The Scottish Government

« Previous | Contents | Next »

Listen

3. Supporting People - Approaches to Benefit Analysis

Introduction

3.1 Conducting a Cost-Benefit Analysis of Supporting People ( SP) calls for the identification, quantification and valuation of the impacts of the activities supported by the programme.

3.2 The fundamental aim of the Supporting People programme is described as follows on the Scottish Government website:

"Supporting People" is a new integrated policy and funding framework for housing support services introduced from April 2003. The aim is to provide good quality services, focused on the needs of users, to enable vulnerable people to live independently in the community, in all types of accommodation and tenure.

3.3 It follows from the above that the impacts and benefits of Supporting People should be sought first in terms of effects on the ability of beneficiaries to maintain an independent life in the community. There may be other, secondary, benefits and if these can be identified they should be brought into consideration.

3.4 In assessing the benefits of the programme it is useful to distinguish two stages in the assessment:

• The first stage is to establish and measure the impacts of the support on beneficiaries
• The second stage is to value those impacts.

3.5 The evidence on the impacts/benefits of SP and on the value of those benefits will be considered below. However, given the aims of SP, the focus will be on the impacts of the programme on the ability of people to maintain independent living and the consequences of failure to maintain that ability. "Failure" to be able to live independently may be absolute, with the alternative being some form of institutional care, or may be relative and imply a poorer quality of life or the existence of adverse effects of unsupported living.

Existing approaches to measuring the benefits of SP

3.6 As noted in Section 1, three other studies have been undertaken in the recent past which have sought to measure the benefits of Supporting People. These are the two studies undertaken by Matrix - one in England and one in Wales - and the study of financial benefits in England undertaken by Capgemini. These studies contain much analysis which is relevant to the present study and set out methodologies which could be applied - at least in part - to the analysis of SP in Scotland. The approaches used in the Matrix studies are reviewed below and conclusions drawn concerning their relevance to analysis of SP in Scotland. (As noted previously, the Capgemini study has not yet been published and, although the project team has seen an early draft, we are not able to offer more than a basic headline on what that draft reports.)

The Matrix Study of SP in England

3.7 The study of the benefits of the SP programme undertaken by Matrix Research and Consulting in England in 2005 was an attempt to produce a comprehensive analysis of the costs and benefits of SP. This study can certainly be regarded as seminal in that it provides the methodological foundations for the Welsh study undertaken by Matrix and for the Capgemini study.

3.8 The then ODPM (now DCLG) commissioned Matrix Research and Consultancy Ltd to undertake a study to identify and measure the benefits of the Supporting People programme across the range of client groups who receive housing related support. The approach taken by the Matrix study was an entirely desk based exercise focussed on reviewing the available literature and the analysis of data.

3.9 The Matrix study results provide an extremely valuable resource in that they contain much useful information on the characteristics of the population groups assisted by SP programmes and on what is known concerning the effects of these programmes. The Matrix study also devoted considerable effort to methodologies for estimating the value of benefits. At the same time, the study also provides lessons in terms of the limitations facing any attempt to produce a comprehensive cost benefit analysis of this programme.

3.10 Although the Matrix study data relate to England, the information and evidence on the effects of intervention and some aspects of the valuation analysis are applicable to Scotland.

Literature Review

3.11 The Matrix study involved a detailed review of the available literature on the impacts and benefits of the various types of project and activity funded by SP. According to the study report, the researchers examined over 300 documents. The review concluded that for most client groups there was reasonably strong qualitative evidence that the types of activity supported by SP programmes created benefits for clients.

3.12 However, a more critical finding of the review was that there was a lack of quantitative evidence on the effects of SP interventions. The report stated:

"One of the key issues identified early on through this assessment was the lack of evidence in relation to the quantification of impact of housing related support. The research team was unable to identify many studies that have used a scientific or quasi-scientific approach measuring impact of housing related support, as research in this area is generally qualitative in nature" ¹⁶

3.13 In effect, this meant that it was impossible from the existing research to derive evidence of the expected or estimated impact of SP programmes on any quantitative indicators. Thus, while it might be concluded that SP support made it less likely that an old person would have to move into residential care, there was no evidence to indicate how much less likely this outcome was as a result of SP.

3.14 As stated above, producing a figure for the value of output requires quantified estimates of impact and a basis for valuing those impacts. The Matrix analysis implies that this process is compromised by lack of quantitative evidence on impacts.

3.15 The fact that the review provided such limited insights appears to have been a key factor in determining the approach to benefits measurement or "modelling" adopted in the Matrix study.

The Matrix modelling approach

3.16 The Matrix study addressed the problem of lack of quantitative evidence on impact by adopting what may be called a "what if" approach. The study created a "baseline" for each client group by collecting a large amount of data on various aspects of the behaviour and experience of client groups - particularly activities which could be regarded as "adverse events" or which resulted from an adverse event. Examples of such activities were the need to seek advice or treatment from a doctor or other practitioner, hospital admissions, experience of crime (as victim or perpetrator) and experience of homelessness. An important feature of these events is that they are considered to give rise to financial costs for the public sector and, in some cases, for individuals. The critical assumption was that the incidence of these events could be reduced by SP programmes.

3.17 The relevance of this approach and of the baselines does thus rest to a large extent on the validity of the assumption that these indicators can be affected by SP interventions - a point which will be considered further below.

3.18 The Matrix study approach can be described by simple hypothetical example. If we consider programmes aimed at older people, one baseline might be that each "older person" in the population would be expected, on the basis of an analysis of relevant data, to spend an average of 2 days per annum in hospital. It might be argued that older persons receiving SP support would, due to the support, experience a reduction in the average number of days spent in hospital both because the support a) helped them avoid episodes which led to hospitalisation and b) enabled earlier discharge if they were admitted to hospital. If it was further assumed that the effect was that people receiving support from SP would, on average, spend 10% less time per annum in hospital than would otherwise be the case for similar people not helped by SP programmes, then a programme supporting 500,000 older people would be expected to bring about a reduction of 100,000 in the number of days spent in hospital.

3.19 The Matrix study applied this approach to a range of "indicators" across the full range of client groups. In general there was fairly sound evidence concerning the baseline level of activity for each indicator. Where there was research evidence on the likely impact of SP support on the baseline indicator then this was used. However, in a great many cases no such evidence existed so the approach adopted was to assume a more or less arbitrary figure for the reduction in the number or incidence of adverse events experienced by people receiving SP interventions as a consequence of those interventions. In many cases a figure of 5% was used - indeed, this became referred to as the "default" value.

3.20 The Matrix study thus estimated or hypothesised impacts of SP on a large number of baseline indicators. The next step was to put values on those impacts. In the Matrix study the values of impacts were mainly conceived of as "avoided" expenditure (mainly public but also private). ¹⁷ For example, if a particular group of people has a "baseline" of 100,000 hospital clinic visits per annum and if SP support means that the effect on the health of that group is such that clinic visits are reduced by 5% then the impact is 5,000 fewer clinic visits. If each visit costs the public purse £200 then the benefit of SP in relation to this group and to this effect is £1 million per annum.

3.21 The major analytical effort in the Matrix study was on the construction of very detailed baselines for client groups and in estimating the impact on public expenditure of predicted changes in the indicators used in the baselines ( e.g. fewer hospital visits, fewer crimes). In contrast, the assessment of the impact of SP on the indicators was, because of the lack of evidence, frequently based on an arbitrary assumption.

3.22 The Matrix study applied the above analysis and valuation of benefits to eight client groups:

• People with learning disabilities
• Older people
• Single homeless
• Homeless families
• People with mental health problems
• Ex-offenders and those at risk of offending
• Women at risk of domestic violence
• Drug users.

3.23 Support to older people is, in terms of resources and client numbers, the largest element of SP. It is useful, therefore, to examine in detail the Matrix approach to this group.

3.24 Support to older people living at home is often held to create a benefit in enabling people to "stay put" rather than to move into sheltered housing or residential care. Thus, if SP enables an older person to remain in his or her own home rather than move into residential care then the cost saving is the avoided cost of residential care (which can be very substantial).

3.25 It is easy to conceive how home based support of the type provided through SP could increase the confidence of someone about living independently and thus their ability to remain at home. This still begs the question of how much impact SP has on the ability of people to remain at home. On that point the Matrix study is clear - there is little hard evidence.

3.26 Moreover, the Matrix study went on to assert the existence of a number of expected or supposed effects of SP on older client groups which were much less clear cut. The study argued that for most older persons the SP generated cost savings involved such effects as reduced need for care from GPs or from other health professionals. The validity of such conclusions must be judged in the light of the evidence presented to support those conclusions - i.e. the evidence that SP projects do reduce the need for medical care.

3.27 With regard to older people, the Matrix study advanced what it termed five "hypotheses" concerning the impact of SP. These hypotheses were, in summary:

• That SP prevented or slowed the deterioration in the ability of clients to live independently - so reducing pressure on more intensively supported forms of accommodation such as residential care;
• That SP support helped maintain the health of clients in various ways - e.g. by preventing accidental injury but also by monitoring their condition - it being argued that this would reduce the overall use of medical services;
• That SP support helped older people avoid being victims of crime;
• That SP support helped prevent older people from becoming homeless;
• That by helping older people maintain their health, SP reduced the need for personal care from other social services.

3.28 To make this framework operational, the Matrix study had to undertake two specific tasks - a) to establish the validity of the above hypotheses and, critically, to quantify the impact of SP on reduced use of those services and b) to provide total and unit costs for the baseline levels of the various types of support and service whose use would be diminished or prevented if the above hypotheses were correct.

3.29 The Matrix study provides a great deal of information and data in relation to the second of these tasks and, indeed, it can be concluded that this task was largely achieved. However, the study was able to provide almost no information on the first matter and, in particular, produced only limited evidence concerning the extent to which SP projects could have the hypothetical effects.

3.30 In the face of this, the Matrix study made considerable efforts to quantify the existing use of various services by older people (the baseline) and then analysed the implied savings on the assumption that SP projects reduced by 5% the requirement for each service. This 5% figure was used for every service - e.g. it was assumed that 5% of old people receiving SP in their own homes would otherwise need supported housing and that without SP there would be 5% more visits to GPs and 5% more contacts with physiotherapists and so on. The report accepts that the 5% figure is essentially an arbitrary assumption.

3.31 During the course of the present study the research team have had the benefit of extensive discussions with staff involved in the organisation and delivery of SP programmes - though the case studies and other consultations. This level of contact was not a feature of the Matrix study.The present study team's view is that there is no clear evidence to support the view that SP projects reduce demand of medical services other than by reducing accidental injury and, possibly, enabling earlier discharge of people from hospital. It is not inconceivable that SP support may encourage people to pay more attention to their health ( i.e. to visit the doctor more readily). Alarm systems - a common SP supported activity - enable people to summon help more quickly, for example in the event of injury. It is likely that early access to help reduces the severity of and complications of such injuries but the effects are not readily quantifiable.

3.32 The Matrix study concludes, on the basis of the above analysis, that SP services to older persons produce £668 million of annual benefits. Well over half the benefit - £380 million is held to be from the reduced requirement for residential or sheltered care, an impact that involves just 41,000 persons (£9,300 per person). The remaining £280 million of benefit comes from such items as the supposed reduced need for health and other care and relates to some 840,000 persons (£333 per person).

3.33 The English cost data indicate that the mean SP cost per client for older people is £356 per person. At first sight, the Matrix results suggest some important conclusions - preventing people from moving into various forms of "care" shows by far the highest return while services which simply reduce demand for social and health services offer much poorer value for money.

3.34 The obvious implication of the Matrix analysis is that the potential for financial impact is greatest where the costs/benefits associated with a marginal change in the behaviour or status of service users are high and that this is the case with measures which may keep people out of care. For example, even if the assumption concerning the impact of SP on keeping people out of care is wrong by a factor of 10 - so that need for residential care or supported housing is reduced by only 0.5% - the benefit per user would still be £930. Benefit would thus still be 2.6 times cost. That SP has the potential to be highly cost effective by keeping people out of care is clear. However, there is no indication in the Matrix report as to whether or how SP funds could be targeted effectively on those persons most at risk of having to move into more supported accommodation.

3.35 Leaving aside people who are assumed to be "kept out of care" by SP, it is notable that most of the quantified benefits estimated for older people in the Matrix study relate to just two main elements.

3.36 The first - valued at £192 million - is based on the assumption that 20% of this group of older people would, if the SP funded support was not available, suffer a rapid deterioration in their ability to cope and would require a home care package costing £1,144 per annum. This represents an avoided cost of £194 million. No evidence is provided to support this assumption. Moreover, the rationale of the argument that a support service with a mean cost of a few hundred pounds a year is an effective substitute for a care package costing over £1,000 per annum is unclear.

3.37 The second main benefit element is reduced need for in-patient care. It is assumed that without SP there would be an increase in the number of old people admitted for in-patient care and also that availability of SP means that older persons can be discharged from hospital at an earlier stage. In both cases the arbitrary "5%" assumption is applied and the combined cost saving is about £58 million pa. Thus £252 million of the estimated cost saving of £280 million comes from these items. This works out at approximately £1,250 of benefit per person whose admission to hospital is prevented and a saving from early discharges worth about £82 per person who is admitted.

3.38 The benefit - in terms of cost savings - created by support to users whose need for residential care, a support package or in-patient treatment is not affected is conversely small. On the basis that there are about 270,000 people who receive SP but do not benefit from a reduced need for residential care, for intensive support or for hospital treatment then this group experiences benefits worth £22 million - £81 per person. Moreover, as stated earlier, a number of the supposed benefits from SP projects for older people are open to question. First, the assumption that these projects will reduce demand for various services by 5% is entirely arbitrary. As stated above, we consider that there is limited evidence to justify even the presumption of a reduction in use of medical services - as one of the hypothetical mechanisms of impact is better monitoring of the health of older people and advice on accessing services, it is at least plausible that SP projects may generate more demand for social and medical services.

3.39 Drawing these points together, we can see that almost all of the calculated benefits of SP for older people according to the Matrix research are either estimates of reductions in the need for some kind of institutional care (accommodation or hospital) or rest on the argument that many people would otherwise have to receive a similar for or support for some other source. In contrast, the estimated benefits to people who would not have moved to more supported accommodation or not required in-patient treatment or not have required a care package had SP not been available are very modest.

3.40 The significance of this segmentation of beneficiaries depends heavily on how far it is possible to identify the various types of beneficiary in advance - i.e. before resources are allocated. If it is possible to identify in advance those older people who are most likely to be candidates for or in need of highly supported accommodation or who are most likely to have an illness or injury which requires hospital admission then resources can be targeted on those groups. A possible implication of the Matrix report would that resources should be targeted on the most fragile older people. We will consider this issue further below.

3.41 We have discussed the Matrix results in some detail with regard to older people. Given that the core methodology is fairly consistent across groups, we can deal with the other groups more briefly.

3.42 What we have termed the "5%" assumption is central to the estimates of benefits for other groups. Thus the modelling of the "offenders" client group combines careful analysis of the costs of various services (including prison) used by offenders with the assumption that SP reduces the requirement for these services by 5%. With very few exceptions there is no evidence base for this "5%" assumption.

3.43 In the various client groups other patterns which appear in the analysis of the results for older people are also evident. In almost every case the financial significance of benefits related to reduced need for accommodation of some form or another - including social housing and emergency housing - far exceeds that of any other type of benefit.

3.44 The most complex analysis is possibly that related to support for women escaping domestic violence. A considerable body of evidence on the costs of domestic violence against women and children is presented. The analysis assumes that by accessing SP services, and thus escaping an abusive partner, women threatened by domestic violence reduce the risk of further assaults by 80%. This seems a reasonably well founded figure derived from British Crime Survey data which show that in only 19% of cases does domestic violence continue when a woman quits the home shared with an abusive partner. The benefit per case for this group is £110 million or £19,000 per client - twice the level of cost without taking any account of the reduction in pain and fear experienced by the women and children affected.

Implications

3.45 In the interpretation of the Matrix findings it is also important to recognise the problems created by the use of averages. In reality SP is a highly diverse programme and unit costs per client vary greatly between areas and projects.

3.46 Wide variation in unit costs is only explicable in terms of variation in the type and intensity of services provided. For example, one important distinction is between services which involve direct "one to one" contact and which are provided in some collective manner - particularly through the use of technology. A "technological" service such as a monitoring system which registers the presence or mobility of the supported person or which enables people to call for help is almost certainly bound to be cheaper than any one to one service.

3.47 The problem is that the benefits may be as divergent as the costs. We have seen that in the Matrix data a very large share of the quantified benefits is accounted for by small groups - e.g. people who do not need supported accommodation or to be admitted to hospital because of SP help. It may be calculated that in the Matrix study around 233,000 older people, or under 26% of older SP clients (people for whom SP means that they do not need more supported accommodation, people not needing to be admitted to hospital thanks to SP and people who would otherwise need a care package), account for 92% of the benefits.

3.48 As already indicated, the significance of these differences depends very much on whether groups with different likelihood of benefiting from SP can be identified in advance. The issue can be explained in terms of insurance. If only one in ten thousand people who insure their homes against fire experience a domestic fire then one would not conclude that all the benefits of the insurance premiums paid by one thousand people accrued to that one household. All of the people taking out insurance benefit from the mitigation of risk precisely because no individual knows whether he or she will suffer a fire.

3.49 If we cannot say in advance which households supported by SP will be the ones who will be enabled to avoid the need to enter residential care or hospital admission then the costs and benefits of the programme can only be considered at an aggregative level and the conclusion may follow that it is right to provide services to a large population even though most will never need to consider more intensively supported accommodation or be admitted to hospital. However, if it is possible to identify the "at risk" groups then different conclusions follow.

3.50 The insurance analogy is again relevant. The more accurately risk can be measured - both in terms of the probability of an event occurring and the costs of the event - the more accurately the benefits of insuring against an outcome can be assessed. Thus if we can identify - by some set of characteristics - people with a high probability of being unable to live independently without support, it makes sense to focus resources on that group. Conversely, providing support to people who are very likely to be able to sustain an independent life without that support may not be justified.

3.51 We know that, according to the Matrix study, the outcome of SP support is that in a year about 600,000 people out of 880,000 who receive SP enjoy benefits valued at little over £80 - under 25% of the unit cost of the programme. What is not clear is whether any practical way exists of "screening out" these people without denying services to the 280,000 persons who (according to the Matrix study) receive an average benefit of £2,200 per annum (seven times cost). All of this, of course assumes that the Matrix assumptions on the impact of SP are robust.

3.52 We have commented critically on the benefit assumptions relating to "reduced demand" for various out-patient health services in the Matrix work on older people. It worth noting that these impacts account for just £12 million out of the estimated £660 million of benefit. For older people, the only sources of really substantial benefit are reduced need for supported accommodation and the replacement by SP of services which would have to be provided in any case (though the assumption that these alternative services would cost three times as much as SP does not appear defensible). Together, these two elements account for 86% of the measured benefit.

3.53 As noted above, in most of the other client groups the greater part of benefit is accounted for by either the costs of different forms of housing/accommodation for people unable to live in the community without SP or the cost of services which, it is argued, would have to be provided in the absence of SP. The notable exception is the case of women escaping domestic violence where benefits are measured in terms of the "cost savings" of reduced levels of assault.

3.54 The final, and highly significant, point concerning the Matrix results is that quantified benefits are only "avoided costs". The report recognises that there are many unquantified benefits (or potential benefits) such as improvements in the quality of life of the client groups and their families but the quantified benefits tend to dominate the presentation of results.

The Matrix Study of SP in Wales

3.55 In 2005 Matrix undertook for the Welsh Assembly Government a study of the costs and benefits of SP in Wales. The operation of SP in Wales has its distinctive features - in particular there are two funding streams. The Supporting People Grant is administered by local authorities and is used to fund chargeable support services. This grant is used mainly for long term support to older people, people with learning disabilities and people with mental health problems. The Supporting People Revenue Grant is administered by the Welsh Assembly Government and is paid directly to Accredited Support Providers to fund non-chargeable services, usually short term in nature.

Methodology

3.56 The Welsh study followed the methodology of the English study quite closely. Aside from updating the literature review and the obvious need to use Welsh data, the Welsh study introduced one main new procedure. This related to the assumed or estimated impact of SP on the various baseline values. It will be recalled that the English study made extensive use of "default" values for the impact of SP which were often essentially arbitrary ( i.e. the assumption of a 5% reduction on baseline values of adverse effects).

3.57 The Welsh study sought to test and "validate" these default values through a process of consultations/surveys of Supporting People lead officers in local authorities, members of the All Wales Criminal Justice Group and provider representatives. The Stakeholders were shown the default impact values (largely based on the English Study) and asked both to comment on those values and to suggest alternatives. The report states that "stakeholders found it difficult to comment on the figures." ¹⁸ Despite this, the views of those consultees who did make a response were incorporated in the study report and had some influence on the actual modelling. The report notes that consultees still generally proposed higher, often much higher, impacts than assumed in the modelling exercise and so concluded that the modelling had been "conservative"

3.58 It is hard to know how much weight to place on the responses of the consultees, not least because the Matrix report provides little information on the quality or number of those responses. The report provides figures for 90 individual indicators of impact and it is impractical to reproduce all those results in the main text of this report. There are, however, some patterns which it is useful to highlight and these can usefully be considered in relation to the client groups used in the analysis.

3.59 Women Fleeing Domestic Violence In the original Matrix study modelled impacts for this group were notably high - Matrix arguing that a good evidence base existed to warrant these assumptions. The levels of impact suggested by consultees were generally in line with the modelled impacts other than that higher impacts were suggested for the number of visits to GPs, the number of outpatient attendances and the number of Community Health contacts.

3.60 People with Learning Disabilities Default values ( i.e. 5%) were used in the modelling of 9 out of 12 indicators on which there was general consultation even although consultees suggested much higher figures than those modelled for impacts on use of medical services ( GP, outpatient and mental health service support). The assumed impact on reduced length of stay in hospital was raised sharply (to 50%) from the default value even although consultees did suggest that the impact on length of stay in hospital for general hospital admissions would be less than actually modelled. In the light of consultations with a more restricted group of consultees the decision was made to add a substantial expected impact on reduced needs for residential care and adult placements as a result of SP interventions.

3.61 People with Mental Health Problems There were 15 indicators for this group and in 10 cases the consultees proposed substantially higher impacts than those used in the modelling. The differences between the modelled values and those proposed by consultees were greatest in the 8 out of 15 indicators where the 5% "default" value was retained by the consultants. These concerned impacts on attendances at and length of stay in hospitals and experience of crime. Consultees also argued for higher than modelled impacts on the need for residential care.

3.62 People with Alcohol Problems This group was not covered in the English study but was included in the Welsh study because of its importance in the programme. No consultation was undertaken on impacts and a set of arbitrary impact values was used - 5% effect for most indicators and 25% for effects related to homelessness.

3.63 People with Drug Problems Consultees proposed much higher figures for all the impact levels, as compared with the modelled figures. The modelled figures for impact were based on a default 5% reduction in "events" for all indicators other than those related to homelessness where the modelled impact was 25%. The impact levels suggested by consultees were generally four times the modelled levels except for the homelessness impacts where the mean suggested impact was just over 40%. The Matrix team evidently felt unable to place much weight on these views and thus retained the default values.

3.64 Young Single Homeless and Children Leaving Care There were 16 impact indicators for this group and in most cases the modelled impact was, at 30%, fairly close to the mean levels suggested by consultees. In three cases (impacts on re-offending, impacts on burglaries and impacts on the ability to secure permanent accommodation, the modelled impact was somewhat below the level proposed by most consultees, while for impacts on tenancy failures the modelled impact was at the maximum value proposed by consultees.

3.65 Ex-offenders The 5% default values were retained for almost all indicators for this group despite the preference of consultees for levels three to four times higher. Modelled impacts on homelessness (at reductions of 33% to 50%) were closer to consultee proposals.

3.66 Homeless or Potentially Homeless Persons The modelled impacts of SP on homeless were relatively high - around 30% for almost all indicators and these were close to the levels proposed by consultees. Only in the case of effects on re-offending was the 5% default retained though the reason for this - a lack of information on the baseline value - is arguably inappropriate.

3.67 Older People For this group the 5% default values were retained for almost all the indicators. Consultees had proposed very large impacts across a range of indicators - for example a 30% reduction in hospital admission rates, a 38% reduction in burglaries, the prevention of a 49% increase in the percentage of people with higher dependency levels among those presently receiving floating support. The matrix team evidently felt unconvinced by these proposals and cited a lack of evidence as grounds for retaining default values.

Results

3.68 The main departure of the Welsh study as compared to the original English work was to incorporate higher estimates of impact across most client groups. The consequence was that the assessment produced a clear excess of financial benefits ( i.e. reduced or avoided costs) over SP expenditure. Table 3.1, overleaf, reproduces the main results.

3.69 The most substantial excess of benefits over cost were for women fleeing domestic violence (as in the English study) and for people threatened with homelessness. The report emphasised that many benefits existed beyond "avoided costs" and that it could not be assumed that services where quantified savings were less than costs represented poor value for money.

Table 3.1 Costs and Benefits of SP in Wales

Source: Costs and Benefits of the Supporting People Programme, Welsh Assembly Government

3.70 Within the groups, particular elements of savings accounted for a high proportion of the total savings - a pattern also observed in the English study. For example, for women fleeing domestic violence the savings in the healthcare costs and costs of crime associated with severe acts of violence accounted for 96% of the savings. For the other groups the main elements of saving and the proportion of the total savings accounted for by those elements were as follows:

• People with Learning Disabilities - reduced need for residential care (80%)
• People with Mental health problems - cost savings due to reduced dependency (50%)
• People with alcohol problems - reduced homelessness (93%)
• People with drug problems - prevention of homelessness/reduced need for emergency accommodation (71%)
• Young single homeless/leaving care - reduced homelessness and reduced offending (80%)
• Ex-offenders - reduced offending and prevention of homelessness (99%)
• Homeless/potentially homeless people - prevention of homelessness and reduced offending (71%)
• Older people - reduced need for care (66%)

3.71 These figures reflect the pattern in the English study that most benefits (in cost saving terms) arise from a limited range of effects - notably reduced need for residential and other care and reduced homelessness. As with the English study, a very large part of the overall benefit is related to services to women (and families) fleeing domestic violence.

Previous Studies - Conclusions

3.72 Despite the limitations of evidence, the two studies produced many valuable insights. Thus, at least so far as the costs avoided approach is concerned, we can see that only by reducing the demand or need for public services which are expensive - institutional care/supported accommodation, in-patient care and, possibly social housing, can SP generate a substantial pay off.

3.73 The second main conclusion is that the effects of SP can be divided into two broad categories - SP support (or support similar in nature) is either essential to enable people to live in the community (so that the alternative is residential care) or it reduces the prevalence of "events" which have adverse effects and which generate costs.

3.74 In relation to the ability of people to "stay put" in the community, one uncertainty is the degree to which one can identify in advance those clients for whom SP is critical to their ability to live in the community. Obviously if one could identify those clients, there would be a strong case for giving them priority in access to SP support. However, it is possible that the effect of SP support is in many cases to reduce the probability that any individual's capacity to live in the community will deteriorate. Looking more widely, to the extent that SP support has the effect of, say, reducing the level of offending or reducing the incidence of repeat homelessness it is not likely to be possible to say which specific individuals will benefit.

Further Evidence on Benefits of SP

3.75 For the present study a further systematic literature search was undertaken. Use was made of the IDOX online library, employing search criteria based on the types of Housing Support Service that can be provided by Supporting People funding (see Appendix B for list of search criteria). Once retrieved from IDOX, the literature was reviewed and the findings from each document were entered into a database, categorised by type of service user. The search covered literature relating to the whole of the UK which had been published since the Matrix/ ODPM report. The findings from this literature review which have not been cited in earlier work are presented below.

3.76 This work echoed aspects of the Matrix analysis in that it revealed the existence of very limited evidence on tangible and quantifiable benefits achieved by the Supporting People programme in Scotland. The focus of the literature is on descriptions of the types of benefit that are expected to be achieved by the programme for service users in Scotland.

3.77 The study also undertook a review of Care Commission Inspection Reports for the four areas which were the focus of case studies. The aim was to identify benefits that have been achieved by Housing Support Services (many of which are in receipt of funding from Supporting People). This element of the literature search provided some evidence of the programme's benefits from the service user's perspective. As might be expected, however, the benefits described in these reports are largely unquantifiable. The findings from the Care Commission Inspection Reports are discussed below.

3.78 In general, the published and "grey" literature relating to Supporting People projects in Scotland is scarce. The literature that does exist on Housing Support Services is also patchy in terms of coverage and quality.

3.79 A summary of the key benefits identified during the literature search is presented below. The Supporting People client group to which the benefit relates is noted in brackets at the end of each finding. The benefits quantified in these findings relate almost entirely to costs that would be avoided if effective Housing Support Services were provided to the service user with the result of avoiding adverse outcomes such as - for example - a failed tenancy or episode of homelessness.

Housing Support Service Type: Assisting the service user to engage with individuals, professionals, and other bodies with an interest in the welfare of the service user

3.80 The evidence on the financial benefits of preventing homelessness is very varied. A recent study produced the estimates of the cost of homelessness for specific household types as summarised below ¹⁹.

3.81 The cost of one 2-year episode of homelessness for a young separated man with a history of mental health problems and drug abuse - which might be avoided if appropriate support were provided to avoid the failed tenancy - was estimated in one study as £83,000. The benefits accruing from avoiding this failed tenancy would be £3,000 to the landlord; £21,000 to local authority for temporary accommodation and £4,500 for support services; £40,000 to the health service; and £14,500 to the police and criminal justice service (client groups: people at risk of homelessness; people with mental health problems; people with a history of drug problems) ²⁰. It is important to note that this is not an "average" figure applicable to all recipients but an example of the costs which could be generated by an individual in what might be termed a "worst case" scenario.

3.82 The same study calculated the cost of one 1-year episode of homelessness for a single young male who left home after having problems with his family - which might be avoided if appropriate support were provided to avoid him becoming homeless - as £15,000. The benefits accruing from avoiding this episode of homelessness included £7,000 and £2,500 to the local authority for temporary accommodation and support services and £4,000 in the cost of unemployment to the general economy (client group: people at risk of homelessness) ²¹.

3.83 The cost of one 1-year episode of homelessness for a middle aged widower who lost his home through alcohol dependency - which might be avoided if appropriate support were available to prevent the episode, was estimated at £24,500. The benefits accruing from avoiding this episode of homelessness would therefore be £3,000 to the landlord in respect of an avoided failed tenancy; £10,500 and £2,000 to the local authority in respect of temporary accommodation and support services; £7,000 in health services cost; £1,500 in police and criminal justice services; and £500 in resettlement (client group: people at risk of homelessness, people with alcohol problems) ²².

3.84 The cost of one 6-month episode of homelessness for a young woman formerly in care and with rent arrears - which might be avoided if appropriate support were available to prevent the episode - was estimated at £4,500. The benefits accruing from avoiding this episode of homelessness would be £3,000 to the landlord in respect of an avoided failed tenancy; £500 in resettlement costs; and £1,000 to the economy in respect of unemployment (client group: young people leaving care; people at risk of homelessness) ²³.

3.85 The cost of one 6-month episode of homelessness for a young single man leaving an institution - which might be avoided if appropriate support were available to prevent the episode - was estimated at £7,500. The benefits accruing from avoiding this episode of homelessness would be £3,000 to the local authority in respect of temporary accommodation; £1,000 to the NHS; £3,000 in resettlement; and £500 to the economy due to unemployment (client group: people with mental health problems; people at risk of homelessness) ²⁴.

3.86 The cost of one 1_-year episode of homelessness for a young woman escaping domestic violence - which might be avoided if appropriate support were available to prevent the episode - was estimated to be £72,000. The benefits accruing from avoiding this episode of homelessness would be £23,500 to the local authority in respect of avoided costs of temporary accommodation and support services; £16,000 by the NHS; £15,000 by the police and criminal justice system; £1,500 in resettlement; and £16,000 to the economy due to unemployment (client group: women escaping domestic violence; people at risk of homelessness) ²⁵.

3.87 This literature gives examples of the benefits in terms of "avoided" public spending which could be achieved by Supporting People projects in this field ranging from £9,000 to £48,000 for preventing a year-long episode of homelessness, depending on the circumstances of the individual. The higher figures do, however, assume substantial effects on such items as crime and health. The benefit to a landlord of avoiding a failed tenancy is approximately £3,000; the remaining benefits are spread between other public agencies.

Housing Support Type: Advising or assisting the service user in dealing with relationships and disputes with neighbours

3.88 One of the main aims of SP support to young people is to help then to maintain tenancies and avoid thus avoid homelessness. Neighbour disputes are a common source of difficulties with tenancies. A Research study found that mediation can be very successful in preventing homelessness in young people, and quoted a prevention of homelessness outcome in between 45% and 96% of all cases referred. The research also indicates that schemes that link young people into wider support networks may be very cost-effective in the long term - but costs vary widely between schemes according to the types of service provided (client group: young people at risk; people at risk of homelessness) ²⁶.

Housing Support Type: Advising or assisting with resettlement of the service user

3.89 The potential gain from helping ex-offenders or those at risk of offending to become more settled and to avoid offending is very large. Most studies show that the costs of individual crimes - both to victims and to the justice system - are very substantial. What is less clear is how effective support has been in these respects.

3.90 A recent English project ('Prison Links') has helped 9 ex-prisoners remain in tenancies and has maintained 48 tenancies where prisoners are on remand. The project has solved Housing Benefit problems for 58 prisoners, and appears to have reduced the number of rough sleepers in the area (although no direct cause and effect link has been shown) from 20 to 3 ²⁷. What is less clear is how far resolving housing problems reduces offending.

Findings - Care Commission Inspection Reports

3.91 The benefits outlined in the Care Commission Inspection Reports are largely qualitative benefits to service users and their carers; however, they are reminders of how fundamentally important these housing support services are to their clients. The study reviewed the Care Commission Inspection Reports for Housing Support Services run in the main cities and towns in the Case Study Areas; Aberdeenshire, Edinburgh, South Lanarkshire, and West Lothian. Not all of the reports contained details of benefits to service users and their carers; those that did are detailed in the table below.

3.92 Again it is not clear from the reports which of these services are in receipt of Supporting People but the table excludes services which would certainly not be eligible for SP funding.

Aberdeen

Edinburgh

South Lanarkshire

West Lothian

Literature relating specifically to supporting people projects in Scotland

3.93 The literature search identified two pieces of literature specifically relating to Supporting People projects in Scotland. The first of these relates to the provision of Telecare services in West Lothian, and the second to a voluntary supported living service in Ayr.

3.94 West Lothian - the Telecare Experience: Research suggests that West Lothian Council has successfully integrated the use of Telecare and Assistive Technology with new service support mechanisms into a new Home Safety Service. Their core safety package consists of a lifeline unit, two flood detectors, two passive infrared detectors, an extreme heat detector, a smoke detector, and one pendant; these are connected to a support centre that is staffed all day every day. To date, the service has been rolled out to around 2,000 older and vulnerable people (including people with mental health problems, people with disabilities or illnesses such as epilepsy, and people experiencing domestic violence). Most of these service users have had the core safety package installed into their homes, but around 10% have additional technology to identify and alert carers to incidents such as epileptic seizures, falls, wandering, medication alerts, and bed and chair occupancy.

3.95 The cost of providing this service, spread over 5 years, has been calculated as around £7 per week per service user. Benefits are already being demonstrated as accruing from the service; for example:

• Delayed discharges from hospital have reduced from 70 (four years ago) to 20.
• The average length of stay in a nursing home has dropped from around three years in 2000 to around 16.4 months in 2004.
• The proportion of people over the age of 65 blocking beds in West Lothian is 1.4 per 1,000, compared with a Scottish average of 2.74 and a Lothian average of over four.
• The mean duration of length-of-stay by someone blocking a hospital bed is 30 days, compared with a Scottish average of 112 days.
• One Consultant has estimated that around 3,000 hospital bed days were saved in the first year of the service's operation, due to reduced admissions and early discharges.

3.96 Ayr - Voluntary Supported Living The Partners for Inclusion ( PFI) project is a voluntary supported living service, which was created with the aim of helping people with learning difficulties and mental health issues to move out of institutional settings and into their own homes. The project has helped 42 people (at the time of the article) using a person-centred and individual approach to supported living. While the article does not indicate any specific quantifiable benefits to the service user or his carers, it does provide evidence of intangible benefits in terms of maintaining the independence of the service user (he is 'ecstatic' in his own home, and is proud of it) and reducing the care burden on his family ²⁸.

Conclusions

3.97 As stated at the beginning of this section of the report, the assessment of the benefits of Supporting People requires that we can:

• Establish what changes in the behaviour or experience of individuals give rise to or constitute benefits
• Quantify the changes caused by SP programmes
• Place a value on those changes/benefits.

3.98 The evidence reviewed in order to address those questions consists of the Matrix studies, a further literature review as reported above and insights from the case studies, consultations and lead officers survey discussed below.

3.99 There is a fairly substantial body of qualitative evidence that support services of the type provided through SP provide benefits to recipients (and to society) in terms of enhanced capacity to maintain an independent life and in terms of an improved quality of life. Officers and organisations involved in Supporting People programmes and in working with client groups value SP and consider that it has very substantial effects on the quality of life, and independence of clients.

3.100 The strongest objective evidence of significant impacts relates to the effects of SP support in:

• Enabling people, especially older people and people with severe disabilities, to avoid the alternative of being moved into supported accommodation or residential care
• Reduced length of hospitalisation for older people suffering illness or injury
• Improved quality of life (reduced anxiety, improved mental well-being) for households of various types.
• Prevention of homelessness and its attendant difficulties.

3.101 For other groups - people with drug and alcohol problems, women escaping domestic violence and people with mental health problems there is certainly a widespread belief among providers that SP creates worthwhile benefits but the literature and research evidence is limited.

3.102 For all benefit types and client groups there is serious lack of quantitative measures of impact. The study of Telecare cited above is striking in containing substantial quantitative evidence - though the cause/effect relationship is not proven.

3.103 It is the face of this uncertainty that the existing studies have adopted what may be termed a "what if" approach in which the financial benefits of SP are estimated on the assumption (more or less well supported) of a given level of impact.

3.104 The approach to the valuation of benefits has almost entirely been in terms of "avoided" costs - though the significance of much of this information is contingent on the view taken of the effects of SP as discussed above. It is very clear that a reduction in the need for intensive forms of support - i.e. residential care - is the factor with the financial "muscle" to generate large financial returns to SP spending.

3.105 There is very little evidence on the scale of or value of quality of life impacts - which is not, of course, to imply that such impacts are not important and socially valuable.

3.106 We reflect on the appropriateness of all the assumptions discussed in this section in the comparative analysis of costs and benefits set out in Section 5 below.

« Previous | Contents | Next »