Tobacco dependency treatment services for people with severe mental illness: A mixed methods rapid evaluation

Smoking in people with severe mental illness (SMI)

Smoking is a significant contributor to the health inequalities experienced by people with severe mental illness (SMI). Individuals with SMI, such as those with schizophrenia, psychosis, and bipolar disorder, are over three times more likely to smoke compared to the general population with a smoking prevalence of approximately 40%, compared to 12%, respectively^1,2. These individuals are almost four times more likely to smoke at a young age (16–19 years), smoke approximately five more cigarettes daily, and extract more nicotine from each cigarette, leading to a greater degree of nicotine and tobacco dependency^1,3–6. This contributes to a heightened risk for people with SMI developing smoking-related illnesses, such as respiratory and cardiovascular conditions, as well as certain cancers, contributing to a reduced life expectancy of approximately 15–20 years⁷. The risks of smoking may be exacerbated for people with SMI who experience intersecting forms of disadvantage that are associated with higher smoking rates, including socioeconomic deprivation, experiencing homelessness, and having a minority sexual and/or gender identity^8–11.

Quitting smoking can have numerous benefits, including both immediate and long-term physical health benefits, improved mental health after the initial nicotine withdrawal phase, and a potential reduction in the dose of certain psychiatric medications^12–14. Despite these benefits, quitting is challenging due to its addictive and habitual nature. Common barriers faced by smokers include managing cravings; using smoking to relieve stress; and the social influence of friends, family, and colleagues who smoke^15,16. People with SMI also face additional barriers compared with those in the general population^1,16–20. For example, smoking is often used as a mechanism to cope with acute instances of poor mental health, even though quitting has been shown to be beneficial for mental health in the long run^15,16,21. Some people with SMI have been exposed to a ‘culture of smoking’ within mental health settings, wherein smoking was condoned as an individual’s freedom of choice and propagated through staff and patient behaviours, beliefs, and interactions^1,22. Moreover, some staff members are less likely to encourage smoking cessation in people with SMI due to beliefs about negative consequences, at least in the short term, such as withdrawal symptoms and the removal of an important coping mechanism^1,15,23.

Tobacco dependency treatment in England

Despite these challenges, evidence-based tobacco dependency treatment can effectively support smoking cessation in the SMI population^20,24–28. These treatments involve a combination of pharmacological and behavioural support tailored to the specific needs of people with SMI^24,25,29,30. Examples of pharmacological support include the use of Nicotine Replacement Therapy (NRT) and nicotine analogue medicines that mimic the effects of nicotine, and behavioural support, which often includes motivational counselling. As part of the smoking cessation intervention for people with severe mental illness (SCIMITAR+) trial, tobacco dependency treatment was delivered by smoking cessation advisors who had received training in mental health. These advisors provided individual in person consultations offering a range of support tailored to people with SMI, including: gradual cigarette reduction before setting a quit date; additional in person consultations post relapse or unsuccessful quit attempts; liaising with healthcare professionals about smoking cessation-related metabolic changes and mental health medication prescribing, and the co-ordination of NRT to help alleviate short-term withdrawal symptoms^31–35. E-cigarettes, or vapes, are also increasingly being used as aids for smoking cessation, with evidence now supporting their effectiveness compared to NRT³⁶.

In 2013, the UK’s National Institute for Health and Care Excellence (NICE) published guidance emphasising the need for innovative approaches to support smoking cessation in people with SMI, to reduce health inequalities³⁷. The guidance urged mental health services to become completely smokefree, prohibiting smoking inside NHS buildings, on NHS grounds and for people with SMI to have full access to evidence-based tobacco dependency treatment. Commitments to deliver smokefree mental health services and provide tobacco dependency treatment services in mental health inpatient services were embedded in policy plans, including the 2014 Five Year Forward View³⁸ and 2019 NHS Long Term Plan³⁹. This was accompanied by data insights, suggested delivery models, best practice guidance, training materials, governance support, and a national improvement programme to support implementation^40,41. Most mental health trusts (NHS providers of mental health services) in England now report having comprehensive smokefree policies and dedicated practitioners to provide tobacco dependency treatment⁴².

Community-based tobacco dependency treatment services tend to be commissioned by local authorities and funded through the Department of Health and Social Care (DHSC) public health grant⁴³. Recent guidance for the delivery of local tobacco dependency treatment services emphasised the need for services to be accessible and appropriate for ‘priority groups’. Priority groups, including people with SMI, are groups with higher smoking prevalences that need a tailored approach to service delivery⁴³. Moreover, the assessment of smoking status and subsequent offering of tobacco dependency treatment support to people with SMI has been financially incentivised in general practice⁴⁴. People with SMI are now more likely than people without SMI to be offered tobacco dependency treatment through general practice^19,45.

These initiatives have significantly enhanced opportunities for smoking cessation among people with SMI. However, barriers continue to limit access to appropriate tobacco dependency treatment and support. About half of mental health trusts in England have reported continued daily smoking on hospital grounds, staff enabling patients to smoke, insufficient capacity to offer tobacco dependency treatment to all patients, and inconsistencies in assessing smoking status on admission⁴². The percentage of prescriptions for pharmacological tobacco dependency treatment for people identified as smokers in primary care has been declining, especially in people with SMI⁴⁶. Tobacco dependency treatment services for the general population are often unable to address the specific needs of people with SMI^17,18,47. Such issues contribute to the high smoking prevalence in people with SMI and consequently, persistent health inequalities⁴⁸.

Piloting community-based tobacco dependency treatment services for SMI

To ensure that tobacco dependency treatment services are more accessible and tailored to the needs of people with SMI, new specialist services have been implemented. In 2022, NHS England commissioned seven early implementer sites to deliver community-based tobacco dependency treatment services tailored to the needs of people with SMI⁴⁹. Service delivery was based on implementation of the evidence-based tobacco dependency treatment intervention in the SCIMITAR+ trial^{29,30,50–52}. The seven early implementer sites were commissioned to pilot one of three distinct referral pathways, each focusing on different NHS settings where people with SMI frequently access care. This study aimed to explore whether some settings presented better or worse opportunities to identify and engage people with SMI in smoking cessation.¹ The three referral pathways are as follows:

The referral pathways were co-designed by NHS England with stakeholders from public health charities, the Office for Health Improvement & Disparities (OHID), academic experts, tobacco dependency treatment service providers, and frontline staff in mental health trusts.² Early implementer sites were encouraged to adapt the pathways based on local context and relationships.

The early implementer sites were commissioned by NHS England to deliver tobacco dependency treatment services between April 2022 and April 2024. The continuation of the services beyond this period was dependent on local budgets and commissioning agreements across the NHS Integrated Care Boards (ICBs).

Aims and objectives

In order to better understand how the pilot services have been operating in practice, and to feed this insight back to local and national stakeholders including NHS England, ICBs, and tobacco dependency treatment service providers, the NHS England Prevention Programme team requested a rapid evaluation to capture insights from the early implementer sites to inform future service delivery and commissioning. The evaluation was delivered by the Birmingham, RAND, and Cambridge Evaluation (BRACE) Centre,³ and data collection took place within a rapid timescale (three months), so that preliminary results could be used to inform decision making ahead of the UK Government’s Spending Review in Spring 2025.

In collaboration with NHS England, and informed by learning captured from scoping work, we defined the following aims and objectives for the evaluation:

Scoping methods

The research team conducted scoping work between April and October 2024 to determine the aims and methodological design of the evaluation. This involved:

In initial discussions with NHS England in early 2024, the original aspiration was to conduct an in-depth evaluation involving large-scale analysis of quantitative data and engagement with service users. To align with the remit of BRACE, as a rapid evaluation centre, this would need to be undertaken within a 12-month timeframe. However, learning from scoping showed meaningful engagement with service users, which was difficult to achieve within this timeline. In addition, the data were fragmented, and the full data set was not as complete as originally envisaged.

Following discussions between NHS England and the research team it was agreed that conducting a rapid evaluation over a shorter timeframe would be of value and would provide insights that could inform decision-making ahead of the Government’s Spending Review taking place in Spring 2025.

Evaluation methods

To ensure that the evaluation could provide timely insights into the Government’s Spending Review, a rapid evaluation methodology was employed. This approach, while operating within a condensed timeframe, involved extensive stakeholder engagement to ensure the robustness of findings. By narrowing the focus of the evaluation, data collection was completed within a matter of weeks, enabling the delivery of preliminary insights by the end of February 2025, demonstrating the appropriateness of rapid evaluation to contribute to time-sensitive decision-making processes. Core research activities were undertaken by William Phillips (WP), Sophie Spitters (SS), Jessica Dawney (JD) and Jennifer Newbould (JN).

Patient and Public Involvement

Scoping work revealed that involving service users within the given timeframe would not be feasible due to the time taken to establish relationships. However, recognising the importance of incorporating lived experience perspectives in health services research, the research team felt it was important to include the views of people with lived experience of smoking and SMI more broadly. A Patient and Public Involvement and Engagement (PPIE) expert, Tina Coldham, with lived experience of SMI, acted as advisor throughout the project and helped to identify people with lived experience of smoking and SMI to take part in workshops. The research team captured the views of those with lived experiences of smoking and SMI through three workshops, which are described in more detail below.

Survey

The research team designed an online survey to be distributed to staff at the early implementer sites. From our scoping work, we learned that the number of staff per site (and in turn participants for the survey) would vary, and for some services, the total number of staff employed was unknown to us. However, we knew that there were three key types of roles that could each provide different perspectives on the service: practitioners, service commissioners, and people with a management or service support role. We aimed to survey at least one person from each role and from each site, resulting in a target of 30 survey responses. The survey was co-designed and distributed in collaboration with Thiscovery⁴, an organisation with expertise in user experience and online platforms that helps researchers to gather knowledge and experience from people who use, deliver and influence health and care services. It was designed to be a short (15 minute) survey, suitable for completion on a laptop or mobile phone. We asked the leads at each of the early implementer sites to circulate a link with the survey and supporting information about the study to tobacco dependency service leads and people with management and other support roles, service commissioners and practitioners.

The survey questions explored topic areas, including service delivery (e.g., setting in which the intervention is delivered and what kind of support options are provided), referral pathways, measures of service success, and how data are collected (and related challenges). Please see the Data Availability section for access to the complete survey questions.

Interviews with site leads

We conducted interviews with service leads across the seven sites, along with interviewees identified as being appropriate to ask about the costs of the service (to meet our fourth aim). Interviews were held online via Microsoft Teams and lasted for approximately 60 minutes. The interviews were recorded and transcribed to enable further analysis. Interviews explored in greater depth topics covered in the survey and captured contextual information to enable greater understanding. The interviews were semi-structured in nature to allow for questions to be tailored to the site and the interviewee. Those who were interviewed regarding costs were asked to share any documentation on the costs of the service with the research team. Please see the Data Availability section for access to the full interview topic guide.

Workshops with people with lived experience of smoking and SMI

After data collection with staff had been completed, and preliminary findings had been collated, the research team and our PPIE advisor facilitated three workshops with individuals with lived experience of smoking and SMI (n=15). At each of these workshops, we presented summary findings of the project to prompt discussion about how the findings resonated with those with lived experience of smoking and SMI and the implications for the future provision of tobacco dependency services. We sought attendee perspectives on the measures of success identified by staff involved in the delivery of the specialist community-based tobacco dependency treatment services and discussed whether those with lived experience of smoking and SMI would identify or prioritise different measures of success. Discussions also considered the service delivery models and referral pathways, reflecting on the potential benefits and limitations from a lived experience perspective. Two workshops were held online, via Microsoft Teams, and lasted approximately two hours, with participants recruited via the PPIE expert’s networks and BRACE PPIE networks. Participants for the third, in-person, workshop were recruited via a lived experience Mental Health Network⁵ who the research team had engaged with during the scoping phase. The workshop lasted approximately 90 minutes. All three workshops were recorded and transcribed to enable further analysis.

Workshop with NHS England and site leads

We also conducted a virtual workshop with national stakeholders and site leads to share summary findings and facilitate a discussion around the implications of these findings for the future provision of tobacco dependency services. The invitation was extended to all sites. Attendees included representatives from the NHS England Prevention Team (n=3) and representatives from four of the seven early implementer sites.

Analysis and reporting

Descriptive statistics were used to summarise the key characteristics of the survey dataset. This included the use of counts, percentages, and cross-tabulations to capture relationships between multiple questions. While most responses were analysed in aggregate across all respondents and sites, some exploration of the data at site level was undertaken. Bar charts were used to visually represent survey responses. These analyses were undertaken by one member of the research team (WP).

Interview transcripts were analysed using the framework method⁵³. The framework method is an approach to thematic analysis that is particularly suited for multi-disciplinary team and mixed-methods research. It involves an iterative process of coding and categorising qualitative data to develop and populate an increasingly sophisticated analytical framework. We used the framework approach within a rapid timeframe to provide a descriptive overview of tobacco dependency treatment services in the early implementer sites. We started data familiarisation during data collection by discussing key themes and insights from conducted interviews during weekly research team meetings (attended by JN, WP, SS, JD). Following this initial familiarisation, one interview was coded in NVivo software (by SS) to form an initial working analytical framework that aligned with the evaluation objectives. This framework was applied to one more interview transcript (by JD) before an analytical framework was agreed by the research team and subsequently applied to the remaining interview transcripts (by SS and JD). More in-depth analysis followed within the categories associated to the evaluation aims (led by SS and JD) to develop a more nuanced understanding of the relevant themes within each category; to understand similarities and differences across sites and to identify qualitative insights that could further describe or explain the quantitative survey findings.

To integrate quantitative and qualitative evidence into a cohesive narrative, the research team employed a triangulation approach that draws on aspects of the Pillar Integration Process⁵⁴. Key insights obtained from the quantitative evidence were listed. A matching process was then carried out, whereby qualitative data on the same theme was matched to the quantitative data, and similarities, differences, gaps, or other ways in which the data sets relate to each other were identified. This was carried out by three members of the research team (JD, SS, WP). Themes that were developed throughout this process were discussed within the research team, with a project advisory group (consisting of a PPIE representative, tobacco dependency practitioner and academic expert on substance dependency), and as part of the workshops with NHS England, site leads, and people with lived experience of SMI and smoking. These discussions informed the process of writing up the research.

The four core members of the research team (JD, JN, SS, WP) met regularly throughout the study to discuss emerging insights and held an internal synthesis workshop. Following this, all findings were written up for this report.

To ensure confidentiality, all findings have been reported anonymously, and we have not assigned details, such as job roles, to quotes.

Consent

Written informed consent was obtained from all participants in the study. Participant information sheets providing details of the project were shared with the participants, and they were given the opportunity to ask questions. Consent was obtained before participation. Survey participants had to opt-in to the survey by clicking a button to confirm that they consented to take part in the survey, which then allowed them to participate fully. The workshop and interview participants also reviewed an information sheet and provided explicit written consent before participating in the interview or workshop.

Ethics

The BRACE Centre and its evaluations are funded by the National Institute for Health and Care Research (NIHR) Health and Social Care Delivery Research (HSDR) program (NIHR 156533 – Birmingham, RAND, and Cambridge Evaluation Centre). The views expressed are those of the authors and not necessarily those of the NIHR or Department of Health and Social Care. Ethical approval for this service evaluation was provided by the University of Birmingham Humanities and Social Sciences Ethics Committee (Reference: ERN_3587-Dec2024) on 11^th December 2024. This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. All data were securely stored and managed in compliance with GDPR.⁶ Identifiable data were stored in password-protected systems accessible only to the research team at RAND Europe and the University of Birmingham. Survey data were initially processed by the THIS Labs using secure services.

Sample

Tobacco dependency services for people with SMI were implemented across seven early implementer sites in England. One of these sites subsequently split out into four satellite sites, so for the purposes of our analysis and to ensure confidentiality, we counted these satellite sites as separate entities. Hence, we refer to a total of ten early implementer sites, for reporting purposes. All ten sites were invited to participate in the evaluation. We were able to engage with nine out of these ten sites through either interviews and/or survey responses. Interviews with staff that held responsibility for more than one site are referred to as “Site 11” in our findings.

We conducted interviews with a total of 11 people, covering staff perspectives from six sites. This included six site leads and five additional staff with oversight of additional management, costs or clinical perspectives.

All but one site (Site 4) submitted at least one survey response, with a total of 42 participants completing the survey. Participation varied by site, ranging from two responses at Site 7 to eight responses at Site 8. It should be noted that the number of staff per site varies and we were unable to ascertain the total number of staff at each site. Additionally, participation by role type varied; respondents identified themselves as either practitioners (n=16), commissioners (n=9), or management (n=17). We received responses from all staff types for Sites 3, 6, 8 and 9. No commissioners participated from sites 1, 2, or 10, and no practitioners took part from Sites 5 and 7. However, we received management perspectives from all sites, except Site 4. After providing basic information (including role type and site) and consenting to participate, all subsequent survey questions were optional. Additionally, certain questions were only asked of certain role types, and most questions were optional. Therefore, not all 42 participants answered every question. The full sample of people that answered a question, and the number of people that gave a particular response are denoted throughout the text by N and n, respectively.

More information can be found in Table 1.

Main findings

In this section, we present the main findings according to the evaluation objectives outlined in the Introduction. The first section presents findings related to the tobacco dependency treatment service delivery models that were implemented across the early implementer sites. This includes the modality of tobacco dependency treatment consultations, the tobacco dependency treatment options being offered and staffing. The second section discusses measurement in tobacco dependency treatment services, including what measures were being collected locally, measurement challenges and perspectives on the most appropriate measures to understand service outcomes. The third section focuses on referral pathways. It addresses how referral pathways were established and adapted across the early implementer sites and the challenges experienced during this process. The final section discusses available data about the costs of delivering tobacco dependency treatment services for people with SMI and local developments in service commissioning.

Service delivery

The early implementer sites were commissioned to deliver tobacco dependency treatment services that are tailored to support people with SMI. Our findings highlight that service delivery models in the early implementer sites were tailored to this group by offering flexibility in the delivery mode of tobacco dependency consultations; in the tobacco dependency treatment options on offer; and flexibility to address service users’ concerns more holistically. Staffing models reflected the expertise required. These findings are discussed below.

Consultation modes provided by sites

The most frequently selected delivery mode was by telephone (85%, n=35), closely followed by in person in the community (78%, n=32). The least frequently selected mode was video call (27%, n=11). All but one person selected more than one delivery mode (98%, n=40). Services were delivered using multiple consultation modes both within and across early implementer sites. A complete breakdown is shown in Figure 1.

Figure 1. Intervention delivery modes, according to all staff (N=41).

This figure shows the breakdown of what delivery mode was used by each survey respondent. Survey respondents could select more than one answer. People who selected “I don’t know” or did not answer the question have been removed for the purposes of this analysis (n=1). This applies throughout. The total number of people that responded to each survey question is displayed in the figure title. This applies throughout.

The most frequently selected modes included by telephone and in person, with video call the least selected option.

There was a common agreement across the early implementer site leads we interviewed that offering tobacco dependency consultations in person was important for service users. Site leads described efforts to encourage in-person consultations. In some sites, these were encouraged through material products that could only be accessed in person, including direct supply of NRT or e-cigarettes, or performing carbon monoxide monitoring to assess smoking status and progress towards cessation. In person consultations were also encouraged by bringing consultations direct to service users – offering consultations in local community settings, in clinical settings where other appointments are attended, or in people’s homes – thereby removing barriers to attend. According to the survey, every site had at least one respondent selecting an in-person service delivery option (hospital, in the community or in service user homes), meaning an f in-person consultation was offered at least to some extent by each participating site.

Despite efforts to encourage in-person delivery, telephone consultations were common. Several site leads noted it was the dominant mode of delivering tobacco dependency consultations in the interviews and telephone was the most frequently selected mode in the survey (85%, n=35). Key reasons for the high rates of telephone consultations identified by site leads included the normalisation of remote consultations resulting from the COVID-19 pandemic, and difficulties in service users being able to make in-person appointments due to living far from the service or due to having conditions that made travelling difficult. This allowed for greater flexibility to accommodate service user needs. Other challenges to deliver in person consultations noted by interviewees, included difficulties securing appropriate community locations; non-clinical advisors being unable to deliver home visits due to safeguarding concerns and in-person consultations being more costly to deliver.

Adopting a flexible service delivery model

Early implementer sites offered various tobacco dependency consultation modes based on their service delivery models, which were split mainly between community and hospital-based services. Community-based services commonly provided in-person consultations, telephone consultations and home visits, with most site leads emphasising telephone-based delivery. Hospital-based services were primarily delivered in in-patient settings. Some hospital-based services also integrated with community mental health teams to offer home visits, community clinics and telephone consultations. Additionally, one site ensured continuity of care through telephone follow-ups after hospital discharge.

Figure 2 depicts the breakdown in what service delivery models were selected by respondents within each site. The figure illustrates a trend across most sites where certain delivery modes are chosen by the majority of (or all) staff, with one or two other modes also selected by a smaller proportion of staff. In some sites, there is significant variation in what modes are selected. For instance, in sites 2 and 8, there was not a single delivery mode option that was selected by all respondents at the site.

Figure 2. Service delivery models, by site (N=42).

This figure shows the breakdown of what delivery mode was used by each survey respondent, further broken down by site. It displays this as both a raw number, and as a percentage of total survey respondents by each site.

Interview data indicates that this variability is likely reflecting a combination of a standard operating model and taking a flexible, service-user-oriented approach. A common theme across the interviews we conducted revealed that despite operating based on a service delivery model with one or two dominant delivery modes, the service was adaptable and tailored to the mental health needs, communication needs, routines and preferences of individual service users. This is corroborated by the survey data given that all but one respondent reported more than one mode for delivering tobacco dependency consultations (98%, n=40).

Consultation modes sometimes changed over time. In person consultations were sometimes followed-up by telephone consultations to support patients who were discharged from a hospital to community setting, or if lower-intensity support was deemed appropriate. For other service users, telephone preceded an in-person consultation to allow relationship building between service users and their stop-smoking advisor, which was seen as an important mechanism to motivate engagement with the tobacco dependency treatment service and smoking cessation.

Tobacco dependency treatment support for SMI needs

Across the early implementer sites, a range of tobacco dependency support options were offered to address the needs of people with SMI that are not commonly addressed in standard tobacco dependency treatment services. The UK standard guidance for tobacco dependency treatment as outlined by the National Centre for Smoking Cessation and Training (NCSCT) involves supported quit attempts of about six consultations and four weeks (28 days) of professional support, or longer if needed, after setting a quit date and developing a structured quit plan²⁸. All early implementer sites introduced adaptations to the standard offer to meet the needs of people with SMI. One site lead described:

Figure 3 outlines the range of support options offered. It shows that 95% of respondents (n=39) reported supporting people to cut down cigarettes before setting a quit date, making ‘cut down to stop’ the most selected option. While ‘relapse support’ was least frequently selected, it was still selected by most staff (66%, n=27).

Figure 3. Support offered by staff (N=41).

This figure shows the types of tobacco dependency support offered to service users. A wide variety of support is offered by staff. Full definitions of each type of support is contained in the survey. To access the full survey questions, please see the Data Availability section.

Responses by site demonstrate that a substantial minority of staff (41%, n=17) selected every tobacco dependency support option. These staff came from seven of the nine sites, indicating that all support options were available in at least seven out of the nine sites.

The SMI-tailored tobacco dependency support options in Figure 3 address particular challenges experienced by service users. More specifically, site leads mentioned issues such as fluctuations in mental health impacting people’s ability and motivation to engage with services; relying on smoking to cope with the experience of poor mental health; de-prioritisation of smoking cessation when faced with poorer mental health or other life challenges; heavier smoking in people with SMI and higher risks of relapse in people with SMI.

Additional time and flexibility in the delivery of tobacco dependency treatment services were highlighted by site leads as the most important accommodations to the challenges faced by service users. This includes more time and flexibility to set a quit date with service users; to offer ongoing support for quit attempts that are temporarily interrupted and to address potential relapse.

Site leads also emphasised the role of e-cigarettes in motivating engagement with the tobacco dependency treatment services. Besides the provision of NRT (either directly or indirectly through referral elsewhere), which was reported by all survey participants, most participants also reported providing access to e-cigarettes or vapes (90%, n=36). E-cigarettes were offered free of charge across the early implementer sites through mechanisms like the national swap-to-stop scheme⁵⁵, store vouchers and direct supply within the service.

In addition to a tailored tobacco dependency support offer, site leads highlighted the need for a relational approach that is understanding and compassionate towards the challenges experienced by service users with SMI.

Holistic support for service users

Tobacco dependency, especially in people with SMI, was recognised as a complex biopsychosocial problem that required a holistic approach. Different types of holistic support options were delivered across the early implementer sites, including tobacco dependency treatment support for carers, family members or friends; treatment for alcohol and drug dependencies; as well as support for health and wellbeing challenges beyond smoking. Several site leads explained that their service was set-up to support service users with additional issues that impacted on their smoking.

All site leads operating community-based tobacco dependency treatment services mentioned that they offered or organised additional support for service users to address challenges beyond smoking. Similarly, 100% of survey respondents (n=39) said they provide additional support to service users beyond smoking cessation alone, by providing information or by signposting to other services. The degree to which this support was integrated within the service or identified externally varied across sites. In one site, tobacco dependency advisors were trained to do a holistic support assessment, followed by coaching support or connecting people with external community services depending on the issues. In another site, the provider organisation offered additional services beyond tobacco dependency treatment that service users could access. Across community-based sites, service users were signposted to other community services as part of opportunistic conversations about health and wellbeing.

Site leads operating hospital-based tobacco dependency treatment services did not usually discuss connections with wider community services. Instead, they discussed connecting with hospital-based services to offer tobacco dependency treatment support alongside mental health and/or clinical care. One site lead highlighted the importance of inter-organisational collaboration to address the full complexity of needs faced by a proportion of service users with SMI.

Tobacco dependency staffing and training

Different types of tobacco dependency advisors were employed across the early implementer sites, including health and well-being coaches, social prescribers, and nurses. Site leads discussed opportunities across these roles for delivering tobacco-dependent treatment to people with SMI. For example, clinical advisors can conduct home visits and often prescribe treatments such as NRT, while health and well-being coaches are trained to support lifestyle-related challenges such as physical activity or healthy eating.

Tobacco dependency advisors received both tobacco dependency and mental health-related training to deliver a service sensitive to the needs of people with SMI. The training involved both NCSCT courses and bespoke training delivered in-house. This is corroborated by the survey, with all staff delivering the intervention reporting having received at least some form of relevant training (100%, n=37). Most site leads operating community-based tobacco dependency treatment services highlighted the in-house delivery of regular professional development updates on tobacco dependency treatment. In one of these sites, the local mental health trust was commissioned to deliver mental health-specific training. The site leads operating hospital-based tobacco dependency treatment services also mentioned offering training to their local community-based tobacco dependency treatment services, either currently or as a future goal. Hospital-based tobacco dependency treatment services, specifically, held a great degree of interdisciplinary expertise.

Several sites had trained all their tobacco dependency advisors to support people with and without SMI, while other sites recruited a limited number of tobacco dependency advisors responsible for the SMI cohort. Importantly, all tobacco dependency advisors who responded to the survey question on staff confidence (n=15) said they felt confident delivering the intervention.

Reflections of people with lived experience of SMI and smoking

The need for tailored tobacco dependency treatment services to address SMI-specific needs resonated with the experiences of the workshop participants. Participants with lived experience of SMI and smoking shared many of the challenges with smoking and smoking cessation that the tobacco dependency treatment services in the early implementer sites aimed to address. For example, several participants shared they had stopped, and started, smoking multiple times – sometimes due to fluctuations in mental health – emphasising the importance of relapse support.

Participants discussed the role that smoking can play in someone’s mental health experience. They described using smoking as a ‘crutch’ in times of poorer mental health, to manage potential side effects of prescribed antipsychotics like lethargy or low energy levels, and to self-medicate when prescribed psychiatric medications weren’t perceived to be effective. The importance of tobacco dependency advisors being trained in mental health and compassionate towards such experiences was stressed.

Moreover, workshop participants agreed that the delivery of tobacco dependency treatment services needs to flexible and adaptable to individual service users to accommodate the diversity of experiences and needs of people with SMI.

Measures of success

Across the early implementer sites, there was variation in the measures were collected by sites to measure success. Most sites collected quit rates as a minimum (e.g., at 28-days or longer), alongside other metrics such as referral numbers and type of intervention. The level of detail and granularity of the measures collected varied. Some sites also sought to collect more qualitative outcomes or feedback, such as service user feedback or case studies of service user experiences. Sites also faced challenges when collecting measures related to data quality, data access, and difficulties in understanding what they should have been collecting. Although quit rate was the most collected measure, feedback from sites and workshop participants highlighted the need for a variety of measures to effectively demonstrate success and service outcomes.

Collected outcomes and measures of success

The majority of sites mentioned that they collected quit rates, with one site describing the four-week quit rate as their “bread and butter”. This is in line with NHS England reporting guidelines, and is the standard measures used by generic (non-SMI tailored) services. However, a common theme from interviews was that quits may take longer than the standard 28 days, so longer term quit rates were also collected.

Other metrics collected as part of reporting included referral data (i.e. how many people have been referred into the service), level of engagement at the service (e.g., how many appointments), pathway selected (e.g. tobacco reduction versus quit date), and the type of intervention used and the mode.

Sites also utilised individual metrics related to their service aims. For example, a key measure of success for one site was being able to see a service user within 24 to 48 hours from referral. Another metric used at a site was the conversion rate, the proportion of service users that successfully quit, which was used to show the quality of the service and the ability to support people to stop smoking. Some sites also had specific key performance indicators set by their local authority, commissioners, and/or the ICB, or had plans for these to be developed and added to the service in the future. While, NHS England had set reporting for sites and provided a template for which measures needed to be reported.

Alongside these measures, some sites collected feedback and case studies to highlight success and understand service user perspectives of the service. This included collating information around service accessibility for service users and overall experience of using the service.

Another prominent pattern that did emerge is that sites were collecting different metrics to one another. Other than the short-term quit rates, there was a wider variety in the number and granularity of data.

Data collection and measurement challenges

There were some challenges identified by sites related to collecting and measuring outcomes of success. Some sites had challenges related to data quality, and the way in which outcomes were recorded was also a challenge in some cases.

Sites also noted some initial difficulty with understanding exactly what they needed to collect and what measures would be important to, for example, NHS England. There were also challenges in translating data into formats from non-NHS organisations (e.g. a local authority) to NHS data fields.

Staff perspectives on measures of success

In the survey, staff were asked about which measures of success they thought it was important to collect in the survey. The smallest proportion of survey respondents 60% (n=21) agreed or strongly agreed that the commonly used 28-day quit rate was important to measure service impact, compared to 100% (n=38) of respondents identifying impact on physical health as an important measure. Service engagement, service user experience of the service and overall physical health were all more frequently deemed to be important measures of success, according to those completing the survey (See Figure 4).

Figure 4. Measures of success that participants consider to be important (N=38).

Survey respondents were asked to rate on a scale how important they thought each measure of success was. Interestingly, quit rates were least likely to be rated as important, and broader measures such as impact on physical health was rated as important by all respondents.

One participant emphasised the importance of collecting data in a consistent manner across sites and encouraged collaboration across teams. Another highlighted the importance of collecting metrics that allow sites to be able to report back to their commissioners, hospital board, or equivalent, to highlight the impact of their service and the benefits that have been achieved.

In summary, the survey data revealed staff felt it was important to collect broader, more long-term outcome measures.

Reflections of people with lived experience of SMI and smoking

Further to this, people with lived experience of SMI and smoking who attended our workshops reflected that whilst quitting was an obvious aim of the service, there should be recognition of the wider benefits the intervention brings, with some participants saying that the service could be viewed as a form of broader health and wellbeing support. This sentiment was echoed across all groups we spoke to, and study participants suggested additional metrics that were important to measure the impact of the service. A compilation of potential measures of success provided in the survey, site lead interviews and lived experience workshops are shown in Box 1:

Referral pathways

The early implementer sites were commissioned to pilot one of three patient referral pathways: the primary care model, the in-patient/outpatient mental health trust model, and the community mental health services model. As discussed below, our findings highlight that most sites had adapted their referral pathways since they were initially commissioned, but this has revealed multiple barriers to expanding pathways.

Referral pathways expanded since initial commissioning

Early implementer sites were initially commissioned to pilot a specific referral pathway. However, according to site lead interviews, most sites have since gone on to expand their pathways. This is corroborated by the survey, where 81% (n=30) of respondents said referral pathways had changed, with at least one participant from each site saying this. Pathways changed most often by taking referrals from a broader array of services (62%, n=23), including other mental health services, drug and alcohol services and community services such as job centres. Some survey participants said their site had expanded their referral pathways by broadening service user eligibility criteria (30%, n=11) or by accepting referrals from a broader geographical region (38%, n=14). Figure 5 depicts this in more detail.

Figure 5. How referral pathways have changed (N=37).

This figure displays the breakdown in how survey respondents thought referral pathways had changed since the service was first established. Most frequently, staff said that referrals were now accepted from a broader range of services than they were when first set-up.

Hence, sites had multiple patient referral pathways in place at the time of data collection. As shown in Figure 6, all sites had at least one person in the survey select referral pathways from community mental health services and self-referral. Eight of the nine participating sites had at least one person selecting inpatient services (a Mental Health Trust), seven sites had selected the outpatient services pathway and six sites had selected GP services. Five out of the nine sites had all referral pathways selected at least once, and all sites had selected at least two different referral pathways. The relationship between individual survey responses and the aggregated site level response is consistent with the patterns observed in service delivery mode, with certain referral pathways selected by all staff within a site, with other pathways often also selected, but by fewer staff. Overall, this demonstrates that staff are utilising a variety of different referral pathways are in place across sites.

Figure 6. Referral pathways, by site (N=40).

This figure illustrates the referral pathways that survey respondents believed were available, categorised by site. It displays this as both a raw number, and as a percentage of total survey respondents by each site. The figure shows that a wide variety of different referral pathways are in place. Most sites had one or two pathways that were selected more consistently by staff, with a smaller number of respondents selecting pathways that other respondents in their site did not select.

Why referral pathways were expanded

According to some participants, the main reason pathways were expanded was to increase the number of referrals coming into their service.

One site lead discussed that it might be more challenging to get referrals for a service that is specifically tailored to people with SMI, as it is a relatively small group of people compared to a service for the general population. Hence, a more targeted and pro-active approach to identify and engage eligible service users was required.

Not all sites expanded their pathways to the same extent. Sites that were operating a hospital-based service and initially used the in-patient referral model had the highest proportion of survey participants selecting that referral pathways had not changed (Sites 1 and 5, see Table 2).

How referral pathways were expanded

The most selected option in the survey regarding how referral pathways were expanded was referrals from a broader array of services (62%, n=23). Frequently, interviewees mentioned awareness raising and relationship building across a range of services that are frequently utilised by individuals with SMI, as common ways to do this in practice. Awareness raising was often done by offering training in the Very Brief Advice (VBA) intervention to encourage quit attempts⁵⁸, offering more bespoke training, and collaboratively addressing barriers and challenges experienced by referrers. These efforts were intended to create trusted and collaborative relationships between tobacco dependency staff and referral partners.

Challenges with establishing new referral pathways

A critical aspect of establishing pathways, is being able to identify people with SMI who smoke and to engage them in smoking cessation. Ways to do this varied, but site leads mentioned the use of NHS databases, building relationships with referrers, awareness raising, and pro-active service user engagement strategies. However, site leads discussed several challenges associated with embedding these mechanisms within their local settings.

One way to identify whether people have an SMI and smoke is through NHS medical records. One site lead shared how they utilised NHS databases within their service to identify people with SMI who smoke and were in-patients at that time, so they could reach out to them pro-actively.

Most community-based services, however, did not have access to NHS medical records, which made it challenging to identify suitable prospective service users. Without access to medical records, sites had to rely on health professionals in other organisations to enquire about the SMI status of people. SMI status is usually not explicitly captured when people are referred for generic smoking cessation support. Site leads explained that referrers are often not aware of someone’s SMI status and may find it inappropriate to enquire about someone’s medical diagnoses, either because they do not feel they have the expertise or because they are worried about damaging relationships. In addition, prospective service users might not want to disclose their SMI status.

Additional barriers were identified regarding interacting with healthcare professionals more broadly. For example, site leads mentioned that some clinical staff do not see referring people as their job or a priority.

Additionally, there is a reported misconceptions amongst healthcare staff about the motivation of people with SMI to stop smoking.

Lastly, site leads revealed that the short-term funding cycles of the service was counter-productive to establishing referral pathways, as many of these pathways were established through relationship building.

Reflections of people with lived experience of SMI and smoking

The workshop discussions with people with lived experience of smoking and SMI confirmed the need for multiple pro-active approaches to reach people with SMI who smoke to engage them in smoking cessation. Participants shared varied experiences with referrals to tobacco dependency treatment services that reflected the challenges shared by site leads. While one participant highlighted that they were consistently encouraged to quit smoking during periodical physical health reviews, another participant’s experience confirmed that some mental health professionals still reinforce misconceptions about the motivation and ability of people with SMI to quit smoking.

Workshop participants also highlighted variable engagement with mental health services and physical health reviews that could cause inequalities in being offered tobacco dependency treatment support if referrals are solely accepted from these services. They especially emphasised the importance of accepting self-referrals.

The diversity in participant experiences validates the widespread efforts made across the early implementer sites to identify and engage potential service users through different referral routes.

3.2.4. Costs

Availability and comprehensiveness of cost information

Four sites did not report any cost information and only three provided a detailed breakdown of costs. Staff costs were the predominant cost of service delivery. Across the three sites that provided a breakdown, staff costs comprised between 70-85% of the total service costs. Multiple site leads highlighted challenges with recruitment and staff turnover.

Other service costs, besides staff, that were reported included staff travel costs, staff training, clinic rooms, NRT and e-cigarettes, carbon monoxide monitoring equipment, educational and promotional materials and contributions to organisational overhead costs.

Some of these were hidden costs. That is, while costs were reported in the interviews, they were not included in the total service costs. Instead, they had been absorbed in the delivery of other services. One interviewee, for example, highlighted that management for the SMI service was absorbed within the budget for the standard tobacco dependency treatment service and delivered by the same provider organisation. Similarly, training was often delivered in-house as part of the continuing professional development activities already in place. Another site lead highlighted that travel costs were heavily sponsored by the local authority, and clinic venues were offered free of charge or at heavily discounted rates by community and NHS organisations. Not all sites that provided a cost breakdown included overhead costs. While commissioners in some sites encouraged service providers to absorb the SMI service within their usual model, some site leads pointed out additional costs related to the SMI service.

Site leads noted the need for tobacco dependency commissioning models to consider the additional time and flexibility required to deliver services for people with SMI. One participant noted that similar accommodations are needed, which then necessitate associated costs, to make tobacco dependency treatment services more accessible and equitable for other groups, such as people with learning disabilities. At the time of writing, local negotiations were taking place to decide if and how community-based tobacco dependency treatment services for SMI will be commissioned in the future.

Understanding the service delivery model of tobacco dependency treatment services for people with severe mental illness

All early implementer sites were delivering evidence-based tobacco dependency treatment services tailored to the needs of people with SMI. The service delivery models that were implemented across sites varied in terms of the service setting (hospital or community-based), the primary mode for delivering tobacco dependency consultations (in person or telephone), and the types of staff employed as tobacco dependency advisors (including health and wellbeing coaches and nurses). Remote (e.g., telephone) and in-person in the community consultations were most frequently utilised, and all sites had at least one participant stating that an in-person consultation mode was available. A wide variety of different types of support were offered by practitioners, many of which included going above and beyond help stopping smoking, including offering support after relapse and connecting with broader health and wellbeing support. Of central importance to most respondents was the value of accommodating a service user-centred approach, in which the exact specificities of the intervention were adapted to the needs of each individual service user.

This user-centric approach was commonly mentioned by site lead interviewees and strongly endorsed by people with lived experience. Specifically, this included flexibility in the delivery mode of tobacco dependency consultations and offering flexibility in the timing and intensity of tobacco dependency treatment support. Many of the different types of support reported by the sites are associated with increased smoking cessation effectiveness compared to usual care in the broader evidence base^25,29,30,59. While this evidence is stronger for some modes and support types than others (there is more evidence for the effectiveness of in-person consultations than for telephone-delivered consultations^60–62, and uncertainty around the benefits of cut-down to quit compared to abrupt quit attempts⁶³), they enabled support for individuals who might not otherwise engage with tobacco dependency treatment support. This service user centric approach aligns with guidance published in 2024 for the commissioning, delivery and monitoring of tobacco dependency treatment services, which highlights the need for such services to offer tailored support not only to people with SMI, but also to other ‘priority groups’ with high smoking rates and additional support needs⁴³.

Our findings demonstrate that a wide variety of delivery models can be offered, and that they can be adapted to the specific needs of people with SMI. Staff and people with lived experience emphasised the value of this, and the broader evidence base also supports this, suggesting that this is a key service aspect that should be retained and prioritised, going forward.

Exploring how success of the specialist mental health tobacco dependency services is defined and measured locally

Different types of measures were collected across the early implementer sites to understand the success of their service, including 28-day quit rates, referral rates, service utilisation, service user characteristics, and service user experience. The 28-day quit rate was consistently measured across all sites. Whilst the majority of survey respondents agreed this was an important measure to understand the success of their service, it was least frequently selected as important amidst other measures of success. Longer term quit rates, specifically, are necessary to evaluate tobacco dependency treatment services for people with SMI where support often extends beyond 28 days. Short-term, 28-day, quit rates are also insufficient to understand whether the health benefits of long-term smoking cessation are being realised⁶⁴. Our findings demonstrate alignment between best-practice recommendations and staff views on the importance of collecting longer term quit rates, suggesting these are a priority to consider in local and national monitoring frameworks^43,64,65.

Site staff and workshop participants with lived experience of smoking and SMI also called for more holistic measures of success that capture wider impacts of tobacco dependency treatment services. Instead of quit rates, service user experience and impacts on physical and mental health were more often considered to be important measures of success. This aligns with the ‘quintuple aim’ for measuring healthcare improvement, which poses that health services should simultaneously aim to improve service user experience, improve health outcomes, reduce costs, improve provider experience and advance health equity⁶⁶.

Current monitoring frameworks and recommendations focus predominantly on service-level measures and quit rates as a proxy measure for improved health outcomes^43,67. However, our findings suggest that other measures are needed to capture the holistic impact of tobacco dependency treatment services for people with SMI. Site staff and workshop participants with lived experience of smoking and SMI pointed out that engagement in tobacco dependency treatment services often has additional benefits beyond smoking cessation, which aligns with our service delivery findings demonstrating that tobacco dependency advisors develop compassionate therapeutic relationships with service users and often support service users with health and wellbeing challenges beyond smoking cessation. Moreover, some sites collected information on service user characteristics to measure how local services impacted specific service user groups and assess advancements in health equity. A consistent measurement and monitoring approach that includes the measures discussed in this report would allow for larger-scale data interrogation to better understand the impacts of tobacco dependency treatment services for people with SMI.

Understanding referral pathways into the specialist mental health tobacco dependency service

When the early implementer sites were initially commissioned, they had one of three distinct referral pathways in place. Since then, most sites have expanded their referral pathways to receive referrals from additional services, including mental health services, public health services, voluntary and community sector services, as well as implementing the option for self-referral. The main reason for doing this was to increase the low referral numbers coming into the service. Sites reported doing this by establishing new ways to identify and engage prospective service users, and by building relationships with other organisations. However, site leads highlighted several barriers facing them in their efforts to increase referrals. Some services reported challenges in getting colleagues in other organisations to make referrals. For example, sites reported there being a misconception amongst healthcare professionals around the motivation of people with SMI to quit, and that community partners were reluctant to on the grounds of disrupting their rapport with service users. In addition, the short-term funding cycle of the service sometimes limited the hard work done in establishing referral pathways, especially when they are based on building personal relationships.

Specifically, methods to establish new pathways included advertising the tobacco dependency treatment services in local communities, raising awareness in local services, and training health and mental health staff in delivering the Very Brief Advice (VBA) intervention to motivate and support people to quit smoking⁵⁸. Staff and workshop participants with lived experience of smoking and SMI emphasised the importance of pro-active and direct efforts to reach and engage service users in the community – both to motivate and empower people directly and to include people with SMI who engage less with mental health services. These efforts to develop referral pathways also align both with national guidance on increasing the reach and access to smoking cessation support for people who smoke⁴³ and with behaviour change models, such as the transtheoretical model of health behaviour change⁶⁸ and the COM-B behaviour change wheel⁶⁹. Misconceptions around smoking and SMI within healthcare has also been found in the wider evidence base. For instance, misconceptions about the ability of people with SMI to quit and that quitting may actually jeopardise a person’s mental health^1,20,50. This all aligns with the findings reported by site leads and was ratified by people with lived experience who also spoke about the broader barriers to engaging with all mental health services generally. Further, the barriers to establishing referral pathways are not unique to this area, but rather a common barrier faced by public health services on short-term funding cycles.

Our findings demonstrate how the implementation of referral pathways involves efforts to inform, educate, persuade and train referral partners to encourage greater referrals. Self-referral is a particularly important new pathway in this regard, empowering service users and ensuring there is a referral route available for people who may not have an SMI diagnosis or prior engagement with mental health services. Hence, developing referral pathways should thus be understood as a critical component of successful tobacco dependency treatment services. However, many barriers exist, including misconceptions amongst staff and short-term funding. Efforts to overcome these barriers should be considered and integrated into the commissioning, implementation and delivery of these services.

Identifying cost data associated with the delivery of different models of specialist mental health tobacco dependency services

This evaluation also aimed to explore what data is available for costs associated with different delivery models of tobacco dependency support for those with SMI. The availability and comprehensiveness of this data varied by site. In sites that were able to provide a breakdown, staff costs were the predominant cost of service delivery accounting for between 70–85% of the total service costs. However we noted that some hidden costs were not captured when finances were reported, for example costs donated in kind by other services or absorbed under the costs of wider services (such as tobacco dependency services for the general population).

The economic evaluation of the SCIMITAR+ trail indicated that a bespoke approach to smoking cessation for those with SMI did not increase overall costs of the service compared with a standard service⁷⁰ and was likely to be cost effective over 12 months compared to usual care⁵¹. Findings from our evaluation however indicate that costs within the pilot sites are currently poorly and not systematically captured. Therefore, we are unable to comment on the relative costs of the service and how this compares to the findings of the SCIMITAR+ evaluation. Some participants noted it was challenging to demonstrate the benefits of the service in terms of outcomes (see previous section) and costs.

To enable a more in-depth estimation of the relative costs and benefits of the service, more systematic capturing of costs and outcomes is needed across sites. However, this could be particularly challenging at this point in time. Since the commissioning of this evaluation, the UK government announced the abolition of NHS England^44,45 and 50% reductions to the running costs of ICBs^46,47. These reorganisations bring further uncertainties to the future of tobacco dependency treatment services for people with SMI and staff working the service, both within the early implementer sites and more widely.

Recommendations for tobacco dependency services for people with SMI

To ensure the future success and value of the early implementer services, the following should be considered by national policy makers and commissioners:

Recommendations for future evaluation and research

Based on our evaluation, we recommend the following to researchers seeking to conduct similar or additional work in this area:

Limitations

This evaluation was conducted within a short timeframe. The period from agreement of the scope of the evaluation to producing preliminary findings to inform the Government's Spending Review, was three months. As a result, we were constrained with what was feasible in terms of scope, design, data collection, analysis and reporting of our learning. Consequently, the views of people that have used the services have not been included. Instead, we sought the views of those with lived experience of SMI and smoking.

The service is heterogeneous and has been adapted to local contexts. We have identified different approaches, referral mechanisms and service user populations being targeted. As such, forming transferable, broad recommendations is challenging due to the differences between the sites included in this evaluation.

Further to this, there were challenges obtaining cost data from the sites in part, due to the difficulty identifying key personnel who had sufficient knowledge of detailed costs and/or would be able to share information with us. Reasons for this may include commercial sensitivities and/or services that are no longer running.