key: cord-0060058-xabvslvh
authors: Loeffler, Elke
title: Co-production in Health, Social Care and Public Safety
date: 2020-10-04
journal: Co-Production of Public Services and Outcomes
DOI: 10.1007/978-3-030-55509-2_4
sha: a39d97d823a558aac64fe7814d149af82f5419df
doc_id: 60058
cord_uid: xabvslvh

This chapter provides a systematic literature review on the extent of co-production in health, social care and public safety at the local and national level and its effectiveness. In particular, we explore in depth the potential and actual role of co-delivery in each sector with regard to the prevention, treatment and rehabilitation from social problems. Furthermore, we discuss the evidence of the effectiveness of co-commissioning, co-design and co-assessment in the three sectors.

In this chapter we explore how co-production actually operates in three sectors of public services-health, social care and public safety. These sectors have been chosen for two reasons-first, they cover a wide range of public services, each of them of high priority to many citizens. Second, they are sectors in which the author has conducted significant research and where there are important research results which can throw light on the strengths-and weaknesses-of co-production.

The chapter looks at co-production in each of these service sectors in turn. In each service sector we begin by exploring the role of citizen action in co-delivery. In order to explore in depth the potential and actual role of co-delivery in each sector, we first present a generic model of public intervention to deliver improved outcomes. The following sections then highlight how co-delivery can, in principle, contribute to each of these types of intervention, sector by sector.

In each sector, we then highlight some of the ways in which citizen voice, through co-commissioning, co-design and co-assessment can also play a role.

The key empirical evidence which we explore in each service sector are:

• The extent of co-production at national and local levels • Evidence on the effectiveness of co-production.

In this way, this chapter presents a much more holistic and joinedup approach to the actual and potential role of co-production in these services than has generally been presented in the literature. Finally, the chapter looks at the conclusions which emerge across these three sectors.

Public interventions are typically undertaken in order to tackle problems which have political salience, the solution to which will improve publicly desired outcomes. In Fig. 4 Loeffler and Bovaird [2019a, 249]) interventions comprise prevention, treatment and support for recovery or rehabilitation from the problem. For each of these main pathways, in turn, we can identify a number of potential contributing pathways. In order to prevent future occurrence of a problem, the public sector can either reduce the prevalence of the social conditions which make the problem more likely to occur; or it can promote behaviour change amongst those likely to suffer from the problem.

To improve treatment, it is important to detect the problem as early as possible and then to find appropriate treatments.

To improve recovery and rehabilitation, appropriate public interventions and community support have to be designed which reduce the negative consequences from problems and/or strengthen the resilience of people experiencing a problem. Then people suffering from the problem have to be convinced to follow the rehabilitation pathway.

In the following sections, we therefore consider, sector by sector, evidence relating to how co-production through citizen action contributes to prevention, treatment and rehabilitation/recovery. In each sector, we then consider how citizen voice, through co-commissioning, co-design and co-assessment can also play a role.

Prevention of health problems: It has long been recognised that people can make a very large contribution to their own health by behaviours which help prevent future health problems. Indeed, the role of citizens in preventing health problems is often much greater than the role of health professionals, even to the extent that some of these activities by citizens and groups of citizens can be seen as essentially self-help or self-organisation, rather than co-production with the professional health service. However, many prevention activities are triggered by public sector action, such as recent public campaigns about the importance of 'hands hygiene' to reduce the risk of passing on Coronavirus infection or, more generally, professionals advising people whose bad habits risk damaging their health to change their lifestyle. Other prevention activities may involve direct collaboration with public service organisations and citizens, as in the falls prevention initiative of the Aberdeen City Health and Social Care Partnership, which involves volunteers working as 'Falls Ambassadors' and health professionals co-delivering so-called 'Stepping Forward Together' sessions in local communities (Thompson and McConnachie 2019) .

In particular, paying attention to a healthy diet, avoiding overconsumption of alcohol and keeping fit are generally considered to be key to better health by public health services. Moreover, these contributions to better health are not simply achieved by individuals acting alone-a person's future health can also be improved by the support of others, both in their family and outside the family (e.g. in encouraging family or friends to eat a healthy diet, limit their alcohol consumption and take exercise).

An indication of how common individual preventative measures are in health is given by our survey of five EU countries, which shows that over 50% of citizens interviewed reported changing to a healthier diet and trying to take more exercise during the previous year (Loeffler et al. 2008, 18) . Rather more serious attention to health prevention, as in getting a health check from a doctor, was reported by over 30%. In the subsequent Australian rerun of this survey, two-thirds of respondents reported trying to exercise, while over half reported changing to healthier diet and almost a half had seen a doctor for a health check (Alford and Yates 2016, 165) .

These results appear a little more optimistic than some UK government research which suggested that in 2008, 70% of adults were not meeting at least two of the government guidelines which were then in place on smoking, alcohol use, diet or physical activity (Buck and Frosini 2012, 13) . Moreover, this UK research was even less optimistic with respect to citizens with lower education and from lower socio-economic groups, who were more likely to fail to meet three or four of these guidelines. Given recent claims by the NHS to be focusing more on prevention, it might be expected that this picture would be improving over time. However, more recently in 2015, 48% of the England population were still not meeting the daily recommended guideline of consuming five portions of fruit and vegetables per day (and this was 57% in the lowest health life expectancy areas), despite a national health campaign (ONS 2017, 55) . The urgent need for more action by citizens to prevent health problems arriving is therefore evident.

Health treatment: Co-production of better health treatment is often called 'self-care'. Self-care has been defined by the Department of Health as the care taken by individuals towards their own health and well-being, in which DoH includes the care extended to the individual's family and the community (Department of Health 2007). People who are practising self-care can also gain from the support of 'experts by experience', since citizens who have suffered from a condition can be particularly helpful in educating and giving skills training to people who have recently contracted the condition and to professionals who need to know how best to support people who are practising self-care (Hairon 2007) .

Self-care comprises two key areas-detection and reporting of illhealth; and contributing to the successful healthcare intervention. In both of these areas, the role of professionals remains strong-but in both areas there is nevertheless scope for significant citizen input. In particular, in the case of highly infectious conditions such as Coronavirus, self-reporting to the NHS and strict adherence to self-isolation is key to containing the disease. Furthermore, early detection of conditions such as breast cancer by performing self-examinations, in combination with other screening methods, significantly improves likelihood of later successful professional interventions.

There is little quantitative evidence on either of these elements of co-production of better health treatment. However, a case study of Highlands Hospital in Jönkoping in Sweden provides evidence of significant improvements to mortality rates when patients are helped to identify and to report more quickly and more clearly on potentially dangerous 'flareups' in their condition (Bovaird and Tholstrup 2010) . Self-monitoring of health conditions can be substantially improved by use of a growing range of technologies, including strap-on apps, such as MyFitnessPal, Fooducate and Instant Heart Rate, which monitor body functions in ways which were previously only available in doctors' surgeries or hospitals. For example, a study of patients with congestive heart failure who used telemonitoring (a weekly nurse phone call to patients, with automated telephone questionnaire) found that over one year there was an 83% decrease in hospital admission rates for patients (Department of Health 2007).

In health care, self-treatment (e.g. through drug injection or dialysis) has become much more common in recent years. Nevertheless, for all the discussion of the potential of self-treatment, its level still appears to be much lower than it might be (Patel and Patel 2014) . Although 90% of people in the UK use over-the-counter medicines to manage minor conditions without going to their GP, and in 2010 973 million over-the counter medicine packs were sold by pharmacists, this was still less than the 1028 million prescription items dispensed (Proprietary Association of Great Britain 2011). Moreover, it has been estimated that GPs currently spend an hour a day (57 million consultations per annum) seeing patients with minor conditions that could be self-treated-and this has been estimated to cost the NHS £2 billion per annum (Proprietary Association of Great Britain 2010). Indeed, according to the Department of Health, 39% of GP consultation time is spent treating patients who present with self-treatable minor ailments (Colin-Thome 2004, 11) . Could this be changed? In 2007, over 90% of people were interested in developing selfcare skills and over 75% believed they would be more confident if they had support from a professional or peer (Department of Health 2007) .

Looking at people living with chronic obstructive pulmonary disease, the Centre for Reviews and Dissemination (2014, 2) summarises the results from three systematic reviews into the effectiveness of self-care as demonstrating that multi-component interventions reduce respiratoryrelated hospital admissions and improve quality of life for people.

Another way in which self-care can be co-delivered is through trained community health volunteers (CHVs). In a very different context, namely rural areas in Kenya, a recent study into training CHVs to deliver integrated preventive and curative packages of care to manage common childhood illness in hard-to-reach communities showed that the proportion of CHVs exhibiting appropriate skills to examine for signs of illness improved from 4% at the baseline to 74% after 6 months of training and the proportion of care-givers who first sought treatment from a CHV increased from 2 to 31% (Shiroya-Wandabwa et al. 2018) .

However, despite these potential benefits, there is evidence that healthcare professionals do not fully explore the potential of self-care. In 2005, a national survey in the UK found that more than 50% of patients who had seen a healthcare professional in the previous six months had not been encouraged to develop self-care skills and one-third said they had never received any advice regarding self-care (Department of Health 2005).

Rehabilitation/recovery from health problems: there are three key coproduction activity areas in rehabilitation and recovery of health-coproduction activities which aim to improve and recover general health, activities aimed to continue (either short-term or long-term) conditionspecific treatment, and activities aimed to protect against recurrence of the condition. The first of these co-production activities is similar to the set of prevention activities which we covered earlier and the second of these activities is similar to the set of treatment activities covered earlier.

However, the third set of co-production activities, which are about the contribution of patients to preventing the occurrence of their health problem may involve some different kinds of inputs. Perkins et al. (2012, 2) define recovery (in a mental health context, but this definition applies more generally) as follows: "It involves making sense of, and finding meaning in, what has happened; becoming an expert in your own self-care; building a new sense of self and purpose in life; discovering your own resourcefulness and possibilities and using these, and the resources available to you, to pursue your aspirations and goals." This immediately highlights the centrality of co-production to the recovery process, with its emphases on a person's own interpretation of the position, the role of self-care and the importance of resourcefulness.

A key role in rehabilitation and recovery is often played by people with similar conditions giving peer support. Mead (2003) defines peer support as a system of giving and receiving help, founded on key principles of respect, shared responsibility and mutual agreement of what is helpful. Summarising systematic reviews of the role of peer support in mental health recovery, Repper and Carter (2011, 400) conclude that "What [peer support workers] appear to be able to do more successfully than professionally qualified staff is promote hope and belief in the possibility of recovery; empowerment and increased self-esteem, self-efficacy and self-management of difficulties and social inclusion, engagement and increased social networks. It is just these outcomes that people with lived experience have associated with their own recovery; indeed, these have been proposed as the central tenets of recovery: hope, control/agency and opportunity."

The Health and Social Care Partnership of Aberdeen City has recognised the potential of peer support for better self-care and has supported co-delivery of more peer support with and for people living with diabetes in Aberdeen. This included the funding of a training course for volunteers who wished to develop skills on how to run peer support groups and provide more effective peer advice. As one volunteer, who had already been working with Diabetes Scotland previously, states: "I learnt a lot about my condition from talking to other people who were living with diabetes as well as health care professionals. The knowledge I gained really helped me self-manage and live well. … I had received a lot of support from the charity [Diabetes Scotland Aberdeen] and other people living with diabetes. I wanted to volunteer and give something back" (Dave Curry, quoted by Murray 2019).

For people living with long-term conditions (including arthritis, asthma, cancer, cardiovascular disease, diabetes, HIV/Aids or pain), Hairon (2007) summarised a Department of Health study which reported that a significant improvement in coping was achieved by multicomponent self-care interventions, such as behavioural therapy, coping strategies and support groups. The Department of Health therefore encouraged the development of psycho-social self-care programmes to increase coping for people with long-term conditions, which would help people deal with stressful situations and improve well-being and quality of life.

Co-commissioning of health care: Co-delivery of prevention, treatment and rehabilitation/recovery are essentially part of the co-production of health care by service users and their communities. However, there are also important roles for citizens in the other three Co's-cocommissioning, co-design and co-assessment-all of which entail the use of citizen's voice. We look first at approaches to co-commissioning of health care.

• Personalisation-a form of micro-commissioning, through which health care patients are able to determine (within predefined guidelines) how the funds available for their health care should be spent, including prioritisation of the outcomes which they seek to achieve (Musekiwa and Needham 2020) . The number of people in England with personal health budgets has been rising year on year since they were launched in 2014, with nearly 23,000 people receiving one in the first nine months of 2017/18 (Madsen 2018) . They have been particularly available to people with long term health conditions but also increasingly to people with ongoing mental health needs, learning disabilities and autism. The personal health budgets initiative was proposed in an official government strategy in 2008 and re-affirmed in a 2010 White Paper, Equity and Excellence (Department of Health 2010). An evaluation of the personal health budgets pilot in England reported that they were associated with significant improvement in patients' care-related quality of life and psychological well-being after 12 months, although they did not appear to have had an impact on health status, mortality rates, health-related quality of life or costs during this period (Exworthy et al. 2017) . Revealingly, qualitative research suggests that it takes time (perhaps a year or more) for health service staff working with patients who have a personal health budget to overcome their initial reluctance to cede control and respect the 'expertise by experience' of patients (Musekiwa and Needham 2020) . A systematic review (Coulter et al. 2015 ) of a range of personalisation approaches to health care summarised the results of 19 randomised trials published before 2013 covering conditions such as diabetes, mental health problems, heart failure, kidney disease, and asthma, finding that involvement in personalised care planning probably led to small improvements in some indicators of physical health (better blood glucose levels, lower blood pressure measurements among people with diabetes, and control of asthma) and also probably reduced symptoms of depression, and improved people's confidence and skills to manage their health. However, it found no effect on cholesterol, body mass index or quality of life. It appeared that the process worked best when it included preparation, recordsharing, care co-ordination and review, involved more intensive support from health professionals, and was integrated into routine care, which indicates the key role of a co-productive approach. The study concluded that personalised care planning was promising and could lead to better health outcomes but that more research was needed into which aspects are most effective for specific patient groups.

• Participatory budgeting (PB)-i.e. giving citizens or groups of service users the chance to decide collectively on priorities and budget allocations between key outcomes sought, services or projects to be delivered. As discussed in Chapter 3, PB has been actively implemented in many countries and in many services since the 1980s. It is often undertaken at neighbourhood level and sometimes focuses specifically on specific groups. The Department of Health (2017, 34) asserts that: 'Making progress on our priorities and addressing the challenges the NHS faces over the next two years cannot be done without genuine involvement of patients and communities". However, at least in England, there have not been many PB examples which have focused specifically on priorities in health care or health outcomes, as a national evaluation shows (Department for Communities and Local Government 2011, 42) . Examples which have taken place include the prioritisation of health care projects by local communities in the neighbourhood of Thornhill in Southampton, following a pilot in 2008 (Department for Communities and Local Government 2011, 60-62) . Similarly, in the PB process in Paris, health was one of the local issues where residents could propose projects and, since PB started in 2014, seven health-related projects have been given high priority in the voting and therefore been implemented-but this was only a low proportion of all projects chosen (DICOM 2020). Clearly, in spite of the positive examples which have occurred, PB is far from becoming embedded within health commissioning processes in most OECD countries-and certainly not in the UK. Indeed, a recent summary of the state of play in English health system planning (the so-called 'Sustainable Transformation Plans' or STPs) concludes that patients and the public have been largely absent from the STP process so far-partly because of the limited time available and partly because national NHS bodies had asked leaders to keep their draft STPs out of the public domain (Exworthy et al. 2017) . Moreover, despite worldwide implementation of PB, a recent review of all the academic evidence finds that the implications of PB for health and wellbeing have not been the focus of attention in public health literature and there is therefore little reliable evidence (Campbell et al. 2018 ). • Appointing service users and other citizens to procurement panelsthis gives citizens a direct role in the choice of providers. For example, when the local NHS England public health commissioning team completed a procurement exercise in Cornwall for school-based vaccination services, they brought in a group of 12-14 year-old students, who actively participated in presentations from bidders and asked very challenging questions, which enabled commissioners to see how providers communicated with their service users and also resulted in the successful provider making some changes to the service delivery models and future service user engagement. Although the students didn't score the procurement bidders, their feedback on the presentations and Q&A session was fed into the final scoring and decision-making about the successful bidder (NHS 2017) . A variant of this approach was used by Bedfordshire Clinical Commissioning Group, where stakeholder panels comprised of service users and care-givers met with the shortlisted bidders for commissioned health services and asked questions around the contents of the bid and how the bidders would deliver the new services, with their views being fed into the procurement process (Hoyle 2014 ).

• Co-planning of services -e.g. deliberative fora such as Citizen Assemblies. A UK variant of this approach was NHS Citizen, which has the overall aim of ensuring that people and communities have an increasing say in health policy development, and how NHS services are commissioned, designed, and delivered. A programme of activities was initiated in 2012/13, including the design of democratic participation methodologies. Over four years, more than over 4000 people were involved from across the country, in both face-to-face and online conversations (NHS 2020). As part of its work, NHS Citizen in 2014 ran a full-scale Citizens Assembly, which consisted of around 200 patients, care-givers, activists, volunteers, third sector and public services workers, coming together in an informal setting with the NHS England Board to discuss agenda items prioritised by citizens. Participants had volunteered, with most having had a history of active participation in the health system, for example as employees or as patients and care-givers. They split into five 'issue groups' which had been selected from over 80 potential issues submitted in advance. After the detailed work in each issue group, all participants and the NHS England Board reconvened in a plenary, where two representatives of each group highlighted the key insights from their discussion and heard a response from a Board member to each group, with ensuing discussion. The organisers described the conversations as rich and constructive, although more time would have been needed to cover the broad scope of issues and to form clear solutions (Local Government Association 2016). A simpler version of co-planning is exemplified by an initiative involving five full-day meetings of a 28-member citizen panel convened to establish hospital restructuring priorities in Ontario (Chan and Benecki 2013) . A survey of participants found they were enthusiastic about the experience (although some patients were anxious about the magnitude and complexity of the task); participants thought the panel had accomplished something important of benefit to the community and the hospital, and that the citizen panel was an effective approach. Moreover, the hospital board approved nearly all the recommendations from the panel, including the closure of 26 beds and two outpatient programmes, contributing to a budget improvement over two years.

• Co-financing of services -an approach which demonstrates a strong commitment to the value for money of public services. The simplest and most direct way this can be done is through payment of a charge for a service. However, as this is usually seen as a coercive mechanism, rather than a voluntary activity by citizens, it is generally not included within the category of 'co-production'. Voluntary financial contributions to public services, projects or infrastructure, however, do constitute co-production. Recently, a more general approach enabled by the internet has emerged in the form of crowdfunding, where a large number of small contributions are pooled to support a specific initiative, a practice used both in the private and public sectors. For example, the London Borough of Lewisham (Hilton and Blake 2011) successfully used crowdfunding to support its annual fireworks celebration, as discussed in Chapter 3. Furthermore, crowdfunding has proven especially effective and popular for supporting self-organisation with charitable causes, particularly around health-many top-earning online crowdfunding campaigns are for medical causes, including campaigns to pay directly for an individual's medical care or to accelerate research on and access to experimental treatments, particularly as sophisticated search tools allow campaigns to locate potential donors and also allow these donors to find campaigns of interest to them (Young and Scheinberg 2017) . However, even simple approaches can be very successful, as Captain Tom Moore demonstrated with his campaign to raise over £30 m for the NHS during the coronavirus crisis (see Chapter 3). While, for the most sympathetic cases, medical crowdfunding can save lives, it likely to be of little help for those cases which evoke less sympathy from typical citizens (Snyder et al. 2016) .

Co-design of health care: Interesting approaches to bringing citizens into the co-design of health care include:

• Designing communication tools (websites, flyers, newsletters) with experts by experience-for example, Diabetes UK worked with young people with type 1 diabetes in Scotland in the co-design of the website JUST.DUK.1T (NIHR 2016). The website gives young people the facts on key subjects about which they need to know-advice on leaving home, going out, travelling abroad, and risks associated with driving, tattoos, eating disorders, drugs and alcohol. • Designing preventative activities with people accessing services and care-givers-e.g. in the Aberdeen City Health and Social Care Partnership, a falls prevention initiative was initiated as part of a Governance International Coproduction Star training programme (Thompson and McConnachie 2019) . After a few meetings between the involved professionals and older people who had experienced a fall, the group made the joint decision to take the falls prevention message to people 'in their own spaces' (lunch clubs, coffee groups, and community clubs). Within less than a year the Stepping Forward Together co-production initiative had grown from two initial volunteers to 14 so-called 'Falls Ambassadors'. They have co-delivered falls prevention sessions to eight community groups. According to Thompson (2019) , who supports this initiative as a Senior Occupational Therapist, "this amounts to over 200 participants who have heard about services, tried some strength and balance exercises, listened to their 'peers' talk about the falls journey they have made and taken away resources to inform and educate them regarding falls prevention." • Designing service improvements -e.g. through Experience-Based Co-Design (ECBD), in which experiences are gathered from patients and staff through in-depth interviews, observation and group discussions, identifying key emotionally significant points, which are then summarised in a short edited film shown to staff and patients, who then explore the findings in small groups together, identifying and implementing activities that will improve the service or the care pathway (Robert et al. 2020 ). Up to 2014, at least 59 EBCD projects had been implemented in six countries covering clinical areas such as emergency medicine, drug and alcohol services, cancer services, paediatrics, diabetes care and mental health services (Donetto et al. 2014) . In an international survey of EBCD users in 2013, 90% of users who responded said that "it really engaged patients", 78% that "it really engaged staff", 54% that "it led to clear improvement priorities" and 51% that "it really made a difference to the way we do things around here" (Donetto et al. 2014, 28) . At the same time, almost 50% of respondents considered that the main weakness of the co-design approach was that "it took too long" (Donetto et al. 2014, 5) . Another example of a comprehensive service redesign process the in Madrid region is the development of personalised medication schemes with elderly patients who need to take multiple medications, so that they don't forget to take them or get confused on what should be taken and when (Gil and Parrado 2012) . Personalised medication plans are co-designed by health professionals and patients, incorporating regular reviews by local primary care centres and pharmacies. To ensure compliance with the personalised medication plan, health visitors and pharmacy staff continue to engage with the patient in several follow-up activities. In the first four years after the start of the programme, more than 100,000 patients benefitted from the programme, with more than 1000 pharmacies participating (35% of all pharmacies in the Madrid region), with very positive improvements in the health of patients. By the end of the second year of the programme, the proportion of patients aged over 74 who were at high risk of adverse effects from their medication (according to the so-called 'Beers criteria') fell from 16.3% in 2006 to 14.4% in 2008. A survey of the 127,206 patients engaged in the programme in 2011 (with response rate of 64%) indicated that 91.6% of respondents knew how they were supposed to take their prescribed medications, 92.4% knew the prescribed dosage and 95.37% exhibited good compliance (indicated by the Morisky test) (Gil and Parrado 2012) .

However, a more sobering indication of the role of co-design in health care is given by the Q Initiative, a long-term UK health initiative supporting individuals and their improvement work, which started in 2015 and by 2018 had grown to include more than 2500 people in a multi-strand initiative, delivered in partnership with NHS Improvement and organisations from across the UK. In this largescale improvement community, only 3% of members were 'patient leaders' or 'experts by experience' (Pereira and Creary 2018, 11) .

Co-assessment of health care: Interesting approaches to bringing citizens into the co-assessment of health care include:

• Filling in surveys on health care-although this is not always a highintensity activity, many surveys do involve considerable time and even thoughtful and imaginative responses on the part of service users or other citizens, in which case they qualify as a form of co-production. For example, since 2007 the UK has conducted an annual GP Patient Survey and published the results, showing patients' experience of GP healthcare services, including access to GP surgeries, satisfaction with opening hours and use of out-ofhours GP services, with over 800,000 returns, a response rate of around 40% (IPSOS MORI 2017). This survey has had an impact on government policy and practice-the national regulator of health care quality has used patient experience scores as part of its performance assessment of primary care provision and the Care Quality Commission includes patient experience surveys as an indicator of standards to be expected of all GP practices. The Department of Health has used the data collected by patient surveys, including the GP Patient Survey, to review the needs of marginalised groups such as ethnic minorities, and to establish teams to respond to GP practices and Primary Care Trusts with the lowest scores (University of Manchester 2014). • Patient-led or community-led health surveys or research-for citizens, this is much more intense as an activity and can be a particularly valuable form of co-production. An interesting example from Italy is provided by a survey led by the third sector organisation Cittadinanzattiva to enable patients to assess rehabilitation services in their region to inform health care policies of the National Ministry of Health. The questionnaire has been co-designed by patient organisations and health care professionals (Cittadinanzattiva 2019). • Web-based rating of health services -this has been a valuable offshoot in recent years from the rapid growth of health care websites. In the UK, all GPs had to register on a website after 2008, so that every visit could be rated by service users (Smith 2009 ). However, this platform was dropped by the new coalition government after 2010. Subsequently, a number of public sector and independent web-based platforms were set up in England, such as NHS Choices, Patient Opinion, and iWantGreatCare, some of them on mobile phone apps such as the iPhone-based Great Care (Patel et al. 2015) , so that users could rate their experience with a general practitioner (service, hospital, dentists, and other health care services). It was hoped that this would give patients a voice so that there could be improvements in transparency and the quality of care. The use of online patient ratings for GP surgeries was later reintroduced by the Care Quality Commission as part of their inspections (CQC 2013). In the USA, there is a range of websites which publish ratings of health care organisations, including the websites of government organisations such as Medicare ('Hospital Compare'), CDC and the Agency for Healthcare Research and Quality (AHRQ). Interestingly, the website WebMD (WebMD, n.d.) in its section on 'How to Use Online Ratings for a Hospital', advises that "information from users on a web site is the least reliable", although it gives no evidence for this statement. (2013) Another is: How are we doing? Compliments, comments, concerns. These booklets, which also appear elsewhere throughout the hospital, explain in user-friendly language how to raise a concern or complaint and include relevant forms. The hospital promises to acknowledge complaints within three days and answer them within 25 days. Recent examples of changes in clinical practice arising from feedback include the redesign of the patients' care pathway in A&E and new procedures in the gynaecology department for women suffering miscarriages. • Citizen or patient inspectors -In October 2014, the Care Quality Commission started a comprehensive programme of inspections of GP practices, often including in the inspection teams an 'Expert by Experience' (someone who uses a GP practice or has a particular experience of this type of care). As part of the inspection, teams spoke with patients to assess the quality of care in a practice (Care Quality Commission 2017a). • Peer review of health services -An example of peer reviews with 'Experts by Experience' is provided by Sheffield South West Primary Care Trust, where a panel of patients and carers was trained to be actively involved in clinical audit and service improvement projects through a clinical audit patient panel (Challans 2008 ). The subsequent evaluation of its impact concluded that the panel was able to assist the Trust in developing and improving services within primary care, although it faced significant challenges in changing the culture towards actively involving patients and carers. • Another form of peer review of health care in the UK is given by Health Impact Assessment (HIA), involving stakeholders and experts (including 'Experts by Experience') who may be affected, involved in the implementation of, or have specialist knowledge of the ways in which policies, programmes and projects impact on the health and wellbeing of the population (Chadderton et al. 2008 ). However, a review of practice in Wales suggests that: "In all of the HIAs used as part of this research, levels of public participation compared to that of the statutory and voluntary sector were low, with on average three or four members of the public involved in each one (excluding the community initiated HIA)" (Chadderton et al. 2008, 82) . Nevertheless, they concluded that "The majority of the [members of the public] who were interviewed as part of this research reported that they had found their involvement in the HIA to be a positive experience", while "public sector representatives interviewed as part of the research focused on the fact that it is members of the public who are affected by the issues or projects relating to the HIA, that the proposed changes would take place within their communities, and that they held the knowledge and value of personal experience to be able to effectively inform the HIA, and highlighted that these positive contributions outweighed any of the more problematic issues" (Chadderton et al. 2008, 81) .

Summing-up, the available evidence suggests that the extent of coproduction with patients and the public in health care is still very limited, at least in the UK. While there is a lot of talk about the need to focus more on prevention, most health care providers only engage with citizens when they require treatment and devote few resources to preventative coproduction approaches. Furthermore, the evidence from the UK shows that only a small minority of healthcare professionals help their patients to learn effective self-care practices.

At the same time, there have been a number of high profile NHS publications and policy-papers with a commitment to give patients and citizens a 'voice' to improve public services and outcomes. However, most co-commissioning, co-design and co-assessment initiatives which are reported in the literature are still not widespread. One partial exception is Experience-Based Co-Design which, at least, in the UK has been applied in a wide variety of health care contexts (Robert et al. 2020) , although admittedly only involving a small minority of NHS patients as yet.

Some authors suggest that the lack of individual and organisational health literacy has been one of the main obstacles to the implementation and limited effectiveness of co-production in health (Palumbo and Manna 2018) . Of course, the lack of (access to) relevant information and the use of inappropriate language are persistent problems encountered by service users of the NHS. Dealing with this miscommunication could involve education of both patients and staff. However, at least in most OECD countries, health professionals have already received formal training, so the priority may be rather to enable them to tap into the tacit knowledge and resources of 'experts by experience'. Finally, there is strong evidence that the culture of health services still privileges the technical knowledge of 'expert staff' over the everyday knowledge of 'experts by experience', and that this cultural barrier only recedes slowly, even where co-production in health is being practiced (Musekiwa and Needham 2020).

Tackling social problems through social care covers a wide spectrum of interventions. A useful taxonomy of these interventions is provided by the Social Work Theory and Methods Comparison Table of the Social Work Learning Forum (n.d.). This identifies eleven different theories and methods of social care intervention (see Table 4 .1) and all of them involve some contributions from people who use social care. Two clear lessons emerge from Table 4 .1. First, these theories and models are, in principle, appropriate to all phases of social care intervention, including prevention and treatment of social and psychological problems, and rehabilitation/recovery from such problems. However, this theoretical position has to be contrasted sharply with current social Assumes people can cope with change User co-production Although crises interrupt normal coping mechanisms, they provide opportunity to improve skills and resilience User co-production A period of disorganised thinking and behaviour is to be expected Crises can reawaken unresolved issues from past-but can also offer chances to correct non-adjustment to past events User co-production Cognitive-behavioural approach/rational emotive behaviour therapy Based on the assumption that thoughts, beliefs, images and attitudes influence our behaviour and if these are changed, our behaviour will change.

User co-production Involves worker helping service user to identify, challenge and reframe 'unhelpful beliefs' User co-production Can involve modifying behaviour using rewards and incentives User co-production Involves using such approaches as DEF (Dispute beliefs, replace beliefs with Effective rational belief, describe the Feelings which will result) Motivational interviewing Worker adopts an empathic and non-confrontational, but nevertheless directive, approach Worker is alert to language used, particularly looking for language of change Worker provides information relevant to the problems of the service user Worker encourages service users to list benefits and costs of lifestyle and alternative lifestyles User co-production (continued) Encourages the person to describe their life in their own words User co-production Opportunity to tell their story, in the process defining their identity User co-production Supports the person to feel in control of the narrative User co-production Draws attention of person to possibility of a different narrative for the future User co-production Source Adapted from Social Work Learning Forum (n.d.) work practice in the UK, especially in recent decades. Here, social work budgets have become ever more constrained in relation to the level of problems presenting to the public sector, so that social work has often focused much more on treatment, and to a lesser extent on rehabilitation/recovery, with less attention to prevention. This has meant that co-production has often been more evident in the treatment phases and, to some degree, in rehabilitation-which, in turn, suggests that the potential benefits of co-production might currently be highest in regard to prevention, if public policy were to give prevention a higher priority.

Second, much social work theory and practice focuses specifically on the individual ('the service user') rather than on wider social relationships. Consequently, where co-production is promoted, it is mainly individual co-production by the service user, rather than collective co-production arising from inputs from the wider community. Indeed, consideration of the potential role of community inputs is largely missing from Table 4 .1this aspect of co-production has not been fully taken on board by most approaches to social work. It is much more characteristic of community development (see Russell 2020) which generally distinguishes itself sharply from social work.

A detailed taxonomy of behaviour change interventions across behavioural domains has been provided and tested by Abraham and Michie (2008) . These interventions all support self-management of adults living in the community (although without specialised histories in relation to the target behaviours, so excluding adults with known mental or physical health problems).

It is clear from this taxonomy that few of these behaviour change interventions rely on highly professional expertise or specialised trainingonly a maximum of six interventions seem to fall into this category. Consequently, 20 of the interventions seem potentially appropriate for co-delivery with service users.

Moreover, although this is intended to be a taxonomy of how care providers can help 'self-management', where co-production from the service user is directly relevant, it is clear that some of the interventions are likely to be benefit from support by members of the service user's network or of the wider community.

We can clearly see from Table 4 .1 that, in principle, there is enormous scope in 'social work' interventions for contributions from service users and members of their networks and communities. We now turn to consider what is actually happening in practice. As with health in the previous section, we now consider the role of co-delivery separately in problem prevention, treatment and rehabilitation/recovery before we consider citizen voice within co-commissioning, co-design and coassessment.

Prevention of social problems: The kinds of problems for which social care is appropriate can be triggered by a very wide variety of external circumstances, most of which are at least partly outside the control of any individual. Preventing these problems therefore requires a correspondingly wide range of inputs. In particular, as well as the support which social care professionals can give, there is a need for a wide range of contributions on the part of the person concerned.

It is also important to recognise that many of the influences which can give rise to mounting social problems (e.g. poor housing, poor health, poor educational attainment, low employability) can only be solved by other public service professionals, not necessarily by social workers. Prevention of social problems therefore has to be seen as a task for joinedup public services, not for social care alone. Moreover, this means that citizen co-production to prevent social problems arising can, and should, occur within all the relevant public services, not simply in social care. This further means that social care services need to have the capacity both to signpost people who are at risk of developing social problems to other services which can help them to avoid such problems, and must also have the skills to work closely with these other services, so that all services are focused on achieving the social outcomes which are desired.

In practice, most social care services only become applicable once a problem has already arisen and a person has been referred for care. (This has become increasingly the case as social services have suffered from resource cuts under 'austerity' budgets). Prevention approaches tend therefore to be under-developed. What is often understood by 'prevention' in social care is prevention of existing problems becoming worse, which we will deal with under the heading of 'rehabilitation/recovery'.

Looking at the wider role of prevention, namely preventing potential social problems developing to the level where they cause actual problems for a person or their 'significant others', it is clear that the role of the person concerned, and their network, in preventing these problems becoming significant is generally much more important than the role of social care professionals. Indeed, some of these activities by citizens and groups of citizens can be seen as essentially self-help or self-organisation, rather than co-production with the professional social services. When self-help or self-organisation works well there may be no need for investing extra public sector resources through co-production. However, in contexts where self-help or self-organising is not sufficient or not effective, co-delivery by professionals may enable much bigger improvements of outcomes.

However, there is clearly a chicken-and-egg issue here-because there are usually too few social care workers to deal with current demands for social care (at least in the UK), only high needs cases can generally be accepted for intervention, with the result that support to prevent problems from developing is under-financed. However, this means that much larger numbers of people will present with major social care needs in the future, although these needs might have been nipped in the bud by earlier intervention through co-production.

This suggests that the role of co-production in social problem prevention is under-used and, indeed, under-recognised. One challenge may be the lack of resources (or willingness to reallocate resources) for problem prevention. Another challenge for social care staff is to how reach out to citizens to encourage them to participate in co-produced prevention initiatives, where they have not yet been referred to social care services. In some cases, the target groups are even hard to define, where the question arises whether co-production should focus on the wider group of people who are a risk or on people already experiencing specific, if still mild, problems. Predictive analytics are likely to inform this debate much more in the future and enable public service commissioners to gather more evidence on the effectiveness of co-production related to prevention in social care.

Where only low levels of professional social care can be devoted to prevention activities, then there should be a high priority on those codelivery initiatives which are likely to achieve a big improvement in outcomes for important target groups.

This is perhaps most vividly demonstrated in relation to the huge social problem presented by isolation and loneliness. Even fifty years ago, it was recognised that a significant proportion of admissions to elderly residential care were for people who did not need the specialist care available in those homes but who were so isolated and lonely in their own homes that residential care seemed the easiest solution (Wager 1972) . In 2016/17, 1.4 m or about 8% of adults over 50 in England reported often feeling lonely (Age UK 2018) . As Age UK (2018, 10) argues: "Tackling loneliness is about building communities with the social and physical infrastructure that can help build resilience; ensuring widespread awareness of and access to organisations, activities and support; creating neighbourhoods that are welcoming and feel safe; enabling people to identify, work with and develop tailored support for lonely individuals. Social activities are a part of this, yet alone they are insufficient."

As this analysis clearly suggests, inputs are needed both from professionals and citizens, with community inputs being particularly important. However, this has been an area where public sector interventions have traditionally been sporadic and patchy. Many initiatives have indeed been undertaken to tackle loneliness and social isolation and a very useful taxonomy from Windle et al. (2011) includes:

• One-to-one support-e.g. through befriending and mentoring programmes • Group-based services-e.g. day centres, social group schemes for people with shared problems (e.g. bereavement or specific health conditions) or interests (e.g. craft or cultural activities). • Wider community engagement-e.g. participation in sport, use of libraries and museums, volunteering, using and joining outreach programmes, using and contributing to timebanks-nowadays often promoted through 'social prescribing' or 'community connector' schemes (Age UK 2018, 13) .

As can be seen, almost all of these initiatives have considerable elements of co-production. Unfortunately, there is very little evidence available on the overall picture of how many lonely and socially isolated people are benefitting from such initiatives, although there is a large literature on specific case studies across the UK and internationally. One of the few insights into the volume of such approaches is provided by a public health study in Kent County Council into the group of older people (over 50) at risk of social isolation and loneliness (as suggested by their socio-demographic characteristics). Of this group, only 10% reported contacts with social care services and only 25% reported contacts with professional community health workers (Abi-Aad and Kennard 2018, 24). These proportions were higher than for the overall population of older people, where the figures were 5 and 19% respectively. However, this highlights that both social care and community health care only get to a minority of older people. (This analysis omits, however, third sector inputs which support public sector services, policies, and outcomes). This concern is reinforced by other research which shows that 17% of older people are in contact with family, friends and neighbours less than once a week and 11% are in contact less than once a month (Victor et al. 2005) . However, there is no research yet which explores the reach of the large number of initiatives by public services which have attempted to tackle loneliness-this remains a serious gap in the literature.

Other prevention initiatives at local and national levels with elements of co-production include interventions such as 'Early Help' programmes and Local Area Coordination (LAC). The former includes different kinds of support provided by professionals to a family when a problem first emerges, with the objective to improve the outcomes of children or young people. According to evidence gathered by the Department for Education (2018) and the Early Intervention Foundation (2018) it is more effective to provide early help when problems first arise than to intervene later. The LAC community connector scheme uses a preventative, personcentred approach toward developing individual and family resilience by offering one-to-one support with a Local Area Coordinator who connects people with local communities, networks and services (Billingham and McEleney 2016; Broad 2015 ). An external evaluation by Derby University of initial LAC schemes in the local authorities of Thurrock and Derby found "diverted or saved costs between £800,000 and £880,000 in the first two locations, in the first 10 months" of the scheme (Broad 2015, 40) .

Treatment of social problems : Given that prevention of social problems has been rather starved of resources in recent times, at least in the UK, most of the theories of social intervention in Table 4 .1 have traditionally been interpreted as approaches to treatment of existing social problems, rather than ways of preventing them.

A useful taxonomy of social work treatment approaches is provided by Popple and Leighninger (2011)-see the first two columns of Table 4 .2.

In the third column of this table, we also highlight some ways in which co-production by service users can contribute to these different treatment approaches. While Popple and Leighninger (2011) focus mainly on one-to-one working, we draw out in the fourth column of the table the potential role of community co-production in all phases of social work intervention. If the social work intervention is successful, social care services may become obsolete, which would end the cycle of co-production as outlined in Table 4 .2.

As with prevention, there is very little empirical evidence on the extent to which local communities contribute to the kinds of social care and support activities which are outlined in Table 4 .2.

However, some interesting evidence comes from a co-production study in Germany, which surveyed over a thousand citizens about the level of support which they gave to either older people or to young people to meet their needs-see proportion of respondents do engage in these support activities, if only a few times per year-although only a relatively small proportion (typically 20% or fewer) engage in them at least once a week. The Figures also show that, both in the case of care activities for older people and for young people, many people provide such support for people outside their family, contrary to suggestions often made that such caring is relatively selfish and largely devoted just to family members (and many people provide such support both for people inside and outside their families).

In terms of lessons on co-production, the fact that so many people do engage in these activities at least occasionally suggests the potential for public services to increase the level of community support for young and older people outside their families. However, a rather different picture emerged when we probed the extent to which people engaged in co-production with their local authority in order to improve the wellbeing of young or older peoplesee Fig. 4 .4. Here, the percentages of respondents who say they have engaged in some way with their local authority on issues connected with older people or young people is, at maximum, 14% and, on some issues, only around 7%. This is where co-production at the local level is evident and these figures suggest that in social issues it is not high.

There is therefore a very large gap between these levels of coproduction and the overall levels of support, through self-organisation, which people give to older and young people outside their families. Fig. 4.4 The extent of co-production between citizens and local authorities to improve the quality of life of young or older people (Source Adapted from Löffler et al. [2015, 8]) This suggests that there is major scope for the public sector, at least in Germany, to tap the energy and commitment of citizens to improve the social life of the more needy of their fellow citizens-but this scope is not currently being tapped. It seems likely that a similar gap-and therefore similar potential for further co-production-also exists in other European countries.

Rehabilitation/recovery from social problems: By their very nature, many social problems are long-term or even permanent, so 'recovery' is generally not an appropriate description and 'rehabilitation' is more about learning to live with the problems, rather than diminishing them. The issues posed by these aspects of social care therefore have much in common with those arising for people living with long-term health conditions (which have been partly dealt with already in Sect. 4.3).

As with health, a key co-production approach to rehabilitation which is widely promoted by social work is peer group support. In Sect. 3.4 we have already given some examples. Empirical evidence supports the potential value of peer group approaches-for example, in relation to depression, Pfeiffer et al. (2011, 35) report the pooled results from randomised controlled trials as indicating that "peer support interventions improve depression symptoms more than usual care alone and that the effects may be comparable to those of group cognitive behavioural therapy". More generally, Davidson et al. (2018, 1) suggest that paid peer support has been around since the birth of psychiatry in the late eighteenth century, and (citing the Institute of Medicine) that various forms of peer support can be found in virtually every branch of medicine dealing with chronic conditions, from asthma and cancer to diabetes and hypertension. They go on to stress that peer support does not 'treat' mental illness-rather, it complements clinical care by instilling hope, engaging patients in self-care and health services, helping them to navigate complex and fragmented systems, and promoting their pursuit of a meaningful life. In a rare cost-benefit analysis, Willis et al. (2018) concluded from their study of peer support groups for people living with dementia that the groups had created social value ranging from 1.17 to 5.18 for every pound (£) of investment. People living with dementia experienced increased mental stimulation and less loneliness and isolation, while care-givers reported less stress and burden of care.

Given the fiscal austerity regime which has characterised the public sector in many OECD countries since 2008, savings in public service costs have been seen as particularly important by many governments. The effect of co-production on costs have been explored in a number of studies. A particularly interesting study by Knapp et al. (2010) estimated cost savings from several co-production initiatives in social care and health. According to their estimates, time banks could bring savings and other economic pay-offs of over £1300 per member, with administration costs of under £450 per year. Befriending schemes, typically costing about £80 per older person to administer, resulted in reduced need for treatment and mental health support, with savings of around £35 per older person in the first year alone. Local community organisers working with hard-to-reach individuals typically cost up to £300 per person but could have economic benefits (e.g. less time lost at work, savings in welfare payments) of £900 per person in the first year alone.

Co-delivery of prevention, treatment and rehabilitation/recovery are essentially part of the co-production of social care by service users and their communities. However, there are also important roles for citizens in the other three Co's-co-commissioning, co-design and co-assessmentall of which entail the use of citizen's voice. Interesting approaches which have been used to bring citizens into the co-commissioning of social care include:

• Personal budgets (often with direct payments) for people eligible for social care, a form of micro-commissioning whereby service users can determine their own priorities (within limits) for how the budget should be spent. There are already many countries using this approach. In 2014-15 88% of UK personal social services users had a personal budget from their local authority, and 22% were receiving direct payments (NAO 2018, 4) . NAO concluded that the evidence from users showed most, but not all, benefited from having a personal budget. Paradoxically, however, there was no link at local authority level between the proportion of users with personal budgets and overall levels of user satisfaction (NAO 2018, 12) . Moreover, there is debate about the extent to which personal budget holders actually play a significant role in deciding how the budget should be spent-it is clear that, in some cases, service staff play a strong role in influencing such decisions or local authorities may not even tell recipients how much money is in their personal budget, so some of the co-production benefits are not gained (Slasberg 2017) . Moreover, a recent survey suggests that the choice that personal budgets are designed to offer is rapidly disappearing, as local authorities clamp down on how this money can be spent, increasingly requiring it to be used only for 'personal care' in the narrowest sense (Abrahams 2017). • Participatory budgeting (PB) to decide priorities in the social care budgets of public agencies (see also Sects. 3.1 and 4.1) has been growing in importance in recent years-for example, in Scotland there are now over 190 PB initiatives, many of them in health and social care (see https://pbscotland.scot/map or Escobar 2020), although there is a debate as to whether many of these initiatives actually contribute to 'core' services provided by local authorities (O'Hagan et al. 2017 ). An interesting case study is provided by Aberdeenshire Health and Social Care Partnership, which since 2015 has tested out participatory methods to engage with communities in budgetary decision making. With support from Scottish Government, £200,000 from the Integrated Care Fund budget was distributed for preventative health and wellbeing approaches through a participatory budgeting process, originally in two communities but then extended in a further 6 PB rounds. One digital element of this programme involved over 2300 participants, who generated over 240 ideas (Democratic Society 2019). However, it should be borne in mind that many of these initiatives take a very wide view of 'wellbeing', so they often involve activities which are quite far away from conventional activities of health or social care. • Identifying social care priorities with specific groups of people who have differing social care needs. This includes approaches such as Citizen Assemblies-for example, a group of 47 representative citizens brought together over two days to prepare a report for the Health and Social Care Committee of the UK House of Commons on how to fund adult social care (Involve 2018). • Giving service users and other citizens a role in the procurement of social care. There are interesting examples from around the world of involving citizens in the public procurement process (see Bohorquez and Devrim 2012). This often involves trained service users or other interested citizens contributing to the selection of service providers or even staff recruitment. For example, in Surrey County Council young people worked with local councillors and staff to select the providers of local prevention services for young people (Bovaird and Loeffler 2014, 22) . • Co-planning of social care services -e.g. deliberative forums, 'Planning for Real'. As an example, in Western Australia, the non-profit organisation Therapy Focus included two people with a lived experience of disabilities on the Board of Governance in order to hear the voice of 'people with lived experience' where decisions had the greatest impact and established a Parent Reference Group with ten parents of children with disabilities who advise the top managers on policy reform and clinical practice (Barrows 2017 ). • Co-financing of services -e.g. crowdfunding. Social service users are often charged for their use of public services, unless they fall within certain categories of disadvantage, and this remains the case today for social services in the UK. However, this typically more resembles a market transaction than a case of co-production. Furthermore, social services delivered by charities are often funded by charitable donations. More recently, citizens generally have also been able to fund social services through a rapidly increasing range of internet platforms offering crowdfunding opportunities. The success of crowdfunding appears to derive from its ability to get the public excited about specific campaigns in a way that a public service could not if simply requesting the public to make direct donations to it as an organisation. There are three main types of crowdfunding, meeting different needs-donations-based, peer-to-peer, and equitybased (which we will not consider here as it is not generally connected to citizen co-production).

-Donations-based crowdfunding involves people funding a project, product or business, without any expectation of seeing their money again. Many local authorities are already promoting this type of crowdfunding successfully to generate funding and some (like Dorset and Plymouth) use it to supplement their own funding for a range of activities, often increasing community engagement and allowing local residents instead of the council to decide which organisations most need grant funding. A different model was used by the London Borough of Lewisham in 2015, where it invited local community groups to bid for a pot of £100,000 by uploading their proposals on the Spacehive website and seeking community monetary pledges-the council then pledged up to £10,000 towards the most popular projects (Glover 2017, 8) . If the target amount of funding is not reached by the deadline, no money is taken from those who have pledged. If the target is reached, donors are often given a small token of appreciation related to the project or activity supported. An interesting variant of this from Alex Tabarrok, known as the dominant assurance contract, involves a core funding offering all those who make a pledge a small payment if the overall target is not reached, so they have a win-win proposition of either funding something they believe in or making a small monetary gainthis is likely to increase the success of crowdfunding efforts quite significantly-see Green (2017) . -Peer-to-peer (P2P) lending, which is more investment than lending, bypasses banks by matching a borrower directly with an individual or organisation with money to lend, allowing higher returns to the lender, and a faster, more flexible (and perhaps cheaper) process for the borrower. Local authorities can use this method to invest in social enterprises, local charities or other organisations, while private citizens can use it to invest in projects from other members of their community. Local authorities could also act as a facilitator rather than an investor, by brokering relationships between local people and community organisations needing funding and those willing to lend, although it is more likely that they will simply act as navigators to the online platforms which is doing the matching process. Of course, many different organisations can provide the underlying platforms which support P2P lending-for example, Triodos Bank, a pioneer of sustainable banking was the first UK bank with a crowdfunding platform enabling people to invest directly in organisations seeking to deliver positive social and environmental impacts (Social Enterprise UK 2020).

Interesting approaches to bringing citizens into the co-design of social care include:

• Designing communication tools (websites, flyers, newsletters) with experts by experience-an interesting example is given by the local authority of Stockport in the UK, which redesigned its adult social care website with a sample of its service users, giving rise to a much greater flow of website visitors, reducing the number of easy-toanswer questions being posed to staff by telephone and face-to-face and, in consequence, saving the council around £300,000 p.a. in staff time (Wells 2011) . A further example is given by the local dementia network in East Dunbartonshire, which enabled people living with dementia to decide on the frequency of its newsletters and on what stories to feature, tell the stories in their own words, and help distribute the newsletters (Brown et al. 2016, 20) . • Designing care plans with service users and care-givers-writing individual care plans is one example of where social care users work with professionals to co-design approaches which will prevent problems from occurring or, at least, from getting worse. For example, in the non-profit Therapy Focus in Western Australia, the therapy team engages in a conversation with service users to find out what they can do and what they need help with. Advice on peer-to-peer service options is now a key element of the Therapy Focus Strategic Plan to co-design services that people need and want (Barrows 2017 ). • Neighbourhood redesign of services and facilities with community groups to improve social outcomes (e.g. Design Labs). An evaluation of 15 neighbourhood-based projects in England to build community capital and help prevent need for future social services concluded that the community can indeed contribute much but that such initiatives "take considerable time to establish and grow, and the scale of development is often extremely modest", requiring investment in support and infrastructure (Henwood 2012, 6) , warning also that community capital can be fragile and reliant on one or two charismatic individuals.

Interesting approaches to bringing citizens into the co-assessment of social care include:

• Surveys completed or led by service users or communities-the evaluations which emerge from such surveys give service commissioners and providers information on which they can base service improvements. For example, the annual national Adult Social Care Survey in England and Wales asks people who are over 18 and who use adult social care about their experiences and how these services are helping people to live safely and independently in their own homes (NHS Digital 2019). Other, more local surveys are often run by service users or communities in order to evaluate local services and check if locally desired outcomes are being achieved. Often assessment methods are not as formal as surveys, comprising such activities as focus groups, gathering of stories from service users and carers, etc. A recent study by Omeni et al. (2014) conducted a cross-sectional survey of service users and professionals within community mental health services over three mental health trusts in England covering 4.5 million residents, of whom around 220,000 were in contact with mental health services. Over 56% of respondents indicated that their local authority social care department involved service users in evaluations of their services-this was, by far, the most common way of involving service users. Furthermore, 64% of those respondents involved in evaluations reported that this involvement had a positive impact (Omeni et al. 2014 , Table 5 ). • Action-oriented complaints systems-as Simmons and Brennan (2013) highlight, "complainants may work alongside service providers in the co-creation of ideas, co-production to put those ideas into practice, or co-partnership to promote the extension of those ideas and practices". They quote an illustrative case study, Dad's Story, where one woman who had a very negative experience of having to navigate health and care services on behalf of her father, a person living with dementia, told her story through a video, created by the Dementia Services Development Centre in Stirling, which was then widely used in training for staff in the NHS and others with responsibilities for dementia care (Simmons and Brennan 2013, 34) . • Citizen inspectors -the drive towards greater transparency in public services (e.g. through the publication of online budget information) has provided citizens with the opportunity to make requests under 'Freedom of Information' legislation. However, as Mckenna (2020) points out there is not a lot of evidence that citizens have become 'armchair inspectors' of public services. One exception is in the co-assessment of public services, including social care services in the Scottish local authority of West Lothian in 2011 (Kelly 2012) and expanded in 2015. The inspections carried out included a coassessment by trained citizen volunteers of a housing service for people with high support and care needs. • Peer review of services with ' experts by experience' -as noted earlier (Sect. 3.6), the Care Quality Commission (2017b, 6), the regulator of health and social care in England, recommends providers "to listen and act on the views and experiences of people who use their services and check on how well they do this". For example, the Customer Engagement Team of Warwickshire County Council recruited and trained 23 Peer Reviewers in 2011 to co-assess the quality of commissioned services, five of whom had learning disabilities and focused specifically on services used by people with learning disabilities (Hawthorne 2013) . Each Peer Reviewer was supported by a Customer Engagement Officer and a member of staff from Changing Our Lives, a local non-profit organisation, supporting social care users. During the process, these support staff gave prompts if reviewers got stuck, but ensured that the Peer Reviewers led the reviews. While none of the reviews uncovered safeguarding or serious concerns, they did highlight various areas across services that could be improved and also some areas to be celebrated as good practice. Following the pilot, providers made regular requests for input from these Peer Reviewers, while the commissioners concluded that Peer Reviewers provided additional information to the qualitative data collected by the Contract Monitoring Team and gave a unique insight from the customer's and carer's perspective. However, it has to be recognised that there is still a lack of evaluations of such 'peer reviews' and, indeed, some local authorities in the UK have discontinued their 'peer review' programmes.

Summing-up, the evidence on the role of co-delivery of social care in problem prevention, treatment and rehabilitation shows a considerable discrepancy between the potential of co-production, based on key social care intervention theories, and the actual extent of co-delivery in this service sector. The comparison of co-production in problem prevention, treatment and rehabilitation pathways in social care also suggests that, at least in the UK, most co-production in social care is still focussed on problem treatment, and to some extent, rehabilitation. While there is evidence of the effectiveness of co-production in problem prevention initiatives, there is little research regarding the extent of such schemes in terms of people reached. In particular, there has been little focus on support of self-management in social care so far. It remains to be seen whether digital technologies in social care will provide professionals with tools to enable their service users to do more self-management within a co-production framework or whether telecare apps will simply promote a shift from a largely professional-led service provision to pure self-help. As far as the extent and effectiveness of co-commissioning, co-design and co-assessment in social care is concerned, the available evidence suggests that, while there is a wide variety of such initiatives, they are often time-limited and remain small-scale in their reach. One exception is personal budget holding in social care (although not all forms of personal budget may be considered as genuine micro co-commissioning). A key barrier to co-commissioning in the social care sector is that needs assessments play a key role in deciding the eligibility of a person for social care services. These needs assessments largely determine which services are commissioned for that person. As a result of reduced budgets in social care, many needs assessments are now only done at a point of crisis-and at this point many service users can contribute little, which reduces the scope for both effective co-assessment and co-commissioning with service users, their family and social network.

This strong link between needs assessments and services commissioned applies also to children and young people who are already in social care. While the benefits of placing children and young people at the centre of assessments is well documented in research (Helm 2011) and the meaningful involvement of children and young people in local authority decisions regarding their care and support has now become a legal imperative in the UK (Bennett et al. 2016, 20) , many social workers find this challenging. Furthermore, it is widely recognised that in addition to children and young people, parent carers and wider family networks should play an important role within assessment. However, in practice, this co-assessment is not always effective. As Bennett et al. (2016) suggest "…although generally the primary carer will be interviewed as part of the assessment process this does not always ensure that parent carers and wider family members feel that their views are listened to or that their experience of the process is positive". 4.5 Co-production of Community Safety Outcomes

Policing and the criminal justice system have two main aims-first to help the community to feel safe, both by reducing crime and anti-social behaviour and by directly tackling unjustified fear of crime; and second, to achieve justice in the community. These aims are both communitywide, helping to improve quality of life in the community for citizens in general but also stakeholder-specific, attempting to help those most intensively involved, i.e. victims and offenders. In this section we look at how co-production can help to achieve these aims. It needs to be said right at the outset that, while the literature contains some evidence that co-production may often be effective in achieving high level outcomes in community safety, the relevant number of studies remains too small for full confidence (Loeffler and Bovaird 2019b) . Moreover, only a narrow range of outcomes are typically addressed in much of the research into community safety production-in particular, much research has focused on crime reduction, perhaps because this is usually a government priority. Consequently, the effect of co-production on the collective outcome of justice in the community and on quality of life outcomes is still under-researched. In addition, many research studies to date have been purely qualitative, illustrating the potential of coproduction but giving only weak indications of the strength of its drivers and impacts.

In Loeffler and Bovaird (2019b, 5) we present an overall model of how different outcomes are brought about by activities in policing and the criminal justice system. This model also highlights a range of coproduction activities which contribute to these pathways. Most of the co-production activities in policing and criminal justice (and this is also true of most of the more detailed examples reported below) mainly involve forms of co-delivery. Other modes of co-production such as cocommissioning, co-design and co-assessment are still relatively rare in this field. This section will summarise the findings of an extensive literature review in Loeffler and Bovaird (2019b) on the extent to which different types of co-delivery impact on outcomes in policing and the criminal justice. Later sections will then consider evidence relating to cocommissioning, co-design and co-assessment of policing and the criminal justice system.

One critically important driver of crime reduction which is highlighted in Loeffler and Bovaird (2019b, 5) is outside of the influence of policing and the criminal justice system, namely the set of 'social causes' of crime (Behn 2014 ) and therefore is not considered further here.

In Loeffler and Bovaird (2019b, 5) we highlight a range of pathways to cut crime through reducing social causes of crime (not further explored here), reducing opportunities for crime, deterring crime, encouraging desistance and removing criminals from the community in line with the model presented in Fig. 4 .1. The first two of these are often labelled 'crime prevention' but actually all five approaches can help to prevent crime. The scope for citizen contributions is particularly strong in reducing opportunities for crime-e.g. through encouraging safe behaviour such as locking all doors and windows when leaving the house. Some of these activities can be seen as pure self-help-but where they have been encouraged or supported by public safety campaigns or other public service initiatives, which is often the case, then we can validly classify them as coproduced. In surveys of citizens in five EU countries, Loeffler et al. (2008, 18) found that a high proportion of respondents in an EU survey claimed to take such steps (e.g. well over 80% reported taking care to lock doors and windows and around 40% kept an eye on their neighbour's house and asked them to do the same). However, our literature search highlighted that relatively few community crime prevention initiatives have been evaluated.

Interestingly, in the EU survey, at the very bottom of the responses on prevention activities was 'seeking advice from the police on safety issues' (Loeffler et al. 2008, 18) . Only 5% of European citizens often asked the police for advice on how to best protect their property, while 14% sometimes do so. UK citizens were most inclined to make use of this free service provided by the police, whereas Danish and Czech citizens were the most reluctant (Loeffler et al. 2008, 18) . The Czech case is particularly interesting-only just over 1% of Czech respondents in the survey often contact the police for crime prevention advice. As the survey also showed, Czech citizens felt relatively unsafe in their neighbourhood, while national crime statistics showed that property-related crimes made up 70% of all crimes in 2004. The Czech focus group suggested that the most important barrier to more active involvement of citizens in community safety was historical: "Most people still feel that the police are not a friend and are not serving the citizen, but rather are a repressive power" and, although a range of government initiatives have been launched to rectify this, trust in the police is still low.

Similarly, opportunities for citizens to contribute to deterring crime are important, especially through citizens alerting the police to potentially criminal activities. In the USA, Groff et al. (2013) found that foot patrols in a controlled experiment achieved a 23% reduction in violent crime in Philadelphia, working on a model which provides more proactive community contacts and more community intelligence. Foot patrols reacted more consistently than car patrols to signs of social disorder in their patch, in line with the 'broken windows' hypothesis that foot patrol officers, by their presence and through tackling social disorder, can (re-)establish normative order in a community and encourage greater informal community control.

Neighbourhood Watch schemes provide another example-in England and Wales they cover 3.8m households with 173,000 volunteer coordinators, and in the USA they cover over 40% of the population. A systematic review of Neighbourhood Watch schemes around the world (Bennett et al. 2008) concluded that Neighbourhood Watch schemes were associated with a reduction in crime of between 16 and 26% (with 19 studies indicating it was effective in reducing crime and only 6 producing negative results).

In Germany, police in the County Council of Mettmann have been training elderly people to provide advice on public safety issues to other elderly people (Löffler et al. 2015, 31) . The training puts strong emphasis on the enabling role of these 'security advisors'-they do not behave as 'experts' who know everything but rather as a partner to help other elderly people to identify risks and potential solutions.

There is a long history of community policing, through which the police make use of community inputs, e.g. through community-based crime prevention, or even civilianization (Skolnick and Bayley 1988, 5) . However, its implementation in the USA has been described by Cordner (2014) as superficial-in 2002 fewer than 25% of agencies had adopted robust elements of community policing, such as involving citizens in recruitment and evaluation of police officers or reviewing complaints against the police. Even more worrying is the accusation by Taylor (2006) that recent US 'zero tolerance' policing practices have undermined policecommunity collaboration, fundamental to 'broken windows' approach, and exacerbating exactly those problems which co-production has been designed to tackle. Community policing can also be the site of conflict in the community-it is sometimes captured by the most active community groups or can become part of the power struggle between community groups representing different interests.

Another co-production approach to crime prevention directly is demonstrated by Mentors in Violence Prevention in Scotland, which operates in more 150 secondary school in over 20 local authority areas, aiming to empower students to safely speak out against all forms of violence from rape and sexual harassment to bullying and abusive behaviour. Based on an original template from the USA, the Scottish Violence Reduction Unit successfully piloted and developed the approach from 2011 onwards. It sees students as a school's greatest resource in tackling violence and trains senior pupils as peer mentors, who then deliver sessions to younger students. Over the last five years over 6000 mentors have been trained and have delivered around 2000 lessons a year, targeting issues such as bullying, gender norms, domestic violence, knife crime and harmful sexual behaviour. More recently it has been adapted to higher education settings, workplaces and even Scotland's night time economy, with pubs and clubs embracing the scheme as a way to help keep customers safe (Violence Reduction Unit, n.d.). More recently similar approaches have been proposed in London and other cities.

Encouraging Desistance

Co-production can also have an important impact on crime by encouraging desistance, which is a process of behavioral change, helping individuals at risk of committing crime to desist sustainably from criminal activity. Weaver and McCulloch (2012, 7) highlight empirical evidence from the seminal work of LeBel (2007 LeBel ( , 2009 , later confirmed by other studies, which established that volunteering or advocacy behaviours by (potential) offenders had a positive correlation with increased self-esteem and life satisfaction, and a negative correlation with criminal attitudes and behaviours.

A meta-analysis of US studies into focused deterrence strategies undertaken by Braga and Weisburd (2012) concluded that they are associated with an overall statistically significant, medium-sized effect on crime reduction. The emphasis in these strategies was not only on decreasing offending (although they did involve intensive police contact with, and pressure upon, known and suspected criminals, such as drug dealers) but also on a wide range of other activities aimed at making crime less likely to take place or to be successful, such as decreasing opportunities for crime, deflecting offenders away from crime (e.g. by focusing services and support on dealers, so that those willing to change their lives have the support they need), increasing the collective efficacy of communities, and increasing the legitimacy of police actions. Indeed, Braga and Weisburd suggest that the large effects observed from these strategies came precisely from its multifaceted influence upon criminals. Moreover, these approaches to focused deterrence had an important impact upon collective efficacy-the belief that the community could make a difference-in the relevant communities, by engaging members of the community in the strategies developed. This is likely to have been highly important, given that self-efficacy has emerged in many studies as a key driver of co-production in all public services, including community safety (Parrado et al. 2013; Alford and Yates 2016; Loeffler and Bovaird 2019b) .

Turning to research on desistance in relation to youth crime and antisocial behavior, mainly from rigorous US studies, Ross et al. (2011, 68-70) identify child skills training as a distinctively effective early intervention programme. By teaching children "social, emotional, and cognitive competence by addressing appropriate effective problem solving, anger management and emotion language", this training prepares them for a greater role in the co-production of their own future welfare.

A different co-production approach to encouraging desistence is highlighted by Bright et al. (2015) , who summarise empirical research evidence from US teen courts. Adolescents hear the cases, which concern low-level crimes, and usually also make decisions about sanctions to be applied to teen perpetrators. The interaction with positive peers, together with the subsequent community involvement which is often part of the sanctions applied, have meant that teen court 'completers' experience relatively low rates of recidivism up to one year after their hearings, suggesting that they have been encouraged to make positive changes to their lives.

Two promising co-production programmes for encouraging desistance are also identified by Ross et al. (2011, 68-70) -in the first, a peer mentor from the community spends time with a young person who is at-risk. The peer acts in a non-judgmental, supportive capacity, but is also a role model. In the second, young people experience and learn skills in a range of non-academic activities in their after-school recreation time, overcoming their low self-esteem through taking part in highly structured activities co-produced with professional support. Weaver and McCulloch (2012, 13) also highlight a successful peer support approach: the Routes Out of Prison project in Scotland employs Life Coaches (who are mainly reformed offenders or ex-substance abusers) to give emotional and practical peer support to short-term prisoners before and after they are released from prison. The Life Coaches also act as advocates, helping to deal with issues like housing debt, benefits advice, health, addiction, training, education and work experience. This peer support has been positively evaluated by all key stakeholders, including prisoners-indeed, 43% of the prisoners involved in 2009-10 signed up to continue their engagement after they were released.

Finally, Weaver and McCulloch (2012, 11 ) discuss a St. Giles Trust project which employs former prisoner peer advisors, after they are released, to work intensively with newly released prisoners to help in their resettlement, focusing in particularly on preparation for work. An evaluation by ProBono Economics (2010) estimated that this peer support approach reduced re-offending by 40%, with savings of £10m in 2009, together with significant quality of life improvements for ex-prisoners.

Removing criminals from the community depends on the court system and, ultimately, on the detection and successful prosecution of crime. A key contribution which citizens make directly to removing criminals from the community is through their role as witnesses in court, which is generally a voluntary choice by citizens. The importance of this role can be seen from the fact that in 2014-15, there were over 11,000 abandoned criminal trials in England and Wales because of the non-appearance of witnesses at court (although this only constituted 2.1% of crown court trials) (BBC 2016) . The work of the Citizens Advice Witness Service in England and Wales also constitutes a form of co-production-it relies on nearly 2500 volunteers to complement its 300 staff, together helping over 168,000 citizens per year who are giving witness and are daunted and worried about this potentially traumatizing or even dangerous process (Citizens Advice 2016).

It is clear from the model in Loeffler and Bovaird (2019b, 5 ) that crime detection is central to several pathways to outcomes, including crime deterrence and, through punishment of crime, to desistance and removing criminals from the community. Co-production can play a major role here, e.g. through neighbourhood patrols (such as Street Watch) or Speed Watch (where local residents capture evidence of speeding motorists in their streets with a laser speedgun/camera and send a photo of the car with its speed recorded to the police). Such activities can not only detect crime but also deter it by encouraging behaviour change. An evaluated example is provided by the Speed Watch scheme of Wiltshire County Council, which mobilised 765 local residents in 140 volunteer teams to monitor motorists' speeds (Milton 2016) . The evaluation found that there was a 35% reduction of fatal and serious injuries from traffic accidents in Wiltshire, compared to a national fall of 22% during this period. Crime detection is also aided by citizens reporting crime through hotlines like 911 in the US and 999 in the UK, both alerting police more speedily to crimes committed and providing richer information for police investigations. More recently, such mechanisms have been reinforced by internet and social media platforms, which give the police extra tools for gathering citizens' inputs for the detection and pursuit of crime. For example, since 2004, Dutch police have used an online system, Citi-zenNet (Burgernet ) (Meijer 2012, 200) , which allows them to call for information from citizens on recent crimes (e.g. burglaries near their house) or even crimes currently taking place (e.g. criminals seen running from scene of crime). The system not only helps citizens to feel safe at home but it also speeds up the tracking of suspects or of missing peopleand it may strengthen trust in the police. An average of 4.6% of citizens in the nine cities signed up for CitizenNet and 9% of all the cases qualified as fit for CitizenNet action were solved with the help of its informationmore than 50% of all successful police actions, so an important addition to existing procedures (Meijer 2012, 200-201) .

Of course, these 'hotlines' and online platforms can have both an 'upside' and a 'downside': on the one hand, they can be seen as a positive example of community organising to reduce harms experienced, but they can also be seen as a form of 'coveillance', where citizens are marshalled to play a role in the surveillance structure of the state. Clearly, it is important to remain on the positive side of the fine line between harnessing 'crowd intelligence' and simply resonating rumour and prejudice, which would be an aspect of the potential 'dark side' of co-production (Steen et al. 2018) .

While legislation in both USA and UK allows for citizens to directly intervene to stop crime, e.g. through citizen arrests (Robbins 2016) , such direct intervention may be dangerous (to themselves, to the assumed offenders, and to bystanders), so police have traditionally discouraged such direct action, unless "it is safe to do so" (Weaver 2018) . Nevertheless, when it does occur, it is often strongly commended, both by the press and the media (Nsubuga 2017) . Moreover, this may be more important than often realised-recent research by Philpot et al. (2019) , using widespread surveillance cameras to record more than 200 incidents in Amsterdam, Cape Town and Lancaster (England) shows that in 9 out of 10 incidents at least one bystander intervened (with an average of 3.8 interveners), with no significant difference across the three cities, in spite of their greatly differing levels of crime and violence.

Taken together, this evidence therefore suggests that community coproduction between citizens and the police can indeed play an important role in detecting and tackling crime. However, public governance principles remain important, so some safeguards are necessary, e.g. to ensure that 'coveillance' does not infringe people's right to privacy and that 'vigilante' behaviour, where citizens take the law into their own hands or engage in inappropriate action against other citizens, does not occur.

There is generally a large gap between officially reported crime levels and the public's perception of crime. For example, in 2015 just over 60% of UK adults perceived crime to have risen in recent years (ONS 2015a), whereas in fact recorded crime had reached its lowest level since 1981 (ONS 2015b). Moreover, Cornaglia et al. (2014) estimate statistically that that the social cost of violent crime is about 80 times the direct impact on the victim, in terms of mental well-being alone (although property crime was found to have no such social cost impact on victims or non-victims).

There is some evidence that co-production can influence this gap between perceived and actual levels of crime. Weisburd and Eck (2004, 59) in their review of well-founded academic studies concluded that community policing strategies did not significantly cut crime or disorder but did cut levels of citizen fear of crime. Similarly, Cordner (2010) concludes that most studies of community policing have found that residents like community policing and feel safer when it is implemented where they live and work-but it is more difficult to tease out which particular elements of community policing have this effect. He goes on to suggest that community engagement, in itself, contributes to reductions in fear of crime, possibly by decreasing social distance between residents, increasing social cohesion, increasing informal social control, and raising public confidence in the police.

Restorative justice, which now plays an important role in the UK criminal justice system, is a co-produced approach to delivering justice (Weaver 2011) , involving some personalisation for the offender (and, indeed, for the victim). Johnstone (2004) suggests that it comprises four common but rather different practices, with different levels of ambition in terms of what is aimed to be achieved:

• offering victims of crime a chance to meet in a safe setting with the person who harmed them, and with a trained mediator, to develop an action plan for repairing the harm caused; • bringing in family members and community supporters of both victims and offenders to support the offender's reparative efforts and behaviour change to reduce their chance of reoffending; • bringing together community residents, justice and social service staff and family and support networks to consider what the community should do to help both victim and offender, linking restorative justice to community regeneration and capacity building, often making recommendations to the judge;

• encouraging community members to form support groups around feared ex-prisoners (particularly sex offenders) to help them change their behaviour, to monitor their behavior and ensure safety of the community. Meadows et al. (2012) concluded from an evaluation of a restorative justice programme in South Yorkshire that victims were generally satisfied with the outcome, feeling empowered by their experience, developing a greater sense of control and reporting increased confidence in the police. Similarly, offenders had experienced a positive effect. It also appeared that participating in restorative justice reduced the likelihood of reconviction amongst offenders, although the results were not quite statistically significant.

In a similar evaluation of a Youth Restorative Intervention (YRI), run jointly by Surrey Police and Surrey County Council's Youth Support Service, Mackie et al. (2014) found that the direct costs of running the YRI were lower than processing the offender through the police (which would have entailed youth cautions, youth conditional cautions and prosecution)-the YRI cost £360 per case, compared to £600 per case through the alternative (police) approach, which also had a 6% higher level of offending. Moreover, they calculated further savings to the public purse of £440 from the restorative justice approach, implying a rate of return of 2.86:1 from investment in this co-production initiative. If the reduced social costs relating to each victim (estimated to be some £200) is added, this brings up the rate of return to 3.41:1.

The figure in Loeffler and Bovaird (2019b, 5) shows that co-production in the policing and the criminal justice system can have an indirect impact on the quality of life of all citizens through reductions in the level of crime and the fear of crime and through the collective benefit it brings from achieving community justice. However, there are two stakeholder groups whose quality of life can be directly improved by co-production of community safety-namely victims and offenders/ex-offenders. However, for offenders many approaches to rehabilitation are not actually conducive to their quality of life. Ward and Brown (2004) characterise the dominant approach to offender rehabilitation in Canada, the UK, Australia, and New Zealand as a risk-need model, in which the risk factors associated with recidivism are systematically targeted, seeking the reduction of maladaptive behaviours, elimination of distorted beliefs, removal of problematic desires, and modification of offence-supportive emotions and attitudes. While they accept that this model does indeed reduce recidivism rates, they point out that it does not promote pro-social and personally more satisfying goals, which would be more effective in motivating offenders. In contrast, they highlight the Good Lives Model of offender rehabilitation, originating in Canada but now used internationally, which is a strength-based approach, seeking to give offenders the capabilities to secure valued aspects of human functioning and living. The Good Lives Model incorporates the co-productive capabilities of prisoners: excellence in play and work (including mastery experiences); excellence in agency (i.e. autonomy and self-directedness); relatedness (including intimate, romantic and family relationships) and community; spirituality (in the broad sense of finding meaning and purpose in life); and creativity.

However, this approach in the justice system is not common- McCulloch et al. (2016, 441) conclude that for offenders "… the predominant experience described was one of punishment, judgment, humiliation, depersonalization and a 'total imbalance of power'. For most this was a distancing, disenfranchising and disorientating experience that, for some, directly impeded their capacity to cope, far less co-produce." In spite of some counter experiences in line with co-production principles, the authors suggest that the criminal justice system in the UK has been drifting from humane, participatory and complex justice practices towards a more punishment-oriented approach, which may be more politically popular but which will not be able so easily to incorporate co-production.

However, some of these counter-experiences suggest the possibility of disrupting this trend-for example, Weaver and McCulloch (2012, 10) report an evaluation of Foundation 4 Life, a London-based programme which engages reformed offenders and former-gang leaders to deliver behaviour modification workshops and programmes to young people who are either still offending or at risk of offending-on exiting the programme, 20% of participants said they would actively make a change, and 91% of reformed offenders acting as facilitators had not re-offended and some had obtained related employment.

In public services like policing and community justice, with a strongly hierarchical governance system, involvement of citizens (whether service users or community members) in commissioning decisions can be expected to be relatively uncommon, compared to services characterised more strongly by network governance (Loeffler and Timm-Arnold 2020) . Indeed, Weaver and McCulloch (2012, 10) comment that involving service users in commissioning processes in public safety or law and order is quite exceptional, particularly in community-based criminal justice services. This is in line with the finding from our EU study of five European countries (Loeffler et al. 2008, 20) that only just under 6% of respondents had participated in groups that were working to improve community safety. Nevertheless, there are some interesting examples of cocommissioning. For example, there is an increasing number of commissioning initiatives working with deprived communities, with a focus on improving public safety, such as the "Listening Events" with local communities in North West Kilmarnock in Scotland (Bone 2012) .

One approach to co-commissioning which has seen several rounds of experimentation in the UK is participatory budgeting (PB). In 2008, the Home Office allocated nearly £500,000 to 24 pilot areas to trial PB at local authority level, with budgets to be allocated to projects contributing towards community safety, although the definition of community safety was broad and was left for each pilot to define (PB Unit 2008). In Scotland there were five Community Safety PB pilots from 2010. In 2013 an online PB platform was used by the North Wales Police and Crime Commissioner to gauge priorities in relation to anti-social behaviour on the part of respondents living in isolated rural communities. Unfortunately, none of these initiatives has been rigorously evaluated.

In the UK, the Police Reform and Social Responsibility Act 2011 placed a statutory duty on chief officers to make arrangements for consulting and informing the public on local crime and disorder 'within each neighbourhood', with mandatory regular meetings between the public and local police officers (although the format and regularity were not prescribed). The initiation of Police and Crime Commissioners in 2012 opened up extra opportunities for bringing citizens into the commissioning process for community safety-one of its main aims was to introduce democratic accountability to the determination of local policing priorities -in words of the Home Office: "We have put policing back in the hands of the public" (House of Commons Home Affairs Committee 2014). For example, the Dorset Office of the Police and Crime Commissioner adopted the approach that no person or group of people are hard to reach. Its Community Engagement Strategy emphasised putting more effort and creativity into reaching these groups and reducing barriers to engagement. Its five 'levels of community engagement' included information giving; consultation; joint decision-making (Community Engagement Forums, Victim Groups and Surgeries); joint action (Partnership Decision-Making); and community empowerment (Cabinet Office 2015). The latter three areas clearly entail co-production.

However, Higgins (2018, 49) reported that, although many PCCs in the UK had set up Police and Communities Together meetings to consider community priorities. many of these had already fallen into disuse or been discontinued and, where they had continued, they were generally reported to be poorly attended, unrepresentative and illustrative of the tension between the 'low level' local issues those attending wanted the police to address, and the 'higher-harm' remits created for neighbourhood teams by their forces. This had resulted in explicit, community-set priorities becoming more marginal to the work of neighbourhood teams.

This UK experience suggests that, in spite of some interesting examples to the contrary, it clearly remains the case that co-commissioning in community safety remains rather weak.

While co-design methods are becoming increasingly popular in health and social care services, harnessing "user experience to improve the design of services in community safety is still not very common" (Loeffler 2018, 216) . However, we will give two examples here. One of the most studied areas is the role of prisoners working on codesign of prison services. The UK-based charity User Voice (2016) is staffed and led mainly by people who have personal experience of the same problems the charity is seeking to solve, namely serving a prison sentence and being on probation. User Voice has promoted the setting up of service user councils in prisons and probation services, so that service users can co-design service improvements. A recent evaluation of prison councils (Barry et al. 2016, 95) concluded that "In terms of impacts on services … the analysis suggests that User Voice activity is associated with benefits well in excess of £500,000 across the five User Voice prisons where the team was able to undertake before and after rate comparisons (with comparators)". Moreover, "community conferencing", an approach to restorative justice which is used in many European Countries, the USA, Canada and Australia, typically involves all key stakeholders, such as offenders, volunteer community representatives and public services, in co-designing solutions to repair relationships with local people. It is typically used where the community has been the victim through being impacted by illegal and inappropriate behaviour-this is discussed in the Crown Prosecution Service guidance on Restorative Justice (CPS 2019). This guidance cites a community conference which was initiated by a group of shopkeepers outraged by the long-standing misbehaviour of a group of young people who congregated in their shopping centre. The community conference resulted in a win-win for all stakeholders-the behaviour of the young people in the shopping centre was much improved; police call-outs were much reduced; the shopkeepers were satisfied that their complaints had been acted upon; and the young people felt they had been treated fairly and their needs considered.

Co-assessment gives service users and communities a voice in the assessment of outcomes achieved and quality of service. This might be seen as particularly relevant in community safety and public order, as objective performance information (e.g. crime figures) does not necessarily reflect citizens' subjective perception of public safety, as we have seen above in relation to the fear of crime. However, in practice, there are relatively few well-documented case studies of co-assessment of community safety. Sabet (2014, 260 ) discussing co-production and police services in Mexico suggests that the 'citizen monitor' model has the greatest potential for impact at the local level, using crime data to evaluate law enforcement performance, as well as designing better civil society and governmental interventions. He instances Ciudad Juárez, which established the Citizens Observatory for Security and Coexistence in 2008, in association with the federal Health Ministry, the Pan American Health Organization, and the Autonomous University of Ciudad Juárez, which housed the observatory and covered its personnel and administrative costs. It has worked with the city's traffic police to analyse data on traffic accidents, publishing in 2012 three geo-referenced diagnostic studies on violent deaths, traffic accidents, and crime, leading to joint work with government officials to develop a master plan for road safety.

In a more narrow sphere, researchers have compared complaints systems of different countries in terms of the degree to which 'civilian control' over the complaints system has been embedded, leading to the conclusion that "internationally there is a slow ongoing trend towards the adoption of many features of the civilian control model" (Prenzler 2011, 285) . However, there has been considerable controversy over the degree of 'civilianness' of these bodies for investigating complaints against the police-Savage (2013) highlights the potential damage to the perceived legitimacy of such oversight bodies from the continued presence of former officers in their ranks and goes on to argue that the formal establishment of a body to provide independent, civilian oversight not only needs members who are clearly independent of the police but also the pursuit of a 'distinctive organizational culture', which avoids being dominated by processes which reflect 'cultural survival' and 'cultural continuance' in terms of embedded police attitudes and orientations, which are inherited in part by the oversight bodies through the employment of former police officers as investigators. Savage (2013) emphasises that, while most agree about the value of police experience in police complaints investigation, virtually all agree also that 'outsider' status enables a degree of distance from police culture, bringing benefits for the investigative process in terms of open-mindedness, a more critical approach to police conduct, and closer proximity to the public and therefore to the concerns of complainants and families of those affected by police conduct.

In the UK, some of the Police and Crime Commissioners (PCCs) set up after 2012 have developed additional scrutiny and monitoring arrangements. The Cleveland PCC uses mystery shoppers and independent custody visitors to report on the local police force, while the Avon and Somerset PCC established a panel of independent residents which reviewed complaint files against the police force and published reports (House of Commons Home Affairs Committee 2014, 22).

Finally, it is notable that police forces are increasingly using social media to engage with citizens, as in the West Midlands Police (Hine-Hughes 2013). As a result, social media are becoming much more important as a two-way communications tool between police and citizens (Ruddell and Jones 2013), including a "soft" co-assessment tool, which complements formal complaints mechanisms. However, a recent study of the use of social media by British police forces (Fernandez et al. 2017) found that police tweets that attracted higher engagement talked about roads and infrastructures, events and missing persons, and aimed to raise awareness about important crimes-there was no mention of seeking feedback on police performance.

There are still relatively few studies on how co-production affects high level outcomes in community safety and only a narrow range of outcomes is typically addressed, with particular attention to crime reduction, perhaps because this is usually a government priority. The effect of coproduction on justice in the community and on quality-of-life outcomes of victims, prospective victims and offenders is still under-researched.

Most of the co-production activities in policing and criminal justice mainly involve forms of co-delivery-other modes of co-production such as co-commissioning, co-design and co-assessment are still relatively rare. More research is needed on how effective user and community co-production can be implemented in services characterised by more hierarchical modes of governance, such as public safety.

This chapter analysed the empirical evidence on the extent of user and community co-production in health, social care and public safety and its effectiveness for improving public outcomes. In particular, it provides a conceptual framework of pathways to outcomes, which is dynamic and locates co-production initiatives as part of prevention, treatment and rehabilitation strategies for dealing with social problems. This generic 'pathways to outcomes-model' has been refined for public safety and backed up with a literature review (Loeffler and Bovaird 2019b) .

The analysis in this chapter provides several key insights: First, there is limited evidence of the extent of co-delivery to achieve better public outcomes through problem prevention, treatment and rehabilitation in the three sectors. Most of the literature focusses on case studies. More quantitative evidence is needed on the service users and communities reached by national or local co-delivery initiatives and their importance within the public sector in terms of budgets spent on co-delivery.

Second, while important public health and social care intervention theories either incorporate elements of co-production or are at least compatible with co-production, the use of co-delivery approaches for problem prevention remains limited. Most co-delivery initiatives in these two sectors focus on problem treatment, and to some degree, on rehabilitation from specific problems or conditions. This is different in public safety where the extent of co-delivery in the detection of crime is limited and most co-delivery focusses instead on prevention or dealing with the consequences of crime. The reason may be that, at least in administrative law countries, problem treatment in public safety is still considered to be the exclusive task of the state. In a German co-production study, a number of focus group participants suggested that co-production is not possible in averting danger, whereas the majority of focus group participants thought that engaging citizens as volunteers plays an important role in the delivery of emergency and preventative services (Loeffler and Timm-Arnold 2020, 18) . Further international comparative research is needed on how administrative traditions impact on co-production across sectors such as health, social care and public safety.

Thirdly, there is also a difference regarding the range of co-production in terms of the Four Co's (co-commissioning, co-design, co-delivery and co-assessment) between the three sectors. As this chapter suggests, there seem to be fewer co-commissioning, co-design and co-assessment initiatives in public safety than in social care and, to some degree, than in health services. As Loeffler and Timm-Arnold (2020, 11) argue "services characterised by governance modes with strong elements of network working are likely to enable a wide range of co-production approaches, including co-commissioning, co-design, co-delivery and co-assessment", contrasting this with public safety services, which are characterised by a more hierarchical mode of governance.

Nevertheless, the fact that a wide range of co-delivery initiatives in public safety could be identified in the literature review by Loeffler and Bovaird (2019b) and the focus group research by Loeffler and Timm-Arnold (2020) shows that co-production, and in particular co-delivery, is possible even within more hierarchical modes of governance. In health services, the mode of governance depends on the intervention pathway. As far as the detection and treatment of conditions are concerned, health services are highly regulated (McMullin and Needham 2018, 157-158) , whereas many health prevention services tend to be more characterised by network modes of governance.

The empirical evidence analysed in this chapter suggests that the key co-production approach identified as most significant in all three sectors is co-delivery. As Loeffler and Timm-Arnold (2020, 20) comment "this is not surprising in public safety, since hierarchical modes of governance are less likely to favour giving a role to external stakeholders in decision making". However, it may be seen as more surprising in the health and social care sectors, where government intervention has often focused particularly on increasing the role of citizen voice. The evidence from the German case study suggests "the paradox that in German local public services there is more talk than action about 'citizen voice'-and more action than is often recognised in relation to citizen action" (Loeffler and Timm-Arnold 2020, 20) . It remains an open question as to what extent this paradox applies also to the UK and other OECD countries.

Using Acorn Wellbeing & the Kent Integrated Dataset (KID) to identify and analyse older people more likely to be experiencing social isolation and loneliness. Maidstone: Kent County Council

A taxonomy of behavior change techniques used in interventions. Health Psychology: Official Journal of the Division of Health Psychology

How 'personal' are personal budgets in 2017? Community care

All the lonely people: Loneliness in later life

Co-production of public services in Australia: The roles of government organisations and co-producers

How Therapy Focus works with young people and their families in Western Australia to co-design therapeutic interventions

Evaluation of the user voice prison and community councils: Final report

Thousands of trials end over no-show witnesses-watchdog

The Performancestat potential

Transforming culture and practice in children's social care assessment-Part 1

The effectiveness of neighborhood watch

Local Area Coordination: Catalyst for a system wide prevention approach. Local Area Coordination Network. Sheffield: The Centre for Welfare Reform

A new role for citizens in public procurement. Mexico: Citizens & Markets (c/o Transparencia Mexicana)

Reducing crime and improving health in NW Kilmarnock using community assets. Governance International Case Study

The new commissioning model of services for young people in Surrey: Evaluation of achievements and its implications. Birmingham: INLOGOV and Governance International

Repositioning Europe and America for economic growth: The role of governance and private actors in key policy areas

The effects of focused deterrence strategies on crime: A systematic review and meta-analysis of the empirical evidence

Perceptions and outcomes following teen court involvement

People, places, possibilities: Progress on Local Area Coordination in England and Wales. Sheffield: The Centre for Welfare Reform

Present': Co-producing improved wellbeing with people living with dementia in East Dunbartonshire. Birmingham and East Dunbartonshire: Governance International and East Dunbartonshire Council

Clustering of unhealthy behaviours over time. Implications for policy and practice. London: The King's Fund

Police and crime commissioners and civil society. London: Cabinet Office

The impact of participatory budgeting on health and wellbeing: A scoping review of evaluations

The state of care in general practice 2014 to 2017: Findings from CQC's programme of comprehensive inspections of GP practices

Care quality commission public engagement strategy

Self care support for people with COPD

Involving the public in HIA: An evaluation of current practise in Wales

Patient involvement in clinical audit at a primary care trust

Evaluating the success of a hospital's community engagement process

Witnesses need support to feel confident giving evidence

Riabilitazione, pazienti inascoltati: il questionario per fare il punto

A review of the NHS hospitals complaints system: Putting patients back in the picture

Direct to patient communication: Patient empowerment or NHS burden?

Reducing fear of crime: Strategies for police

The Oxford handbook of police and policing

Crime and mental well-being

Personalised care planning for adults with chronic or long-term health conditions

Restorative justice

Online feedback to drive CQC inspections

Revisiting the rationale and evidence for peer support

Learning report: Digital participatory democracy in Scotland

Working together to safeguard children: A guide to inter-agency working to safeguard and promote the welfare of children. London: Department for Education

Paris Budget Participatif

Using experience-ased co-design to improve the quality of healthcare: Mapping where we are now and establishing future directions

About early intervention: Why it matters

Transforming lives, communities and systems? Coproduction through participatory budgeting

The governance of integrated health and social care in England since 2010: Great expectations not met once again? Health Policy

An analysis of UK policing engagement via social media

Don't forget to take your medicine!" Improving the use of medicine by elderly patients in the region of Madrid. Governance International Case Study

Guide to crowdfunding for local authorities. London: Local Government Information Unit

Feeling the burden of social care and the NHS? Crowdfunding can save us

Exploring the relationship between foot and car patrol in violent crime areas

NICE public health guidance promote lifestyle changes

I can see what you can't see"-How Warwickshire County Council involves people with learning disabilities as peer reviewers

Judgements or assumptions? The role of analysis in assessing children and young peoples needs

Empowering Communities: Community skills development and neighbourhood workforce planning

The future of neighbourhood policing. London: Police Foundation

How West midlands police engages with communities

Saving Blackheath Fireworks Night! What a successful public fundraising campaign looks like. Governance International Case Study

Involving service users and carers in a procurement": How Bedfordshire Clinical Commissioning Group involved service users, carers and stakeholders in their procurement of mental health, learning disability and CAMH services

Citizens' assembly on social care: Recommendations for funding adult social care

Latest GP patient survey results released

The idea of restorative justice

An inspector calls: Citizen-led service inspection by West Lothian Council. Governance International Case Study

Building community capacity: Making an economic case

An examination of the impact of formerly incarcerated persons helping others

Formerly incarcerated persons' use of advocacy/activism as a coping strategy in the reintegration process

How offenders transform their lives

Case study: NHS England Citizens' Assembly

Co-production and co-creation: Engaging citizens in public services

Co-commissioning of public services and outcomes in the UK: Bringing co-production into the strategic commissioning cycle

Assessing the impact of co-production on pathways to outcomes in public services: The case of policing and criminal justice

If you want to go fast, walk alone: If you want to go far, walk together": Citizens and the coproduction of public services

Comparing user and community co-production approaches in local 'welfare' and 'law and order' services: Does the governance mode matter? Public Policy & Administration

Koproduktion in Deutschland. Studie zur aktuellen Lage und den Potenzialen einer partnerschaftlichen Zusammenarbeit zwischen Kommunen und Bürgerinnen und Bürgern

Youth restorative intervention evaluation final report

More people to access personal health budgets under Government plans

Coproducing justice sanctions? Citizen perspectives

Co-assessment through citizens and service users in audit, inspection and scrutiny

Co-production and co-creation: Engaging citizens in public services

Defining peer support

Evaluation of the South Yorkshire Restorative Justice Programme

New public governance, the third sector and co-production

Community Speedwatch Scheme in Wiltshire to reduce speeding and empower residents

Dave's in the picture as portrait initiative raises awareness of diabetes. Evening Express

Co-commissioning at the micro-level: Personalized budgets in health and social care

Personalised commissioning in adult social care. London: National Audit Office

Framework for patient and public participation in public health commissioning

NHS Citizen

Social care user surveys

T1 Resources

Police thank the heroes of London Bridge who risked their lives to help victims

Evaluating participatory budgeting activity in Scotland-Interim report year 2

Service user involvement: Impact and participation: A survey of service user and staff perspectives

Crime statistics, focus on public perceptions of crime and the police, and the personal wellbeing of victims

Crime in England and Wales: Year ending

An overview of lifestyles and wider characteristics linked to Healthy Life Expectancy in England

What if things go wrong in co-producing health services? Exploring the implementation problems of health care coproduction

Correlates of co-production: Evidence from a five-nation survey of citizens

Promoting self care in the community: The evidence and why it is important

General practitioners' concerns about online patient feedback: Findings from a descriptive exploratory qualitative study in England

Q: The journey so far: Connecting improvement across the UK-Insights and progress three years in. London: Health Foundation

Implementing recovery through organisational change: 1. Recovery Colleges. London: Centre for Mental Health

Efficacy of peer support interventions for depression: A metaanalysis

Would I be helped? Cross-national CCTV footage shows that intervention is the norm in public conflicts

Social work, social welfare, American society

The evolution of police oversight in Australia

St Giles Trust's through the gates: An analysis of economic impact

The campaign for real self care

Get Well, Feel Well and Stay Well: A vision for self care in the United Kingdom

A review of the literature on peer support in mental health services

Vilifying the vigilante: A narrowed scope of citizen's arrest

The Palgrave handbook of co-production of public services and outcomes

Prevention and reduction: A review of strategies for intervening early to prevent or reduce youth crime and anti-social behaviour. London: Institute for Fiscal Studies

Social media and policing: Matching the message to the audience

Getting to authentic co-production: An asset-based community development perspective on co-production

Co-production and oversight: Citizens and their police (Working Paper-WP Series on Civic Engagement and Public Security in Mexico)

Seeking 'civilianness': Police complaints and the civilian control model of oversight

Coaching community health volunteers in integrated community case management improves the care of sick children under-5: Experience from

Grumbles, gripes and grievances: The role of complaints in transforming public services

Theme and variation in community policing

Personal budgets: Why do councils still call the shots on care? The Guardian

Patients can post online comments about GP practice

Fund my treatment!: A call for ethics-focused social science research into the use of crowdfunding for medical care

Pioneering sustainable bank first to launch crowdfunding platform

Social work theory and methods comparison table

Coproduction and co-creation: Engaging citizens in public services

Incivilities reduction policing, zero tolerance, and the retreat from coproduction: Weak foundations and strong pressures

Police innovation: Contrasting perspectives

Stepping Forward Together. AHPScot Blog

Stepping forward together-Co-production for falls prevention in Aberdeen. Governance International Case Study

User VOICE establishes new Service User Councils with the help of Sodexo

The prevalence of, and risk factors for, loneliness in later life: A survey of older people in Great Britain

Violence Reduction Unit. (n.d.). Mentors in violence reduction

Care of the elderly: An exercise in cost benefit analysis. London: Institute of Municipal Treasurers and Accountants

The good lives model and conceptual issues in offender rehabilitation

Co-producing community justice: The transformative potential of personalisation for penal sanctions

Co-producing criminal justice: Executive summary

Met police caution against public helping officers under attack

How to use online ratings for a hospital

What can police do to reduce crime, disorder and fear? The Annals of the

Stockport Council's new adult social care website 'My Care, My Choice': A business case for service co-design

Quantifying the benefits of peer support for people with dementia: A Social Return on Investment (SROI) study

Preventing loneliness and social isolation: Interventions and outcomes. Social Care Institute for Excellence Research Briefing 39

The rise of crowdfunding for medical care: Promises and perils