Living with diabetes alongside a severe mental illness: A qualitative exploration with people with severe mental illness, family members and healthcare staff

Diabetes is two to three times more prevalent in people with severe mental illness, yet little is known about the challenges of managing both conditions from the perspectives of people living with the co‐morbidity, their family members or healthcare staff. Our aim was to understand these challenges and to explore the circumstances that influence access to and receipt of diabetes care for people with severe mental illness.

barriers to adopting behavioural changes that form a major part of successful diabetes management. 24,25 Little research, however, has explored the experiences of this co-morbidity from the viewpoint of those living with severe mental illness and diabetes, 24 and no studies have included the perspectives of those who provide both formal and informal care.

| Study aim
This qualitative study aimed to (a) explore the experiences of living with severe mental illness and diabetes and managing both conditions and (b) understand the circumstances influencing access to and receipt of diabetes care.
The research was part of the EMERALD project which is a mixed-methods study designed to increase our understanding about the increased risk of poor diabetes outcomes for people with severe mental illness. 26 The COREQ checklist was completed to aid transparent reporting of methods. 27

| Sampling strategy
2.1.1 | People with severe mental illness and diabetes Eligibility criteria: • Aged 18 years or over, with capacity to provide informed consent. • Recorded diagnosis of severe mental illness (schizophrenia, schizoaffective disorder, bipolar disorder, depression with psychosis), excluding those experiencing an acute relapse. • Diagnosis of type 1 diabetes or type 2 diabetes. • Living in the community (including supported housing, but not those admitted to acute hospital settings).
To capture a broad range of experiences, a maximum variation sampling strategy 28 was adopted (Table 1). Efforts were made to construct an ethnically diverse study sample, which was important because of the increased risk of diabetes in Black and South Asian populations. 8 Participants were recruited from six general practices (from 10 that volunteered to assist with recruitment) with a range of deprivation scores, from 1 for most deprived through to 10 for least deprived (a score combining seven domains of deprivation at small area level recorded on the National General Practice Profile 29 ), six mental health trusts in the North and North West of England, and via a research database containing details of participants with severe mental illness and diabetes who had previously consented to contact for future relevant research studies. 30 Staff from the recruiting sites initially invited any eligible person they identified and we continually monitored sample characteristics to aim for maximum variation, asking sites to target people with under-represented characteristics towards the end of recruitment. For example, the ethnic profile of our sample of people with severe mental illness and diabetes lacked diversity so we asked sites to focus specifically on recruiting people from a Black or South Asian ethnic group.
We aimed to recruit a minimum of 30 people with severe mental illness and diabetes, after which we monitored data saturation, adding new participants until no new information relevant to the study aim was forthcoming [S31].

| Family members and friends
We asked participants with severe mental illness and diabetes to identify family and friends who provided informal support who might be willing to take part in an interview, and identified additional family member participants through the existing research database. As people with severe mental illness and diabetes are less likely than the general population to have adequate social support [S32], a lower target sample size of 15-20 people was set.

| Healthcare staff
Healthcare staff involved in the commissioning or provision of mental or diabetes healthcare for people with severe mental illness (in primary and secondary care) were identified through university networks or the sites that were recruiting people with severe mental illness and diabetes. To gain a broad range of perspectives, we initially planned a Presence of other co-morbidities Whether they received treatment for severe mental illness from primary and/or secondary care Whether they received treatment for diabetes from primary and/or secondary care purposive sample of 15-20 people from several staff groups: commissioners/managers, clinicians and other staff, with an aim to continue recruiting participants until we achieved data saturation.

| Contact and consent
All potential participants were provided with written information about the study and asked to return a response form or contact the research team if they wished to participate. Those expressing an interest were contacted by SB (a female postdoctoral social scientist with qualitative research training and experience), who explained the study and arranged and conducted the interview. Written or audiorecorded verbal consent was given by all participants prior to data collection.

| Data collection
In-depth semi-structured interviews were employed using topic guides (Appendix 1) developed with reference to existing literature and with input from, and pilot testing with, members of a Patient and Public Involvement panel, DIAMONDS VOICE. Topic guides were designed to cover key areas of interest while minimising participant burden (important for people with severe mental illness-related cognitive and attentional difficulties [S33]), and employed flexibly allowing more or less time for participants as required.
Interviews were conducted between April and December 2018, in participants' homes for people with severe mental illness and family members, and at places of work or by telephone for staff. Interviews were audio-recorded and transcribed for analysis.

| Data analysis
The Framework method [S34], a form of thematic analysis, was used to analyse the data. This method combines the exploration of a priori concepts with the generation of themes derived inductively from the data. This approach enabled us to explore similarities and differences between the different participant groups. A coding framework was developed and applied to the data (undertaken by JL and LK with input from SB and JT), and themes and the relationships between them were identified and developed (undertaken by JL, JT, SB and CEWK). To explore differences between participant groups, the analysis for each group was conducted separately yet with a dialogue between them to enable cross-group comparison; for example, the coding framework (Appendix 2) for participants with severe mental illness and diabetes was tested against family member and staff data and adapted to be responsive to differing perspectives. To enhance the rigour of the analysis, emerging themes were challenged by data from divergent accounts. This process enabled the development of a more nuanced understanding of the data. NVivo software [S35] was used to manage and code the data. Themes were developed through team discussions and regularly checked against the codes to ensure that, while abstracted to a more conceptual level, they captured and represented the accounts of participants. The analytic process involved regular discussion with the broader study team including a representative from DIAMONDS VOICE. To assure authenticity of study findings, themes were discussed at two workshops involving people with severe mental illness and diabetes and family members, with one also attended by staff.

| RESULTS
In all, 78 people were interviewed; 39 people with severe mental illness and diabetes, 9 family members and 30 healthcare staff. The majority of people with severe mental illness and diabetes (n = 30) were recruited from NHS mental health trusts; 7 were recruited from general practices and 2 were recruited from the research database.
Duration of interviews with people with severe mental illness and diabetes was between 17 and 98 min (median =45), between 23 and 97 minutes (median =52) with family members and between 26 and 75 minutes (median =43) with staff. Six participants (4 staff and 2 people with severe mental illness and diabetes) declined audio-recording, so detailed notes were made during and immediately after interviews. All interviews were conducted in English apart from one, where a translator was used for an interview with a Punjabispeaking participant.
The sample of people with severe mental illness and diabetes (Table 2), included 22 men (56%) and 17 women (44%). Schizophrenia was the most common severe mental illness diagnosis (n = 22, 56%) followed by bipolar disorder (n = 13, 33%). Although the majority (n = 36, 92%) had type 2 diabetes, the ordering and duration of diabetes and mental illness diagnoses varied across the sample, as did participants' experiences of care and treatment. The age of participants ranged from 28 to 71 years (mean age: 53 years), and the sample included seven people from a minority ethnic group. Two participants (5%) were employed, five were retired (13%) and 32 (82%) were unemployed, although several of the retired and unemployed participants had previously been employed.
Although friends were included in the definition of a person providing support for someone with severe mental illness and diabetes, all nine participants (see Table 3) were family members (spouses, n = 6; parents, n = 2; adult children, n = 1). Six were women and three men; all male participants were spouses.
Healthcare staff participants were from varying disciplines and roles spanning mental health, primary care and diabetes services. Nursing was the most represented profession (n = 13; see Table 4).
Five overarching themes were identified from the analysis that featured across participants' accounts of living with severe mental illness and diabetes, and were reflected in the perspectives of many of the staff and family members who took part in the study. These are described below, with illustrative quotes for each theme/sub-theme presented below and in Table 5 (those with type 1 diabetes have been identified with the designation T1). We have also indicated where relevant how participants varied in their experiences, for example drawing attention to instances that were only talked about by a few participants or most. This is to guide the reader through our data and is not meant to infer that similar relationships may exist in the wider population of people living with these conditions, which is not the aim of qualitative research.

| Theme 1: Severe mental illness governs everyday life
Most participants from all three groups provided detailed accounts of the pervasive effect of mental health problems on daily lives; affecting people with severe mental illness' ability to leave the house, work, retain a driving licence, engage in personal care or household management, make or maintain friendships, and manage appointments. Severe mental illness was also occasionally reported to affect personal finances, although only family members and staff explicitly highlighted ways in which this affected diabetes management, for example by limiting access to transport and healthy foods.
All three participant groups commonly reported that severe mental illness overshadowed the importance of diabetes, the treatment of which was often 'governed by mental health' [ES-T2-02, person with bipolar disorder]. Indeed, one spouse noted that 'I know you're looking at diabetes and mental health… but from my point of view, the issues are, without a doubt, the mental health issues' [ES-T2-17, family member].
The precedence of severe mental illness over diabetes was often most apparent when participants described periods of poor mental health, where the focus was on immediate needs and survival: 'If part of your mind is thinking of suicide, it's like very difficult then, to be particularly panicky when someone says your blood sugar's gone up by two points' [ES-G2-01, person with bipolar disorder]. For a few participants with severe mental illness and diabetes, there was a sense that diabetes could be addressed in the future, in part because of the less immediate impact on their lives compared to severe mental illness. One staff participant understood this as a lack of optimism from people with severe mental illness that they would live long enough to be affected by diabetes complications, and this sentiment was reflected in the accounts of several participants with severe mental illness and diabetes as they spoke about their health.
A further example of the foregrounding of mental health was found in the majority of participants' accounts of their diabetes care, which they discussed with less specificity and engagement than when describing their mental healthcare. Descriptions by those with type 2 diabetes were often perfunctory, and even when further questions were asked, participants tended to give little more than brief descriptions of regular checks they received such as blood tests or blood pressure checks. The three participants with type 1 diabetes spoke in more detail about their diabetes care, although they Taking medication could itself exacerbate these feelings, and a few participants perceived a lack of control over their severe mental illness and its treatment: 'the side effects of the drugs are not good … you wouldn't take them if you had the choice. I mean, I am desperate, so I take my medication' [ES-T2-16, person with schizophrenia]. Linked to feelings of powerlessness was an acceptance among many participants from all three groups that when mental health problems dominated daily life, diabetes would be neglected, as this participant with bipolar disorder explained, 'When you're mad as a hatter, you don't take any notice… I was offered that [diabetes self-management education] by the GP service, and I went, "no thanks, I'm too busy being mad"' [ES-T2-03]. People with severe mental illness and diabetes "the sort of GP service had tried to do a diagnosis, tried to tell me it (diabetes) was important, which was never going to have any impact, 'cause when you're mad as a hatter, you don't take any notice. They're just noise in a corner" [ES-T2-03] "Whereas, I think for somebody who's diabetic, it would be, your main focus would be on the diabetes, and you'd be trying to work out what you're going to eat, and make sure, whereas for me, I'm always worried about the mental health side, and if I am in a really bad place, I don't care anyway." [ES-G2-01] "But when you're having a real bad episode all that goes out the window.
And you're not looking after yourself. You're not looking after your diabetes either. And unless somebody is really on top of it with you, you can get into a mess." [ES-T3-07] "Because you can't think clearly, that's…not being able to think clearly is hard, because then you're not applying the right tools to address your diabetic situation." [ES-T7-04, T1]

Family members
Staff "I've got another patient when her mental health deteriorates she often goes into crisis and it's often related actually to benefits and things like that so her self-harm increases, her diet increases but she does just stop taking her tablets as well just because she can't be bothered with it really, so a lack of motivation you know I don't want to be here therefore why should I bother about this because I don't care, I forget the complications because I might not be here long enough to get them." [Staff ES-PC−05] "I think it's very easy to focus on physical side of things but the mental health is just a massive side of it and I think if a patient has I think if we were more trained and more aware of the mental health of a patient I think it could probably alleviate a lot of the problems for both patients and the staff. Because somebody's proper clinically depressed and flat, they don't even eat let alone go and see the GP or even get out of bed…" [Staff ES-T5-03] "They're relapsing but they're not risky. They're like, well…just see your GP, would be the answer. In the meantime, they're not eating or they're eating chocolate for five meals a day, because it's easy and they're not sleeping anymore. Their weight is ballooning or shrinking, they're getting more physically unwell, their blood sugars are raging. Staff "when it's not managed well, when they're not taking their insulin or taking their metformin when they should. When the blood sugars are high, they become more aggressive, argumentative, that's often a sign that they're not managing it properly and it might not just be a sign of their mental health deteriorating because it could be physical cause like diabetes and it's a good thing that we have these clinics where this can be monitored a bit, you know, blood glucose." [Staff (Community Mental Health Nurse) ES-T1-01] "And on the other hand, if people are diabetic and they're not maintaining their blood sugars right, it can give rise to symptoms which are synonymous with anxiety and low mood, so there is a big interface between diabetes and mental illness." [Staff ES-T1-01] "And you know, sometimes, the irritability that comes with a low blood sugar could be interpreted as part of somebody's mental illness. And it's diagnosed and mistreated." [Staff ES-T5-03] "I mean certainly when they are acutely unwell their mental health will really slide and often that can be either that their diabetes is not well controlled or and then that leads on to picking up more infections and being susceptible to things. Quite often we will see them acutely because of a behavioural change that has happened as a result of an acute infection that has usually happened as a result of their diabetes not being brilliantly controlled.

Illustrative participant quotation (f = female, m = male, numbers indicate age at interview)
Family members "you know, that's keeping him reasonably fit except that the diabetes and the drugs -all the drugs say that they're likely to put the blood sugar up and as a result of taking the antipsychotics and the antidepressants or whatever the other one is -he takes metformin to keep the sugar down." [Family member ES-D1-02] Staff "On the flip side, the patient can begin as not diabetic, but is known to the services for a long time with severe and enduring mental illness and is on a range of antipsychotic medication, which then gives rise to… not, as such, direct weight gain, but an increase in diet, an increase in weight. Then, at some point, they can develop diabetes or poor blood sugar control, and that happens quite commonly. We call it Metabolic syndrome." [Staff ES-T1-01] "I think the difficulties that have been experienced often is that because of the mental health medication they are on they have got a very big appetite so they end up eating the wrong foods.

Illustrative participant quotation (f = female, m = male, numbers indicate age at interview)
Family members "if psychiatry caused the diabetes, why must the GP be doing it? Why is mental health not interested enough to see what's going on…because very seldom does your GP talk to your psychiatrist. Maybe once in a while. They don't even share the same computer system" [Parent ES-T2-08] "Like in three minutes he'll drink three milkshakes, and then he'll vomit.
And then he can't understand where it came from. And then they make us appointments with the stomach doctor. And I said the head doctor should be working on this, it's got nothing to do with the stomach doctor. It's because the psychiatrist does nothing. We're ending up at the stomach doctor, who thinks I'm stupid because do you know, with that kind of intake you should be vomiting. I do know that. I know that, but what do you want me to do about it?" [Family member ES-T2-08] "But now since it's out of hospital we have to get one dosette box from the GP and a separate dosette box from the psychiatry, with different medicine in it. "I do find that information -even though technology is a wonderful thing and should make everything easier sometimes things don't get passed on that should get passed on. Certainly, it takes ages for any decisions about medication from the psychiatrist doctor at [mental health centre] -he changes his medication. It seems to take ages until they readjust." [Family member ES-D1-02] Staff "What is our role in supporting them… I think at the moment it appears that we are the only ones that are actually supporting them in managing their diabetes because I think there is very little support from secondary care in the management of Type 2 diabetes." [Staff ES-PC−02] "I think it's very easy to focus on the physical side of things but the mental health is just a massive side of it and… I think if we were more trained and more aware of the mental health of a patient I think it could probably alleviate a lot of the problems for both patients and the staff" [Staff (Practice Nurse) ES-PC-04] "the credit will probably have to go to the health care assistants that we've got who often see these patients almost week in week out, monitoring weight and blood pressures…and able to develop really really nice relationships of trust and really plugging away at the health understanding behind some of the lifestyle management with diabetes and we've had some really good successes in terms of really impressive weight loss" [ES-PC−06, GP] "We would have to liaise which can be really frustrating because it's not a quick answer that you get, you've got the single point of access entry into mental health and so if somebody isn't under services at the moment and you are wanting to get some information on medication review or whatever you have to go through a single point of access and they get assessed by the wellbeing team and then they get passed on to whoever they feel is most appropriate person to see them or the team and it's very long winded. It can take up to 4/6 weeks to get a simple medication review or even to get some advice sometimes which is not brilliant really.

| Theme 2: Mood influences diabetes self-management
Most of the participants with severe mental illness and diabetes described the fluctuating moods they experienced as part of daily life, distinguishing these from the more extreme symptoms that occur during a relapse. Feelings of low mood, depression, stress or anxiety, which several participants linked with worrying about diabetes, were reported by many participants (including several family members and staff) to derail participants' attempts to manage their diabetes through exercise and diet, leading to lethargy and lack of motivation, and, frequently, comfort eating. One family member noted, for instance, that when her mother is 'frustrated or angry or just not feeling 100%, she won't be disciplined, especially with diet' [ES-T3-08, family member]; similarly, a GP had observed people with severe mental illness 'who, when their mental health deteriorates, their eating deteriorates, so they may start to comfort eat… and so they lose their diabetic control' [ES-PC-05]. More extreme declines in mood were described by several participants to engender a sense of abject helplessness where any inclination to manage diabetes could dissipate entirely: Several participants from all three groups discussed how both conditions could affect mood, and described the impact that low mood or anxiety could have on both conditions. However, very few participants, including staff, talked about managing mood as a priority, even though, as one psychiatrist explained, the consequences could be significant: 'six months down the line and they're acutely psychotic … their blood sugars are all over the place … then you've given yourself two big problems to manage, what would have been one small problem' [ES-T2-12].

| Theme 3: Cumulative burden of managing multiple physical conditions
Problems relating to mental health were often not the only challenge to diabetes management. Nearly all of the participants with severe mental illness and diabetes reported additional health problems such as chronic obstructive pulmonary disease (COPD), asthma, cardiovascular disease, obesity, musculoskeletal problems, sleep problems or pain. Participants from all three groups described two common ways in which this impacted on diabetes management. First, by limiting their ability to engage in physical activity or leave the house to see others or attend appointments: 'well I've got my knee that's difficult for a start off … It stops me from doing a lot of walking what I used to like doing' [ES-T2-04, person with schizophrenia].
The second challenge was deciding which condition to prioritise, and several participants pointed out that the invisibility and perceived lack of immediate consequences of diabetes could mean that, like severe mental illness, physical health Theme Sub-theme Participant group

Illustrative participant quotation (f = female, m = male, numbers indicate age at interview)
"Everyone who's involved with that has responsibility to understand both sides. The psychiatrist needs to understand the physical impact of both the condition and the treatment. The GP needs to understand the interactions between the two. Any physician, medical professional, acute hospital needs to understand that having the two together is going to make things more complicated. There isn't a uniform answer that one side fits all approach." [Staff ES-T2-12] "it's not tailored to people with SMI so you know the courses aren't necessarily run in places that are local to them, they are not run in a way that's kind of made in any way enticing to them. I mean for example I sent a chap on a Desmond course recently you know the Diabetes Education course and he got kicked off because they didn't like his behaviour, well you know he's a chap with schizophrenia and actually even getting him there was massive and unfortunately you know he was never going to sit there and behave like everybody else so yeah kind of getting them to access stuff is very multi layered I suppose, it's not just getting them in the door there are lots of other bits that need to happen to really make that work.

| Theme 4: Interacting conditions and overlapping symptoms
Many, but not all participants across the three groups reported an overlap in severe mental illness and diabetes symptoms, and several participants (including staff) noted that it could be difficult to identify the underlying cause of symptoms such as fatigue, low mood, agitation or anxiety: 'I've noticed there's quite an overlap between feeling mentally low and feeling unwell because your blood sugar is up' [ES-G8-01, person with bipolar disorder and T1 diabetes]. Several staff participants observed that fluctuating levels of blood glucose could manifest as symptoms which could be readily misinterpreted as psychological symptoms and 'give rise to symptoms which are synonymous with anxiety and low mood' [ES-T1-01, psychiatrist]; or lead to people becoming more 'aggressive, argumentative… That's often a sign that they're not managing (diabetes) properly and it might not just be a sign of their mental health deteriorating' [ES-T1-02, community mental health nurse].
The overlapping nature of symptoms was occasionally reported to have implications for diabetes management, as one participant explained, 'if I go on a high, sometimes, they've got to check my blood sugars, because they don't know if it's the blood sugars that are causing me to go a bit loopy. Or it's my mental illness' [ES-T4-10, person with bipolar disorder]. This led a few participants with severe mental illness and diabetes to question why they did not receive more regular diabetes checks: In diabetes what you're missing is the physiological feedback and a consultation can, to a degree, give you some of the feedback, even if it's only three times a year … Because you can lose the plot over the course of a year, whereas I think if you have a horizon of four months, that gives you an end point in sight.
[ES-G8-01, person with bipolar disorder and T1 diabetes] However, there were divided opinions and some uncertainty about the exact relationship between severe mental illness and diabetes, with most participants across the three groups describing the conditions as interacting, while a few participants with severe mental illness and diabetes saw no link and several (including staff) described a 'direct correlation' [ES-G2-01, person with bipolar disorder] or even a causal pathway: 'anxiety makes [blood sugar levels] go up as well' [ES-T3-03, person with bipolar disorder]. As one staff participant noted, this uncertainty could potentially lead to misdiagnosis: 'sometimes, the irritability that comes with a low blood sugar could be interpreted as part of somebody's mental illness. And it's diagnosed and mistreated' [ES-T5-03, nurse]. Linked to this, some staff identified a training need: Many participants in all groups perceived a relationship between the adverse effects of medications prescribed for severe mental illness and the development and management of diabetes. Commonly reported adverse effects were increased hunger, lethargy and weight gain, which were reported to impact on people's motivation, mood and capacity to live healthily and manage their diabetes. As one participant explained, 'I'm not really a greedy person but when you get the hunger from the tablets … you can't ignore it' [ES-T2-16, person with schizophrenia].

| Theme 5: Support for everyday challenges
Many participants across the study sample highlighted the central role of family members, friends or healthcare staff in providing practical or emotional support for the everyday challenges that impact on diabetes management. Several people with severe mental illness talked about the importance of having someone to accompany them shopping, to appointments, for exercise or to a café while others valued help at home, for example with finances, cooking or organising medications. Talking to friends, other people with mental health problems or a known professional helped some participants too. For a few participants, engaging in social activities such as visiting a day centre or place of worship could provide 'another motivation (to) carry on' [ES-G9-01, person with schizophrenia]. For several participants with good family support, the supporting family member could also act as a sentinel, watching for symptoms to emerge, 'my daughter is the one, she can tell by my voice when I'm not right' [ES-T3-04, person with bipolar disorder].
Although in most cases this type of care was provided by a family member or mental healthcare co-ordinator, a few participants with severe mental illness and diabetes expressed a desire for more intensive support, for example assistance with budgeting, dietary planning and exercise regimes: 'It should be a lot more help. Not just from GPs and nurses but there should be teams going out into communities and people what are really overweight and really obese, they should be sitting them down and going over a budget plan and a plan to lose weight' [ES-G3-01, person with schizophrenia].
However, despite the value placed on this type of support, many participants with severe mental illness and diabetes did not feel well supported. Several had lost informal support due to a breakdown in relationships, not feeling able to talk to others, or the illness or death of a family member. Staff occasionally highlighted potentially negative effects of informal support, citing examples of family members encouraging unhealthy behaviours or not understanding mental illness, and identified a need for education for those in a supporting role.
A lack of continuity of care was identified by several participants across the groups as a key barrier to accessing personalised support from healthcare staff: 'there is a lot of difference between somebody knowing you and just seeing different people each time' [ES-D1-02, family member]. Perceptions among some people with severe mental illness and their family members that healthcare staff's time was limited and their roles prescribed were also barriers to them seeking support, and could lead to them using physical and mental health services according to the traditional divide between them. For example, one participant would not discuss her paranoia with a GP 'because it would take up an hour of a GP's time, and that's not fair, not fair on the GP and not fair on the other patients' [ES-T2-16, person with schizophrenia]. This division was also acknowledged by several staff participants from primary care, 'our annual [diabetes] review it isn't really to do with the mental illness' [ES-PC-01, GP], and mental health services, 'they don't tend to ask about it [diabetes] because they don't see it as part of a mental health nurse's role' [Staff ES-T1-02, mental health nurse].

| Summary of key findings
This qualitative study provides important insights into how co-morbid severe mental illness and diabetes is experienced. This often occurs in the context of multiple other health conditions and against the backdrop of additional challenges relating to employment and social support. Notably, the complex interaction between the two conditions highlights the important role of mood and of severe mental illness medications in diabetes self-management, the difficulties of differentiating between overlapping symptoms, the limited or variable prioritisation of diabetes care and management within the context of severe mental illness and other co-morbidities, and the barriers to accessing support for everyday challenges.

| What this study adds
This study adds to growing evidence that diabetes management is overshadowed by the many competing mental and physical health needs experienced by people with severe mental illness, which when experienced together can be overwhelmingly pervasive [10, 21, 24, 25, S36]. By including the perspectives of people living with severe mental illness and diabetes, family members and healthcare staff who support them, this study offers new insights. It confirms that diagnostic overshadowingthe attribution of physical symptoms to coexisting mental illness, leading to under-recognition of physical conditions such as type 2 diabetes [S37]-extends beyond the diagnostic period and affects diabetes management. The study also suggests that the separation of diabetes and mental healthcare, and the difficulties people experience in distinguishing between certain mental illness and diabetes symptoms both contribute to the overshadowing of diabetes. The latter is not commonly reported as a barrier to diabetes self-management [S38], but in the context of severe mental illness can contribute to diabetes being continually deprioritised until more distinguishable symptoms, changes in treatment, complications or related conditions occur, leading to increased morbidity and mortality. 14 The persistent and fluctuating nature of depression and anxiety among the people with severe mental illness and diabetes in this study, which impacted on their capacity and motivation for diabetes self-management, offers another potential explanation for the poor outcomes in this population. Research consistently shows that co-morbid depression increases the risk of mortality in people with diabetes [S39], and presents challenges for diabetes self-management [S40]. While depressive symptoms are common in people with severe mental illness [S41], they are often overlooked in the literature about co-morbid diabetes, which tends to focus on psychotic symptoms and their treatment. This study suggests that, like diabetes, depression and anxiety may also be overlooked in practice because of the focus on managing and preventing psychotic symptoms.

| Study strengths and limitations
Robust qualitative methods, aided by extensive patient and public involvement, allowed us to develop a rich, detailed understanding of how experiencing severe mental illness and diabetes alongside each other can impact on diabetes management. We gained additional insights by including people who support this population, enhancing our interpretation of study findings. However, while representing a diverse range of views and experiences, the study excluded people who were not in contact with healthcare services and people experiencing a psychotic episode. Additionally, the experiences of people with type 1 diabetes, those who were very unwell with diabetes and people from a minority ethnic group were under-represented, and our sample of family members and friends was smaller than planned. Care should therefore be taken when considering the transferability of findings to these groups, although by including participants with type 1 diabetes or diabetic complications we have importantly highlighted the similar ways in which severe mental illness impacts on diabetes management regardless of diabetes type, treatment or severity.

| Implications for clinical practice and research
Approaches to tackle the systemic overshadowing of diabetes are needed to ensure that it is afforded appropriate priority in the context of severe mental illness. In particular, more intensive and tailored support is needed to help people overcome the multiple barriers to self management, especially when their mental or physical health deteriorates. Clinical guidelines for multimorbidity provide a useful starting point [S42], recommending an individualised approach to care that takes account of how a person's conditions interact and impact on their lives. However, while these guidelines acknowledge the importance of care co-ordination, more needs to be done to tackle the traditional 'silo-working' of mental and physical health services, which in this study led people with severe mental illness and diabetes as well as healthcare staff to focus on one condition at a time.
Collaborative care models may offer some potential here, enhancing co-ordination between mental and physical health services, providing patients with personalised and more regular diabetes care, and ensuring staff have access to training and specialist knowledge which this study found was lacking. However, while this model is effective for managing depression and co-morbid long-term physical health conditions including diabetes [S43], there is very limited, albeit promising evidence about its value for co-morbidity in severe mental illness [S44].
Offering better continuity of care may also help. This was highly valued by the people with severe mental illness and diabetes in this study, and for some helped to compensate for the lack of informal support we observed in many participants. Although challenges to this were identified, including lack of time and continuity of staff, recent evidence from a large observational cohort study in England found that greater continuity in primary care for people with severe mental illness was associated with a reduction in unplanned hospital use, leading the authors to suggest that better relational continuity (i.e. seeing the same physician) may improve the management of physical health in this population [S45].
Finally, providing bespoke diabetes education for people with severe mental illness, their family members and healthcare staff may help to address the unique barriers to diabetes management this population experience, regardless of diabetes type, treatment or severity. For example, introducing strategies to manage the impact of low and fluctuating mood and the side effects of severe mental illness medication, and helping people to distinguish between overlapping mental illness and diabetes symptoms, may contribute to improvements in diabetes self-management.

BELLASS Et AL.
Committee C (Mental Health, Women and Children's Health) until July 2019. DS is an expert advisor to the National Institute for Health and Care Excellence (NICE) centre for guidelines and a member of the current NICE guideline development group for Rehabilitation in adults with complex psychosis and related severe mental health conditions; a Board member of the National Collaborating Centre for Mental Health (NCCMH); a Clinical Advisor (paid consultancy basis) to the National Clinical Audit of Psychosis (NCAP); these are the personal views of DS and not those of NICE, NCCMH or NCAP. DS has received personal fees from Wiley Blackwell publication 'Promoting Recovery in Early Psychosis' 2010, ISBN 978-1-4051-4894-8, joint editor in receipt of royalties, outside the submitted work; personal fees received as member of the current NICE guideline development group for Rehabilitation in adults with complex psychosis and related severe mental health conditions. SB, JL, CEWK, LK, TD, LH, RJ, SLP, NS and JT declare no conflicts of interest.

ORCID
Charlotte Emma Wray Kitchen https://orcid. org/0000-0002-9323-0061 Family history 22 29 First port of call for concerns 26 27 Good days and bad days 5  Timing of care received 6 14 Understanding of care received 6 9 Wishes for and thoughts on improvements 28 56 Worries about healthcare 7 12

Experience of mental illness 39 855
Behaviours associated with mental illness 6 12 Burden of mental illness 4 5 Coping mechanisms 10 19 Crisis points 20 31 Current state of mental health 31 87 Disclosing mental illness 10 28 Duration or timings of illness and treatment 38 75 Effect of outside influences 8 18 First port of call for concerns 33 41 Good days and bad days 26 39 Impact of mental illness 27 97 Not feeling in control 3 7