates that IPS is not well developed in the UK, so this will require strategic planning. Below, we make a number of recommendations about how to achieve change. We found that the quality of the support provided is as important as its organisational features. Developing IPS to ensure clients’ needs are fully met will entail: Combining IPS with psychological preparation aimed at motivation and confidence building within a supportive relationship Ensuring that co-working involves the client as a full participant in all discussions Ensuring the provision of accurate, up to date financial advice Enabling links with wider networks such as job centres Supporting clients to make their own informed choice about disclosure to a potential employer and to frame their experiences positively should they choose to disclose Liaising with occupational health professionals so that screening procedures do not present a last minute hurdle. The needs of people from Black and minority ethic groups are not currently well addressed by supported employment services. At 42% of service users, women also appear to be under-represented. 7 2 3       4 12. Implications and recommendations For commissioners: 1 The Department for Work and Pensions as a funder of large scale programmes geared to job placement or job retention should ensure that the providers of pan-disability programmes are equipped to work effectively with people with mental health problems. The Supported Employment Fidelity Scale should be used to monitor development of employment services for people with mental health problems. Equal access to employment support for women and for people from Black and minority ethnic groups should be a matter of urgent concern. 2 3 For providers: 4 Access to vocational services should not be restricted on the grounds of the severity or duration of an individual’s mental health problems, nor on the grounds that working may be harmful to them. Training in supported employment skills should be made available to clinical staff in community mental health settings and allied services. It should focus on the quality of employment support from a service user perspective as well as on organisational features. Psychological preparation for work should be provided for those individuals with low levels of motivation and confidence as an integral aspect of employment support, alongside practical job preparation and job search activities. 5 6 For future research 7 The development and promotion of employment support services for people with mental health problems in England should be accompanied by comprehensive research and evaluation. The needs of women and people from Black and minority ethnic groups should be a specific focus for study. 8 8 1. Introduction In the UK, over half a million people of working age have a mental health related disability and people with mental health problems are less likely to be employed than other disabled people, with only 20% in work compared to people with learning disabilities, for example, at 23% and people who have visual problems at 52% (Disability Rights Commission, 2006). For people with severe and enduring problems such as schizophrenia, the proportion in work may be lower, with some studies reporting rates as low as 4% (Marwaha and Johnson, 2004). . The deeply embedded assumption within mental health services is that people with mental health problems do not want or are unable to work (Cunningham et al., 2000). Yet surveys indicate that most of this group do aspire to paid employment (Bates, 1996; Rinaldi and Hill, 2000; Secker et al., 2001; South Essex Service User Research Group et al., 2006), while international evidence indicates that, with appropriate support, many can achieve their goal (Drake et al., 1996). . SESAMI (Social inclusion through Employment Support for Adults with Mental Illness) was a two-year study designed to contribute to knowledge about how people with severe mental health problems can be helped to find and keep open employment. The study was financed from Higher Education European Social Fund Objective 3 resources and six partner agencies, designated here as Agencies A to E, made contributions in kind. To our knowledge, the partner agencies were amongst the largest providers of employment support to people with mental health problems in the UK. Agency A differs from the other partner agencies in that it operates within the mental health sector. Agencies B, C and D operate within the voluntary sector, by far the most common location for UK vocational services. Agency E is a pan-disability organisation with contracts to implement two government programmes, Workstep and New Deal for Disabled People (NDDP), while Agency F is an arm of the Department for Work and Pensions (DWP) that contracts out the Workstep programme under a range of brokerage arrangements. (See Section 2 for information about the government programmes). SESAMI comprised five main strands: 1. Assessment of the extent to which partner agencies were working in line with the international evidence base and exploration of factors influencing this. 2. Identification of the outcomes achieved for their clients by partner agencies over a period of 12 months. 3. Identification of predictors for moving into work. 4. Exploration of clients’ and service providers’ views about what works in supporting clients towards employment. 5. In-depth study of the perceptions and experiences of those who had moved into work. Ethics approval was granted for the study by the North West Multi-Centre Research Ethics Committee. Research governance approvals were granted by the relevant NHS Trusts where applicable. 9 This final report from the SESAMI study draws together results from across the five strands. The following section provides background information about the policy, practice and research contexts of the study. Section 3 describes the methods used and the results are then presented in Sections 4 to 9. Section 10 builds on those results to develop an emerging theory about both what works in enabling people with severe mental health problems to find and keep a job, and how. The final two sections of the report provide a discussion of the results, followed by our conclusions and recommendations. 10 2. 2.1 Policy, practice and research context Policy Since the New Labour party came to power in 1997, government policy has been dominated by the Welfare to Work agenda. Reducing the numbers of social security benefit claimants of working age has been an important economic and policy target, with a series of programmes targeted at disabled people, particularly those in receipt of Incapacity Benefit (IB), together with lone parents and the young unemployed. The Green Paper outlining plans for welfare benefit reforms, A New Deal for Welfare: Empowering People to Work (DWP, 2006) and the subsequent Welfare Reform Bill laid before Parliament in July 2006 are the latest in a series of policy initiatives and programmes designed to move people off welfare and back into the labour market. The first series of programmes introduced by the incoming Labour government were the New Deals. One of these - the New Deal for Disabled People (NDDP) - is delivered through advisers known as job brokers to individuals with health or disability related barriers to labour market entry. It aims to provide career advice, practical support and guidance through the benefits maze. Another programme, Workstep - successor to the Supported Placement Scheme - is designed to help people who face more complex barriers, and can provide training and ongoing support to the employer as well as the employee. A more recent initiative, Pathways to Work , is aimed mainly at new claimants of Incapacity Benefit (IB) and supplements the personal advisor approach with condition management advice provided by the NHS and a cash incentive of £40 per week ‘return to work’ payment. NDDP, Workstep and all other Jobcentre Plus commissioned programmes are included in the ‘Choices’ package in Pathways to Work. Pathways to Work was an instant success and the government aims to build on this success with a national roll out of the programme, linked to other changes to Incapacity Benefit set out in the Welfare Reform Bill. A significant proportion (40%) of current IB claims are attributed to mental health problems and one third of claimants cite mental ill health as the reason for their disability. The majority of these claims will be due to the most prevalent mental health problems, mild to moderate depression or anxiety disorders, which can often be transient. While the published evaluations of existing government programmes seldom differentiate outcomes for people with specific types of disability, there is an indication from the NDDP evaluation that people with mental health problems, once placed, were more likely than other disability groups to lose their jobs (DWP, 2004). Early findings from Pathways to Work seem to show that the programme’s immediate positive impact on the employment of people with mental health problems may be less marked than for people with other diagnoses (Adam et al., 2006) The Workstep programme benefits relatively few people with mental health problems. DWP statistics indicate that in April 2005 only 7% of the total number of 14,000 places were taken up by people whose main impairment was mental ill health (www.afse.org.uk/conference/presentations/research.ppt ). In relation to clients’ views of the benefits of Workstep, Meah and Thornton (2005) found that people ‘with mental health conditions also indicated that it was important to have the continuity of a designated support worker’ (p.38). A qualitative study of the role of Pathways to Work personal advisers found that ‘moderate to severe’ mental health issues were judged to be among the more difficult cases and that advisors sometimes found it 11 desirable to waive or defer a work focused interview on grounds of mental illness (Dickens et al., 2004). Using approaches based on the Pathways to Work model, the government proposes to reduce the number of people dependent on benefits due to disability by 1 million over ten years. All disabled people, including those with mental health problems, will face a stricter benefits environment, structured to reward efforts made towards gaining employment.  Existing claimants will retain their benefits but their situations will be reviewed periodically for employment potential  If they try work and cannot sustain it, benefit levels will be protected  From 2008, the Employment & Support Allowance will be paid to new claimants in return for work related activity (job hunting, work experience etc.)  There should be no assumption that any person is exempted from work by their diagnosis alone  People whose functioning is severely affected will receive benefits without ‘conditionality’ – i.e. without having to demonstrate work-related activity. The last two points encapsulate a dilemma for people with severe mental health problems and their carers: whether to claim their entitlements to the support necessary to get and keep a job or to seek exemption on grounds of severity of impairment. The support needed to get work may be considerable because of the particular disadvantages faced by people with severe mental health problems, many of which have not been altered by policy. These disadvantages include stigma, ignorance and fear of mental illness, which may make employers reluctant to take on people with mental health problems, and low expectations within mental health services of what people are able to achieve, which may undermine their confidence, distort their perception of their own capacity to work, and discourage them from seeking a job. The Disability Discrimination Acts of 2000 and 2005 go some way towards addressing attitudes amongst employers by establishing the employment rights of people with disabilities, including those related to mental health, and specifying employers’ obligations to ensure equality of opportunities. In the 2005 Act the definition of mental health problems was changed to eliminate the requirement for the condition to be ‘clinically well recognised’ thus emphasising function rather than diagnosis. The Act was also extended to cover all employers whatever the size of the workforce. Potentially this means that more employees with mental health problems could claim protection under the Act and be entitled to reasonable adjustments in the workplace. However, the net effect could also be to encourage employers to find ways of avoiding engaging people whom they know to have a history of mental ill health. The Americans with Disabilities Act on which the DDA was based has not improved the employment rates of disabled people in the United States (www.postgazette.com/pg/05196/538181.stm). 2.2 Practice Anti-discrimination legislation, however comprehensive, can only come into play once a person has applied for a job, and low levels of confidence reinforced by low 12 expectations within mental health services therefore also need to be addressed. A recent emphasis on social inclusion, given impetus by publication of a major report on mental health and social exclusion by the Office of the Deputy Prime Minister (2004a) and the subsequent guide to promoting social inclusion (Office of the Deputy Prime Minister, 2004b), is intended to address this. The initial report on social exclusion was based on a wide-ranging study that included reviews of the literature, a consultation exercise that received over 900 responses from across stakeholder groups, seven consultation events around England involving 500 people with mental health problems and carers, and visits to schemes that were already tackling the issues. The report sets out a 27-point action plan to bring together the work of government departments and other organisations in a concerted effort to challenge attitudes, enable people to fulfill their aspirations, and significantly improve opportunities and outcomes. The actions required were divided between six categories, one of which was employment, with the National Institute for Mental Health in England (NIMHE) charged with producing guidance for service commissioners. That guidance was finally published two years later (National Social Inclusion Programme, 2006) with a key recommendation for the implementation of evidencebased practice within vocational services, in particular the Individual Placement and Support (IPS) approach. The IPS approach has six features that distinguish it from other approaches (Becker and Drake, 2003): 1. A clear focus on competitive employment rather than on more sheltered types of work. 2. Rapid job search as opposed to lengthy training and job preparation. 3. Integration with mental health care. 4. Responsiveness to service user preferences. 5. Continuous and comprehensive assessment before and during employment. 6. Time-unlimited support in work. Thus within IPS, employment and clinical care are seen as linked strands of a ‘recovery’ process, acting separately and together to improve well-being and quality of life. The goal is competitive employment in integrated work settings, with minimal prevocational training or assessment. There is an emphasis on clients’ needs and preferences, with follow up support, both vocational and clinical, being provided for as long as necessary. 2.3 Research The evidence on which the recommendation to implement the IPS approach was based is extensive. In a review of the research, Schneider (2005) identified 11 evaluations, including five randomised controlled trials, two meta-analyses, a systematic review and three other cohort or case studies. All 11 studies conclude that when compared to other vocational interventions IPS is more effective in enabling people to get real jobs with real wages, even service users who experience multiple disadvantages in addition to their mental health problems. IPS has such unparalleled 13 evidence in its favour that it is becoming known as evidence-based supported employment. However, the great majority of the evidence emanates from the United States. Although a Canadian study indicates that IPS is transferable to that country (Latimer et al., 2006) and results from a randomised controlled trial currently underway in the UK and Europe (Burns, 2005) will shed further light on the issue, little is yet known about its transferability to the UK. This is important because the UK differs from the US with respect to social security (Turton, 2001) and health insurance entitlements (MacDonald Wilson et al., 2003), both important considerations in a disabled person’s choice to work. The overall aim of our study was therefore to advance knowledge about employment support by evaluating the approaches being taken in the UK, in the light of what we know about what helps people with mental health problems to get jobs. We set out explicitly to look at the ‘real world’ operation and outcomes for people with mental health problems of established as opposed to experimental programmes. In the following section we describe the methods we used to achieve our aim. 14 3. Methods The SESAMI study methods comprised a combination of quantitative and qualitative approaches designed for use in the different strands of the study. 3.1 Assessing fidelity to the IPS approach and exploring influencing factors Sample Five of the six partner agencies assisted with this aspect of the study. Because Agency F contracts out the Workstep programme under a range of brokerage arrangements it was not meaningful to include the agency in this aspect of our study. For the other five agencies, we interviewed 21 representatives from 16 sites where the agencies operated. Where possible both a manager and an employment support worker (ESW) were interviewed at each site. Interviews varied in length from 45 to 90 minutes and were tape-recorded and transcribed verbatim. Table 3.1 shows the distribution of the interviews across the five agencies. Table 3.1 Distribution of interviews across five agencies Agency A Agency B Agency C Agency D Manager 1 2 3 1 ESW 3 3 2 3 Measures A schedule was developed to focus the interviews on employment support practice within the agencies. This was based on the Supported Employment Fidelity Scale developed by Becker and Drake (2003) to assess fidelity to the IPS approach. The fidelity scale comprises 15 items drawn from IPS principles and core practice components. Each item is scored on a scale of 1 to 5 using mainly quantifiable descriptive anchors. A score of five indicates the provider has fully implemented the IPS standard, while a score of one suggests that the standard has not been met. (See page 22 for the items and the standard required for a full adherence rating for each items). The total score for the scale is calculated by summing the item scores. Scores greater than 65 indicate good adherence, scores between 56 and 65 fair adherence and scores lower than 56 non-adherence. Although the scale has not been formally validated in the UK, Becker and Drake (2003) recommend that the format for eliciting information should be conversational rather than structured, thus allowing sufficient flexibility to accommodate differences between the English and North American terminology and programme delivery contexts. Data analysis The first stage of the analysis involved two researchers independently examining each transcript and identifying where on the 5-point scale the participant’s response to the 15 items fitted. These provisional scores were then compared to identify any differences in the ratings. This occurred for only 11 responses, and the differences 15 Agency E 1 2 Total 8 13 were readily resolved through discussion. In the process of agreeing scores, it became clear that in the case of four interviews insufficient information was available to score one item, and in one case there was insufficient information to score two items. In order to account for these missing data, summated scores for each interview were calculated as a percentage of the total possible for that interview. Thus in the two cases where data on one item were missing, the total possible value was taken as 70 rather than 75. In the one case where data on two items were missing the total possible value was taken as 65. There was little difference in individual interview scores within agencies and the average percentage score across interviews for each agency was therefore taken as the overall score for that agency. To allow comparison with the adherence bandings for the fidelity scale, these were also calculated as percentages of the total possible. Thus good adherence was represented by a score of 87% or more, fair adherence by a score between 75% and 86%, and non-adherence by a score of less than 75%. A thematic qualitative analysis of the interview transcripts was also carried out, following the steps advocated by Miles and Huberman (1984) in order to understand the factors that influenced the extent of agencies’ adherence to the IPS approach. 3.2 Identifying outcomes and predictors for moving into work Sample In order to recruit participants to the SESAMI study we provided letters of invitation to the partner agencies and asked them to forward these to all their current clients who were identified as having mental health problems. The invitation included brief information about the aims of the study, and a short screening questionnaire to identify the severity and endurance of people’s mental health problems (Box 3.1). Of the 888 people sent an invitation, 292 responded, agreeing in principle to participate, and returning their completed screening form. Two hundred and seventy one met our inclusion criteria of scoring at least 3.5 on the severity scale, and were contacted by a researcher with further information. In total, 212 participants were successfully recruited, from the north of England, the midlands, the south and the south east. 16 Box 3.1: Severity screening scale Number of years had mental health problems: 3 years = 2 Description of mental health problems: Nerves, depression, anxiety, substance misuse not combined with anything else = 1 Co-morbid substance misuse, schizophrenia, psychosis, hearing voices, bi-polar disorder, manic depression = 2 Hospitalisation history: Never admitted = 0 Ever admitted to psychiatric hospital = 1 Admitted in last two years = 1.5 Ever admitted for more than 6 months = 2 Medication: Not prescribed in last 2 years = 0 Prescribed in last 2 years = 1 Treatment by a professional: Not received in last 2 years = 0 Received in last 2 years = 1 Interviews were conducted at two time points (time 1 and time 2), with approximately 12 months between each. Six people were interviewed by telephone where it proved logistically difficult to meet face to face. Others were interviewed at their agency’s premises, their local job centre, or an alternative place of their choosing. At the time 1 interview, we found that 30 people had not been actively involved with their agency for a period of over a year, and they were subsequently excluded from the study. Of the 182 people included at time 1, 155 were interviewed at time 2, a response rate of 85%. Of the 27 who were not interviewed at time 2, 22 could not be contacted, three declined to participate, and we were asked by their agency not to interview two clients whose mental health had deteriorated. We analysed these missing cases to see if any categories were over represented and found no bias in the time 2 sample. Measures The questionnaire designed for this strand of the study included items to elicit details on demographics (age, sex, ethnicity, educational qualifications and housing circumstances), and on mental health, employment, and financial circumstances. Standardised measures of outcomes and predictors included:  Self esteem – self efficacy factor from the Empowerment Among Users of Mental Health Services Scale (Sciarappa et al, 1994)  Herth Hope Index (Herth, 1989, revised 1999) 17     Minnesota [job] Satisfaction Questionnaire – MSQ (short form, Weiss et al., 1967) Behaviours indicating nearness to the labour market (Box 3.2) Perceived obstacles to work (Box 3.2) Selected items from the Carers and Users Expectation of Services questionnaire – CUES User Version (Lelliot et al., 2001). The MSQ short form consists of 20 questions measuring overall job satisfaction, intrinsic (inherent in the job) satisfaction and extrinsic (external to the job) satisfaction. Examples of intrinsic factors are level of responsibility and security. Examples of extrinsic factors are opportunities for advancement and company policy. The CUES items had two components. A first question asks the individual how their life compares with a given standard (as good as this, worse than this, very much worse). A supplementary question then asks if they are satisfied with this aspect of their life. We asked about three issues of particular relevance to our study: daily activities, relationships and social life. Box 3.2: Items from scales Perceived obstacles statements I haven't got the skills and experience I'm unlikely to find job because of my age An employer would be put off by my mental health problems I'd be worse off financially if I started work My confidence about working is low My mental health makes it hard for me to work My family do not like the idea of me working Data analysis Our analysis of outcomes and predictors for moving into work was based on the responses from the 155 people who were interviewed at both time 1 and time 2. For outcomes we focused on:  Comparing results on the outcome measures at the two points in time to identify changes over the 12 month period  Comparing scores at time 2 for those who were in work with those who were not in work  Correlations between self esteem and hope scores and weekly earnings, type of job and stage of employment support In order to identify predictors for moving into work we looked at differences in the time 2 results between people who were in work and those who were unemployed. We then excluded people who were already in work at time 1 and compared results at time 2 for those who had moved into work over the course of the study and those who remained unemployed. Nearness to labour market behaviours In the past 6 months, have you: … looked at job adverts in the newspaper' … visited a job centre' … registered with a private employment agency' … done any kind of work related training course' … applied for a job' … looked for work in other ways' 18 Having examined the distribution of the variables used, we tested for statistically significant differences using Pearson Chi-square tests for dichotomous variables, Mann-Whitney U and Wilcoxon tests for non-parametrically distributed variables, and t tests for normally distributed variables. 3.3 Exploring perceptions of what works During the interviews with partner agency staff used to assess fidelity to the IPS approach, participants were also asked what they thought worked well about what they did and what they thought clients found helpful. The structured questionnaire used to identify outcomes and predictors for moving into work also included a preliminary question inviting agency clients to comment on their experiences of working with their employment support agencies, in what ways the support provided was helpful and how it could be improved. Handwritten notes were made to capture the responses and subsequently typed up. The notes and relevant data from the staff interviews were analysed using similar thematic techniques to those used to identify factors influencing fidelity to the IPS approach. 3.4 Exploring perceptions and experiences of people in work Sample At the end of the baseline SESAMI interviews, participants in work were asked if they might be interested in taking part in an additional in depth interview. Approximately six months later an information sheet describing what would be involved was sent to those who had expressed interest and 20 people consented to an interview. The interviews were carried out mainly at agency premises and lasted between 40 minutes and one and quarter hours. All the interviews were tape recorded with participants’ consent and fully transcribed for analysis. Interview schedule A semi-structured schedule designed to guide the interviews covered: • Motivation to engage • Expectations of employment support • Assessment process • Job preparation/funding/matching • Follow up support • Overall evaluation of the service provided • Service links with employers and mental health services • Disclosing mental health problems to employer • Social networks • Barriers to getting back to work • Impact of working • Hopes for the future. 19 Data analysis As for the analysis of factors influencing fidelity to the IPS approach, a thematic qualitative analysis of the interview transcripts was carried out following the steps advocated by Miles and Huberman (1984). Nvivo software was used to aid the organisation and retrieval of data. The following sections of our report present the results from each strand of the study. For clarity, participants’ background characteristics at time 1, outcomes at time 2 and predictors of moving into work are presented separately, in sections 5, 6 and 7 respectively. 20 4. Fidelity to IPS and influencing factors Results for the five partner agencies included in this strand of the study are presented in two sections describing agency scores on the Supported Employment Fidelity Scale and the influencing factors identified through our thematic analysis. 4.1 Agency item scores Table 4.1 shows the five agencies’ percentage scores, averaged across interviews, for each of the 15 items included in the fidelity scale. The standard for full adherence is shown in brackets. The average percentage fidelity score for each of the five agencies is shown in the final row. As can be seen, only Agency A attained a good adherence rating. Agencies B, C and D all attained a fair adherence rating, while the rating for Agency E indicated that this provider was not delivering supported employment based on the IPS approach. Table 4.1 Agency scores (% of total possible) Items (standard for full adherence) Caseload size (25 or fewer) % time on vocational work (100%) Continuity of worker (Each worker carries out all aspects) Contact with mental health services (Attached to & work closely with CMHT) Workers form a vocational unit (With at least weekly group supervision; provide services for each others’ clients & back up for each other) Eligibility & referral criteria (All service users encouraged to participate, no eligibility criteria used) How and where clients assessed (Assessment ongoing in work, minimal pre-job search testing) Time from referral to job search (Within 1 month) Who decides which job (Based on client preferences) % clients in same job setting / type (10% or less) % permanent competitive jobs (Virtually all) Help to find another job when one breaks down (Almost always) % clients receiving follow up support / how long for (Time unlimited support for most) % time workers spend in community (At least 70%) Outreach if don’t engage / drop out (At least monthly, time unlimited) Average agency percentage fidelity score A 95 95 100 100 75 80 100 100 100 95 85 95 60 90 80 90 B 92 100 96 48 80 60 96 52 88 85 80 72 52 65 56 75 C 80 96 88 20 80 60 100 60 100 76 80 100 72 76 56 76 D 60 90 80 20 60 60 100 80 90 90 100 85 60 85 75 76 E 60 100 80 20 80 60 20 87 60 60 100 87 60 53 27 67 Agency A secured good adherence scores on ten of the 15 fidelity scale items and fair adherence scores on four, with one non-adherence score for follow up support. Agencies B, C and D were very similar in terms of the distribution of good, fair and non-adherence scores received. All three agencies received good adherence scores on five items, fair scores on three to five items and non-adherence scores on five to seven items. The agencies attained good scores on three of the same items: time on vocational work, assessment and job choice. All three also received non-adherence 21 ratings on three of the same items: contact with mental health services, eligibility criteria and follow up support. However, there were differences between the three agencies’ adherence ratings on two items: caseload size, on which Agency B attained a good adherence rating and Agency D a non-adherence rating; and providing assistance when a job ends, on which Agency C attained a good rating and Agency B a non-adherence rating. Agency E attained a good adherence rating on only four items (time on vocational work, time from referral to job search, focus on permanent competitive jobs, assistance to find another job), a fair rating on only two and a non-adherence rating on the other nine. 4.2 Influencing factors Thematic analysis of the interview transcripts revealed constraints on agencies’ capacity to provide an evidence-based IPS service. These revolved around funding, the values espoused within agencies and organisational policy. Funding constraints Funding constraints accounted for several of the low fidelity scale scores attained by Agencies B, C D and E. At the three voluntary sector agencies (B, C and D) three issues emerged. Firstly, the funding arrangements for all three agencies entailed the use of eligibility criteria to determine access to services. These included clients being claimants of certain benefits and living in specific areas. Secondly, funding arrangements constrained the extent of follow-up support the three agencies could offer clients who secured paid employment. At most sites, a limit of between 12 and 26 weeks was imposed. However, at one Agency B site no funding provision was officially available to offer any follow up support: With [the funder] what they want us to do is exit clients once they actually find employment. From their first day they’ve actually found employment, and then the thing that is a criteria is that person has to stay in that job for 13 weeks for it to be regarded as a job outcome. Clearly, when follow up support was time limited this had implications for agencies’ capacity to support clients to find a new job if their current job broke down, since agency staff would be unlikely to be aware of the situation once the designated follow up support period ended. Thirdly, funding constraints impacted on outreach at both Agency B and Agency C. The limited resources available again meant this activity was time limited, with in most cases only three attempts through phone calls and letters made to contact clients who did not engage or dropped out: What we’d do is write to them and try and contact them and say we had noticed that they haven’t been engaging. Cause obviously we always have a 22 waiting list as well so if somebody is not engaging then they’re taking up a place somebody else could be using. So we’d inform them that if they don’t contact us then we would have to archive their file. At Agency E, funding arrangements required all new clients to undergo basic skill testing that was antithetical to the minimal testing advocated for IPS. In addition, funding constraints influenced the extent to which the organisation offered follow-up support to clients. Depending on the government programme under which clients were funded, support in work would either be limited to contact every three months after the first month, or to telephone contact: If it’s Workstep then we’d visit them in the first month and then every three months but if it’s NDDP it would be more telephone support. Value based constraints In some respects the values espoused by ESWs at Agencies B, C and D prompted them to circumvent some of the funding constraints outlined above and unofficially support their clients in ways that adhered more closely to IPS standards. For example, ESWs reported that although they weren’t supposed to, they would provide follow up support to clients in work by making regular phone calls and would also be available to help clients if their job ended. To compensate, administrative tasks would be completed at home outside normal working hours. However, value based constraints also contributed to the three agencies’ low fidelity scores. In particular, embedded within the accounts from these agencies was a perceived conflict of values with mental health services. Some ESWs judged mental health services to be dominated by a medical model of mental health, which they could not reconcile with their own, preferred social model. Thus, contrary to IPS practice standards, these workers saw benefits in limited contact with mental health services: I think people that use the mental health services find it is still dominated by the medical model of health and I think we offer more of a real sort of model and I think that’s quite a refreshing change for a lot of people. I think that’s one of the advantages as well, that we don’t double up with mental health services, we complement by offering something different, more social if that makes sense. Clearly, a more integrated approach would require commitment from both mental health services and employment support agencies, and some staff did describe attempts to make contact that were met with no response. Nonetheless, the value base of the agency staff themselves, which appeared to reflect a wider agency worldview, did emerge from the interviews as a significant constraint. In addition, values espoused by staff at Agencies B and C led to extensive pre-job search preparation and training, as opposed to focusing on job finding within a month with ongoing assessment and support as required. At these agencies, staff perceived rapid entry to work as potentially too stressful for clients and as a threat to their mental health, rather than as integral to their recovery. Thus a staged process of training was provided, often beginning with training for a National Vocational Qualification (NVQ), followed by an unpaid work placement or voluntary work, 23 before beginning the job search. In some cases the period between referral and job search could be as long as two years: We would like to do an initial interview to find out their desired path into employment which could be two years…we do an assessment of their skills, assessment of the jobs they are wanting to do and we’d try and find the correct training that they need. This could be something like basic skills to start leading towards an NVQ and we’d then find them a work placement. At Agency B, staff appeared particularly protective towards their clients and would encourage them to look for jobs in the public sector on the grounds that the public sector would provide a more supportive environment than the private sector: Well there’s two reasons for [focusing on public sector jobs], I think the first reason is because we don’t want people to be in a situation where there’s the possibility of relapse so we try and sort of avoid the situations but also companies don’t support [clients] anyhow. There isn’t like a job retention programme which supports them they’ve got nowhere to turn to they can’t turn to their unions as they won’t get any support there. As a result higher proportions of clients were likely to be placed in similar job settings and types than required for adherence to the IPS approach Organisational policy constraints As the following extract illustrates, organisational policy within Agency E was to work with ‘key account employers’ with whom clients were encouraged to apply for jobs if possible: We do have what we call pipelines which we try and encourage perhaps people to look at. What we call vocational pipelines which is where a lot of our key accounts do come through, but that doesn’t mean to say that if somebody comes through who has got their own job goal and it’s realistic and they have got the skills to do it we are not going to give them the support in achieving that, but that’s part of a broader strategy, that then we could look at supermarkets and there are so many different job types within that that it’s huge within its own right really. As a result, not only were a high proportion of clients placed in the same job setting and type of work, but job choices were also often based on vacancies with key account employers rather than client preferences. Although less constraining on Agencies B, C and D than funding arrangements and the values espoused within the agencies, in some cases these agencies’ own policies could also be antithetical to IPS standards. In particular, at one Agency D site ESWs were expected to carry a caseload of 75 people, three times greater than the recommended maximum of 25. And at one Agency B site risk assessment procedures could limit eligibility for the service: 24 We need quite a comprehensive risk assessment on them to make sure we’re not putting them in a situation where they’d be vulnerable or putting the public at any risk. Because if they have a tendency, I mean, maybe to be violent or something like that you know when they become ill. In contrast, the good adherence score attained by Agency A stemmed directly from the fact that at this agency organisational policy was to a large extent explicitly based on the IPS approach and standards. Working practices were grounded in the IPS approach and all ESWs underwent training in IPS. In turn, the values espoused within the agency were consonant with IPS standards. Both managers and ESWs expressed the belief that it is both possible and desirable for people who have severe and enduring mental health problems, and who express a desire to work, to achieve this in a short period of time, with the right support: Work is fundamental to recovery, it’s really important as opposed to traditionally mental health services that have discouraged people from going into work. It’s been seen as being stressful and so you’ll meet clients who often say ‘Oh I can’t work because my doctor told me I can’t’ and it was his GP who was reliant on what he’s learnt in a text book that was written 20 years ago. So this poor guy is not working and not thinking about working on the basis of what was accepted medical attitude towards mental health and employment from 100 years ago, rather than then evidence that has recently been published in the BMJ, so it’s really important that that changes. (Manager). It’s open to anyone, that’s the IPS motto. (ESW). However, organisational policy at Agency A also accounted for the non-adherence scores received for follow up support. Once a client was discharged from mental health services they were no longer eligible for employment support. It was assumed that if a client was well enough to be discharged from services, then lower levels of support were more appropriate. It was also the case that clients who found work risked being discharged from mental health services, resulting in their employment support ending: If someone gets employment they get discharged back to primary care cause it’s seen that they are stable. 25 5. Participants’ background characteristics Of the 155 people interviewed at both time 1 and time 2, 90 (58%) were male and 65 (42%) female. The majority (84%) described themselves as of White British ethnicity. Nine percent described themselves as Asian (a category including Bangladeshi, Chinese, Indian and Pakistani) and 2% as Black British. Three percent identified themselves as belonging to ‘other ethnic groups’. Their average age was just over 42 years old, ranging from 22 to 67 years (standard deviation 9.874). Information about education, housing circumstances, mental health, employment and income follows. Unless otherwise stated, there were no significant differences relating to age, sex or ethnicity. 5.1 Educational qualifications Figure 5.1 shows participants’ highest educational qualifications. Twenty percent had a higher education degree, while 6% had no educational qualifications. The most frequently reported qualifications were technical, a category including National Vocational Qualifications. Figure 5.1 Highest educational qualification Technical 24.52% Higher Education degree 20.0% A level or equivelant 16.77% GCSE or equivelant 12.9% Further Education dip 9.68% IT or office qualification 8.39% None 5.81% Professional qualification 1.29% 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% Percent 5.2 Housing circumstances Ninety two percent (142) of participants had lived at their current address for over a year. Eighty-eight percent (137) lived in their own home, and the majority (73%) 26 were satisfied with the place where they lived. Thirteen people were living in sheltered housing, and all but one were satisfied with their accommodation. As Figure 5.2 illustrates, participants were more likely to live alone than with family members or others. Figure 5.2 Participants’ living circumstances Alone 44.52% With partner 14.84% With partner and children 12.26% Single parent 4.52% With other family 15.48% With non family 7.74% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% Percent 5.3 Mental health At time 1, the length of time over which people reported experiencing mental health problems ranged from one to 50 years, with a mean of 13.51 years (standard deviation 10.08). Seventy percent had a history of being admitted to psychiatric hospital, 25% in the previous two years. Their summary severity scores on our screening questionnaire ranged from 3.5 to 8, with a mean of 6.07 (standard deviation 1.236). This was significantly higher for people from Black and minority ethnic groups (mean = 6.56, standard deviation 1.2855, compared to mean = 5.97 standard deviation 1.2139 for White British participants, p=0.004) 5.4 Employment Eighty eight participants (57%) were unemployed at time 1, 45 (29%) were in paid employment, 14 (9%) were in a work experience placement and eight (5%) were in sheltered work. Although two of the work experience placements were waged, the others offered only expenses. 27 Of the 100 people who were not in paid employment or a waged placement, two had never worked, although both had been in unpaid work experience placements in the past. Ninety-four of those who did have previous experience of paid employment were able to estimate the length of time since they had last been employed. There was a wide range, from one month to 34 years with a mean of 5.63 (standard deviation 6.118). Over a third (36%) had been unemployed for up to two years, and nine had been out of work for over 15 years. Regardless of whether participants were currently in work, we asked how long they had spent in paid employment since age 18. Again, there was a wide range, with nine people (6%) reporting work experience of 12 months or less and 23 (14%) reporting experience of 25 years or more (mean 14.05, standard deviation 10.419). 5.5 Financial circumstances Two participants declined to give financial information. Table 5.1 details the welfare benefits and tax credits received by the rest of the sample. The majority of those who were unemployed claimed incapacity benefit (60%), housing benefit (68%) and council tax benefit (70%). Half also claimed Disability Living Allowance. Fewer people in paid work received benefits, but Disability Living Allowance and Incapacity Benefit was paid to 28% and 20% of the sample respectively. Fifty percent were on a sufficiently low income to qualify for working tax credit. Table 5.1 Benefits and Tax Credits % claiming In paid work (n=53) Job seekers allowance Income Support Statutory Sick Pay Incapacity Benefit Disability Living Allowance Carers Allowance Disabled Person’s Tax Credit Working Tax Credit Child Tax Credit Housing Benefit Council Tax Benefit Unemployed (n=100) 0 6 0 20 28 0 12 50 15 15 24 5 40 1 60 50 1 0 0 5 68 70 28 6. Outcomes The results from our outcome measures are presented first for employment and related outcomes and then for self esteem and hope. Statistical tables relating to the results can be found in Appendix 3. 6.1 Employment and related outcomes These outcomes are considered under five headings:  Employment status  Occupational categories  Hours worked per week  Satisfaction with the aspects of life from the CUES measure  Income. Employment status Figure 6.1 illustrates the changes in employment status reported by participants between time 1 and time 2. As can be seen, of the 88 people who were unemployed at time 1, 22 had moved into work during the course of the study, and three of the 14 who had been in a placement at time 1 also moved into work. Six people moved out of paid employment, and seven who were in a placement at time 1 were unemployed at time 2. Figure 6.1 Change in employment status Time 1 (n=155) Working (53) Time 2 Still working (47) Not working (6) Time 2 Still unemployed (64) Working (22) Placement (2) Unemployed (88) Time 2 Still in placement (4) Working (3) Not working (7) Placement (14) Occupational categories The occupations reported by the 54 participants in work at time 1 and the 72 participants in work at time 2 were categorised in line with the Standard Occupational Classification 2000 (Office of National Statistics, 2005). As Table 6.1 illustrates, the most commonly reported jobs at both points in time were in the personal services, 29 sales and customer service category and in the associated professional, technical, administrative and secretarial category. A slightly lower proportion of people were in entry level or elementary occupations. Over the course of the study, there was no movement in the proportion of participants in the highest occupational category (managers, senior officials and professionals). However, five people moved into the associated professional, technical, administrative and secretarial category and seven people moved into the personal service, sales and customer service category. Fewer people (3) moved into entry level or elementary occupations. Table 6.1 Occupational category Managers, senior officials and professionals Associate professional, technical, admin and secretarial Skilled trade occupations Personal services, sales and customer services Process, plant and machine operatives Entry level occupations Placement Unemployed Total Time 1 Time 2 Time 2 Time 2 Frequency Percent Frequency Percent 6 3.9 6 3.9 14 3 15 2 13 14 88 155 9.0 1.9 9.7 1.3 8.4 9.0 56.8 100.0 19 2 23 6 16 6 77 155 12.3 1.3 14.8 3.9 10.3 3.9 49.7 100.0 Analysis of participants’ responses to the Minnesota Satisfaction Questionnaire revealed results relevant to the types of job participants were doing. Although there were no significant differences in any aspect of job satisfaction between time 1and time 2 for the whole sample of participants in work at the two points in time, at time 2 there was a correlation between overall job satisfaction and being in a higher occupational category that just reached statistical significance (p=0.05). Hours worked per week In terms of the hours people were working each week, the proportion working 15 hours or fewer increased from 18% to 25% between time 1 and time 2 (Table 6.2). Across the whole sample, however, the average number of hours worked increased from 25 to 30 hours per week (Table 6.3). 30 Table 6.2 Hours worked at time 1 and time 2 Time 1 Time 2 Hours Frequency Percent Cumulative Frequency worked percent 1-5 5 9.1 9.1 7 6-10 3 5.5 14.5 5 11-15 2 3.6 18.2 6 16-20 10 18.2 36.4 6 21-25 10 18.2 54.5 12 26-30 5 9.1 63.6 5 31-35 7 12.7 76.4 7 36-40 10 18.2 94.5 17 >40 3 5.5 100.0 7 N=55 N=72 Table 6.3 Average hours worked at time 1 and time 2 N Min Max T1 hours worked per week 54 3 70 T2 hours worked per week 70 1 60 Satisfaction with aspects of life (CUES) Percent 9.7 6.9 8.3 8.3 16.7 6.9 9.7 23.6 9.7 Cumulative percent 9.7 16.7 25.0 33.3 50.0 56.9 66.7 90.3 100.0 Mean 24.93 29.50 St dev 13.066 13.204 Participants’ responses to the three items taken from the CUES measure enabled us to examine differences in satisfaction between those who were unemployed and those who were in work. Table 6.4 indicates that there was a small increase in the proportion of participants who judged their lives to be ‘as good as’ the three CUES statements between baseline and follow up interviews. At the individual level this was only statistically significant for social life (p=0.010). Closer inspection showed that this improvement was due to a significant improvement in social life perceived by people who were in employment at time 2 (p=0.004) Table 6.4 CUES comparative component, time 1 and time 2 % ‘as good as this’ Time 1 Time 2 Daily activities 67 68ns Relationships 66 70ns Social life 51 61** Table 6.5 compares the responses of participants who remained unemployed at time 2 with the responses of those who had moved into work. This analysis revealed that participants who had moved into work were significantly more likely to rate both daily activities and social life ‘as good as’ the CUES statements (daily activities p