a. Behavioural interventions.

Of the four reported behavioural interventions, the Targeted Training in Illness Management (TTIM) found no significant difference in HbA_1c (glycosylated haemoglobin) between the intervention and treatment-as-usual groups after 60 weeks, but did find a significant improvement in diabetes knowledge in the intervention group compared to usual treatment [28,32]. A bespoke cardiovascular risk reduction intervention (PRIMROSE), with emphasis on personalised risk factor management, guiding patients to relevant services, and initiating and continuing indicated pharmacotherapy such as statins, showed no improvements in total blood cholesterol over 12 months compared to usual care [30]. Shaw et al. described an adapted PRIMROSE-A intervention, with further focus on statin prescribing, care integration, and patient mental health, and found it to be positively viewed by patients and staff [33]. Lee et al.et al. described an adjunct mentor support intervention for SMI patients partaking in a mainstream weight loss program, and found it to be feasible and acceptable [29].

b. Modified primary care models.

A cohort study by Gertner et al.et al. described an enhanced primary care model, structured as a patient-centred medical home (PCMH), which provided tailored care for SMI patients, with care coordination, peer support and self-management programme provision, where patients were seen, on average, 6 times annually in 30–40 minute appointments [26]. Further modifications included smaller patient panels, as well as specific staff training on treating people with SMI [26]. Primary care providers also had close working relationships with patients’ behavioural care providers in this model, with monthly meetings for discussion of patient needs [26]. This model was associated with decreased HbA_1c and systolic blood pressure, as well as increased HbA_1c, LDL and blood pressure screening rates, compared to standard primary care [26].

c. Provider education.

An RCT by Evins et al.et al. evaluated a combined intervention of primary care provider education and community health worker (CHW) support for smoking cessation [25]. The combined intervention improved long-term tobacco abstinence study, highlighting the value of personalised support and primary care provider education [25].

d. Clinical tool implementation.

An RCT by Rossom et al.et al. described the implementation of a clinical decision support (CDS) tool that provided individualised CVD risk summaries and treatment recommendations for SMI patients, which was associated with a 4% relative CVD risk reduction compared to control at 12 months, but few significant reductions in individual risk factors such as smoking [31]. A cohort study by Dronavalli et al.et al. found that financially incentivised care plan use was associated with a reduction in daily tobacco smoking rates and excess body weight compared to lack of care plan use [24]. Other studies also incorporated clinical tools, such as care-plans, or information booklets or leaflets for patients, but as components of more complex interventions [28–30].

a. Staff training.

Staff training was a key component or prerequisite in seven included studies. Evins et al.et al. reported a smoking cessation intervention where education of primary care provider (PCPs) including physicians, nurse prescribers and physician assistants, aimed to increase smoking pharmacotherapy use [25]. PCPs were educated by trained doctoral level staff about the risks and benefits of first-in-line prescriptions, particularly varenicline, for SMI patients through in-person educational outreach and written material provision [25]. The PRIMROSE intervention reported by Osborn et al.et al. and Hassan et al.et al. contained a 2-day training package and provision of a manual for general practice nurses outlining behavioural change interventions informed by the behavioural wheel strategy, regarding risk factor management individualised for patients [27,30]. Nurse practitioner training was also a key component of the PRIMROSE-A intervention described by Shaw et al. [33]. In the TTIM intervention nurses received two days of initial training on a diabetes management intervention for SMI patients from the principal investigators and other team members, including peer educators, as well as communication training from a psychiatrist [28,32]. The provided education was continuous, with quarterly training sessions for discussion of progress and development of skills delivered via in-person meetings, phone calls and email [28]. Peer educator involvement was important for modelling positive SMI and diabetes related behaviours during training in this intervention [28]. In the PCMH described by Gertner et al.et al., staff received specialised training on working with SMI patients, to better understand aspects of care unique to SMI patients, and to learn how to de-escalate crisis situations [26].

b. Engagement and fidelity.

Patient attendance was mentioned in three studies and was found to be generally acceptable. Osborn et al.et al. reported that 46% of patients attended 6 or more appointments out of 12, Lee et al. that 12 out 16 patients attended over half of the weekly check-ins, and Lawless et al. that 61% of patients attended all sessions, for each respective intervention [28–30]. Staff fidelity to intervention delivery was mentioned in two studies. Osborn et al. reported moderate adherence, with 67.7% of PRIMROSE intervention components delivered to protocol, though statin initiation was still low in both the treatment and usual care group, and Lawless et al.et al. reported near complete fidelity to the manualised TTIM intervention content, apart from some sessions lasting longer than the allocated 90 minutes [28,30].

c. Cost efficiency.

Cost efficiency was reported in only one study, with Osborn et al.et al. reporting that PRIMROSE was on average £895 cheaper per patient than usual treatment over 12 months, primarily due to reduced psychiatric admission costs [30]. Rossom et al. reported that CDS tool implementation incurred a much smaller cost per patient than more intensive interventions, potentially due to its automated nature, requiring minimal maintenance when guidelines are updated, but no figures were provided [31].

a. Nurse practitioners and healthcare assistants.

Nurse practitioners were integral to intervention delivery in five included studies, with health care assistants also delivering interventions in two of the aforementioned studies. The TTIM intervention involved trained nurse educators who delivered group-based sessions, supported personalised care plan formation, educated patients about diabetes self-management, and communicated with patients’ care providers [28,32]. The PRIMROSE and PRIMROSE-A interventions respectively were primarily provided by nurse practitioners, as well as health care assistants, who delivered behavioural change interventions during appointments with SMI patients [30,33]. They aided in aspects such as personalised care plan formation, goal setting, treatment adherence monitoring, guiding patients to relevant supportive services, and initiating pharmacological therapies were indicated [30,33]. A registered nurse was also part of the multidisciplinary team of the PCMH reported by Gertner et al., but their roles were not elucidated further [26].

b. Peer support.

The provision of peer support for patients, by people with lived SMI experience, was reported in four included studies. The TTIM intervention involved peer educators who shared personal experiences about diabetes management and SMI with patients, worked closely with nurse educators and modelled effective self-management interventions [28,32]. In the PRIMROSE-A intervention, coaching sessions were offered by a peer-coach with lived experience of SMI, involving discussion of personal experiences and goals [33]. Two trained peer specialists were also involved in the multidisciplinary team of the enhanced primary care model described by Gertner et al., though their roles were not elucidated [26].

c. Mentor support.

The WHEEL intervention described by Lawless et al. involved knowledgeable mentors of different professions, such as a graduate psychologist, who offered weekly scheduled calls with SMI patients enrolled in a mainstream weight management programme [29]. Mentors provided adjunct practical support to patients, and helped locate and activate memberships for the mainstream programme [29].

a. Patient-level barriers and facilitators.

Lee et al. identified patient fear of judgement and unfamiliar or social situations due to hallucinations negatively impacting their desire for social interaction as a barrier to patient attendance, with paranoia also identified as a barrier [29]. Fatigue related to antipsychotics was also reported to negatively impact patient ability to attend sessions [29]. Hassan et al. reported that health care plan use was viewed as both a facilitator and barrier by patients, with some reporting problems with repetitiveness or challenges filing them in, and others finding them helpful to track progress [27]. A strong therapeutic relationship between patient and carer was identified as key facilitator for patient engagement across three studies, by increasing patient willingness to engage [27,29,33]. The availability of a telehealth option was identified as a facilitator of patient engagement by Shaw et al., especially during periods of SMI symptom exacerbations, though some patients disliked the impersonal feel [33]. Notably, there was limited reporting on other potential patient-level barriers such as cognitive impairment, financial constraints, or broader access-related challenges, suggesting these factors may be underexplored in the current primary care intervention literature.

b. Staff-level barriers and facilitators.

Hassan et al. identified prior staff experience of working with SMI patients as a facilitator, as it enabled knowledgeable, confident and open interactions with patients, while a lack of SMI knowledge and experience among some staff led to anxiety and negatively impacted intervention delivery [27]. Staff training was identified as a facilitator, with Lawless et al. reporting that nurses generally found the TTIM manual useful for intervention delivery, Hassan et al. that training was essential in increasing staff confidence and competence interacting with SMI patients and Shaw et al. reporting that staff perceived provided training as useful, especially due to addressment of initial education and knowledge disparities [27,28,33]. Inadequate training was seen as a barrier, with Shaw et al. reporting that a lack of training content depth reduced staff confidence in intervention delivery [33]. The use of care plans was viewed as both a barrier and a facilitator by staff, similar to patients. Hassan et al. reported that some staff found written care plan implementation into routine practice difficult and time-consuming, however Lawless et al. mentioned that personalised care plan use helped nurses better understand individual patient’s learning requirements [27,28]. Mental health stigma among staff was another barrier to care delivery, with Hassan et al. reporting that preconceptions among some staff negatively impacted their perceptions of intervention delivery capability, and reduced willingness to participate, though most staff were motivated and had a positive attitude [27].

c. Logistical barriers and facilitators.

A key barrier identified across two studies was the need for increased time and resource allocation. Hassan et al. reported that implementation of the PRIMROSE intervention into busy primary care settings resulted in staff concerns and doubts regarding its feasibility, particularly in relation to increased workload, time constraints within routine consultations, and the practicality of integrating longer, intervention-specific appointments into existing schedules [27]. Contextual integration challenges also impaired staff intervention delivery, with staff facing challenges dedicating sufficient time to facilitate patient engagement and accessibility, and fitting longer appointment times around their daily schedules [27]. Similarly, Shaw et al. found that time and resource allocation for PRIMROSE-A was a common staff concern, especially regarding long-term integration into practice outside the trial setting [33]. However, extended appointment times and regular contact were identified as key facilitators, though nurses emphasised that this extra time must be utilised effectively [33]. Further barriers identified by Shaw et al. included problems with reimbursement, general organisation, and communication problems between healthcare and managerial staff, for example with the lack of provision of peer mentor contact details to healthcare staff in certain instances [33]. Lawless et al. identified inconvenient class times, transport limitations, inconsistent attendance, inconsistent peer behavioural modelling due to heterogeneity of symptoms, and challenges contacting patients as logistical barriers to intervention delivery [28]. Facilitative solutions implemented by nurses to tackle such problems were also reported, including changing of class times based on patient feedback, and the nomination of a secondary contact by patients whom nurses can contact if required [28]. Additionally, the availability of external resources, including local referral services, was identified as a facilitator to care by Hassan et al. [27].

a. Therapeutic relationship.

The therapeutic relationship between patient and healthcare professionals was reported as an important aspect in four studies, valued by both patients and staff. Lee et al. reported that patients valued psychological space and protected time provided by trained mentors (e.g., healthcare or support professionals), enabling discussion about their concerns, and validating their experiences [29]. PRIMROSE-A also emphasised the fostering of positive therapeutic relationships, addressing additional patient needs and non-judgemental staff attitudes, contributing to positive patient experiences with this intervention [33]. The establishment of positive therapeutic relationships, with long term follow-up care and goal setting was also viewed as positive and rewarding by staff, as reported by Lawless et al. and Shaw et al. [28,33]. Hassan et al. reported that staff understood the value of therapeutic alliances in instilling confidence and trust in patients, which they cultivated, and resulted in increased patient willingness to engage [27].

b. Perceived intervention purpose and value.

Clarity of understanding of intervention purpose and value was mentioned in four studies. Hassan et al. reported generally good understanding of the purpose of PRIMROSE among patients, facilitated by clear explanations by general practitioners and information sheet provision [27]. This intervention was also perceived as valuable by staff, as they understood the potential improvements in patients’ physical health and quality of life, and by patients as they recognised the opportunity to improve their own health [27]. Patients also viewed the longer appointments in PRIMROSE as more holistic than usual GP appointments [27]. Shaw et al. mentioned that PRIMROSE-A was greatly valued by patients, especially due to addressment of their social needs during the COVID-19 pandemic through connection with peer educators and carers, though some patients incorrectly understood the intervention to be regarding mental health purely, and not cardiovascular health also [33]. Staff also found the experience of PRIMROSE-A rewarding, for example due to receival of positive patient feedback [33]. Lee et al. reported that patients felt optimistic prior to initiation of WHEEL, potentially fostered by information booklet provision, and appreciated the opportunity to share and learn about weight management in the context of living with SMI [29]. Sajatovic et al. reported that the vast majority of patients surveyed post intervention agreed that TTIM was useful, and that it covered issues most important to them and in general, regarding diabetes self-management in an SMI context [32].

c. Feedback for improvement.

Patients feedback for intervention improvement was mentioned in three studies. Lee et al. mentioned that patients recommended peer educator involvement in adjunct support to address feelings of isolation [29]. More structured mentor meetings were also requested to help patients with disorganised thoughts better share their experiences [29]. Requests were also made for further tailoring of intervention information booklets to address individual patient challenges and goals, and for sending of email appointment reminders to patients [29]. Patient preferences for appointment frequency and duration were mixed, with some preferring more frequent and rigid timings, and others more flexible scheduling [29]. Shaw et al. mentioned that some patients desired more follow-up appointments, longer appointments and further peer educator involvement [33]. Hassan et al. mentioned that staff recommended creation of designated timeslots for intervention delivery, to facilitate clinical implementation and longer appointments, with a need for further external resources and staff training also highlighted [27]. Patients also requested better integration of digital technology to track intervention progress [27].

Primary care interventions and modified primary care models.

Prior literature reviews and guidelines have identified, evaluated and explored CVD risk management interventions for SMI patients [9,16,34]. In contrast to our study, the aforementioned reviews did not exclusively focus on primary care interventions, by including studies set in secondary care, and they described pharmacological therapies, including metformin for weight management and statins for cholesterol management, which our review did not emphasise. A systematic review was also performed by Osborn et al. to identify pharmacological and behavioural interventions to inform the creation of the PRIMROSE intervention, which two studies in our scoping review covered [27,30,35]. This systematic review mentioned that most included trials were set in secondary care, highlighting a potential lack of research set in primary care, which our review corroborates as only 10 relevant primary care studies published in English between 2015 and 2025 were identified [35].

A study by Murphy et al. theorised a care continuum for CVD risk management in SMI patients, describing strategies such as clinical tool use and modified primary care models, including patient centred medical homes, which our review also identified [26,31,36]. Murphy et al. refers to the value of peer mentors in helping patients engage with chronic condition self-management, which was also identified by four studies in our review [26,28,32,33,36]. Furthermore, the issue of mental health stigma and unconscious bias among staff was identified, as in our review, with exploration of potential solutions, such as setting up of contacts between stigmatised individuals and healthcare workers, though the issue of limited data on methods to address this stigma was highlighted [27,36].

Only one study in our review explored the PCMH primary care model, though existing literature describes another PCMH called SMI PACT with similar features, including tailored SMI care provision, frequent follow-ups, and implementation of a similar collaborative team, excepting peer mentors, but not with cardiovascular risk reduction as a primary outcome [37,38]. SMI PACT was associated with improved overall chronic disease care and positive patient experiences, in part due to considerate and personalised communication, fostering therapeutic relationships [37,38]. This promise of the PCMH model in chronic disease management echoes Gertner et al. in our review, and the value of therapeutic relationships was a consistent finding across four studies in our review [26–29,33].

Qualitative interventional aspects.

Qualitative studies by Jakobs et al. and Burton et al. identified barriers and facilitators to existing CVD risk management for SMI patients in primary care, with similar features emerging in the context of novel interventions in our review [39,40]. For example, facilitators such as staff training and doctor-patient relationships, and barriers such as lack of staff knowledge, increased workload and mental health stigma among staff, also emerged in our review [27,29,32,33,39,40]. While prior literature has highlighted lack of information exchange between primary care and psychiatric services [39,40], this review more frequently identified communication challenges within primary care teams and between clinical and managerial staff [26,28,32,33]. This distinction suggests that both system-level integration issues and organisational or interpersonal communication barriers may impact implementation.

Burton et al. also described patient and staff identified strategies for care improvement, including enhanced care continuity and achievable goal setting [40]. Similar strategies also emerged in our review, both as considerations implemented in included studies, such as lengthened appointment times and personalised goal setting in PRIMROSE, and as participant feedback for intervention improvement, such as a desire for longer term follow-ups identified by Shaw et al. [27,30,33]. Costa et al. also found that primary care providers desired more information on suitable local services, such as weight loss referral programmes, echoing findings from Hassan et al. in our review [27,41].